Eneloop XX Vs Turnigy 2400 Cycle Testing

[bcwang]

Power Me Up,

You had good answers to all my points so I'm not going to bother quoting them. I'm still wondering about the huge inconsistency in the XX cells, even if the slightly higher capacity is the reason they don't outlast the turnigy cells in number of cycles. I have some questions that may jog more conversation.

- You are terminating the discharge at 0.9v but is this as measured under load or during a pause? If during a pause, the cells may be over discharged.

- This version of the charger isn't able to do constant discharge current I thought, isn't it only able to drain through a resistor? Or am I mistaken and you can control the current very well? If the current is just based on cell voltage, if the XX cells hold higher voltage under discharge than the Tunirgy, they are also discharging at a higher rate and thus under harder conditions.

- To answer one of your questions, the voltage profile of eneloops seems to be higher than some other nimh cells in the regard that they are at a higher voltage. In the MH-C9000 charger for instance, some brands of cell can terminate on -delta V, while Eneloops always end on peak V limit of 1.47v. Maybe most Nimh behave like eneloop and it's the few outliers that are lower voltage. But it is known that the C9000 never terminates on -delta V for eneloop.

- The graphs you generated of the voltage and temp during charging, are those voltages measured under charge load or during pauses between charge pulses? The reason I ask is the C9000 measures during a pause and ends charge termination at 1.47v. Your charts show voltages up around 1.52v. If they were measured the same way, the inflection point termination is definitely ending much further beyond where the C9000 ends it's charge. I'm not saying it's over charging since the c9000 is undercharging eneloops and has to depend on a top-off 100ma 2 hour charge to fill up the rest of the cell. But I wonder if handling the top off gently like that leads to prolonged cell life.

-The cell temperature in that graph, where is the measurement point? If it's a thermistor only contacted by conduction through a charge terminal, the actual cell temp may be much hotter. I'm just trying to get a gauge of how much earlier your inflection point algorithm is ending the charge compared to when the cell temps start to really rise.

-Does the charger/discharger contribute any heat that may be influencing cell temperature?

- With this analyzer you have - I think a potentially valuable test would be to compare life cycles if you only drain to 1v rather than 0.9v under load and see if it improves cell life greatly. You could also have variations which may showcase how nimh life can be greatly improved (or not) by avoiding full charge and full discharge:
0.9v discharge cutoff - full charge
1.0v discharge cutoff - full charge
0.9v discharge cutoff - 90% charge (maybe with peak voltage termination to hopefully hit before any of the other termination methods)
1.0v discharge cutoff - 90% charge

-An unrelated topic but for your firmware - Maybe a mode that can end charge earlier (maybe by customizable peak voltage) with hopes that it can increase cycle life. Kind of like li-ion where partial charge and discharge net far better life. I'm not sure with Nimh chemistry if the full charge chemical process helps keep the chemicals more balanced for longer life. But if there is no benefit to a full charge and only risk with extra heat and wear, a charge less than full may be a great contributor to battery life and I can believe many people don't need the full capacity of the battery and would prefer longer cycle life. Test evidence that this helps batteries last would really seal the deal for the need for this kind of feature in a charger.

I'm actually a backer of the lcd version of the charger so I'm really looking forward to using it!

 

[Bright+]

It seems that increased internal resistance is the cause of these failures and in this case it has happened before much capacity has been lost.

The cells are still usable with good capacity for low current devices.

Yeah, that's my experience too. Once this problem develops, they'll usually get rejected by the charger. If normal chargers will no longer charge them, I consider them done.

 

[Power Me Up]

I'm still wondering about the huge inconsistency in the XX cells, even if the slightly higher capacity is the reason they don't outlast the turnigy cells in number of cycles. I have some questions that may jog more conversation.

Yes, I'm still wondering about the inconsistency as well - I think it's worth running more tests for!

