Constant Current Loads

[Mark]

I've been doing a little work lately on making up a constant current load that can be used with the UltraSmartCharger.

At the moment, I've got a circuit partly working on a breadboard - I've just got to change some components to get it working properly, then I'll be building it on a vero board.

I've taken a look at the costs for doing short runs of printed circuit boards and I'm thinking about eventually heading down that path. I'm currently wondering if anyone else would also be interested in having such a board?

Current planned features are: ability to discharge both channels at the same time. Trim pots to adjust the current - adjustable from basically 0 through to 2 amps. Probably have a three position switch (and 3 trim pots) so that 3 different currents can be dialed in and you can switch between them without having to further adjust the trip pots. If there is sufficient interest, I'm thinking about 4 options for others to get them

1. Make the PCB design available for others to order directly through their own choice of PCB manufacturer.
2. Offer pre made PCBs ready to post to anywhere in the world.
3. Offer kits including the PCB and all of the required parts.
4. Offer a complete board ready for use.

At this stage, I'm just using through hole components, so it should be easy enough for anyone to assemble if they've already got some soldering experience.

The last option would be better for those that aren't confident with a soldering iron, but keep in mind that header pins would still need to be soldered to the charger to be able to do external discharges. A total of 5 solder joints would need to be made - 2 for each channel plus the raw 5V header pin to power the circuitry on this board - the ground connection would be made through the FTDI pins already on the board.

I haven't done up an estimate of costs at this stage - I'll wait to see if this is of interest to anyone first!

[RLBrooks]

Mark,

How would the average user (me) use this load? Why would I need to? Doesn't the charger have a way to discharge batteries prior to recharging them?

Cheers... Russ

[Mark]

You would need to connect 6 wires between the board and the charger - 2 each for each channel, plus 5V and Ground.

You don't need to have this - the charger has a 3.9 ohm resistor on it for discharging cells. The reason to have an external load is for a couple of reasons: The internal load only allows a relatively low discharge rate and being a purely resistive load, the current drops as the voltage of the cell drops.

With the internal resistive load, it only allows for about a 300mA discharge rate on average. The charger can use PWM to reduce the current through the resistor to make it more constant, but it can't make it go higher than the resistor would naturally allow, so you're really limited to about 200 mA if you want a constant current discharge. Discharging into a PWM load also isn't the same as a constant current discharge - the cell will actually heat up more when doing a PWM discharge compared to the same rate with a constant current.

Normally, batteries are tested with a 0.2C discharge rate - for Eneloops, that works out to 400mA, so discharging them at a slower rate isn't quite the same and will give a (marginally) higher capacity result.

When doing cycle testing, I want to be able to charge and discharge the cell as fast as is reasonably possible so that it doesn't take forever - the only way to do that is with an external load. The testing that I have been doing has been with 5W wire wound resistors, but this has the problem of the current still reducing as the cell voltage drops and this makes it harder for others to reproduce the same test results or compare with testing done by others. Unfortunately, the current charger can't do PWM on the external load either.

If you haven't seen the cycle testing that I've already done so far, you can find it here:
viewtopic.php?f=5&t=69

[RLBrooks]

Thank you for the explanation. It sounds like the external load while flexible and maybe nice to have, is mainly for cell testing purposes and won't provide the end user any significantly useful advantage in the normal day to day of charging and maintaining batteries.

I don't do any tracking of my batteries or even try to keep pairs together. I don''t know which have been used more than others and which are better performers than others, nor which is starting to fail. This is the why the idea of a 'smart' charger, that treated cells individually attracted my attention. I was concerned by mixing up pairs and maybe starting pair charges with cells at different levels, I wasn't getting my money's worth from the cells. It doesn't appear that your more flexible discharge load will improve the charger for what I want it for, although I understand your need and that it will likely improve the charger design so I'll benefit anyway.

Cheers... Russ

[Mark]

No worries.

This is something that is likely to appeal to only a small subset of users with this charger, so that's why I'm posting this up to see just how small it is!

It's one of those things that if you don't already want it, it's probably not something that you would be likely to need.

[InHisName]

Yes, I find having a constant current load would be useful for my tests at home. Not sure how much more I can stand to 'pour' into my costs. I have already put in > 2x what I spent on a C9000.

By the way, is there a LOW constant current (no pulsing) choice ? i.e. 0.1C speeds.
Be nice if it was no higher than 100ma so I could diddle with lots of old junk AAAcells. I've even got 3 duraloops moved into that category, now. Those gold magnets seem so simple and elegant solution to charging AAAs.

[Mark]

At this stage, I'm only really guessing on a final parts cost, but I would estimate that it would be in the region of about $20 including the PCB. Postage would have to be added on top though and if you were wanting a fully assembled version, I'd have to charge a bit for my time as well.

You can have the charger discharge at low rates such as 100 mA already - if you use the configuration utility, you can set the discharge rate in there - make sure that you set the discharge mode to "Internal" as well.

Keep in mind that this will use PWM with the internal load resistor - this should still be close enough to being a constant current especially at low rates that it shouldn't matter too much. The problem with the C9000 is that even when you set it to discharge at 100 mA it's really still discharging at 1000 mA with a very slow PWM frequency - if a cell can't handle a 1 amp discharge rate without falling below 0.9V under load, you can't discharge it on the C9000. That shouldn't be a problem when discharging with the UltraSmartCharger since it uses a much higher frequency for the PWM. In the end, the only real difference is that I would expect that the PWM discharge on the UltraSmartCharger will cause more heat generation in the cell under test - at low rates though, I don't think it would be significant.

With the next firmware version, I'm making some changes so that it should be able to discharge into both the internal load and an external load at the same time and still use PWM to keep the current essentially constant - by my estimates, you should be able to get a constant current discharge up to about 600 mA with an appropriately sized external load resistor.

[Mark]

I've finished the first prototype:





It's a bit crude, but it works!

In initial testing, I've found that the current isn't held quite as constant as I would have liked - I suspect that the values of the wire wound resistors are drifting as they warm up.

I've made some modifications to the firmware so that it can now use the internal discharge resistor to compensate for this - for a 1 amp discharge, I've set the external load to about 950 mA and the other 50 mA is handled with PWM through the internal resistors. This is now allowing it to maintain within +/- 1 mA of the current that I want. This also means that I don't have to be too precise when setting the trim pot.

I'm planning to add a 3 position switch and another couple of trim pots so that I can easily switch currents without having to fiddle with the trim pot, but I'm still waiting for the PCB mount switches to arrive.

In the mean time, I've now got it running long term cycle testing to compare an Eneloop XX cell to a Turnigy 2400 mAh cell to see how their longevity compares.