This is not the version I am working on right now. However if things keep going well on Kickstarter then I will hopefully be able to make this a reality. The 4 battery version will be able to handle both AA and AAA sized NiMh rechargeable batteries. It will also be able to handle a wider range of charge and discharge currents.
It's up on Kickstarter! You can check it out here: Open Source NiMh Battery Charger/Analyzer
If you came to my blog from Kickstarter and you are bored, you could click on the link above and just keep going back and forth between my blog and Kickstarter.
If you watch the video, make sure you watch until the very end.
For those of you who like Raw Data, here is a single charge followed by a discharge that I did last night. I exported it as Tab delimited, so just pull it into your favorite spreadsheet program to check it out, make plots from it, what ever you like.
As a result of my extensive testing I have now cycled a large set of rechargeable batteries, varying in age, many many times. In the process I have collected a lot of data and have come across my fair share of anomalies. Here are a couple examples with accompanying explanations.
A sudden change in the temperature plot that goes up in the winter and down in the summer?
I remember the very first time I saw this (my design was still young and untested, so it was easy to question the electronics first) I thought it was a blip in power or an instability in the temperature sensor. The way I finally verified that it was in fact the AC (actually at the time it was winter, so it was the heater) turning on and off, I ran the charger through several cycles while inside a food cooler with a blanket over it, just in case. I looked through the data and... no blips!
This next one puzzled me for a while, a sudden fluctuation in the voltage while charging?
It was not until I started adding a short discharge every so often between charging, that I came to a solution... a Bubble in the battery. When the bubble forms it effectively increases the resistance of the cell. Charging with a constant current means that the current delivered to the battery does not change even as the voltage or resistance does. An increase in resistance while holding the current the same results in an increase in voltage (the upper blue line is the voltage while current is being applied). Now this bubble also effects the ability of the battery to deliver as much voltage into a load. The orangish bottom line is the voltage across the battery during the short drain. So Voltage up while Charging and Down while Discharging means a bubble in the battery!
Work has been keeping me very busy lately and away from my battery charger, so I appreciate the patience of all those who have been following this project. I recently looked through past data I had collected and found a charge plot from the same battery on the old 1.8VREF charger (although many cycles ago) to compare with the new 18-bit charger. It doesn't necessarily look like much of a difference, but it is tremendous. Now that I have it all working and have ran it through a few cycles, I am consistently getting resolution around 0.016mV. Readings from the temperature sensor have also improved (compare the temperature plot below with the one from my very first post), but I am past the noise floor of the temperature sensor (by a lot) and therefore can't improve it much more.
Old 10-bit Charger with 1.8V external voltage reference ( about 1.8mV resolution )
Charge plot from 18-bit charger ( 0.016mV effective resolution )
Temperature plot from 18-bit charger ( ≅ 0.1°C usable resolution )
I occasionally stop by Harbor Freight on my way home from work. I like to use their rechargeable batteries for testing just in case something goes wrong. Then if something does go wrong, at least it is not one of my eneloops. Well today I decided to buy two packs of their D batteries, which come two in a pack, for a total of 4 batteries. When I got home my curiosity got the best of me and I decided to open one of them. I had seen/read about the Energizer D and knew eneloop uses AAA's and AA's in their C's and D's (as seen in the pictures below).
So I was not surprised to find an inside similar to the energizer D.
I was surprised however to see it held in place by rosin (not even a plastic holder like in the energizer). I decided I didn't actually need the D size so why not open another one and have a pair of the smaller inside batteries. Another Surprise!!
This next one did not match the first. It had 4 AAA's in parallel (noticed in the second image above how close the metal tab is to shorting to the outer casing! It is less than a mm. Not very safe! And considering they are only held in place with rosin, which softens when it heats up). I figured they must have switched the design at some point and I must have ended up with 2 D's (one pack) of this style and 2 D's of the other. I decided to open them all up. Yet another Surprise!!!
3 AA's in parallel. Talk about inconsistency! I only bought 4 batteries and yet 3 of them were not the same. On top of that, they are all supposed to be D's. Another thing is the 2500mAh advertised capacity. It makes sense for the first one I opened (and maybe even for the 4 AAA's because they would show a decreased capacity under the loads normally seen by D batteries) but the 3 AA's would actually have a capacity much greater than the advertised 2500mAh. Use all these in series and one of them is bound to be driven into reversal.
Here are two zoomed in crops from a discharge cycle so you can compare between the old 10-bit and the new 18-bit.
After a lot of frustration the new 18-bit ADC is finally working! Right now I am consistently showing about 16 bits of actual resolution without any software filtering. That equates to a resolution of about .03mV. This is really exciting and much better than the 1.8mV resolution of before. I will post pictures and data when I have more time.