Battery Performance Tester

Stuff I am working on
Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Sun 13 Sep 13 2009 10:15 pm

Finally! The GP1600 test results with the new ISR routine handling are in. The GP1600 batteries went in after taking the following charge from the charger:-

BATT1 1388mAh
BATT2 1053mAh
BATT3 1199mAh
BATT4 1330mAh

The discharge curve is as follows:-
GP-1600-1234ABCD-10.jpg
GP-1600-1234ABCD-10.jpg (75.53 KiB) Viewed 4968 times
Right off the bat you can see that the CH3 behaviour is much better now. All the curves are also better matched as well. The reported capacities are:-

BATT1 1148mAh 1315mWh
BATT2 1069mAh 1222mWh
BATT3 1063mAh 1152mWh
BATT4 1118mAh 1243mWh

This is looking to be pretty closely matched to the charge taken since some of the charge is usually dissipated as heat and each cell may have had slightly different state of discharge upon start of charge cycle. Basically, I am quite pleased with this performance right now.

Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Mon 14 Sep 14 2009 3:13 am

The Sanyo2700 test with the revised ISR code, and 200Hz PWM frequency is done. The batteries took the following charge going in:-

BATT1 1973mAh
BATT2 1894mAh
BATT3 1650mAh
BATT4 1621mAh

The graphs are as follows:-
Sanyo-2700-1234ABCD-6.jpg
Sanyo-2700-1234ABCD-6.jpg (68.01 KiB) Viewed 4966 times
The reported capacities are:-

BATT1 1614mAh 1776mWh
BATT2 1252mAh 1442mWh
BATT3 1334mAh 1502mWh
BATT4 1243mAh 1385mWh

The general capacity report follows the charge intake. The overall capacity reported of these cells are pretty poor though so we will put the cells through an analyze cycle to see just how close we are. The graph, otherwise, exhibits no major anomalies although CH2 seems to be showing a tendency to have a steeper knee. I'm not entirely sure why this is the case at the moment though. This was also the case with the previous GP1600 test with PWM at 500Hz. Mitigating this, however, is the tendency of BATT2 to show this type of behaviour regardless of which slot it is occupying so it may simply be a battery problem.

Daniel

Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Mon 14 Sep 14 2009 7:42 am

Okay, I think we're there in terms of calibration. This set of results from our GP1600 went in with the batteries taking the following charge:-

BATT1 1361mAh
BATT2 1328mAh
BATT3 1380mAh
BATT4 1328mAh

The resultant graph is nearly identical to the previous test set, which is to be expected since the batteries were occupying the same slots.
GP-1600-1234ABCD-11.jpg
GP-1600-1234ABCD-11.jpg (76.25 KiB) Viewed 4965 times
The reported capacities for this test were:-

BATT1 1163mAh 1341mWh
BATT2 1057mAh 1217mWh
BATT3 1065mAh 1166mWh
BATT4 1116mAh 1250mWh

Again, these were remarkably close to the previous test set results to within about 1% or so. In short, we have consistency nailed here. What we now need is to get some agreement between our reported capacities and the C-9000's analysis results. It should be noted that while similar, the reported capacities right now are not totally correlated to the C-9000's report about 11-cycles ago. We will have to do a fresh analysis of the current set to get verification of the numbers and to see if we need to scale the results some. On the average, however, actual charge capacities are about 85% of charge taken. Using this as a guideline, we estimate the charged capacities of the battery from the charge taken to be about:-

BATT1 1157mAh
BATT2 1129mAh
BATT3 1173mAh
BATT4 1129mAh

These estimated results corresponds very closely to our reported capacities so our results are in fact very credible. We just need to do one more test with this set with the batteries in different slots.

Daniel

Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Mon 14 Sep 14 2009 3:02 pm

I finally decided to profile a set of new ReCyko's after running them down a bit in my audio recorder. The batteries only took about 800mAh so it probably wasn't really empty. In any case, I ran it on the tester CH1 and CH2 and this is what we got:-
ReCyko-1.jpg
ReCyko-1.jpg (54 KiB) Viewed 4965 times
The reported capacities were:-

ReCyko1 2041mAh 2330mWh
ReCyko2 2050mAh 2347mWh

A couple of things to notice - first of all, notice how closely matched the two curves are - which is a very good sign as this shows our tester to be repeatable and reliable. Secondly, the ReCyko batteries have a 2050mAh typical capacity rating so we are spot on there.

