EmbeddedRelated.com
Forums
Memfault Beyond the Launch

64 ADCs on PXA255 !!

Started by RR November 28, 2006
I need to interface a minimum of 32 (maximum of 64) 14-bit ADCs to an
Intel PXA255 XScale processor.
Each ADC is connected to an individual linear Hall Effect sensor.
The sensors are distributed over an area of approx 2 sq metres. I want
an ADC near each Sensor (for signal integrity reasons), so
multi-channel ADCs are not suitable.

The PXA255 has I2C and SSP ports available.

Does anyone have any reccomendations on how I might best do this
(taking into account signal integrity, loading, addressing issues,etc)
?

RR

"RR" <richardrooney@icecomms.net> wrote in message 
news:1164731384.051639.309660@l39g2000cwd.googlegroups.com...
>I need to interface a minimum of 32 (maximum of 64) 14-bit ADCs to an > Intel PXA255 XScale processor. > Each ADC is connected to an individual linear Hall Effect sensor. > The sensors are distributed over an area of approx 2 sq metres. I want > an ADC near each Sensor (for signal integrity reasons), so > multi-channel ADCs are not suitable. > > The PXA255 has I2C and SSP ports available. > > Does anyone have any reccomendations on how I might best do this > (taking into account signal integrity, loading, addressing issues,etc) > ?
Sounds to me like you might want to drive the ADCs from an FPGA and then access the results with a memory-like interface to the processor.
RR wrote:
> I need to interface a minimum of 32 (maximum of 64) 14-bit ADCs to an > Intel PXA255 XScale processor. > Each ADC is connected to an individual linear Hall Effect sensor. > The sensors are distributed over an area of approx 2 sq metres. I want > an ADC near each Sensor (for signal integrity reasons), so > multi-channel ADCs are not suitable. > > The PXA255 has I2C and SSP ports available. > > Does anyone have any reccomendations on how I might best do this > (taking into account signal integrity, loading, addressing issues,etc) > ?
You'll need Serial ADCs - and some sort of a MUX scheme. One ADC candidate would be TI's new ADS1230 - specs 5MHz SLCK Ideal would be SPI ADCs that can be wired in a daisy-chain loop, which would work very well with the SSP, but I don't recall seeing an ADC that is quite that clever. -jg
Tom Lucas wrote:
> "RR" <richardrooney@icecomms.net> wrote in message > news:1164731384.051639.309660@l39g2000cwd.googlegroups.com... >> I need to interface a minimum of 32 (maximum of 64) 14-bit ADCs to an >> Intel PXA255 XScale processor. >> Each ADC is connected to an individual linear Hall Effect sensor. >> The sensors are distributed over an area of approx 2 sq metres. I want >> an ADC near each Sensor (for signal integrity reasons), so >> multi-channel ADCs are not suitable. >> >> The PXA255 has I2C and SSP ports available. >> >> Does anyone have any reccomendations on how I might best do this >> (taking into account signal integrity, loading, addressing issues,etc) >> ? > > Sounds to me like you might want to drive the ADCs from an FPGA and then > access the results with a memory-like interface to the processor. > >
I am doing precisely this in a design. It gives a great deal of control over the access sequence and fixes up the broken behaviour of the PXA255 SPI implementation (yes, it's broken). I make the FPGA a simple asynchronous memory target. Cheers PeteS
On Tue, 28 Nov 2006 19:01:25 GMT, PeteS <peter.smith8380@ntlworld.com>
wrote:

>Tom Lucas wrote: >> "RR" <richardrooney@icecomms.net> wrote in message >> news:1164731384.051639.309660@l39g2000cwd.googlegroups.com... >>> I need to interface a minimum of 32 (maximum of 64) 14-bit ADCs to an >>> Intel PXA255 XScale processor. >>> Each ADC is connected to an individual linear Hall Effect sensor. >>> The sensors are distributed over an area of approx 2 sq metres. I want >>> an ADC near each Sensor (for signal integrity reasons), so >>> multi-channel ADCs are not suitable. >>> >>> The PXA255 has I2C and SSP ports available. >>> >>> Does anyone have any reccomendations on how I might best do this >>> (taking into account signal integrity, loading, addressing issues,etc) >>> ? >> >> Sounds to me like you might want to drive the ADCs from an FPGA and then >> access the results with a memory-like interface to the processor. >> >> > >I am doing precisely this in a design. It gives a great deal of control >over the access sequence and fixes up the broken behaviour of the PXA255 > SPI implementation (yes, it's broken). I make the FPGA a simple >asynchronous memory target. > >Cheers > >PeteS
Hi Pete I'm using SPI with the PXA255 - can you tell me what issues you have found with it? Cheers Vic

Memfault Beyond the Launch