Hey,
After some problems with the UART of 115200 with 32kHz (Yes, I know, it's
horrible) I am designing a new circuit to have a 8MHz crystal (or so) since from
what I understand that's the only way to really have very low errors. I
experienced horrible errors using just the 32kHz crystal.
I would like to have your opinions, since many of you are very experienced, what
would be the best:
1) Just an 8MHz crystal on the MSP430F1611 is enough? if so, which kind (perhaps
you know a part number or some technical specifications such as load
capacitance, accuracy, etc)
2) Should I just opt for another MSP430 derivative with an FLL or more advanced
clocks? Would the MSP430F5xx family be good for this?
I really don't want to worry much about errors.
The reason I am hesitant to leave the MSP430F1611 is because I am using the BSL
like a telos mote does: It uses a switch so that I can both use the BSL and UART
(it switches between the two). Therefore, would upgrading to another MSP430
change my capabilities to use MSPGCC's invert pins and all that that
enables to program with BSL and use UART?
I realize this might be a bit complex, but any help from you guys is
appreciated.
Also, any pointers to any literature discussing this is appreciated as well.
GB
Low error UART communication MSP430F1611
Started by ●January 28, 2009
Reply by ●January 28, 20092009-01-28
If you want to let it sleep but be able to communicate: use 9600 bit/s from
the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
internal oscillator via SMCLK. No extra crystals needed -- the high speed
types do consume a lot of power.
If power consumption is not an issue (which is likely in your case): leave the
SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had practically no
errors here so far.
BTW: errors at 115200 bit/s may also be due to the interrupt latency, a telos
can not use the USART as duplex channel. Go for 57600 bit/s instead, or ensure
simplex.
Best, Andreas
GB wrote:
> Hey,
>
> After some problems with the UART of 115200 with 32kHz (Yes, I know, it's
> horrible) I am designing a new circuit to have a 8MHz crystal (or so) since
> from what I understand that's the only way to really have very low errors.
> I experienced horrible errors using just the 32kHz crystal. I would like to
> have your opinions, since many of you are very experienced, what would be
> the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
> which kind (perhaps you know a part number or some technical specifications
> such as load capacitance, accuracy, etc) 2) Should I just opt for another
> MSP430 derivative with an FLL or more advanced clocks? Would the MSP430F5xx
> family be good for this?
>
> I really don't want to worry much about errors.
>
> The reason I am hesitant to leave the MSP430F1611 is because I am using the
> BSL like a telos mote does: It uses a switch so that I can both use the BSL
> and UART (it switches between the two). Therefore, would upgrading to
> another MSP430 change my capabilities to use MSPGCC's invert pins and all
> that that enables to program with BSL and use UART?
>
> I realize this might be a bit complex, but any help from you guys is
> appreciated. Also, any pointers to any literature discussing this is
> appreciated as well.
>
> GB
>
>
the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
internal oscillator via SMCLK. No extra crystals needed -- the high speed
types do consume a lot of power.
If power consumption is not an issue (which is likely in your case): leave the
SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had practically no
errors here so far.
BTW: errors at 115200 bit/s may also be due to the interrupt latency, a telos
can not use the USART as duplex channel. Go for 57600 bit/s instead, or ensure
simplex.
Best, Andreas
GB wrote:
> Hey,
>
> After some problems with the UART of 115200 with 32kHz (Yes, I know, it's
> horrible) I am designing a new circuit to have a 8MHz crystal (or so) since
> from what I understand that's the only way to really have very low errors.
> I experienced horrible errors using just the 32kHz crystal. I would like to
> have your opinions, since many of you are very experienced, what would be
> the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
> which kind (perhaps you know a part number or some technical specifications
> such as load capacitance, accuracy, etc) 2) Should I just opt for another
> MSP430 derivative with an FLL or more advanced clocks? Would the MSP430F5xx
> family be good for this?
>
> I really don't want to worry much about errors.
>
> The reason I am hesitant to leave the MSP430F1611 is because I am using the
> BSL like a telos mote does: It uses a switch so that I can both use the BSL
> and UART (it switches between the two). Therefore, would upgrading to
> another MSP430 change my capabilities to use MSPGCC's invert pins and all
> that that enables to program with BSL and use UART?
>
> I realize this might be a bit complex, but any help from you guys is
> appreciated. Also, any pointers to any literature discussing this is
> appreciated as well.
>
> GB
>
>
Reply by ●January 28, 20092009-01-28
You didn't explain the source/type of the errors you are seeing. Software?
Hardware? Environment? (e.g temperature, noise, distance)?
Unless you are going very long distances in a very noisy environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
As A.K. suggested if power is not a concern and for whatever reason you must use an external Xtal, I would suggest going with a 7.3728MHz (or multiple) Xtal for most accurate baud rate generation.
C.
________________________________
From: Andreas Kke
To: m...
Sent: Wednesday, January 28, 2009 4:35:29 PM
Subject: Re: [msp430] Low error UART communication MSP430F1611
If you want to let it sleep but be able to communicate: use 9600 bit/s from
the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
internal oscillator via SMCLK. No extra crystals needed -- the high speed
types do consume a lot of power.
If power consumption is not an issue (which is likely in your case): leave the
SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had practically no
errors here so far.
BTW: errors at 115200 bit/s may also be due to the interrupt latency, a telos
can not use the USART as duplex channel. Go for 57600 bit/s instead, or ensure
simplex.
Best, Andreas
GB wrote:
> Hey,
>
> After some problems with the UART of 115200 with 32kHz (Yes, I know, it's
> horrible) I am designing a new circuit to have a 8MHz crystal (or so) since
> from what I understand that's the only way to really have very low errors.
> I experienced horrible errors using just the 32kHz crystal. I would like to
> have your opinions, since many of you are very experienced, what would be
> the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
> which kind (perhaps you know a part number or some technical specifications
> such as load capacitance, accuracy, etc) 2) Should I just opt for another
> MSP430 derivative with an FLL or more advanced clocks? Would the MSP430F5xx
> family be good for this?
>
> I really don't want to worry much about errors.
>
> The reason I am hesitant to leave the MSP430F1611 is because I am using the
> BSL like a telos mote does: It uses a switch so that I can both use the BSL
> and UART (it switches between the two). Therefore, would upgrading to
> another MSP430 change my capabilities to use MSPGCC's invert pins and all
> that that enables to program with BSL and use UART?
>
> I realize this might be a bit complex, but any help from you guys is
> appreciated. Also, any pointers to any literature discussing this is
> appreciated as well.
>
> GB
>
>
Unless you are going very long distances in a very noisy environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
As A.K. suggested if power is not a concern and for whatever reason you must use an external Xtal, I would suggest going with a 7.3728MHz (or multiple) Xtal for most accurate baud rate generation.
C.
________________________________
From: Andreas Kke
To: m...
Sent: Wednesday, January 28, 2009 4:35:29 PM
Subject: Re: [msp430] Low error UART communication MSP430F1611
If you want to let it sleep but be able to communicate: use 9600 bit/s from
the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
internal oscillator via SMCLK. No extra crystals needed -- the high speed
types do consume a lot of power.
If power consumption is not an issue (which is likely in your case): leave the
SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had practically no
errors here so far.
BTW: errors at 115200 bit/s may also be due to the interrupt latency, a telos
can not use the USART as duplex channel. Go for 57600 bit/s instead, or ensure
simplex.
Best, Andreas
GB wrote:
> Hey,
>
> After some problems with the UART of 115200 with 32kHz (Yes, I know, it's
> horrible) I am designing a new circuit to have a 8MHz crystal (or so) since
> from what I understand that's the only way to really have very low errors.