Daniel

Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Mon 14 Sep 14 2009 10:40 pm

As a last check of the system consistency, I charged the GP1600's in the ReCyko+ charger and put it through the tester again, this time with the batteries in reversed order (in all in different slots). The results are as follows:-
GP-1600-1234ABCD-12.jpg
GP-1600-1234ABCD-12.jpg (79.99 KiB) Viewed 4964 times
The reported capacities are:-

BATT1 1199mAh 1370mWh
BATT2 1109mAh 1271mWh
BATT3 1175mAh 1269mWh
BATT4 1200mAh 1327mWh

As you can see, notwithstanding the fact that the batteries now occupied different slots, the curves are pretty much identical in profile. The reported capacities are a little higher because the ReCyko+ charger has a higher termination voltage than the more conservative C9000 charger (without the 2-hour top-up charge). The capacities are still close enough though such that we can trust in these readings.

Daniel

Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Tue 15 Sep 15 2009 8:06 pm

These are some of the batteries that are lined up for the test. So far, I'm not including regular Ni-MH rechargeables, on the LSD types. A couple more that I want to include are the Panasonic Oxyrides, SONY LSD's if there are any, perhaps Kodak, several other cheap house brands, Duracell Ultras, and possibly some dry cells. For the moment though, these are what we have:-
IMG_0982a.JPG
IMG_0982a.JPG (353.75 KiB) Viewed 4962 times
I'll have to go battery hunting again soon but these things are costing me a fortune to acquire and to test.

Daniel

Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Wed 16 Sep 16 2009 6:19 pm

After 2-days of break-in, the ReCyko's were showing the following capacities:-

ReCyko1 1996mAh
ReCyko2 2041mAh

Putting this through the tester we got:-
ReCyko-2.jpg
ReCyko-2.jpg (56.38 KiB) Viewed 4960 times
Reported capacities were:-

ReCyko1 2115mAh 2449mWh
ReCyko2 2059mAh 2415mWh

Bearing in mind that we are discharging to 0.8V rather than 0.9V could explain the slightly inflated figure for ReCyko1 but it is otherwise quite reasonable. A bit of scaling may help.

ReCyko2 seems to be spot on, however, but is breaking the trend. In particular it looked to be running at a higher terminal voltage which could suggest that it's coefficient was a bit on the low side, or that CH1-CH4's are on the high side. Unfortunately it looks like CH5 is out of step with CH1-CH4 so we should avoid using it where possible and stick mainly to CH1-CH4 for our testing purposes.

Daniel

Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Thu 17 Sep 17 2009 1:14 pm

We charged the ReCyko's again, taking the following charge:-

ReCyko1 2005mAh
ReCyko2 2110mAh

The actual charge will be higher since we allow for 2-hours (and more) of topping up. The test curve is as follows:-
ReCyko-3.jpg
ReCyko-3.jpg (68.51 KiB) Viewed 4959 times
Reported capacities were:-

ReCyko1 2136mAh 2435mWh
ReCyko2 2107mAh 2432mWh

The curves no longer seem as closely matched as they were before we ran the break-in cycles although capacities seems to be higher. If we look at the mWh ratings though, we can see that they are still very close - reinforcing the idea that mWh is actually a better measure of capacity than mAh is. Nevertheless 29mAh of difference is practically neglible for all intents and purposes. I think for some of the other cells, I will start without break-in cycles. Repeatability is high but we will do one more test with the batteries in reversed holders to confirm that the behaviour is due to the cells and not the holders.

Daniel

Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Fri 18 Sep 18 2009 1:32 am

Okay, we re-tested the ReCyko's in different holders to see if the general trend holds. Results were:-
ReCyko-4.jpg
ReCyko-4.jpg (55.6 KiB) Viewed 4958 times
Reported capacities were:-

ReCyko1 2103mAh 2443mWh
ReCyko2 2121mAh 2441mWh

Seeing as to how the CH1 (blue line) is still higher, we may have a calibration type of issue. I think it might be better to do all the tests using just the best matched channels. This could be CH2 and CH4 or CH1 and CH3. We will have to look at the measured bits to ascertain this. It is also possible that this kind of behaviour is impossible to solve without better matched load resistors. The problem here is that a channels lower load resistance is being made up for by a lower PWM duty-cycle so that it draws more current but for a shorter period. The end result should be the same but the chemical impact is not. As a result it is going to be difficult to match the curve profile.

Daniel

Daniel Wee
Site Admin
Posts: 2070
Joined: Wed 25 Feb 25 2009 8:00 pm

Re: Battery Performance Tester

Post by Daniel Wee » Fri 18 Sep 18 2009 3:11 pm

We repeated the test using CH2 and CH4 which are the most closely matched channels and we did get a more closely matched graph as follows:-
ReCyko-5.jpg
ReCyko-5.jpg (54.9 KiB) Viewed 4955 times
The reported capacities were also very much closer, expectedly:-

ReCyko1 2110mAh 2468mWh
ReCyko2 2124mAh 2470mWh

This gives us good reason to use better matched loads.

Daniel

Post Reply