ADS8568: Measurement spikes

Hello Winston,

Thank you for your post.

Can you share some examples of the measurement spikes you're observing? If you can plot some of this data which shows the spikes, it would help us understand whether it is valid ADC conversion data with some added noise transients, or perhaps the data is corrupted along the SPI bus.

Regards,

Ryan

The plot shows ADC raw data from channel 7 with input driven to zero V. Channels 1 - 6 look similar while channel 0 is clean. ADC is sampling continuously at 720 Hz.

Thank you,

Winston

Hi Winston,

The images did not upload correctly. Can you try again, or perhaps attach a file with the raw data and plots?

Regards,

Ryan

How do I upload a file? The menu only asks for a URL.

Basically the measurement is close to zero with spikes to ~16000, -16000 and -32000 every few seconds. I looks like two MSBs are corrupted.

Can I just email the file?Jpeg, tiff and png formats all seem to have the same problem.

Is there any reason for the MSB of each channel to be more prone to errors. We are just clocking out data with a 128 12.5 MHz clocks as a continuous stream. Does the MSB have a longer output delay than the other bits?

Hi Winston,

To upload a photo, you can either take a screenshot and paste directly into the message box, or follow the menu options shown below (the gray "Upload" button is not so obvious, but it will launch File Explorer for you to navigate and select a saved image):

Based on your description, this sounds like a digital data capture issue and not some analog input transient. The MSB does have a different propagation delay than the other bits and is launched by the falling edge of FSn ("Frame synchronization"). Ensure that you are waiting at least 12 ns before bringing SCLK low, which captures the MSB.

Using Figure 1 as a reference, can you capture a scope image of the interface for one frame? We can review FSn, SCLK, SDO_x, and BUSY.

Regards,

Ryan