- You are terminating the discharge at 0.9v but is this as measured under load or during a pause? If during a pause, the cells may be over discharged.

Termination is at 0.9V under load.

- This version of the charger isn't able to do constant discharge current I thought, isn't it only able to drain through a resistor? Or am I mistaken and you can control the current very well? If the current is just based on cell voltage, if the XX cells hold higher voltage under discharge than the Tunirgy, they are also discharging at a higher rate and thus under harder conditions.

The non LCD charger can't control the current into an external load, but can use PWM to control the current into the internal load. Unfortunately, the internal load can't maintain a high current as the cell voltage decreases, so what I've done is to build a constant current load (actually 2 of them) for use with the non LCD charger. Unfortunately, the fact that op amps in the real world don't work exactly the same as an ideal op amp should, the current isn't perfectly constant, so I'm using PWM through the internal load to regulate the discharge current. I'm able to maintain the current with fairly good consistency this way.

The LCD charger is able to use PWM on both the internal and external loads, so I'm just using power resistors on that charger.

I have had some problems with contact resistance on the connections for the external loads, so there have been times where the actual current has been lower than desired.

- To answer one of your questions, the voltage profile of eneloops seems to be higher than some other nimh cells in the regard that they are at a higher voltage. In the MH-C9000 charger for instance, some brands of cell can terminate on -delta V, while Eneloops always end on peak V limit of 1.47v. Maybe most Nimh behave like eneloop and it's the few outliers that are lower voltage. But it is known that the C9000 never terminates on -delta V for eneloop.

I agree that Eneloops charge to higher voltages than at least some other NiMH rechargeables. I don't see that as a reason to think that inflection termination wouldn't work correctly though.

- The graphs you generated of the voltage and temp during charging, are those voltages measured under charge load or during pauses between charge pulses? The reason I ask is the C9000 measures during a pause and ends charge termination at 1.47v. Your charts show voltages up around 1.52v. If they were measured the same way, the inflection point termination is definitely ending much further beyond where the C9000 ends it's charge. I'm not saying it's over charging since the c9000 is undercharging eneloops and has to depend on a top-off 100ma 2 hour charge to fill up the rest of the cell. But I wonder if handling the top off gently like that leads to prolonged cell life.

If I recall correctly, someone else did some cycle testing on regular Eneloops in a C9000 and they didn't actually get that many cycles out of them before the C9000 rejected them. I'll have to see if I can find that post - I think that it was here on CPF. Might even be worth repeating the test myself. I suspect that even though the main charge on the C9000 undercharges Eneloops, the top off charge ends up overcharging them a little - it has been speculated that any overcharge - even at low rates is bad for low self discharge cells...

-The cell temperature in that graph, where is the measurement point? If it's a thermistor only contacted by conduction through a charge terminal, the actual cell temp may be much hotter. I'm just trying to get a gauge of how much earlier your inflection point algorithm is ending the charge compared to when the cell temps start to really rise.

Yes, the temperature sensor is mounted on the PCB near the negative terminal of each cell, so the heat is transferred by conduction from the cell holder. I agree that the cell temperature is most likely somewhat higher than the sensor is able to read.

-Does the charger/discharger contribute any heat that may be influencing cell temperature?

That's a very good question! Unfortunately, yes, it does. I don't think that any charger (commercial or otherwise) will be able to charge cells without transferring some heat from the charging circuitry to the cell being charged - short of a specialised test setup at least...

It is a good point that I should have thought of when I was making my list of possible differences between the test that I can run and how Sanyo/Panasonic/FDK run their tests. If I was running this sort of test commercially and wanted to get the maximum number of charge cycles out of the test, I'd be doing my best to isolate the cells from any external heat source such as the charging circuitry. I suspect that they're probably doing that - would be interesting if Rosoku could confirm!

For the testing that I have been doing, the heat transfer from the charging circuitry could be one of things causing the most affect on the results! At least if I was to run testing according to the IEC standard, there would be a lot less heat being generated at 0.25C rates...

