Forums

Can I give 2 different voltage to DVCC and AVCC?

Started by chowkamfai July 23, 2003
Dear all,

     My situation is to build a battery-powered sensoing device using 
MSP430F449 and I would like to use the build-in ADC to measure the 0-
3V analog data.

    We use a DC-DC convertor to provide a stable 3V to all the analog 
ICs and also I would like to fed it to AVCC and also use it as ADC 's 
reference.
    
    But the efficiency of the DC-DC convertor is not so good, so, I 
don't want to connect its output to DVCC to save power. And the DVCC 
uses the main battery supply (3V) which may be droped (~to even 2.6V) 
as the time passed.

    When I give 2.6V to DVCC and stable 3V to AVCC (AGND and DGND 
connected togather), the C-SPY cannot program the code to the 
internal flash memory.

    Does it mean that we MUST connect DVCC and AVCC together?
We can't give 2 different 2 voltage level to them? Please comments.

Thanks!!

Danny Chow
Electronic Engineer



Beginning Microcontrollers with the MSP430

chowkamfai wrote:
> Dear all,
> 
>      My situation is to build a battery-powered sensoing device using 
> MSP430F449 and I would like to use the build-in ADC to measure the 0-
> 3V analog data.
> 
>     We use a DC-DC convertor to provide a stable 3V to all the analog 
> ICs and also I would like to fed it to AVCC and also use it as ADC 's 
> reference.

Clean supplies are as important as stable ones. The internal 1.5V 
reference would seem a better choice for you. then you can run on 
batteries quite happily, use the internal power supply monitor to keep 
tabs on the supply voltage and still get good A/D results as the battery 
fades.

>     
>     But the efficiency of the DC-DC convertor is not so good, so, I 
> don't want to connect its output to DVCC to save power. And the DVCC 
> uses the main battery supply (3V) which may be droped (~to even 2.6V) 
> as the time passed.

get a better one. You also need to study the data sheet more closely. It 
speciifes that no voltage applied to any pin should exceed Vcc by more 
than 0.3V, and that this is a maximum rating. ie you shouldn't design it 
with AVcc > Vcc.

> 
>     When I give 2.6V to DVCC and stable 3V to AVCC (AGND and DGND 
> connected togather), the C-SPY cannot program the code to the 
> internal flash memory.

If you look at the specs for the MSP430Fxxx you will see that the 
minimum voltage for flash programming is 2.7V. personally I would 
recommend a minimum of 3.0V. Get another Dc/Dc converter. there are 
hundreds of different types out there, some rated up to 1A in tiny 
packages. Most of these are very efficient.

> 
>     Does it mean that we MUST connect DVCC and AVCC together?
> We can't give 2 different 2 voltage level to them? Please comments.

You can supply AVcc < DVcc but not the other way round.

Al



Thanks for your comment, onestone~

> Clean supplies are as important as stable ones.
The internal 1.5V 
> reference would seem a better choice for you. then you can run on 
> batteries quite happily, use the internal power supply monitor to 
keep 
> tabs on the supply voltage and still get good A/D
results as the 
battery 
> fades.

The analog output is 0V-3V. So, using 1.5V internal reference is not 
suitable I think...

> 
> get a better one. You also need to study the data sheet more 
closely. It 
> speciifes that no voltage applied to any pin
should exceed Vcc by 
more 
> than 0.3V, and that this is a maximum rating. ie
you shouldn't 
design it 
> with AVcc > Vcc.

The device in fact is a electronic watch which uses a normal 2032 
battery. So, I need try my best to minimize the power consumption. 
The low noise DC-DC convertor (TI TPS60243) is quite low in 
efficiency when the load is low. So, when the analog sensing part is 
off (most of the time), the watch is used as a normal clock. A 
minimal current is needed (e.g. <20 uA) at normal clock but it cannot 
be satisfied using this convertor. So, I prefer to use the battery 
directly as DVCC. Of course, I don't want to use too expensive DC-DC 
convertor.

Datasheet said that ADC input can accept upto AVCC+0.3V. I just think 
that if we keep the AVCC at 3V even the DVCC is only 2.7V, my circuit 
can still accept the 0-3V analog ADC data accurately.

