How does AIRDATA calculate cell voltage deviations?

Deleted member 94047 · Aug 4, 2019

One of my batteries (Mavic Air) has been consistently under performing as compared to the others and I headed to AirData to see if there was an issue with the battery. According to AirData, the maximum voltage difference between the three cells was 0.029v. I downloaded the data and plaid with it in excel. I calculated the maximum difference between all three cells at each data point (maximum of the three minus minimum of the three at each data point). Now that gives a very different result, and a scarier one at that. So I went back to the AirData HD Flight Player tab and grabbed this screen.

If you take the highlighted entry as an example, the difference between the cells with the maximum and minimum voltage would be 0.071v. The difference between cell 1 and cell 2 would be 0.043v. The difference between cell 2 and cell 3 would be 0.028v; which is what AIRDATA gives as the deviation at that data point. I am somewhat confused by this. I thought cell deviation was about the maximum discrepancy between the three cells. Why is AIRDATA using the difference between cells 2 and 3 only? Admittedly, I have no clue about how cell deviations are calculated. So if I am being stupid, please do tell me so. Thanks for any insights.

sar104 · Aug 4, 2019

Your understanding of the meaning of cell deviation in this context is the same as mine, so that looks like an error in the AirData calculations.

Kilrah · Aug 4, 2019

Makes no sense indeed, they might have a bug.

sar104 · Aug 4, 2019

I checked some other AirData records - they do seem to calculate the cell deviation as the difference between two cells, but not always the same two cells and not always the highest and lowest cells.

Deleted member 94047 · Aug 4, 2019

Thanks guys! @sar104, thank you for looking in to it. There has got to be some system to it though, wouldn't you think? It can't be random. Or is taking two cells at random the whole point? That way they approximate the average deviation between cells? Just thinking out loud here. Ok, I will stop before I make a fool of myself.

sar104 · Aug 4, 2019

Doppler said:
Thanks guys! @sar104, thank you for looking in to it. There has got to be some system to it though, wouldn't you think? It can't be random. Or is taking two cells at random the whole point? That way they approximate the average deviation between cells? Just thinking out loud here. Ok, I will stop before I make a fool of myself.

The commonest method is to take the difference between the highest and lowest voltage of all the cells.

The Editor · Aug 4, 2019

I would also not be concerned with a 43mV difference under load.

Anything under 200mV (0.2v) is nothing to be concerned about.

Deleted member 94047 · Aug 4, 2019

@The Editor, I get that. My confusion is over how the difference is calculated. If I take AirData's calculation, the maximum deviation between cells that was registered during the flight was 0.029v. If I understand deviations to mean the maximum difference between any two cells (across all three cells) at a given data point and calculate that myself, the maximum deviation is actually 0.08v. I understand that DJI claims any difference under 0.1v is still normal and I will take them at their word. But 0.08 is way closer to 0.1 than 0.029. I just wanted to understand what cell voltage deviation actually means since the figures AirData gives me disagree with the figures I get when I do the calculation myself.

RayOZ · Aug 4, 2019

From what I can tell, it's the difference between the lowest and 2nd lowest cell value. I am guessing the logic behind it would the the one with the lowest might be the problematic one, and the 2nd lowest is the worse case of the other 'ok' cells.

Deleted member 94047 · Aug 4, 2019

RayOZ said:
From what I can tell, it's the difference between the lowest and 2nd lowest cell value. I am guessing the logic behind it would the the one with the lowest might be the problematic one, and the 2nd lowest is the worse case of the other 'ok' cells.

Thanks Ray, that would explain a lot. I just run the numbers for my other batteries and they all registered about 10 or so instances of a 0.08v deviation between the cells when I calculate the max voltage difference between any two of the three cells. It would make a lot of sense for the deviation to be calculated in the manner you indicate. Great insight, thanks again!

sar104 · Aug 4, 2019

RayOZ said:
From what I can tell, it's the difference between the lowest and 2nd lowest cell value. I am guessing the logic behind it would the the one with the lowest might be the problematic one, and the 2nd lowest is the worse case of the other 'ok' cells.

Good spot - that does appear to be what they are doing.

RayOZ · Aug 4, 2019

I was just thinking, another way of seeing this: assume 1 bad cell, that’s not doing the work. Of the remaining healthy cells, the one with lowest voltage is the one carrying the most load. So cell deviation is the difference between the laziest and most hardworking cells. And if the laziest one fails, it’ll drag the hardest working cell down with it.

Ralph thompson · Aug 5, 2019

I thought I’d read on one of these forums that the RTH trigger was based on the lowest cell so a lazy cell may surprise u one day. With the Inspire we discharge the whole battery every 20 cycles (DJI says 10) to get the cells balanced. A big deviation would, to me indicate that you need to cycle the battery to zero and fully recharge it. I agree that it’s important to monitor this data.

Deleted member 94047 · Aug 5, 2019

Ralph thompson said:
I thought I’d read on one of these forums that the RTH trigger was based on the lowest cell so a lazy cell may surprise u one day. With the Inspire we discharge the whole battery every 20 cycles (DJI says 10) to get the cells balanced. A big deviation would, to me indicate that you need to cycle the battery to zero and fully recharge it. I agree that it’s important to monitor this data.

There seems to be some disagreement here in this forum about the benefit/harm of deep discharging. Some say it is pointless at best and might harm the batteries even. As someone who knows no better, I am ambivalent about doing this ? I wish the experts could come to some sort of consensus over this for the benefit of the rest of us. For now, I will continue to monitor this particular battery because it is under performing both in flight time and voltage as compared to my other batteries - which have the same number of cycles.

Ralph thompson · Aug 5, 2019

I’ve followed the procedure (using the Phantom Angel) for 4 years with the Inspire, which has 6 cells, it seems to be a religion with the Inspire, discussed as absolutely necessary On InspirePilots. But yr right, there doesn’t seem to be the enthusiasm with Mavics. Maybe better technology, fewer cells, not sure.

WithTheBirds · Dec 14, 2019

Doppler said:
One of my batteries (Mavic Air) has been consistently under performing as compared to the others and I headed to AirData to see if there was an issue with the battery. According to AirData, the maximum voltage difference between the three cells was 0.029v. I downloaded the data and plaid with it in excel. I calculated the maximum difference between all three cells at each data point (maximum of the three minus minimum of the three at each data point). Now that gives a very different result, and a scarier one at that. So I went back to the AirData HD Flight Player tab and grabbed this screen. View attachment 78888

If you take the highlighted entry as an example, the difference between the cells with the maximum and minimum voltage would be 0.071v. The difference between cell 1 and cell 2 would be 0.043v. The difference between cell 2 and cell 3 would be 0.028v; which is what AIRDATA gives as the deviation at that data point. I am somewhat confused by this. I thought cell deviation was about the maximum discrepancy between the three cells. Why is AIRDATA using the difference between cells 2 and 3 only? Admittedly, I have no clue about how cell deviations are calculated. So if I am being stupid, please do tell me so. Thanks for any insights.

I think your on the money. The airdata numbers didn’t make sense to me either last time I looked.

Ralph thompson · Dec 14, 2019

Why don’t you quiz Airdata? Maybe they have an explanation, there is a feeling that we are a little sensitive to the numbers, maybe Airdata feels it’s ok to dumb down the data.