> I experienced horrible errors using just the 32kHz crystal. I would like to
> have your opinions, since many of you are very experienced, what would be
> the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
> which kind (perhaps you know a part number or some technical specifications
> such as load capacitance, accuracy, etc) 2) Should I just opt for another
> MSP430 derivative with an FLL or more advanced clocks? Would the MSP430F5xx
> family be good for this?
>
> I really don't want to worry much about errors.
>
> The reason I am hesitant to leave the MSP430F1611 is because I am using the
> BSL like a telos mote does: It uses a switch so that I can both use the BSL
> and UART (it switches between the two). Therefore, would upgrading to
> another MSP430 change my capabilities to use MSPGCC's invert pins and all
> that that enables to program with BSL and use UART?
>
> I realize this might be a bit complex, but any help from you guys is
> appreciated. Also, any pointers to any literature discussing this is
> appreciated as well.
>
> GB
>
>
Reply by ●January 28, 20092009-01-28
OK. I will explain. perhaps I can correct some things easier.
I use an MSP430F1611 with a 32kHz crystal.
My UART is @ 115200.
When I initialize the system I make the clock initialize to 4MHz:
TI_SetDCO(TI_DCO_4MHZ);
This is using TI's code, which I'm pretty sure you know. So it's not 4MHz exactly, it's
DELTA*(4096) = 4001792 Hz.
I used this in the MSPGCC calculator for baud rate , and it tells me max error is 1.37%, but it's more than that.
One issue is that I also don't have ROSC (and can't install one either so I'm making a new board) to lower the temperature dependence of the DCO.
My UART routine is:
void UART_0_115200_Setup()
{
P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
UCTL0 |= CHAR; // 8-bit character, SWRST = 1
UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
// For UART 0 Calculated using Baud rate generator found in mspgcc page
UBR00 = 0x22;
UBR10 = 0x00;
UMCTL0 = 0xDD;
UCTL0 &= ~SWRST; // Initialize USART state machine
IE1 |= URXIE0; // Enable USART0 RX interrupt
}
My circuitry is not very close to the MSP430. There are a couple of inches of distance.
Can this affect it? I guess crosstalk and other things could. Building a new circuit would allow me to eliminate all of this I guess, but I am always looking to improve my code also.
RIght now I am not exactly using the telos configuration. I use the FT2232D with its two channels, one for BSL and another for UART.
Since this board would be USB powered, power consumption isn't a concern.
Do you have any other info about the 7.3728MHz crystal. Do I need a specific load capacitance or ESR?
Thanks a lot for your replies
--- On Wed, 1/28/09, Candemir Orsan wrote:
From: Candemir Orsan
Subject: Re: [msp430] Low error UART communication MSP430F1611
To: m...
Date: Wednesday, January 28, 2009, 4:45 PM
You didn't explain the source/type of the errors you are seeing. Software? Hardware? Environment? (e.g temperature, noise, distance)?
Unless you are going very long distances in a very noisy environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
As A.K. suggested if power is not a concern and for whatever reason you must use an external Xtal, I would suggest going with a 7.3728MHz (or multiple) Xtal for most accurate baud rate generation.
C.
____________ _________ _________ __
From: Andreas Kke
To: msp430@yahoogroups. com
Sent: Wednesday, January 28, 2009 4:35:29 PM
Subject: Re: [msp430] Low error UART communication MSP430F1611
If you want to let it sleep but be able to communicate: use 9600 bit/s from
the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
internal oscillator via SMCLK. No extra crystals needed -- the high speed
types do consume a lot of power.
If power consumption is not an issue (which is likely in your case): leave the
SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had practically no
errors here so far.
BTW: errors at 115200 bit/s may also be due to the interrupt latency, a telos
can not use the USART as duplex channel. Go for 57600 bit/s instead, or ensure
simplex.
Best, Andreas
GB wrote:
> Hey,
>
> After some problems with the UART of 115200 with 32kHz (Yes, I know, it's
> horrible) I am designing a new circuit to have a 8MHz crystal (or so) since
> from what I understand that's the only way to really have very low errors.
> I experienced horrible errors using just the 32kHz crystal. I would like to
> have your opinions, since many of you are very experienced, what would be
> the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
> which kind (perhaps you know a part number or some technical specifications
> such as load capacitance, accuracy, etc) 2) Should I just opt for another
> MSP430 derivative with an FLL or more advanced clocks? Would the MSP430F5xx
> family be good for this?
>
> I really don't want to worry much about errors.
>
> The reason I am hesitant to leave the MSP430F1611 is because I am using the
> BSL like a telos mote does: It uses a switch so that I can both use the BSL
> and UART (it switches between the two). Therefore, would upgrading to
> another MSP430 change my capabilities to use MSPGCC's invert pins and all
> that that enables to program with BSL and use UART?
>
> I realize this might be a bit complex, but any help from you guys is
> appreciated. Also, any pointers to any literature discussing this is
> appreciated as well.
>
> GB
>
>
>
>
>
>
I use an MSP430F1611 with a 32kHz crystal.
My UART is @ 115200.
When I initialize the system I make the clock initialize to 4MHz:
TI_SetDCO(TI_DCO_4MHZ);
This is using TI's code, which I'm pretty sure you know. So it's not 4MHz exactly, it's
DELTA*(4096) = 4001792 Hz.
I used this in the MSPGCC calculator for baud rate , and it tells me max error is 1.37%, but it's more than that.
One issue is that I also don't have ROSC (and can't install one either so I'm making a new board) to lower the temperature dependence of the DCO.
My UART routine is:
void UART_0_115200_Setup()
{
P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
UCTL0 |= CHAR; // 8-bit character, SWRST = 1
UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
// For UART 0 Calculated using Baud rate generator found in mspgcc page
UBR00 = 0x22;
UBR10 = 0x00;
UMCTL0 = 0xDD;
UCTL0 &= ~SWRST; // Initialize USART state machine
IE1 |= URXIE0; // Enable USART0 RX interrupt
}
My circuitry is not very close to the MSP430. There are a couple of inches of distance.
Can this affect it? I guess crosstalk and other things could. Building a new circuit would allow me to eliminate all of this I guess, but I am always looking to improve my code also.
RIght now I am not exactly using the telos configuration. I use the FT2232D with its two channels, one for BSL and another for UART.
Since this board would be USB powered, power consumption isn't a concern.
Do you have any other info about the 7.3728MHz crystal. Do I need a specific load capacitance or ESR?
Thanks a lot for your replies
--- On Wed, 1/28/09, Candemir Orsan wrote:
From: Candemir Orsan
Subject: Re: [msp430] Low error UART communication MSP430F1611
To: m...
Date: Wednesday, January 28, 2009, 4:45 PM
You didn't explain the source/type of the errors you are seeing. Software? Hardware? Environment? (e.g temperature, noise, distance)?
Unless you are going very long distances in a very noisy environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
As A.K. suggested if power is not a concern and for whatever reason you must use an external Xtal, I would suggest going with a 7.3728MHz (or multiple) Xtal for most accurate baud rate generation.
C.
____________ _________ _________ __
From: Andreas Kke
To: msp430@yahoogroups. com
Sent: Wednesday, January 28, 2009 4:35:29 PM
Subject: Re: [msp430] Low error UART communication MSP430F1611
If you want to let it sleep but be able to communicate: use 9600 bit/s from
the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
internal oscillator via SMCLK. No extra crystals needed -- the high speed
types do consume a lot of power.
