The reference voltage setting of lpcom does not match the calculation

[pkr2258]

 HI

.

It was configured to detect low voltage of the battery using lpcomp.

This configuration is already being used in mass production of other products and there have been no major problems.

.

However, during this new development, a phenomenon occurred where the reference voltage setting of lpcomp did not match the calculation.

(Circuit and SW configuration are the same as those currently in mass production)

.

1.  circuit configuration

- Internal LDO output and nRF52832 VDD: 3.3V

-Battery voltage range: 4.5V~3.0V

- Low voltage detection voltage: 3.3V

.

2.  lpcom configuration

- reference voltage = 3.3V x 11 / 16 = 2.27V

- AIN5 input            = 3.3V x 2M / (1M + 2M) = 2.27V

.

.

    

   

.

.

3. Problem occurred

- The reference voltage of lpcomp is 11/16 according to the above calculation.

   However, if it actually operates at 11/16, it does not detect 3.3V and detects around 2.9V.

  .

   So, I adjusted the reference voltage to 15/16 and it was detected at 3.3V normally.

 .

- If you measure the input voltage of lpcomp, it is normal.

. 

  

  

[ketiljo]

The setup and calulations should be correct. Are you sure that VDD is 3.3 V?

[pkr2258]

HI

ketiljo said:

The setup and calulations should be correct. Are you sure that VDD is 3.3 V?

.

The VDD of the reference voltage is the LDO 3.3V voltage.

When actually measured, it measures around 3.2989V.

.

.

What's even more strange is that this configuration is already being used in products that are being mass-produced using the same nRF52832.

However, this phenomenon occurred while developing a new PCB.

[ketiljo]

The information you provide seems to be correct. And since you already have a product where this works as expected, there must be an error somewhere in the new design. Double check everything again.

[pkr2258]

HI

ketiljo said:

Double check everything again.

I read your post and checked the whole thing again, but failed to find the problem.

.

However, I found one thing that needed confirmation.

If we rearrange it, the structure is as follows.

.

.

.

Assuming that there is a case where detection is not possible at the detection voltage of 3.3V, the LDO voltage is lowered as well.

.

If the battery voltage is lower than the LDO output voltage of 3.3V, the LDO voltage is also lowered.

The LDO we used has an Iq of 25nA, so there is almost no voltage drop. (Actual measurement results)

- LDO : ://www.ti.com/lit/ds/symlink/tps7a02.pdf

So the table below was created.

Even if the battery voltage becomes lower than the LDO output voltage and the LDO output voltage becomes lower, lpcomp is always in the sensing condition.

.

Of course, it was also confirmed that it was detected well in previous mass-produced products through experiments.

- Detected when the battery is gradually reduced in a fully charged state.

- When the battery is detected in a discharged state from the beginning

.

Could this be a problem?