Wrong battery cell count and compensation

[GoFaster]

My 4S packs keep getting detected as 3S. Off the charger, the packs measure 16.8V @ 4.2V/cell. This is well above the 12.8V threshold in the code to be detected as 3S. Firmware is v6.6r21.

Also there is nifty compensation in the code to hold hover at midstick as the pack discharges. Is there a documented process to profile a pack discharge curve for the AQ? I can't reverse engineer it from looking at the default CRV values.

Thanks,
- Felix

[bn999]

There is a ground program called "batCal.cc" in the repo. It takes a flight log file and estimates the CRV parameters. The logged flight should be a constant hover with little disturbance from a freshly charged battery all the way down to a voltage below the lowest you would want to fly with. Best to do it indoors if there is wind outside.

I don't know why your cell count detection is wrong. If you hook up a SWD debug link, you could check the adcData.vIn variable as it goes through the cell parsing code in adc.c.

[GoFaster]

[Kisssys]

Felix,

Plug in a fresh battery and do a short log.

Plot LOG_VOLTAGE7 and see what you get. Anything above 15 volts shows 3.3 on my board's. Divide the value by .22 and it should be close to your battery voltage.

[GoFaster]

Hi Steve -

Very odd, I logged 4S, 3S and 2S packs and everything is detected as 3S.
I've verified that the voltage at the AQ power input is indeed pack voltage.

4S pack - initial voltage 16.74V
Initial LOG_VOLTAGE7 approx 2.459

4S pack - LOG_VOLTAGE7

4S pack - LOG_ADC_VIN

3S pack - initial voltage 12.56V
Initial LOG_VOLTAGE7 approx 2.453

3S pack - LOG_VOLTAGE7

3S pack - LOG_ADC_VIN

2S pack - initial voltage 8.35V
Initlal LOG_VOLTAGE7 approx 2.428

2S pack - LOG_VOLTAGE7

2S pack - LOG_ADC_VIN

Can someone with a board from the public beta batch who is running more than 3S please take a look at your LOG_VIN and LOG_VOLTAGE7?

Thanks,
- Felix

[epyonxero]

Just took a look at the logs from a flight I did last night. Started with a full 4S 5000mA battery. Anything above 15v always shows about 14.9 v on the GCS but it looks accurate below 14.9

[GoFaster]

@epyonxero
Thank you very much for taking the time to look and respond!

@Anyone
Are there any test points or component locations where I can probe to see what the voltages are before and after the divider? Also the pin# on the MCU where Vin is sampled?

[Kisssys]

epyonxero

The analog input can't read above 15 volts with the current implementation so what your last post is 100 percent correct with a 4 cell battery. After it got below 15volts it gave a true representation.

The Recom regulators onboard are only good to 18 volts so it has not been an issue. We are aware of it and some changes to make it 5 cell compatible are in the works.

This should not effect detecting of a 4 cell battery.

GoFaster

You have a board problem, it could be the voltage divider for the voltage in made up of R3 R4 or your reference voltage is not 3.3volts.

If your calibration values come out close to what you've seen posted then I would guess it to be the voltage divider. If your not very comfortable pokeing around you might see about a board swap. If you find R3 And R4 next to the I2C connector the picture will show you where to measure. If you read Vin on the Vin marked side of R4 you should get .22 times the value on the LOG_VOLTAGE7 marked side.

Do you have a diode in series with the battery to prevent reverse voltage that could be causing a problem?

[GoFaster]

Steve, you are right. It is a board problem. I'm missing R3, the 1.3K(?) resistor half of the divider.

missingR3.JPG (86.96 KiB) Viewed 12078 times

Measured:
LOG_VOLTAGE7 = 16.52
VIN = 9.37

Do you happen to know (or BillN/Jussi) the Digikey or Mouser part number for this resistor on the public beta boards? I don't know the spec'ed precision and type or even the correct SMD size.
Your pic has R3 as a 138, so R4 on your board would be a 478? My R4 is a 472.

Thanks again for your help Steve.

[Kisssys]

Here ya go. Should be a easy fix.

311-1.33KHRCT-ND