Cell Deviation on the start

[t2adze]

Hey All

Yesterday had this flight - DJI Flight Log Viewer - PhantomHelp.com

Right from the begginign the battery voltage was good, after couple of seconds of flying it started decreasing drastically.
You can probably see from the log.

But after a while it was back to normal and i succesfully did the flight i had planned.

Is there anything i can do about this?

 

[tevek]

I didn't scroll right the way through your log (difficult on an iPad), I suspect you may be looking at retiring that battery as you are getting many excessive variations in your cell voltages, particularly cell #4. I generally disregard the % value as that doesn't indicate what the current draw or individual cell performance is like. I have received a number of > 70mV variations but nothing like the number of occurrences you have logged. No idea how old your battery is but maybe keep it as a spare unit for FW upgrades and very local flying/testing.

 

[t2adze]

Is there a way to fix it?

Can someoe please check log and let me know what can be done
Thx

 

[With The Birds]

How long had that battery been sitting before you flew it (after last charge). What is your days to auto discharge setting on the pack. Has the pack ever been left at a low charge % for an extended period? How old is the pack? How many cycles? When you look at the battery details in GO app what is the total capacity reported?

 

[sar104]

AirData analysis shows some deviations, but does not list them as very significant at this stage.

Airdata UAV - Flight Data Analysis for Drones

 

[With The Birds]

AirData analysis shows some deviations, but does not list them as very significant at this stage.

Airdata UAV - Flight Data Analysis for Drones

When we consider the flight log only represents 4 minutes and we have 10 deviations greater than 70mv depicted on one cell I would be watching this pack like a hawk. If the total capacity reported was more than 15% below rated and it performed similarly after charging I would consider this pack unreliable. I can only see one page on airdata from the link you posted.

 

[sar104]

When we consider the flight log only represents 4 minutes and we have 10 deviations greater than 70mv depicted on one cell I would be watching this pack like a hawk. If the total capacity reported was more than 15% below rated and it performed similarly after charging I would consider this pack unreliable. I can only see one page on airdata from the link you posted.

I agree - I'd certainly watch it closely to see if it gets worse. That's one advantage of using AirData to track batteries. The link I posted was only for the battery cells analysis since that was the subject of the discussion.

Airdata UAV - Flight Data Analysis for Drones

 

[With The Birds]

I agree - I'd certainly watch it closely to see if it gets worse. That's one advantage of using AirData to track batteries. The link I posted was only for the battery cells analysis since that was the subject of the discussion.

Airdata UAV - Flight Data Analysis for Drones

In my view what might be drawn from the cell analysis is better informed having regard to the other airdata analysis, specifically the initial capacity (in this case seemingly close to 15% below that of rated) and the depicted amp draw in flight. In this case there is nothing to suggest the pack was heavily loaded when the subject deviations were reported. While it might be hoped to get more than 97 cycles out of a pack I would be taking this one out of service, it is unlikely to improve and may well degrade quickly and unexpectedly.

 

[sar104]

In my view what might be drawn from the cell analysis is better informed having regard to the other airdata analysis, specifically the initial capacity (in this case seemingly close to 15% below that of rated) and the depicted amp draw in flight. In this case there is nothing to suggest the pack was heavily loaded when the subject deviations were reported. While it might be hoped to get more than 97 cycles out of a pack I would be taking this one out of service, it is unlikely to improve and may well degrade quickly and unexpectedly.

Agreed - I should have included the other battery analysis results.

 

[With The Birds]

Agreed - I should have included the other battery analysis results.

I don't know about should have SAR- it all helps. What only the OP might provide is a response to my earlier questions- specifically days to discharge setting and time since last charged. While reportings here are infrequent it seems on the limited instances discussed that a battery that has entered the auto discharge routine might perform similarly to what we have seen here. My simple hypothesis being that the bleed resistors employed in the discharge routine are not closely matched in tolerance and therefore load the cells in an unequal fashion which must provide they are out of balance at the commencement of flight.

 

[sar104]

I don't know about should have SAR- it all helps. What only the OP might provide is a response to my earlier questions- specifically days to discharge setting and time since last charged. While reportings here are infrequent it seems on the limited instances discussed that a battery that has entered the auto discharge routine might perform similarly to what we have seen here. My simple hypothesis being that the bleed resistors employed in the discharge routine are not closely matched in tolerance and therefore load the cells in an unequal fashion which must provide they are out of balance at the commencement of flight.

That's possible if that's how the bleed resistors are arranged. But why would the auto-discharge system bleed the individual cells independently, rather than discharging them in series?

 

[With The Birds]

That's possible if that's how the bleed resistors are arranged. But why would the auto-discharge system bleed the individual cells independently, rather than discharging them in series?

Design economy is my guess. The bleed resistors are already switched in during charging to provide balancing, matching precision isn’t a critical consideration here as the PLC can simply vary duty cycle to perform the intended function. To perform discharge simply switch the resistors into circuit. They are already configured to individual cells through the balance cell taps. Series connection would require a seperate load and power mosfet or even greater complexity in modifying the switching scheme for the existing resistors.

