OK,

A couple of you are aware that I have been implementing 'variable profile' discharge testing of my own. Well now it is ready, and the calibration complete. I now have the capability to replay DPR-50 in-flight log files as a discharge profile 'on the bench', as well as setting arbitrary profiles.

For my initial two studies I selected a 2S PQ2100XP and a 2S 'Impulse Power' (Saehan ProLite V2) 950 pack. The discharge profile I chose to put on them was one that I had seen for one of my electric aerobats: 8 seconds low C discharge (~6C) followed by 2 seconds of high discharge at ~20C - all on a repeat.

The graphs of delivered power, voltage and temperature is shown for these two packs, and I invite some interesting discussion.

Rod

Forgot to add: cell tracking with burst profile.

Both cells in pack tracked during discharge.

Rod

Rod,

Hope you continue posting these graphs as another standard test to the same protocol for all LiPos. It will give us something to compare our "standard" protocol results to, particularly as regards temperatures reached @ depletion.

BTW, have you seen this?

http://www.rcgroups.com/forums/showp...1&postcount=58

Best regards,

- RD

Rod,

First of all can I say that I like the graphs and thank you for doing them. The profile that you have chosen is not dissimilar (in average discharge and peaks) to that in one of my aerobatic aeroplanes. I also particularly like the individual cell monitoring. I guess you will be developing a set of profiles to represent different types of flying (Aero, glider, edf etc.)? It will be interesting to see how temperature and capacity delivered vary from a constant current discharge at the average current.

I have never flown a pack at full throttle for the duration so I think this type of discharge is more useful than the constant current method. Capacity delivered, temperature and maximum power available throughout the flight is very useful. For example, it is good to see that a pack can deliver 90% of peak power at the end of the discharge as opposed to say 75%. (Both of these packs seem to manage about 90% of peak power at the end which is, I think, remarkable). I guess in an aeroplane with these packs, power would be down to perhaps 80% of peak (full throttle) at the end of the discharge because the current will have dropped due to the reduced voltage. However, the profile still gives a very good approximation to a real flight.

I have quickly compared the plots with the constant current discharges for the PQ2100 and a couple of things spring to mind.

1) The temperature profile is slightly higher than the constant 9C discharge as expected but not too much so that's good.
2) The capacity delivered by the PQ2100 is disappointing to say the least. Far less than the constant 9C discharge. Were these cells known to be degraded in any way?

I look forward to your future tests.

Thanks for the data Rod.
The most telling data points for my applications is the voltage delta early in the discharge, before the cell temperatures increase substantially. At an average of nearly 9C, the overall voltage profiles are most applicable to those with setups that provide about 6 minutes of flight time available and limited cooling. For those of us flying systems with significantly more available flight time (10-12 minutes, in my case), and/or with substantial cooling, the later part of the voltage profiles will likely be far less favorable than what you're seeing in these tests. A 20C burst beyond the 50% discharge point on a cell that has not substantially increased in temperature can be brutal. Based on the voltage profile shapes, I think I can guess which of these two cells would have the highest burst capability late in the discharge and at more modest operating temperatures; and I don't think it would even be close.

I look forward to seeing more data from this type of tests. Thanks again for your efforts.

Brad

Originally Posted by Giz

<snip>The profile that you have chosen is not dissimilar (in average discharge and peaks) to that in one of my aerobatic aeroplanes. I also particularly like the individual cell monitoring. I guess you will be developing a set of profiles to represent different types of flying (Aero, glider, edf etc.)? It will be interesting to see how temperature and capacity delivered vary from a constant current discharge at the average current.

If you (or others) have some interesting measured current profiles then why not export them to text or .csv and attach them to a post here in this thread outlining what they were recorded in and under what configuration. If there are some interesting ones I would certainly be interested in running them through on the bench!

Originally Posted by Giz

<snip>I have quickly compared the plots with the constant current discharges for the PQ2100 and a couple of things spring to mind.

1) The temperature profile is slightly higher than the constant 9C discharge as expected but not too much so that's good.
2) The capacity delivered by the PQ2100 is disappointing to say the least. Far less than the constant 9C discharge. Were these cells known to be degraded in any way?

To the best of my knowledge the PQ cells were not damaged; the behaviour may be linked to the problems I had iun running a high C burst for this PQ pack; for the same profile, but at a peak of 45 or 43A (~22C) the discharge would stop due to hitting the LVC. This poorer performance at high discharge may account for the lower capacity.

This question may soon be answered as Enerland have just supplied some 'later' cells to RD and myself and I will certainly repeat this burst profile.

Rod

Originally Posted by Brad S

Thanks for the data Rod.
<snip>
I look forward to seeing more data from this type of tests. Thanks again for your efforts.

Thanks Brad!

Great work, Rod! Inductance and temperature profiles are the most important parameters to assess how much load a battery can take without deteriorating too fast, IMHO.

Jesper

The discharge profile I chose to put on them was one that I had seen for one of my electric aerobats: 8 seconds low C discharge (~6C) followed by 2 seconds of high discharge at ~20C - all on a repeat.

The graphs of delivered power, voltage and temperature is shown for this pack, and can be compared to the previous two.

Rod

I rather like the look of these cells given that I attach quite a lot of importance to light weight. Shame about the cell mismatch though.

Rod, do you have any idea if these Hyperion cells are the ProLite V2? They could be the old ones I suppose but is there any obvious way to tell them apart from the V1 (TP) 2100s other than the cell markings?

Originally Posted by Giz

I rather like the look of these cells given that I attach quite a lot of importance to light weight. Shame about the cell mismatch though.

The cell mismatch may have been down to my sample size of one! For comparison the 950 does not show the same effect.

Originally Posted by Giz

Rod, do you have any idea if these Hyperion cells are the ProLite V2? They could be the old ones I suppose but is there any obvious way to tell them apart from the V1 (TP) 2100s other than the cell markings?

I have no idea - there are three ways to tell:
1. Run the discharge and look at the volts / capacity
2. Measure the temperature during discharge
3. Look at the cell ID - you would have to dis-assemble the pack for this!

All three ways require you to buy the pack!

Rod

4. I'll ask NipponDave and see what he says!

This pack is a relatively new 3S KOKAM 2000 HD 3S 'CellPro' that has seen <30 flights as part of a 3S installation. It has been subjected to some 'abuse' - fully charged I connected a 2S to this 3S pack in parallel for several minutes (by mistake) and generated some 'heat'. So I cannot state for sure how representative it is. It is however useful whilst I'm getting together my thoughts on discharge profiles

I again applied the same profile to this pack (up to its stated 'burst' rating of 20C): 8 seconds low C discharge (~6C) followed by 2 seconds of high discharge at ~20C - all on a repeat.

The graphs of delivered power, voltage and temperature is shown for this pack, and can be compared to the previous three.

Rod

OK,

I noticed the lower temperatures on the last run so decided to 'stress' this pack a little more and match current to the two 2100 packs I carried out; to give a 6.5C/21C burst profile (1C higher than stated max 'burst').

Attached below are the usual graphs for a 13A/42A burst profile in 80%/20% ratio.

Rod

Hi Rod,

Very good work.

But why the hell are your graphs marked in FAHRENHEIT ??

Why aren't you using S.I. units ?????

This is the 21st century you know.