ADS131M02-Q1: Root cause of the offset issue

Hi Yoshikazu-san,

I will look into the details and get it back to soon.

Thanks&regards,

Dale

Hi Yoshikazu-san,

are the results mentioned in the Excel sheet with global-chop mode enabled or disabled?
With global-chop mode disabled the ~200uV offset for the current shunt measurement would be within the expected specification.

For the voltage measurement, the 0.75mV offset look to large.

What data rate (or OSR) are they using?

Could they provide the raw conversion results from the ADC for the measurements with 0A current and 0V voltage?
Could you ask the customer to measure the offset with the inputs internally shorted by setting MUX0[1:0] = 01b and MUX1[1:0] = 01b? And do the measurement both with global-chop mode enabled and disabled.
That way we can potentially identify if the offset is coming from the device itself or rather the external circuitry.

Note that ADC1 offers better offset performance than ADC0. Therefore it would be preferable to use ADC1 for the current shunt measurement. 

I wanted to mention that we also offer ADS131B02-Q1. This device has improved offset performance compared to ADS131M02-Q1 and also integrates an oscillator so that an external clock would not be required. The ADS131M02-Q1 and ADS131B02-Q1 are pin-to-pin compatible.

Regards,
Joachim Wuerker

Hello Joachim-san,

Thank you very much for your quick reply.

The results in the excel sheet were with global-chop mode disabled.  I thought the offset value I got aligned with the global-chop mode disabled, so I asked my customer to check the offset again with global-chop mode enabled.  Then they enabled the mode, but it seems the results didn't improve.

I'll check what you suggested and get back to you here.  I'll also offer ADS131B02-Q1, but it is strange that the older device ADS131B02-Q1 has better offset performance than the newer device ADS131M02-Q1.

Best Regards,

Yoshikazu Kawasaki

Hi Yoshikazu-san,

it might also be good to verify that global-chop mode is really enabled. You could verify that by monitoring the DRDYn pin. The pin should transition low with a data rate which is ~1/3 of the data rate when global-chop mode is disabled.

The ADS131B02-Q1 is actually the newer device.
We released the non-automotive version of ADS131M02 first to the market. We then developed ADS131B02-Q1 for the automotive BMS market as an evolution of the ADS131M02. We increase the input impedance by adding input buffers, improved the offset performance, and added an internal oscillator besides other things.
Later on we then also qualified the ADS131M02 for automotive due to customer requests.

Regards,
Joachim Wuerker

Hello Joachim-san,

Thank you for your advice.  I understand you think the offset value should be different(lower) if the global-chop mode is enabled, so you suggested we should ask them to confirm if the global-chop mode is really enabled.  It would take ~1/3 of the data rate when the mode is disabled for DRDY pin to go low, but would you please tell me how long it would be if the mode is enabled?  If I know both timings, I can ask them to confirm if the mode is disabled or enabled.

I understand the background of ADS131M02-Q1 and the reason why ADS131B02-Q1 has better offset performance than ADS131M02-Q1.

Best Regards,

Yoshikazu Kawasaki

Hello Yoshikazu-san,

when global-chop mode is disabled, then the period between DRDYn pin falling edges equals 1/Data rate. The data rate depends on the clock frequency and the configured OSR setting: Data rate = (f_CLK / 2) / OSR. You can also check table 8-2 in the datasheet.

Regards,
Joachim Wuerker

Hello Joachim-san,

I got their replies as shown below.  It seems they didn't look at the data well and they finally confirmed the offset improved by enabling the global-chop mode this time, so they apologized their previous comments.

What data rate (or OSR) are they using?

-> OSR=256

Could they provide the raw conversion results from the ADC for the measurements with 0A current and 0V voltage?

-> 800~1000 at gain=1 and 80000~100000 at gain=64.

Could you ask the customer to measure the offset with the inputs internally shorted by setting MUX0[1:0] = 01b and MUX1[1:0] = 01b? And do the measurement both with global-chop mode enabled and disabled.

-> They didn't change the MUXx registers, but they confirmed the data had changed by global-chop mode disable/enable.  They apologized the previous comments about it.

-100~200 at gain=1 and 8000~10000 at gain=64

Now they confirmed the offset values were aligned with the datasheet, but would you please tell me the reason why the global-chop mode is disabled in default?  They think it should be enabled in default unless it has negative impacts on the device performances, so they're worried if enabling the mode might have negative impacts to their system.

Best Regards,

Yoshikazu Kawasaki

Hello Yoshikazu-san,

thank you very much for this update. Those numbers look indeed much better now.

In all our ADCs we usually chose "global-chop mode = disabled" as the default setting. I agree that global-chop mode does only improve the measurement performance (offset, noise), however it has disadvantages when it comes to data rate and digital filter settling/latency times.
As you saw, the data rate for the ADS131M02-Q1 is 3x lower when enabling global-chop mode and also the digital filter settling time becomes longer.
Why are they concerned that global-chop mode = enabled is not the default setting? They have to change the OSR, gain and other settings anyway. So why is changing the global-chop mode setting concerning?

Regards,
Joachim Wuerker