LDC1101: DIG_CONF Measurement Response Time

Arthur Thompson27

Expert 1755 points

Part Number: LDC1101

Tool/software:

Hello,

We are seeing our Rp/L measurements take longer than predicted by Formula 10 in Section 9.1.8 of the datasheet.

Our sensor frequency is 4 MHz.

Here are the values we measured experimentally, by using an oscilloscope to measure the time between the sensor waveform turning on and the LDC generating the INTB signal to indicate the measurement is ready.

RESP_TIME Measured Conversion Time

768 270 usec

1536 538 usec

3072 1064 usec

6144 2128 usec

The expected measured conversion time, e.g. for a RESP_TIME of 6144, should be 6144/(3*4e6) = 512 usec.

Can TI explain why my measured conversion time is taking longer than predicted by the datasheet formula?

Thanks,

Arthur

2 months ago

0 Patrick Simmons 2 months ago

TI__Mastermind 23095 points

Hey Arthur,

Thanks for considering to use Texas Instruments. Could you provide the scope capture you used for determining the conversion time discrepancy? We would like to see if the sensor waveform profile aligns with what we would expect.

0 Arthur Thompson27 2 months ago in reply to Patrick Simmons

Expert 1755 points

Hello Patrick,

Pink is the interrupt pin, and Blue is the single-ended sensor waveform (just half of the total differential signal) with a setting of 6144.

0 Patrick Simmons 2 months ago in reply to Arthur Thompson27

TI__Mastermind 23095 points

Hey Arthur,

Thanks for providing that. Will try to have an update later today regarding debug steps or what we think might be wrong.

+1 Patrick Simmons 2 months ago in reply to Patrick Simmons

TI__Mastermind 23095 points

Hey Arthur,

We did some testing and it looks like we duplicated your issue. From our tests, we startup the device from sleep mode and then start taking samples. Looks like the initial conversion period runs ~4x longer then all subsequent conversions. As of now, we do not have a work around for this behavior. Are you planning to use only one conversion cycle when waking from sleep mode?

0 Arthur Thompson27 2 months ago in reply to Patrick Simmons

Expert 1755 points

Hi Patrick,

Yes, for maximum power saving we are planning to just use one conversion and then go to sleep. We have already shown that setting RESP_TIME to 1536 allows the measurement to complete fast enough for our application. I had been wanting to increase RESP_TIME to 6144 in order to maximize resolution while staying within our time budget, but due to the first conversion being longer than predicted in the datasheet, we can just leave RESP_TIME at 1536.

Thanks for your support,

Arthur

Sensors

Sensors forum

LDC1101: DIG_CONF Measurement Response Time