If power consumption is not an issue (which is likely in your case): leave the
SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had practically no
errors here so far.
BTW: errors at 115200 bit/s may also be due to the interrupt latency, a telos
can not use the USART as duplex channel. Go for 57600 bit/s instead, or ensure
simplex.
Best, Andreas
GB wrote:
> Hey,
>
> After some problems with the UART of 115200 with 32kHz (Yes, I know, it's
> horrible) I am designing a new circuit to have a 8MHz crystal (or so) since
> from what I understand that's the only way to really have very low errors.
> I experienced horrible errors using just the 32kHz crystal. I would like to
> have your opinions, since many of you are very experienced, what would be
> the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
> which kind (perhaps you know a part number or some technical specifications
> such as load capacitance, accuracy, etc) 2) Should I just opt for another
> MSP430 derivative with an FLL or more advanced clocks? Would the MSP430F5xx
> family be good for this?
>
> I really don't want to worry much about errors.
>
> The reason I am hesitant to leave the MSP430F1611 is because I am using the
> BSL like a telos mote does: It uses a switch so that I can both use the BSL
> and UART (it switches between the two). Therefore, would upgrading to
> another MSP430 change my capabilities to use MSPGCC's invert pins and all
> that that enables to program with BSL and use UART?
>
> I realize this might be a bit complex, but any help from you guys is
> appreciated. Also, any pointers to any literature discussing this is
> appreciated as well.
>
> GB
>
>
>
>
>
>
Reply by ●January 28, 20092009-01-28
If you already have crystal controlled 32768Hz ACLK, you could use
software to FLL the DCO to be 225*ACLC=7.3728MHz.
Alternatively, you could use a 7.3782MHz crystal.
In general, it is a good idea to migrate from F1xx to F2xx. But you
have to evaluate case by case.
I think F5xx opens new possibilities. But it looks like TI's effort in
developing F5xx slowed down somewhat.
--- In m..., GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_DCO_4MHZ);
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: m...
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
software to FLL the DCO to be 225*ACLC=7.3728MHz.
Alternatively, you could use a 7.3782MHz crystal.
In general, it is a good idea to migrate from F1xx to F2xx. But you
have to evaluate case by case.
I think F5xx opens new possibilities. But it looks like TI's effort in
developing F5xx slowed down somewhat.
--- In m..., GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_DCO_4MHZ);
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: m...
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
Reply by ●January 28, 20092009-01-28
> When I initialize the system I make the clock
initialize to 4MHz:
> TI_SetDCO(TI_DCO_4MHZ);
> This is using TI's code, which I'm pretty sure you know.
I don't know which code you're referring to. Does it use calibration
data stored in flash or does it use the 32,768 Hz crystal to generate
the 4,001,792Hz? If it is the latest, did you measure the accuracy of
the 32kHz crystal? If you're not using the exact same capacitor load
and/or have not taken the 6pf load of the MSP430F16x into account, you
won't get the ppm that the crystal vendor specifies, making your baud
rate error will be much higher.
Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
crystal (or add it). You should be able to generate 115200 with 1.6%
error easyly.
I seem to remember there are USB-232 chips that can optionaly be
interfaced by an SPI bus instead of the UART. They are still virtual
COM ports on the PC. Some even have JTAG interfaces. Will look into it
when I get the time.
Michael K.
--- In m..., GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_DCO_4MHZ);
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: m...
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> TI_SetDCO(TI_DCO_4MHZ);
> This is using TI's code, which I'm pretty sure you know.
I don't know which code you're referring to. Does it use calibration
data stored in flash or does it use the 32,768 Hz crystal to generate
the 4,001,792Hz? If it is the latest, did you measure the accuracy of
the 32kHz crystal? If you're not using the exact same capacitor load
and/or have not taken the 6pf load of the MSP430F16x into account, you
won't get the ppm that the crystal vendor specifies, making your baud
rate error will be much higher.
Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
crystal (or add it). You should be able to generate 115200 with 1.6%
error easyly.
I seem to remember there are USB-232 chips that can optionaly be
interfaced by an SPI bus instead of the UART. They are still virtual
COM ports on the PC. Some even have JTAG interfaces. Will look into it
when I get the time.
Michael K.
--- In m..., GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_DCO_4MHZ);
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: m...
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
Reply by ●January 28, 20092009-01-28
The code I use uses the 32768Hz crystal to set the DCO. The calibration
constants are (I believe) on MSP430F2Xxx or so derivatives.
I am going to start from the bottoms up, checking my crystal that I do have.
I did take into account the 6pF load capacitance (I knew one day it would be useful to be extra vigilant about things), but I will measure the frequency using a good oscilloscope.
However, I seem to be set on using the 7.3728MHz crystal which seems to be the best solution since power isn't an issue. I haven't been able to find information about the load capacitance. I checked the MSP430F1611 datasheet but all it says is to use capacitors as recommended by crystal manufacturer and that the mp430f1611 has 2pF input capacitance.
Does anyone know the load capacitance? mouser offers 10,16,18 and 20 pF .
Do I need any external capacitors?
Thanks a lot for all your responses.
--- On Wed, 1/28/09, tintronic wrote:
From: tintronic
Subject: [msp430] Re: Low error UART communication MSP430F1611
To: m...
Date: Wednesday, January 28, 2009, 6:07 PM
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know.
I don't know which code you're referring to. Does it use calibration
data stored in flash or does it use the 32,768 Hz crystal to generate
the 4,001,792Hz? If it is the latest, did you measure the accuracy of
the 32kHz crystal? If you're not using the exact same capacitor load
and/or have not taken the 6pf load of the MSP430F16x into account, you
won't get the ppm that the crystal vendor specifies, making your baud
rate error will be much higher.
Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
crystal (or add it). You should be able to generate 115200 with 1.6%
error easyly.
I seem to remember there are USB-232 chips that can optionaly be
interfaced by an SPI bus instead of the UART. They are still virtual
COM ports on the PC. Some even have JTAG interfaces. Will look into it
when I get the time.
Michael K.
--- In msp430@yahoogroups. com, GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_ Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: msp430@yahoogroups. com
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
I am going to start from the bottoms up, checking my crystal that I do have.
I did take into account the 6pF load capacitance (I knew one day it would be useful to be extra vigilant about things), but I will measure the frequency using a good oscilloscope.
However, I seem to be set on using the 7.3728MHz crystal which seems to be the best solution since power isn't an issue. I haven't been able to find information about the load capacitance. I checked the MSP430F1611 datasheet but all it says is to use capacitors as recommended by crystal manufacturer and that the mp430f1611 has 2pF input capacitance.
Does anyone know the load capacitance? mouser offers 10,16,18 and 20 pF .
Do I need any external capacitors?
Thanks a lot for all your responses.
--- On Wed, 1/28/09, tintronic wrote:
From: tintronic
Subject: [msp430] Re: Low error UART communication MSP430F1611
To: m...
Date: Wednesday, January 28, 2009, 6:07 PM
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know.
I don't know which code you're referring to. Does it use calibration
data stored in flash or does it use the 32,768 Hz crystal to generate
the 4,001,792Hz? If it is the latest, did you measure the accuracy of
the 32kHz crystal? If you're not using the exact same capacitor load
and/or have not taken the 6pf load of the MSP430F16x into account, you
won't get the ppm that the crystal vendor specifies, making your baud
rate error will be much higher.
Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
crystal (or add it). You should be able to generate 115200 with 1.6%
error easyly.