 

[sar104]

Design economy is my guess. The bleed resistors are already switched in during charging to provide balancing, matching precision isn’t a critical consideration here as the PLC can simply vary duty cycle to perform the intended function. To perform discharge simply switch the resistors into circuit. They are already configured to individual cells through the balance cell taps. Series connection would require a seperate load and power mosfet or even greater complexity in modifying the switching scheme for the existing resistors.

That makes sense.

 

[solentlife]

Right ....

Question : What was wind and weather conditions like ?

Why ? I have 3 good packs and all will do the same ..... if I am flying in moderate / strong wind - the power required by the motors commanded by the FC varies significantly in what you want to do. The wind will be buffeting the AC and FC will compensate with hefty doses of power.
I have seen my packs voltage drop (Go shows an average cell voltage on screen under %) shortly after take off to indicate at ~3.7V. That's with pack still in the high 80 - 90% state. As the pack goes down - that voltage drop can get even worse. BUT it is not the pack faulty or bad ... its the work needed to hover / stay / climb / fight wind.

Go out and fly in calm conditions and voltage drop is significantly less ... and you see 3.9 - 4.0V shortly after take off with a steady decline as pack goes down.

OK - so one cell drops more than another ... even though factorys match cells as much as possible to build a pack - they still exhibit different voltage drops under serious load. So one cell in a pack can when pack subjectyed to high demand - drop more than the others.

Nigel

 

[solentlife]

That's possible if that's how the bleed resistors are arranged. But why would the auto-discharge system bleed the individual cells independently, rather than discharging them in series?

You have hit the nail square on the head ... it doesn't !

Nothing on that Battery board balances out the pack properly. It bleeds of the energy as best it can with simple resistive discharge.

The main function of the balance lead hidden away inside the pack is to relay cell voltage via telemetry to you on the ground. It's SECONDARY function is to try balance - but it fails miserably at it.

If DJI was to allow us to access the balance lead as with most other LiPo items - we would have proper cell by cell balancing based on bleeding high cells down to match the lowest ...

Nigel

 

[With The Birds]

B

You have hit the nail square on the head ... it doesn't !

Nothing on that Battery board balances out the pack properly. It bleeds of the energy as best it can with simple resistive discharge.

The main function of the balance lead hidden away inside the pack is to relay cell voltage via telemetry to you on the ground. It's SECONDARY function is to try balance - but it fails miserably at it.

If DJI was to allow us to access the balance lead as with most other LiPo items - we would have proper cell by cell balancing based on bleeding high cells down to match the lowest ...

Nigel

The DJI circuitry fails miserably to balance cells in what respect?

Below is the schematic for the DJI phantom 3 battery. It employs TI MSP430 MCU and a seperate IC to implement cell management and protection.

The resistors that provide for, what you call simple resistive discharge, are depicted bottom centre (R1 to R4). They are 100ohm and switched in as needed on individual cells to perform balance/discharge. At nominal cell voltage we are looking at 160mw/cell dissipation (600mw for the pack). By way of comparison something like an IMAX B6 might provide 5w total discharge capacity (presumably a similar load is available in balancing) however the B6 will handle higher capacity packs and can charge at greater than 1C rates also. It is likely that 160nw/cell is adequate in the current implementation for the DJI Phantom packs.

It would seem that the DJI implemwnted Texas Instruments battery management solution provides for significantly more than ineffective balancing and telemetry. None of my packs have ever come off the charger out of balance.

 

[solentlife]

Problem is cutoff occurs before balancing can be effective.

Too many instances of the group I was communicating and flying with saw non balancing ....

You can have the worlds greatest chip - but if the cutoff occurs too early - it is voided literally.

WtB - you are lucky that you never see in balance ... when it comes and it will - the DJI system will not solve it.

Nigel

 

[t2adze]

How long had that battery been sitting before you flew it (after last charge). What is your days to auto discharge setting on the pack. Has the pack ever been left at a low charge % for an extended period? How old is the pack? How many cycles? When you look at the battery details in GO app what is the total capacity reported?

1. Around a week
2. 4 days
3. Not longer then couple of days
4. Around 2 years old
5. up to 100 i think. will check
6. will check and report

 

[solentlife]

1. Around a week
2. 4 days
3. Not longer then couple of days
4. Around 2 years old
5. up to 100 i think. will check
6. will check and report

If pack is left 4 days before use and even with Auto Discharge set at longer - my opinion is to charge up anyway before flight. ALL batterys suffer Self Discharge. I agree that LiPo's and particularly LiFe suffer far less than other batterys - even LSD NiMh.

BUT here we have a battery pack that is NEVER completely switched off ... the 'intelligent board' slapped onto the front is powered permanently ... actually it may power down in Hibernation - but is still connected.
I also agree that board monitoring will be a few mA ... but its still there.

Answers 3, 4, 5, 6 are interesting but IMHO have less impact than 1 and 2.

IMHO of course.

Nigel

 

[Numone]

I didn't scroll right the way through your log (difficult on an iPad)...

As an aside, here's a quick tip Tevek. If you tap the flight warning points on the map, it skips through the log to that point, thus avoiding wearing out your fingerprint...!