> 
> If you look at the specs for the MSP430Fxxx you will see that the 
> minimum voltage for flash programming is 2.7V. personally I would 
> recommend a minimum of 3.0V. Get another Dc/Dc converter. there are 
> hundreds of different types out there, some rated up to 1A in tiny 
> packages. Most of these are very efficient.
> You can supply AVcc < DVcc but not the other way round.
>



When I turn off the ADC part of the device, I would like to turn off 
also the 3V given to AVCC. Will the MSP430 stills operate?
It seems that the AVCC also supplies to FLL+ and port 1...
I think it may cause a problem.

--- In msp430@msp4..., "chowkamfai" <chowkamfai@y...> wrote:
> Thanks for your comment, onestone~
> 
> > Clean supplies are as important as stable ones. The internal 1.5V 
> > reference would seem a better choice for you. then you can run on 
> > batteries quite happily, use the internal power supply monitor to 
> keep 
> > tabs on the supply voltage and still get good A/D results as the 
> battery 
> > fades.
> 
> The analog output is 0V-3V. So, using 1.5V internal reference is 
not 
> suitable I think...
> 
> > 
> > get a better one. You also need to study the data sheet more 
> closely. It 
> > speciifes that no voltage applied to any pin should exceed Vcc by 
> more 
> > than 0.3V, and that this is a maximum rating. ie you shouldn't 
> design it 
> > with AVcc > Vcc.
> 
> The device in fact is a electronic watch which uses a normal 2032 
> battery. So, I need try my best to minimize the power consumption. 
> The low noise DC-DC convertor (TI TPS60243) is quite low in 
> efficiency when the load is low. So, when the analog sensing part 
is 
> off (most of the time), the watch is used as a
normal clock. A 
> minimal current is needed (e.g. <20 uA) at normal clock but it 
cannot 
> be satisfied using this convertor. So, I prefer to
use the battery 
> directly as DVCC. Of course, I don't want to use too expensive DC-
DC 
> convertor.
> 
> Datasheet said that ADC input can accept upto AVCC+0.3V. I just 
think 
> that if we keep the AVCC at 3V even the DVCC is
only 2.7V, my 
circuit 
> can still accept the 0-3V analog ADC data
accurately.
> 
> > 
> > If you look at the specs for the MSP430Fxxx you will see that the 
> > minimum voltage for flash programming is 2.7V. personally I would 
> > recommend a minimum of 3.0V. Get another Dc/Dc converter. there 
are 
> > hundreds of different types out there, some
rated up to 1A in 
tiny 
> > packages. Most of these are very efficient.
> > You can supply AVcc < DVcc but not the other way round.
> >


chowkamfai wrote:

> Thanks for your comment, onestone~
> 
> 
>>Clean supplies are as important as stable ones. The internal 1.5V 
>>reference would seem a better choice for you. then you can run on 
>>batteries quite happily, use the internal power supply monitor to 
> 
> keep 
> 
>>tabs on the supply voltage and still get good A/D results as the 
> 
> battery 
> 
>>fades.
> 
> 
> The analog output is 0V-3V. So, using 1.5V internal reference is not 
> suitable I think...

I use it to monitor 12V incoming supplies, simple resistro divider.

> 
> 
>>get a better one. You also need to study the data sheet more 
> 
> closely. It 
> 
>>speciifes that no voltage applied to any pin should exceed Vcc by 
> 
> more 
> 
>>than 0.3V, and that this is a maximum rating. ie you shouldn't 
> 
> design it 
> 
>>with AVcc > Vcc.
> 
> 
> The device in fact is a electronic watch which uses a normal 2032 
> battery. So, I need try my best to minimize the power consumption. 
> The low noise DC-DC convertor (TI TPS60243) is quite low in 
> efficiency when the load is low. 

Check its current when in shut down. You will need to take that into 
consideration as well.

> So, when the analog sensing part is 
> off (most of the time), the watch is used as a normal clock. A 
> minimal current is needed (e.g. <20 uA) at normal clock but it cannot 
> be satisfied using this convertor. So, I prefer to use the battery 
> directly as DVCC. Of course, I don't want to use too expensive DC-DC 
> convertor.