I seem to remember there are USB-232 chips that can optionaly be
interfaced by an SPI bus instead of the UART. They are still virtual
COM ports on the PC. Some even have JTAG interfaces. Will look into it
when I get the time.
Michael K.
--- In msp430@yahoogroups. com, GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_ Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: msp430@yahoogroups. com
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
Reply by ●January 29, 20092009-01-29
I'm still not convinced the 32K crystal is the source of your problems. You
didn't explain the nature of the errors..(bit errors, frame errors?)
One project that I finished recently has exactly the same setup you have: MSP430 hooked up to 32K crystal used to lock the DCO to ~4MHz and UART comm at 115.2K to a USB-to-VCOM IC. No noticable errors. This system streams out data continiously; as well as being able to upload MBs of logged data to the PC.
Have you confirmed the DCO frequency out of the micro by bringing out the MCLK signal to one of the pins?
When I needed to use the extarnal XTAL2 crystal, I've used the 7.3728 MHz part from DigiKey. I don't remember the part number, but the spec'ed load capacitance was 20pf. And I used 33pf caps on both sides of the actual xtal.
C.
________________________________
From: GB
To: m...
Sent: Wednesday, January 28, 2009 6:22:40 PM
Subject: Re: [msp430] Re: Low error UART communication MSP430F1611
The code I use uses the 32768Hz crystal to set the DCO. The calibration constants are (I believe) on MSP430F2Xxx or so derivatives.
I am going to start from the bottoms up, checking my crystal that I do have.
I did take into account the 6pF load capacitance (I knew one day it would be useful to be extra vigilant about things), but I will measure the frequency using a good oscilloscope.
However, I seem to be set on using the 7.3728MHz crystal which seems to be the best solution since power isn't an issue. I haven't been able to find information about the load capacitance. I checked the MSP430F1611 datasheet but all it says is to use capacitors as recommended by crystal manufacturer and that the mp430f1611 has 2pF input capacitance.
Does anyone know the load capacitance? mouser offers 10,16,18 and 20 pF .
Do I need any external capacitors?
Thanks a lot for all your responses.
--- On Wed, 1/28/09, tintronic wrote:
From: tintronic
Subject: [msp430] Re: Low error UART communication MSP430F1611
To: msp430@yahoogroups. com
Date: Wednesday, January 28, 2009, 6:07 PM
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know.
I don't know which code you're referring to. Does it use calibration
data stored in flash or does it use the 32,768 Hz crystal to generate
the 4,001,792Hz? If it is the latest, did you measure the accuracy of
the 32kHz crystal? If you're not using the exact same capacitor load
and/or have not taken the 6pf load of the MSP430F16x into account, you
won't get the ppm that the crystal vendor specifies, making your baud
rate error will be much higher.
Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
crystal (or add it). You should be able to generate 115200 with 1.6%
error easyly.
I seem to remember there are USB-232 chips that can optionaly be
interfaced by an SPI bus instead of the UART. They are still virtual
COM ports on the PC. Some even have JTAG interfaces. Will look into it
when I get the time.
Michael K.
--- In msp430@yahoogroups. com, GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_ Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: msp430@yahoogroups. com
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
One project that I finished recently has exactly the same setup you have: MSP430 hooked up to 32K crystal used to lock the DCO to ~4MHz and UART comm at 115.2K to a USB-to-VCOM IC. No noticable errors. This system streams out data continiously; as well as being able to upload MBs of logged data to the PC.
Have you confirmed the DCO frequency out of the micro by bringing out the MCLK signal to one of the pins?
When I needed to use the extarnal XTAL2 crystal, I've used the 7.3728 MHz part from DigiKey. I don't remember the part number, but the spec'ed load capacitance was 20pf. And I used 33pf caps on both sides of the actual xtal.
C.
________________________________
From: GB
To: m...
Sent: Wednesday, January 28, 2009 6:22:40 PM
Subject: Re: [msp430] Re: Low error UART communication MSP430F1611
The code I use uses the 32768Hz crystal to set the DCO. The calibration constants are (I believe) on MSP430F2Xxx or so derivatives.
I am going to start from the bottoms up, checking my crystal that I do have.
I did take into account the 6pF load capacitance (I knew one day it would be useful to be extra vigilant about things), but I will measure the frequency using a good oscilloscope.
However, I seem to be set on using the 7.3728MHz crystal which seems to be the best solution since power isn't an issue. I haven't been able to find information about the load capacitance. I checked the MSP430F1611 datasheet but all it says is to use capacitors as recommended by crystal manufacturer and that the mp430f1611 has 2pF input capacitance.
Does anyone know the load capacitance? mouser offers 10,16,18 and 20 pF .
Do I need any external capacitors?
Thanks a lot for all your responses.
--- On Wed, 1/28/09, tintronic wrote:
From: tintronic
Subject: [msp430] Re: Low error UART communication MSP430F1611
To: msp430@yahoogroups. com
Date: Wednesday, January 28, 2009, 6:07 PM
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know.
I don't know which code you're referring to. Does it use calibration
data stored in flash or does it use the 32,768 Hz crystal to generate
the 4,001,792Hz? If it is the latest, did you measure the accuracy of
the 32kHz crystal? If you're not using the exact same capacitor load
and/or have not taken the 6pf load of the MSP430F16x into account, you
won't get the ppm that the crystal vendor specifies, making your baud
rate error will be much higher.
Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
crystal (or add it). You should be able to generate 115200 with 1.6%
error easyly.
I seem to remember there are USB-232 chips that can optionaly be
interfaced by an SPI bus instead of the UART. They are still virtual
COM ports on the PC. Some even have JTAG interfaces. Will look into it
when I get the time.
Michael K.
--- In msp430@yahoogroups. com, GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_ Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: msp430@yahoogroups. com
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
Reply by ●January 29, 20092009-01-29
The errors that I get are simply bytes with erroneous information in them. I
haven't checked exactly for the source. Perhaps it was premature to blame
the clocks but I do think that lacking an ROSC could be the main problem.
Yesterday I measured the 32kHz crystal and the oscilloscope gave me 32.7 and 32.8 kHz (alternating) so I guess it is close (I am trying to improve the resolution).
A Possible problem could be the somewhat long connections I use for connecting to the device. It's not on the same board and I guess there could be interference there.
I will measure MCLK and if it's pretty close to what it should I'll assume it's an interference issue.
Thanks a lot
--- On Thu, 1/29/09, Candemir Orsan wrote:
From: Candemir Orsan
Subject: Re: [msp430] Re: Low error UART communication MSP430F1611
To: m...
Date: Thursday, January 29, 2009, 12:08 PM
I'm still not convinced the 32K crystal is the source of your problems. You didn't explain the nature of the errors..(bit errors, frame errors?)
One project that I finished recently has exactly the same setup you have: MSP430 hooked up to 32K crystal used to lock the DCO to ~4MHz and UART comm at 115.2K to a USB-to-VCOM IC. No noticable errors. This system streams out data continiously; as well as being able to upload MBs of logged data to the PC.
Have you confirmed the DCO frequency out of the micro by bringing out the MCLK signal to one of the pins?
When I needed to use the extarnal XTAL2 crystal, I've used the 7.3728 MHz part from DigiKey. I don't remember the part number, but the spec'ed load capacitance was 20pf. And I used 33pf caps on both sides of the actual xtal.
C.
____________ _________ _________ __
From: GB
To: msp430@yahoogroups. com
Sent: Wednesday, January 28, 2009 6:22:40 PM
Subject: Re: [msp430] Re: Low error UART communication MSP430F1611
The code I use uses the 32768Hz crystal to set the DCO. The calibration constants are (I believe) on MSP430F2Xxx or so derivatives.