- With this analyzer you have - I think a potentially valuable test would be to compare life cycles if you only drain to 1v rather than 0.9v under load and see if it improves cell life greatly. You could also have variations which may showcase how nimh life can be greatly improved (or not) by avoiding full charge and full discharge:
0.9v discharge cutoff - full charge
1.0v discharge cutoff - full charge
0.9v discharge cutoff - 90% charge (maybe with peak voltage termination to hopefully hit before any of the other termination methods)
1.0v discharge cutoff - 90% charge

Yes, I think that makes a lot of sense - I've actually already been thinking along those lines myself. The first test I'm planning to do along these lines is to terminate both charging and discharging early and see how much of a difference that makes. Later on, I'll run separate tests where only either the charge or discharge is incomplete. My current thoughts are to terminate charging at something like 1.40V to 1.44V and to terminate the discharge at 1.1V - it'll be interesting to see how much of a difference this makes on the total number of cycles that can be achieved.

I've got high expectations that doing this will make a significant difference. From what I've heard, Toyota was limiting both the discharge and charge levels on the NiMH cells in the Prius to increase cell longevity. I would expect that they would only do this if they knew that is was going to significantly increase the cell life of their packs, otherwise, they'd just use smaller packs and use their full capacity. Using smaller packs would not only reduce production costs, but it would also reduce total weight which would improve fuel efficiency even further, etc. It might be though that their reason for doing that is purely to make it less likely that individual cells in the packs aren't being reverse charged and overcharged since that would kill the packs relatively quickly and that the benefit for individual cells isn't particularly great at best.

-An unrelated topic but for your firmware - Maybe a mode that can end charge earlier (maybe by customizable peak voltage) with hopes that it can increase cycle life. Kind of like li-ion where partial charge and discharge net far better life. I'm not sure with Nimh chemistry if the full charge chemical process helps keep the chemicals more balanced for longer life. But if there is no benefit to a full charge and only risk with extra heat and wear, a charge less than full may be a great contributor to battery life and I can believe many people don't need the full capacity of the battery and would prefer longer cycle life. Test evidence that this helps batteries last would really seal the deal for the need for this kind of feature in a charger.

Yes, you can actually already do that with the current firmware if you want to.

I'm actually a backer of the lcd version of the charger so I'm really looking forward to using it!

Thanks for being a backer! Hopefully it won't be too much longer before Paul is able to send a charger your way!

 

[Rosoku Chikara]

...I'd like to know what the actual temperature was kept at for their testing? Also, did they use forced air to help keep the cells cool? As I mentioned earlier, it wouldn't surprise me if running at the minimum temperature and with constant air cooling would help the cells to last longer, so I'd be curious if they actually did that...<snip>

I was told that their tests were conducted under laboratory conditions with an aim point of 20 degrees C, and certainly well within the +/- 5 degree standard. He said however, that he believes the actual temperature was much closer to +/- 1 or 2 degrees (or between 18 - 22 degrees C).

Temperature was controlled by environmental air conditioning only. No fans or "constant air cooling" was used.

 

[Power Me Up]

I was told that their tests were conducted under laboratory conditions with an aim point of 20 degrees C, and certainly well within the +/- 5 degree standard. He said however, that he believes the actual temperature was much closer to +/- 1 or 2 degrees (or between 18 - 22 degrees C).

Temperature was controlled by environmental air conditioning only. No fans or "constant air cooling" was used.

That's good to know - thanks!

Did they keep the cells separate from any potential heat sources - e.g. charging circuitry?

 

[Viking]

Temperature was controlled by environmental air conditioning only. No fans or "constant air cooling" was used.

Interesting.
Only one more thing I would have liked to know is if they allow cell temperature to rise above 35 degrees during the test. As far as I understand , forced ventilation is just an option to keep cell temperature under 35 degrees celsius if necessary.