Forums

I2C vs SPI

Started by Joe G (Home) February 8, 2010
Hi All,

SPI  is more simplistic in internal IC hardware design than I2c. If the SPI 
data bus receive garbage (spikes or noise) you can pull the chip select high 
and start clocking the data in/out again.

SPI has
/CS  chip select
Clock
Data out
Data in

I2C has
Clock
Data

Question
========
If an I2C bus receives  garbage, how does one tell a I2C receiver to "reset" 
and start again, as there is only a clock and data line.

Thanks in Advance.

Joe
 


"Joe G (Home)" <joe.g@optusnet.com.au> wrote in message
news:4b6fecc1$0$3003$afc38c87@news.optusnet.com.au...
> Hi All, > > SPI is more simplistic in internal IC hardware design than I2c. If the
SPI
> data bus receive garbage (spikes or noise) you can pull the chip select
high
> and start clocking the data in/out again. > > SPI has > /CS chip select > Clock > Data out > Data in > > I2C has > Clock > Data > > Question > ======== > If an I2C bus receives garbage, how does one tell a I2C receiver to
"reset"
> and start again, as there is only a clock and data line.
By generating a premature stop condition. Meindert
Meindert Sprang schrieb:

>> If an I2C bus receives garbage, how does one tell a I2C receiver to > "reset" >> and start again, as there is only a clock and data line. > > By generating a premature stop condition.
...and if one of the devices eventually blocks the bus with SDA=low (this can happen when a read access is aborted), apply SCL pulses until SDA goes high. Tilmann
Tilmann Reh wrote:
> Meindert Sprang schrieb: > >>> If an I2C bus receives garbage, how does one tell a I2C receiver to >> "reset" >>> and start again, as there is only a clock and data line. >> By generating a premature stop condition. > > ...and if one of the devices eventually blocks the bus with SDA=low > (this can happen when a read access is aborted), apply SCL pulses until > SDA goes high. > > Tilmann
The correct way to RESET the I2C buss, is to do a repeat-START, then a STOP. Works every time. don
don schrieb:

> The correct way to RESET the I2C buss, is to do a repeat-START, then a STOP.
Please note that you simply can't apply a (repeated) start condition when a device continuously holds SDA low. In such cases, you first have to provide SCL pulses until SDA gets high again, then do start-stop. Tilmann
Tilmann Reh wrote:
> don schrieb: > >> The correct way to RESET the I2C buss, is to do a repeat-START, then a STOP. > > Please note that you simply can't apply a (repeated) start condition > when a device continuously holds SDA low. In such cases, you first have > to provide SCL pulses until SDA gets high again, then do start-stop. >
_If_ it lets go at all. I've seen cases where it didn't and a power cycle was required :-( -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
In article <7tedm4Fam1U1@mid.individual.net>, invalid@invalid.invalid 
says...
> Tilmann Reh wrote: > > don schrieb: > > > >> The correct way to RESET the I2C buss, is to do a repeat-START, then a STOP. > > > > Please note that you simply can't apply a (repeated) start condition > > when a device continuously holds SDA low. In such cases, you first have > > to provide SCL pulses until SDA gets high again, then do start-stop. > > > > _If_ it lets go at all. I've seen cases where it didn't and a power > cycle was required :-(
Especially if the slave device is doing a clock stretch, holding SCL low. -- Paul Carpenter | paul@pcserviceselectronics.co.uk <http://www.pcserviceselectronics.co.uk/> PC Services <http://www.pcserviceselectronics.co.uk/fonts/> Timing Diagram Font <http://www.gnuh8.org.uk/> GNU H8 - compiler & Renesas H8/H8S/H8 Tiny <http://www.badweb.org.uk/> For those web sites you hate
Paul Carpenter wrote:
> In article <7tedm4Fam1U1@mid.individual.net>, invalid@invalid.invalid > says... >> Tilmann Reh wrote: >>> don schrieb: >>> >>>> The correct way to RESET the I2C buss, is to do a repeat-START, then a STOP. >>> Please note that you simply can't apply a (repeated) start condition >>> when a device continuously holds SDA low. In such cases, you first have >>> to provide SCL pulses until SDA gets high again, then do start-stop. >>> >> _If_ it lets go at all. I've seen cases where it didn't and a power >> cycle was required :-( > > Especially if the slave device is doing a clock stretch, holding SCL low. >
Sort of the eternal stretch :-) I like SPI a lot more although I2C does have the advantage of being able to hang it all onto the same serial bus. But that can at the same time cause hang-ups while SPI never hung on me. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
Paul Carpenter schrieb:

>> >> The correct way to RESET the I2C buss, is to do a repeat-START, then a STOP. >> > >> > Please note that you simply can't apply a (repeated) start condition >> > when a device continuously holds SDA low. In such cases, you first have >> > to provide SCL pulses until SDA gets high again, then do start-stop. >> > >> >> _If_ it lets go at all. I've seen cases where it didn't and a power >> cycle was required :-( > > Especially if the slave device is doing a clock stretch, holding SCL low.
I have always carefully selected only such I2C slaves that don't do clock stretching... :-) Tilmann
"Joe G (Home)" <joe.g@optusnet.com.au> wrote in message 
news:4b6fecc1$0$3003$afc38c87@news.optusnet.com.au...
> Hi All, > > SPI is more simplistic in internal IC hardware design than I2c. If the > SPI data bus receive garbage (spikes or noise) you can pull the chip > select high and start clocking the data in/out again. > > SPI has > /CS chip select > Clock > Data out > Data in > > I2C has > Clock > Data > > Question > ======== > If an I2C bus receives garbage, how does one tell a I2C receiver to > "reset" and start again, as there is only a clock and data line. > > Thanks in Advance. > > Joe > >
Thanks all for your comments. Joe