I am going to start from the bottoms up, checking my crystal that I do have.
I did take into account the 6pF load capacitance (I knew one day it would be useful to be extra vigilant about things), but I will measure the frequency using a good oscilloscope.
However, I seem to be set on using the 7.3728MHz crystal which seems to be the best solution since power isn't an issue. I haven't been able to find information about the load capacitance. I checked the MSP430F1611 datasheet but all it says is to use capacitors as recommended by crystal manufacturer and that the mp430f1611 has 2pF input capacitance.
Does anyone know the load capacitance? mouser offers 10,16,18 and 20 pF .
Do I need any external capacitors?
Thanks a lot for all your responses.
--- On Wed, 1/28/09, tintronic wrote:
From: tintronic
Subject: [msp430] Re: Low error UART communication MSP430F1611
To: msp430@yahoogroups. com
Date: Wednesday, January 28, 2009, 6:07 PM
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know.
I don't know which code you're referring to. Does it use calibration
data stored in flash or does it use the 32,768 Hz crystal to generate
the 4,001,792Hz? If it is the latest, did you measure the accuracy of
the 32kHz crystal? If you're not using the exact same capacitor load
and/or have not taken the 6pf load of the MSP430F16x into account, you
won't get the ppm that the crystal vendor specifies, making your baud
rate error will be much higher.
Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
crystal (or add it). You should be able to generate 115200 with 1.6%
error easyly.
I seem to remember there are USB-232 chips that can optionaly be
interfaced by an SPI bus instead of the UART. They are still virtual
COM ports on the PC. Some even have JTAG interfaces. Will look into it
when I get the time.
Michael K.
--- In msp430@yahoogroups. com, GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_ Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: msp430@yahoogroups. com
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
Yesterday I measured the 32kHz crystal and the oscilloscope gave me 32.7 and 32.8 kHz (alternating) so I guess it is close (I am trying to improve the resolution).
A Possible problem could be the somewhat long connections I use for connecting to the device. It's not on the same board and I guess there could be interference there.
I will measure MCLK and if it's pretty close to what it should I'll assume it's an interference issue.
Thanks a lot
--- On Thu, 1/29/09, Candemir Orsan wrote:
From: Candemir Orsan
Subject: Re: [msp430] Re: Low error UART communication MSP430F1611
To: m...
Date: Thursday, January 29, 2009, 12:08 PM
I'm still not convinced the 32K crystal is the source of your problems. You didn't explain the nature of the errors..(bit errors, frame errors?)
One project that I finished recently has exactly the same setup you have: MSP430 hooked up to 32K crystal used to lock the DCO to ~4MHz and UART comm at 115.2K to a USB-to-VCOM IC. No noticable errors. This system streams out data continiously; as well as being able to upload MBs of logged data to the PC.
Have you confirmed the DCO frequency out of the micro by bringing out the MCLK signal to one of the pins?
When I needed to use the extarnal XTAL2 crystal, I've used the 7.3728 MHz part from DigiKey. I don't remember the part number, but the spec'ed load capacitance was 20pf. And I used 33pf caps on both sides of the actual xtal.
C.
____________ _________ _________ __
From: GB
To: msp430@yahoogroups. com
Sent: Wednesday, January 28, 2009 6:22:40 PM
Subject: Re: [msp430] Re: Low error UART communication MSP430F1611
The code I use uses the 32768Hz crystal to set the DCO. The calibration constants are (I believe) on MSP430F2Xxx or so derivatives.
I am going to start from the bottoms up, checking my crystal that I do have.
I did take into account the 6pF load capacitance (I knew one day it would be useful to be extra vigilant about things), but I will measure the frequency using a good oscilloscope.
However, I seem to be set on using the 7.3728MHz crystal which seems to be the best solution since power isn't an issue. I haven't been able to find information about the load capacitance. I checked the MSP430F1611 datasheet but all it says is to use capacitors as recommended by crystal manufacturer and that the mp430f1611 has 2pF input capacitance.
Does anyone know the load capacitance? mouser offers 10,16,18 and 20 pF .
Do I need any external capacitors?
Thanks a lot for all your responses.
--- On Wed, 1/28/09, tintronic wrote:
From: tintronic
Subject: [msp430] Re: Low error UART communication MSP430F1611
To: msp430@yahoogroups. com
Date: Wednesday, January 28, 2009, 6:07 PM
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know.
I don't know which code you're referring to. Does it use calibration
data stored in flash or does it use the 32,768 Hz crystal to generate
the 4,001,792Hz? If it is the latest, did you measure the accuracy of
the 32kHz crystal? If you're not using the exact same capacitor load
and/or have not taken the 6pf load of the MSP430F16x into account, you
won't get the ppm that the crystal vendor specifies, making your baud
rate error will be much higher.
Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
crystal (or add it). You should be able to generate 115200 with 1.6%
error easyly.
I seem to remember there are USB-232 chips that can optionaly be
interfaced by an SPI bus instead of the UART. They are still virtual
COM ports on the PC. Some even have JTAG interfaces. Will look into it
when I get the time.
Michael K.
--- In msp430@yahoogroups. com, GB wrote:
>
> OK. I will explain. perhaps I can correct some things easier.
>
> I use an MSP430F1611 with a 32kHz crystal.
> My UART is @ 115200.
>
> When I initialize the system I make the clock initialize to 4MHz:
> TI_SetDCO(TI_ DCO_4MHZ) ;
> This is using TI's code, which I'm pretty sure you know. So it's not
4MHz exactly, it's
> DELTA*(4096) = 4001792 Hz.
> I used this in the MSPGCC calculator for baud rate , and it tells me
max error is 1.37%, but it's more than that.
>
> One issue is that I also don't have ROSC (and can't install one
either so I'm making a new board) to lower the temperature dependence
of the DCO.
>
> My UART routine is:
> void UART_0_115200_ Setup()
>
> {
> P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> UCTL0 |= CHAR; // 8-bit character, SWRST = 1
>
> UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> // For UART 0 Calculated using Baud rate generator found in mspgcc page
>
> UBR00 = 0x22;
>
> UBR10 = 0x00;
>
> UMCTL0 = 0xDD;
>
> UCTL0 &= ~SWRST; // Initialize USART state
machine
> IE1 |= URXIE0; // Enable USART0 RX interrupt
>
>
>
> }
>
> My circuitry is not very close to the MSP430. There are a couple of
inches of distance.
> Can this affect it? I guess crosstalk and other things could.
Building a new circuit would allow me to eliminate all of this I
guess, but I am always looking to improve my code also.
>
> RIght now I am not exactly using the telos configuration. I use the
FT2232D with its two channels, one for BSL and another for UART.
>
> Since this board would be USB powered, power consumption isn't a
concern.
> Do you have any other info about the 7.3728MHz crystal. Do I need a
specific load capacitance or ESR?
>
> Thanks a lot for your replies
>
> --- On Wed, 1/28/09, Candemir Orsan wrote:
> From: Candemir Orsan
> Subject: Re: [msp430] Low error UART communication MSP430F1611
> To: msp430@yahoogroups. com
> Date: Wednesday, January 28, 2009, 4:45 PM
>
>
>
>
>
>
>
>
>
>
>
> You didn't explain the source/type of the errors you are
seeing. Software? Hardware? Environment? (e.g temperature, noise,
distance)?
>
>
>
> Unless you are going very long distances in a very noisy
environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
>
>
> As A.K. suggested if power is not a concern and for whatever reason
you must use an external Xtal, I would suggest going with a 7.3728MHz
(or multiple) Xtal for most accurate baud rate generation.