Then don't use any, follow my first suggestion. use the battery 
directly. monitor it using the Vcc/2 multiplexed A/D signal then use the 
1.5V internal reference for your analog reference. You still need to 
follow normal analog power supply rules, decouiple the AVcc supply 
separately from DVcc etc etc. I monitor Vin using nearly 1M voltage 
divider. It works quite happily and ensures low current.

> 
> Datasheet said that ADC input can accept upto AVCC+0.3V. I just think 
> that if we keep the AVCC at 3V even the DVCC is only 2.7V, my circuit 
> can still accept the 0-3V analog ADC data accurately.

Look in the absolute ratings it very clearly states that no voltage on 
any pin should exceed DVcc+0.3V That includes AVcc. All other ratings 
are subordinate to that.

>>If you look at the specs for the MSP430Fxxx you
will see that the 
>>minimum voltage for flash programming is 2.7V. personally I would 
>>recommend a minimum of 3.0V. Get another Dc/Dc converter. there are 
>>hundreds of different types out there, some rated up to 1A in tiny 
>>packages. Most of these are very efficient.
>>You can supply AVcc < DVcc but not the other way round.

Al


chowkamfai wrote:

> When I turn off the ADC part of the device, I
would like to turn off 
> also the 3V given to AVCC. Will the MSP430 stills operate?
> It seems that the AVCC also supplies to FLL+ and port 1...
> I think it may cause a problem.

Throw away the DC/Dc converter. You don't need it. All you need to do is 
ensure that you have 3V available for flash programming. DON'T TRY AND 
PROGRAM ON ANYTHING BUT BRAND NEW BATTERIES. In fact I would have a 
separate battery like a Li-Ion or LTC for programming.

Al


> 
> --- In msp430@msp4..., "chowkamfai" <chowkamfai@y...>
wrote:
> 
>>Thanks for your comment, onestone~
>>
>>
>>>Clean supplies are as important as stable ones. The internal 1.5V 
>>>reference would seem a better choice for you. then you can run on 
>>>batteries quite happily, use the internal power supply monitor to 
>>
>>keep 
>>
>>>tabs on the supply voltage and still get good A/D results as the 
>>
>>battery 
>>
>>>fades.
>>
>>The analog output is 0V-3V. So, using 1.5V internal reference is 
> 
> not 
> 
>>suitable I think...
>>
>>
>>>get a better one. You also need to study the data sheet more 
>>
>>closely. It 
>>
>>>speciifes that no voltage applied to any pin should exceed Vcc by 
>>
>>more 
>>
>>>than 0.3V, and that this is a maximum rating. ie you shouldn't 
>>
>>design it 
>>
>>>with AVcc > Vcc.
>>
>>The device in fact is a electronic watch which uses a normal 2032 
>>battery. So, I need try my best to minimize the power consumption. 
>>The low noise DC-DC convertor (TI TPS60243) is quite low in 
>>efficiency when the load is low. So, when the analog sensing part 
> 
> is 
> 
>>off (most of the time), the watch is used as a normal clock. A 
>>minimal current is needed (e.g. <20 uA) at normal clock but it 
> 
> cannot 
> 
>>be satisfied using this convertor. So, I prefer to use the battery 
>>directly as DVCC. Of course, I don't want to use too expensive DC-
> 
> DC 
> 
>>convertor.
>>
>>Datasheet said that ADC input can accept upto AVCC+0.3V. I just 
> 
> think 
> 
>>that if we keep the AVCC at 3V even the DVCC is only 2.7V, my 
> 
> circuit 
> 
>>can still accept the 0-3V analog ADC data accurately.
>>
>>
>>>If you look at the specs for the MSP430Fxxx you will see that the 
>>>minimum voltage for flash programming is 2.7V. personally I would 
>>>recommend a minimum of 3.0V. Get another Dc/Dc converter. there 
> 
> are 
> 
>>>hundreds of different types out there, some rated up to 1A in 
> 
> tiny 
> 
>>>packages. Most of these are very efficient.
>>>You can supply AVcc < DVcc but not the other way round.
>>>
> 
> 
> 
> 
> .
> 
>  
> 
> ">http://docs.yahoo.com/info/terms/ 
> 
> 
>