You have probably realized that charging NiMH is fairly difficult. Charging lithium batteries is very straightforward, so it is worth considering using a single lithium cell instead of 2xAA if it is possible.
Is this a circuit to charge batteries in a portable device, or are you just building a general purpose AA cell charger?
If it is a general purpose charger, then with two AA NiMH cells, you really need to monitor the dV/dT and temp rise of each cell separately - perhaps two separate charging circuits. Lots of cheap chargers just charge a few NiMH cells in series based on dV/dT only and that is a perfect way to damage NiMH cells. I have seen this type of charger often overcharge cells and NiMH is not tolerant to overcharging.
If I were building a charger like this, I probably would use a cheap buck converter to reduce the 5V 500mA USB down to 2V at over 1A, and then use this as the charging source with separate circuits for each cell. Both the IC's you have found are probably fine to base a charger on, but you have picked one of the hardest battery types to charge. There will be some head scratching to understand the IC documentation.
Neither dT/dt nor dV/dt are good cutoff criteria for NiMH or NiCd.
Detecting increasing dT/dt means you're already overcharging the cells.
Zero or negative dV/dt detection also places your batteries in the overcharge range.
The additional timers that your ICs have are just for insurance.
A proven criteria that works is d2V/dt2.
A forgotten, but extremely good charger IC for NiMH and NiCd batteries was the Temic U2402B, which I've used with great succes. Even with a 3C charge cycle, batteries would only get luke-warm.
Unfortunately, it disappeared in the Atmel/Vishay takeovers, and the group designing and supporting charger ICs was disbanded as well.
But it's operating principles align completely with the battery manufacturers recommendations, as well as the textbooks on battery charging.
Today, the functionality of the U2402B is normally firmware in a microcontroller with A/D, being part of the device where the battery is placed.
Links:
Textbook excerpt from Google books:
https://books.google.de/books?id=6CCOAwAAQBAJ&pg=PA112&lpg=PA112&dq=d2v/dt2+battery&source=bl&ots=p90Gc8Kus6&sig=V7x5kzveBz1b9r2wpYFGLTQidTA&hl=en&sa=X&ved=0ahUKEwifxPDCr5HVAhWGLVAKHTdKBLwQ6AEIRTAH#v=onepage&q=d2v%2Fdt2%20battery&f=false
U2402B datasheet:
https://cdn-reichelt.de/documents/datenblatt/A200/U2402B%23TEM.pdf
I hope this is an inspiration, at least it gives an idea on how to fast charge NiXX cells.
You mentioned fast charge. I wonder if speed of charging is not important, can I just make a simple and safe charger by slow charging it at C/10 and just terminate by time? Or perhaps drain the battery first.
I've read this is a pretty simple yet safe algorithm if the battery is to be charged overnight (as opposed to fast charging)?
You mentioned fast charge. I wonder if speed of charging is not important, can I just make a simple and safe charger by slow charging it at C/10 and just terminate by time? Or perhaps drain the battery first.
I've read this is a pretty simple yet safe algorithm if the battery is to be charged overnight (as opposed to fast charging)?You have to be careful with timed C/10 charging. If the battery is fully charged and you continue, it does increase temperature and pressure in the cells and probably shortens the life. You might find a lot of manufacturers do not actually state anywhere that it is OK.
It is not a bad idea using a micro for NiMH charging. It doesn't need a big program. This site contains a suggested protocol.
https://www.powerstream.com/NiMH.htm
If you added engineheat's suggested d2V/dt2 protocol, it would be even better. With two cells in series, you will have the situation of one cell indicating full charge while the other isn't so it will make the detection twice as hard. Probably want to have the thermistor positioned so it gets the average of the two cells. From my experience, the thermistor is still a great idea.
If you are charging the cells while there is load on the cells, you make voltage curve methods much harder. In this case, definitely have a thermistor or other kind of temperature sensor.
Overall, NiMh is "safer" than Li-ion in that it's not as explosive or dangerous when things go wrong?
I remember thinking that the U2402B-C was pretty neat but then it became unobtainable. Implementing the same thing with a microcontroller is straightforward if not as easy.Overall, NiMh is "safer" than Li-ion in that it's not as explosive or dangerous when things go wrong?
NiMH and NiCd batteries cannot fail as catastrophically but most rechargeable lithium failures are do to poor design. If you use a single cell and LiFePO4, then I think the ease of charging will make the lithium rechargeable a better choice *unless* you want to use standard NiMH for ease of user replacement. I would also consider multiple LiFePO4 cells in series with class-B analog charge balancing acceptable also at low power levels.
Even though NIMH can't fail catastrophically, it can still be troublesome if the user puts in an alkaline battery or Lithium battery (there are Li battery the same size as AA).
This is a device I eventually want to commercialize, and I feel it's better to keep it simple while starting out. It also brings down the cost. It's a toy device so I don't think people will mind. Some people actually prefer it since they can always swap in fresh batteries in an instant as opposed to waiting for the device to charge via an USB. Any thought?