>
>
>
> C.
>
>
>
> ____________ _________ _________ __
>
> From: Andreas Kke
>
> To: msp430@yahoogroups. com
>
> Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> Subject: Re: [msp430] Low error UART communication MSP430F1611
>
>
>
> If you want to let it sleep but be able to communicate: use 9600
bit/s from
>
> the 32kHz and initiate a handshake, then switch to 115200 from the MSPs
>
> internal oscillator via SMCLK. No extra crystals needed -- the high
speed
>
> types do consume a lot of power.
>
>
>
> If power consumption is not an issue (which is likely in your case):
leave the
>
> SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
practically no
>
> errors here so far.
>
>
>
> BTW: errors at 115200 bit/s may also be due to the interrupt
latency, a telos
>
> can not use the USART as duplex channel. Go for 57600 bit/s instead,
or ensure
>
> simplex.
>
>
>
> Best, Andreas
>
>
>
> GB wrote:
>
> > Hey,
>
> >
>
> > After some problems with the UART of 115200 with 32kHz (Yes, I
know, it's
>
> > horrible) I am designing a new circuit to have a 8MHz crystal (or
so) since
>
> > from what I understand that's the only way to really have very low
errors.
>
> > I experienced horrible errors using just the 32kHz crystal. I
would like to
>
> > have your opinions, since many of you are very experienced, what
would be
>
> > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough? if so,
>
> > which kind (perhaps you know a part number or some technical
specifications
>
> > such as load capacitance, accuracy, etc) 2) Should I just opt for
another
>
> > MSP430 derivative with an FLL or more advanced clocks? Would the
MSP430F5xx
>
> > family be good for this?
>
> >
>
> > I really don't want to worry much about errors.
>
> >
>
> > The reason I am hesitant to leave the MSP430F1611 is because I am
using the
>
> > BSL like a telos mote does: It uses a switch so that I can both
use the BSL
>
> > and UART (it switches between the two). Therefore, would upgrading to
>
> > another MSP430 change my capabilities to use MSPGCC's invert pins
and all
>
> > that that enables to program with BSL and use UART?
>
> >
>
> > I realize this might be a bit complex, but any help from you guys is
>
> > appreciated. Also, any pointers to any literature discussing this is
>
> > appreciated as well.
>
> >
>
> > GB
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
Reply by ●January 29, 20092009-01-29
> I haven't been able to find information about
the load capacitance.
> I checked the MSP430F1611 datasheet but all it says is to use
> capacitors as recommended by crystal manufacturer and that the
> mp430f1611 has 2pF input capacitance.
Load capacitance is a value given by the crystal manufacturer, not by
the MSP. It is the capacitance the crystal needs to operate in the
given frequency. Crystal manufacturers usually provide an application
note explaining how to calculate the right capacitor values for any
crystal giving how they are connected (serial or parallel capacitors).
> Do I need any external capacitors?
Yes, you almos always do. Some chips provide an internal capacitance
which is high enought for some crystals. For example, the MSP430x1xx
provides two 12pF capacitors connected serially to the Xin Xout pins
when in Low Frequency mode, for watch crystals. If your capacitor
needs a higher value, you have to connect more capacitors.
But for high frequency crystals the MSP does not provide internal
capacitors, so you need to connect the right value for the crystal
you're using, considering track capacitance and crystal port input
capacitance.
I saw an application note that explains this very clearly, but I can't
find it now. I'll see if I find it tomorrow.
Michael K.
--- In m..., GB wrote:
>
> The code I use uses the 32768Hz crystal to set the DCO. The
calibration constants are (I believe) on MSP430F2Xxx or so derivatives.
>
> I am going to start from the bottoms up, checking my crystal that I
do have.
>
> I did take into account the 6pF load capacitance (I knew one day it
would be useful to be extra vigilant about things), but I will measure
the frequency using a good oscilloscope.
>
> However, I seem to be set on using the 7.3728MHz crystal which seems
to be the best solution since power isn't an issue. I haven't been
able to find information about the load capacitance. I checked the
MSP430F1611 datasheet but all it says is to use capacitors as
recommended by crystal manufacturer and that the mp430f1611 has 2pF
input capacitance.
>
> Does anyone know the load capacitance? mouser offers 10,16,18 and 20
pF .
> Do I need any external capacitors?
>
> Thanks a lot for all your responses.
>
> --- On Wed, 1/28/09, tintronic wrote:
> From: tintronic
> Subject: [msp430] Re: Low error UART communication MSP430F1611
> To: m...
> Date: Wednesday, January 28, 2009, 6:07 PM
>
>
>
>
>
>
>
>
>
>
>
> > When I initialize the system I make the clock
initialize to 4MHz:
>
> > TI_SetDCO(TI_ DCO_4MHZ) ;
>
> > This is using TI's code, which I'm pretty sure you know.
>
>
>
> I don't know which code you're referring to. Does it use calibration
>
> data stored in flash or does it use the 32,768 Hz crystal to generate
>
> the 4,001,792Hz? If it is the latest, did you measure the accuracy of
>
> the 32kHz crystal? If you're not using the exact same capacitor load
>
> and/or have not taken the 6pf load of the MSP430F16x into account, you
>
> won't get the ppm that the crystal vendor specifies, making your baud
>
> rate error will be much higher.
>
>
>
> Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
>
> Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
>
> the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
>
> crystal (or add it). You should be able to generate 115200 with 1.6%
>
> error easyly.
>
>
>
> I seem to remember there are USB-232 chips that can optionaly be
>
> interfaced by an SPI bus instead of the UART. They are still virtual
>
> COM ports on the PC. Some even have JTAG interfaces. Will look into it
>
> when I get the time.
>
>
>
> Michael K.
>
>
>
> --- In msp430@yahoogroups. com, GB wrote:
>
> >
>
> > OK. I will explain. perhaps I can correct some things easier.
>
> >
>
> > I use an MSP430F1611 with a 32kHz crystal.
>
> > My UART is @ 115200.
>
> >
>
> > When I initialize the system I make the clock initialize to 4MHz:
>
> > TI_SetDCO(TI_ DCO_4MHZ) ;
>
> > This is using TI's code, which I'm pretty sure you know. So it's not
>
> 4MHz exactly, it's
>
> > DELTA*(4096) = 4001792 Hz.
>
> > I used this in the MSPGCC calculator for baud rate , and it tells me
>
> max error is 1.37%, but it's more than that.
>
> >
>
> > One issue is that I also don't have ROSC (and can't install one
>
> either so I'm making a new board) to lower the temperature dependence
>
> of the DCO.
>
> >
>
> > My UART routine is:
>
> > void UART_0_115200_ Setup()
>
> >
>
> > {
>
> > P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> >
>
> > BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> >
>
> > ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> >
>
> > UCTL0 |= CHAR; // 8-bit character,
SWRST = 1
>
> >
>
> > UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> >
>
> > // For UART 0 Calculated using Baud rate generator found in mspgcc
page
>
> >
>
> > UBR00 = 0x22;
>
> >
>
> > UBR10 = 0x00;
>
> >
>
> > UMCTL0 = 0xDD;
>
> >
>
> > UCTL0 &= ~SWRST; // Initialize USART state
>
> machine
>
> > IE1 |= URXIE0; // Enable USART0 RX
interrupt
>
> >
>
> >
>
> >
>
> > }
>
> >
>
> > My circuitry is not very close to the MSP430. There are a couple of
>
> inches of distance.
