ADS8568 Repeated Samples

Steve MacLeod

Other Parts Discussed in Thread: ADS8568

I am operating the ADS8568 in parallel mode with all 4 channel pairs enabled. All four CONVST pins are tied together and driven from a single source. Channel A0 is the only channel being used for measurement. The data on channel A0 randomly produces a repeated sample; in other words, a sample is being replaced with data from the previous conversion. For a test, I grounded CONVST B,C,D and only drove CONVST_A. In this case, the data is as expected. I need to operate all 8 channels. Is there an explanation for the repeated samples?

over 10 years ago

0 Tom Hendrick over 10 years ago

TI__Guru**** 180931 points

Hi Steve,

I've seen this once before in a situation where the ADS8568 was being run at full speed using the internal conversion clock. Is that by chance what you are doing here? What are you using as a host controller, a state based FPGA or some sort of DSP/micro controller?

0 Steve MacLeod over 10 years ago in reply to Tom Hendrick

Prodigy 190 points

Hi Tom,

Thanks for the response. It is being run with the internal conversion clock and a sample rate of 100kHz. The host is a state-based FPGA.

0 Tom Hendrick over 10 years ago in reply to Steve MacLeod

TI__Guru**** 180931 points

Hi Steve,

Take a look at the BUSY output signal. Trigger on the rising edge and compare the falling edge to the point in time where you start reading the conversion results. You should see the BUSY dither a little because the CONVST is not coincident with the conversion clock. What was happening in the case I mentioned yesterday is that the reads were taking place just ahead of the final data buffer update. Can you add a wait state of a conversion clock cycle (~85ns) to your read routine and see if that takes care of the repeated data issue?

0 Steve MacLeod over 10 years ago in reply to Tom Hendrick

Prodigy 190 points

Tom,

The falling edge of BUSY and the rising edge of the first RD pulse varies from 82ns to 88ns, approximately. We can add a delay to the read routine and retest.

-Steve

0 Steve MacLeod over 10 years ago in reply to Steve MacLeod

Prodigy 190 points

Hi Tom,

Adding a delay as you suggested solved the problem. Thank you very much!

-Steve

0 Tom Hendrick over 10 years ago in reply to Steve MacLeod

TI__Guru**** 180931 points

Great to hear! Thank you Steve for letting me know!

Data converters

Data converters forum

ADS8568 Repeated Samples