>
> > Can this affect it? I guess crosstalk and other things could.
>
> Building a new circuit would allow me to eliminate all of this I
>
> guess, but I am always looking to improve my code also.
>
> >
>
> > RIght now I am not exactly using the telos configuration. I use the
>
> FT2232D with its two channels, one for BSL and another for UART.
>
> >
>
> > Since this board would be USB powered, power consumption isn't a
>
> concern.
>
> > Do you have any other info about the 7.3728MHz crystal. Do I need a
>
> specific load capacitance or ESR?
>
> >
>
> > Thanks a lot for your replies
>
> >
>
> > --- On Wed, 1/28/09, Candemir Orsan wrote:
>
> > From: Candemir Orsan
>
> > Subject: Re: [msp430] Low error UART communication MSP430F1611
>
> > To: msp430@yahoogroups. com
>
> > Date: Wednesday, January 28, 2009, 4:45 PM
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> > You didn't explain the source/type of the errors you are
>
> seeing. Software? Hardware? Environment? (e.g temperature, noise,
>
> distance)?
>
> >
>
> >
>
> >
>
> > Unless you are going very long distances in a very noisy
>
> environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
> >
>
> >
>
> >
>
> > As A.K. suggested if power is not a concern and for whatever reason
>
> you must use an external Xtal, I would suggest going with a 7.3728MHz
>
> (or multiple) Xtal for most accurate baud rate generation.
>
> >
>
> >
>
> >
>
> > C.
>
> >
>
> >
>
> >
>
> > ____________ _________ _________ __
>
> >
>
> > From: Andreas Kke
>
> >
>
> > To: msp430@yahoogroups. com
>
> >
>
> > Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> >
>
> > Subject: Re: [msp430] Low error UART communication MSP430F1611
>
> >
>
> >
>
> >
>
> > If you want to let it sleep but be able to communicate: use 9600
>
> bit/s from
>
> >
>
> > the 32kHz and initiate a handshake, then switch to 115200 from the
MSPs
>
> >
>
> > internal oscillator via SMCLK. No extra crystals needed -- the high
>
> speed
>
> >
>
> > types do consume a lot of power.
>
> >
>
> >
>
> >
>
> > If power consumption is not an issue (which is likely in your case):
>
> leave the
>
> >
>
> > SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
>
> practically no
>
> >
>
> > errors here so far.
>
> >
>
> >
>
> >
>
> > BTW: errors at 115200 bit/s may also be due to the interrupt
>
> latency, a telos
>
> >
>
> > can not use the USART as duplex channel. Go for 57600 bit/s instead,
>
> or ensure
>
> >
>
> > simplex.
>
> >
>
> >
>
> >
>
> > Best, Andreas
>
> >
>
> >
>
> >
>
> > GB wrote:
>
> >
>
> > > Hey,
>
> >
>
> > >
>
> >
>
> > > After some problems with the UART of 115200 with 32kHz (Yes, I
>
> know, it's
>
> >
>
> > > horrible) I am designing a new circuit to have a 8MHz crystal (or
>
> so) since
>
> >
>
> > > from what I understand that's the only way to really have very low
>
> errors.
>
> >
>
> > > I experienced horrible errors using just the 32kHz crystal. I
>
> would like to
>
> >
>
> > > have your opinions, since many of you are very experienced, what
>
> would be
>
> >
>
> > > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough?
if so,
>
> >
>
> > > which kind (perhaps you know a part number or some technical
>
> specifications
>
> >
>
> > > such as load capacitance, accuracy, etc) 2) Should I just opt for
>
> another
>
> >
>
> > > MSP430 derivative with an FLL or more advanced clocks? Would the
>
> MSP430F5xx
>
> >
>
> > > family be good for this?
>
> >
>
> > >
>
> >
>
> > > I really don't want to worry much about errors.
>
> >
>
> > >
>
> >
>
> > > The reason I am hesitant to leave the MSP430F1611 is because I am
>
> using the
>
> >
>
> > > BSL like a telos mote does: It uses a switch so that I can both
>
> use the BSL
>
> >
>
> > > and UART (it switches between the two). Therefore, would
upgrading to
>
> >
>
> > > another MSP430 change my capabilities to use MSPGCC's invert pins
>
> and all
>
> >
>
> > > that that enables to program with BSL and use UART?
>
> >
>
> > >
>
> >
>
> > > I realize this might be a bit complex, but any help from you guys is
>
> >
>
> > > appreciated. Also, any pointers to any literature discussing this is
>
> >
>
> > > appreciated as well.
>
> >
>
> > >
>
> >
>
> > > GB
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> I checked the MSP430F1611 datasheet but all it says is to use
> capacitors as recommended by crystal manufacturer and that the
> mp430f1611 has 2pF input capacitance.
Load capacitance is a value given by the crystal manufacturer, not by
the MSP. It is the capacitance the crystal needs to operate in the
given frequency. Crystal manufacturers usually provide an application
note explaining how to calculate the right capacitor values for any
crystal giving how they are connected (serial or parallel capacitors).
> Do I need any external capacitors?
Yes, you almos always do. Some chips provide an internal capacitance
which is high enought for some crystals. For example, the MSP430x1xx
provides two 12pF capacitors connected serially to the Xin Xout pins
when in Low Frequency mode, for watch crystals. If your capacitor
needs a higher value, you have to connect more capacitors.
But for high frequency crystals the MSP does not provide internal
capacitors, so you need to connect the right value for the crystal
you're using, considering track capacitance and crystal port input
capacitance.
I saw an application note that explains this very clearly, but I can't
find it now. I'll see if I find it tomorrow.
Michael K.
--- In m..., GB wrote:
>
> The code I use uses the 32768Hz crystal to set the DCO. The
calibration constants are (I believe) on MSP430F2Xxx or so derivatives.
>
> I am going to start from the bottoms up, checking my crystal that I
do have.
>
> I did take into account the 6pF load capacitance (I knew one day it
would be useful to be extra vigilant about things), but I will measure
the frequency using a good oscilloscope.
>
> However, I seem to be set on using the 7.3728MHz crystal which seems
to be the best solution since power isn't an issue. I haven't been
able to find information about the load capacitance. I checked the
MSP430F1611 datasheet but all it says is to use capacitors as
recommended by crystal manufacturer and that the mp430f1611 has 2pF
input capacitance.
>
> Does anyone know the load capacitance? mouser offers 10,16,18 and 20
pF .
> Do I need any external capacitors?
>
> Thanks a lot for all your responses.
>
> --- On Wed, 1/28/09, tintronic wrote:
> From: tintronic
> Subject: [msp430] Re: Low error UART communication MSP430F1611
> To: m...
> Date: Wednesday, January 28, 2009, 6:07 PM
>
>
>
>
>
>
>
>
>
>
>
> > When I initialize the system I make the clock
initialize to 4MHz:
>
> > TI_SetDCO(TI_ DCO_4MHZ) ;
>
> > This is using TI's code, which I'm pretty sure you know.
>
>
>
> I don't know which code you're referring to. Does it use calibration
>
> data stored in flash or does it use the 32,768 Hz crystal to generate
>
> the 4,001,792Hz? If it is the latest, did you measure the accuracy of
>
> the 32kHz crystal? If you're not using the exact same capacitor load
>
> and/or have not taken the 6pf load of the MSP430F16x into account, you
>
> won't get the ppm that the crystal vendor specifies, making your baud
>
> rate error will be much higher.
>
>
>
> Can't you use the FT2232 Xout signal and supply it into the MSP Xin?
>
> Or maybe better: Use the MSP with the 6MHz crystal and output MCLK to
>
> the FT2232. But if cost is not an issue, change the 32kHz to an 8MHz
>
> crystal (or add it). You should be able to generate 115200 with 1.6%
>
> error easyly.
>
>
>
> I seem to remember there are USB-232 chips that can optionaly be
>
> interfaced by an SPI bus instead of the UART. They are still virtual
>
> COM ports on the PC. Some even have JTAG interfaces. Will look into it
>
> when I get the time.
>
>
>
> Michael K.
>
>
>
> --- In msp430@yahoogroups. com, GB wrote:
>
> >
>
> > OK. I will explain. perhaps I can correct some things easier.
>
> >
>
> > I use an MSP430F1611 with a 32kHz crystal.
>
> > My UART is @ 115200.
>
> >
>
> > When I initialize the system I make the clock initialize to 4MHz:
>
> > TI_SetDCO(TI_ DCO_4MHZ) ;
>
> > This is using TI's code, which I'm pretty sure you know. So it's not
>
> 4MHz exactly, it's
>
> > DELTA*(4096) = 4001792 Hz.
>
> > I used this in the MSPGCC calculator for baud rate , and it tells me
>
> max error is 1.37%, but it's more than that.
>
> >
>
> > One issue is that I also don't have ROSC (and can't install one
>
> either so I'm making a new board) to lower the temperature dependence
>
> of the DCO.
>
> >
>
> > My UART routine is:
>
> > void UART_0_115200_ Setup()
>
> >
>
> > {
>
> > P3SEL |= 0x30; // P3.4,5 = USART0 TXD/RXD
>
> >
>
> > BCSCTL1 |= DIVA_3; // ACLK= LFXT1CLK/8
>
> >
>
> > ME1 |= UTXE0 + URXE0; // Enabled USART0 TXD/RXD
>
> >
>
> > UCTL0 |= CHAR; // 8-bit character,
SWRST = 1
>
> >
>
> > UTCTL0 |= SSEL1 + SSEL0 ; // UCLK = SMCLK, start edge detect
>
> >
>
> > // For UART 0 Calculated using Baud rate generator found in mspgcc
page
>
> >
>
> > UBR00 = 0x22;
>
> >
>
> > UBR10 = 0x00;
>
> >
>
> > UMCTL0 = 0xDD;
>
> >
>
> > UCTL0 &= ~SWRST; // Initialize USART state
>
> machine
>
> > IE1 |= URXIE0; // Enable USART0 RX
interrupt
>
> >
>
> >
>
> >
>
> > }
>
> >
>
> > My circuitry is not very close to the MSP430. There are a couple of
>
> inches of distance.
>
> > Can this affect it? I guess crosstalk and other things could.
>
> Building a new circuit would allow me to eliminate all of this I
>
> guess, but I am always looking to improve my code also.
>
> >
>
> > RIght now I am not exactly using the telos configuration. I use the
>
> FT2232D with its two channels, one for BSL and another for UART.
>
> >
>
> > Since this board would be USB powered, power consumption isn't a
>
> concern.
>
> > Do you have any other info about the 7.3728MHz crystal. Do I need a
>
> specific load capacitance or ESR?
>
> >
>
> > Thanks a lot for your replies
>
> >
>
> > --- On Wed, 1/28/09, Candemir Orsan wrote:
>
> > From: Candemir Orsan
>
> > Subject: Re: [msp430] Low error UART communication MSP430F1611
>
> > To: msp430@yahoogroups. com
>
> > Date: Wednesday, January 28, 2009, 4:45 PM
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> > You didn't explain the source/type of the errors you are
>
> seeing. Software? Hardware? Environment? (e.g temperature, noise,
>
> distance)?
>
> >
>
> >
>
> >
>
> > Unless you are going very long distances in a very noisy
>
> environment, even at 115.2Kb you shouldn't experience 'horrible' errors.
>
> >
>
> >
>
> >
>
> > As A.K. suggested if power is not a concern and for whatever reason
>
> you must use an external Xtal, I would suggest going with a 7.3728MHz
>
> (or multiple) Xtal for most accurate baud rate generation.
>
> >
>
> >
>
> >
>
> > C.
>
> >
>
> >
>
> >
>
> > ____________ _________ _________ __
>
> >
>
> > From: Andreas Kke
>
> >
>
> > To: msp430@yahoogroups. com
>
> >
>
> > Sent: Wednesday, January 28, 2009 4:35:29 PM
>
> >
>
> > Subject: Re: [msp430] Low error UART communication MSP430F1611
>
> >
>
> >
>
> >
>
> > If you want to let it sleep but be able to communicate: use 9600
>
> bit/s from
>
> >
>
> > the 32kHz and initiate a handshake, then switch to 115200 from the
MSPs
>
> >
>
> > internal oscillator via SMCLK. No extra crystals needed -- the high
>
> speed
>
> >
>
> > types do consume a lot of power.
>
> >
>
> >
>
> >
>
> > If power consumption is not an issue (which is likely in your case):
>
> leave the
>
> >
>
> > SMCLK always on at around 1MHz. 1MHz is fine for 115200 -- I had
>
> practically no
>
> >
>
> > errors here so far.
>
> >
>
> >
>
> >
>
> > BTW: errors at 115200 bit/s may also be due to the interrupt
>
> latency, a telos
>
> >
>
> > can not use the USART as duplex channel. Go for 57600 bit/s instead,
>
> or ensure
>
> >
>
> > simplex.
>
> >
>
> >
>
> >
>
> > Best, Andreas
>
> >
>
> >
>
> >
>
> > GB wrote:
>
> >
>
> > > Hey,
>
> >
>
> > >
>
> >
>
> > > After some problems with the UART of 115200 with 32kHz (Yes, I
>
> know, it's
>
> >
>
> > > horrible) I am designing a new circuit to have a 8MHz crystal (or
>
> so) since
>
> >
>
> > > from what I understand that's the only way to really have very low
>
> errors.
>
> >
>
> > > I experienced horrible errors using just the 32kHz crystal. I
>
> would like to
>
> >
>
> > > have your opinions, since many of you are very experienced, what
>
> would be
>
> >
>
> > > the best: 1) Just an 8MHz crystal on the MSP430F1611 is enough?
if so,
>
> >
>
> > > which kind (perhaps you know a part number or some technical
>
> specifications
>
> >
>
> > > such as load capacitance, accuracy, etc) 2) Should I just opt for
>
> another
>
> >
>
> > > MSP430 derivative with an FLL or more advanced clocks? Would the
>
> MSP430F5xx
>
> >
>
> > > family be good for this?
>
> >
>
> > >
>
> >
>
> > > I really don't want to worry much about errors.
>
> >
>
> > >
>
> >
>
> > > The reason I am hesitant to leave the MSP430F1611 is because I am
>
> using the
>
> >
>
> > > BSL like a telos mote does: It uses a switch so that I can both
>
> use the BSL
>
> >
>
> > > and UART (it switches between the two). Therefore, would
upgrading to
>
> >
>
> > > another MSP430 change my capabilities to use MSPGCC's invert pins
>
> and all
>
> >
>
> > > that that enables to program with BSL and use UART?
>
> >
>
> > >
>
> >
>
> > > I realize this might be a bit complex, but any help from you guys is
>
> >
>
> > > appreciated. Also, any pointers to any literature discussing this is
>
> >
>
> > > appreciated as well.
>
> >
>
> > >
>
> >
>
> > > GB
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> > >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
> >
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
