Homemade Battery Recharger w/ Regulator -- HELP!

[swimfan24k]

Hey, I was hoping I could get some help with a project that I hope to finish in the next couple of months... I have searched the depths of the internet, and have not been able to find any help on this at all!

Simply, I am looking to build a recharging circuit for 3x1.2V NiMh batteries (3.6V total) which are fueled by a single 5V solar cell (700mA output max).

The one function on this recharger that I want, and am not sure if this is even possible, is to install a regulator which would begin charging the batteries once they got to a certain low voltage (i.e. 2V), and then would stop charging the batteries once they got to a certain high voltage (i.e. 4V). Is there a name for this type of IC? or could anyone give me some guidance as to how this could be done?

Would be greatly appreciated!!

S

 

[crutschow]

That would be a comparator.

But you can't reliably determine when a NiMH battery is charged by its voltage. You can only do that with lead-acid batteries. To reliably charge NiMH batteries requires a dedicated charger circuit the monitors the voltage change with time and also possibly the battery's temperature. Several companies, such as Linear Technology and Maxim, make such chips. Just Google "nimh charger ic".

 

[audioguru]

Why do you simply guess about voltages when manufacturers and users of Ni-MH cells talk about them on the web?

One cell is 1.4V to 1.5V when fully charged and fresh out of the charger. Its voltage quickly drops to about 1.2V or 1.25V for most of its discharge then slowly drops to 1.0V or 1.1V when it is almost dead. If a battery of a few cells are in series and you discharge it past an average of 1.0V or 1.1V per cell then the weakest cell will become completely discharged and charged backwards by the other still discharging cells which destroys it.

 

[swimfan24k]

I am not simply guessing. The minimum (2V) is the constraint of the device that the batteries are being fed to, where it wont work anymore because of insufficient voltage. As for the 4V, this was an estimated fully charged value for 3 batteries wired in series.

 

[bountyhunter]

When we were designing IC's for gates Energy and they first brough out the NI-MH cells (back around 1992), they gave me some things called tafel curves for their cells. It showed the full charge voltage values for the cells as a function of temperature, when being charged by small current values. If you charge the batteries at a very low current (like <C/10) to prevent cell heating, and they are at a temp of about 20 - 25C, you can use a final charge cell voltage of about 1.35V/cell and they won't overcharge. At that voltage, the final charging current is very tiny.

I'll try to find that old gates Energy book and post the curves. Using CC/CV chargers for NI-CD and NI-MH isn't optimal, but you can do it if the cell temp is fairly constant.

 

[audioguru]

Each manufacturer uses different chermicals and amounts in Ni-MH battery cells so their charging voltages are different.
Energizer's (made in Japan maybe by Sanyo?) voltage is almost 1.35V per cell at 0.1C when fully charged but another Japanese manufacturer (Panasonic maybe?) has a much higher voltage of 1.43V.
German manufacturers are probably also different.

Try charging Ni-Cad and Ni-MH cells in the same charger at the same time. The Ni-MH cells get warm but the Ni-Cad cells get cooler than the surroundings before they are fully charged.

 

[ronv]

If everything happens at room temperature the easiest way might be to detect the temperature rise of the batteries when the charge is complete. Then you could switch to a safe tricke charge to equalize them. You can use the spare comparators to sense the low voltage to disconnect the load when the voltage goes to low..

 

[bountyhunter]

If everything happens at room temperature the easiest way might be to detect the temperature rise of the batteries when the charge is complete.

That works for NI-CD but not for NI-MH. The internal charge reaction for NI-CD is endothermic which means the cell gets slightly cooler as it charges, until it is fully charged then it starts to get warmer.

The NI-MH charge reaction is exothermic, so it gets warmer as it charges. When fully charged, it also gets warmer but it gets warmer faster so the rate of temp rise is sometimes used to detect end of charge on NI-MH.

 

[bountyhunter]

Each manufacturer uses different chermicals and amounts in Ni-MH battery cells so their charging voltages are different.
Energizer's (made in Japan maybe by Sanyo?) voltage is almost 1.35V per cell at 0.1C when fully charged but another Japanese manufacturer (Panasonic maybe?) has a much higher voltage of 1.43V.
German manufacturers are probably also different.

I have a bunch of the high capacity NI-MH "9V" cells made in China and I use a charger that holds a voltage of 1.4V/cell across the battery continuously. I measured the amount of trickle current going into the cell at that voltage on a fully charged battery and it was less than a micro Amp.

 

[ronv]

That works for NI-CD but not for NI-MH. The internal charge reaction for NI-CD is endothermic which means the cell gets slightly cooler as it charges, until it is fully charged then it starts to get warmer.

I think you meant it works with NI-MH but not with NI-CD?
He specified NI-MH

 

[bountyhunter]

I think you meant it works with NI-MH but not with NI-CD?
He specified NI-MH

I posted it correctly. I was pointing out that simply measuring temp rise is a good way to terminate charge with a NI-CD charger, but it will NOT work with NI-MH batteries. It's also why sometimes people burn up their NI-MH batteries by using a NI-CD charger. It doesn't detect end of charge and keeps cooking them.

Here is actual data on a NI-CD cell showing the temp profile:

 

[ronv]

Take a peek at figure 2. Coincides almost perfect with delta V.



https://www.electro-tech-online.com/custompdfs/2012/01/dn380f.pdf

 

[ronv]

I missed the fact that you want to use a solar panel to charge them. In that case neither method will work because both -delta v or temperture depend on a fairly constant charge. To allow it to stop for clouds, etc. would mess it up.

 

[bountyhunter]

This underlines what I posted earlier: the NI-MH gets hot all the time it charges, it just gets hot a little faster when it's fully charged. You have to detect that change in temp rise rate to get the correct end of charge point.

Notice in the curves: if you wait to shut off the charge current until the cell voltage dips, it will have been significantly overcharged with the cell voltage near 40C. That shortens battery life. That's why the preferred EOC detect is to trigger on the temp rise rate increase. That occurs ar about 34C cell temp.

 

[ronv]

I posted it correctly. I was pointing out that simply measuring temp rise is a good way to terminate charge with a NI-CD charger, but it will NOT work with NI-MH batteries. It's also why sometimes people burn up their NI-MH batteries by using a NI-CD charger. It doesn't detect end of charge and keeps cooking them.
:

I think the reason NICD chargers toast NMHD is because most use minus delta V. Minus delta V can be very subtle in NMHD so most NMHD chargers use zero delta V to terminate the charge. Temperature will work for both.

 

[MrAl]

Hi,

If you charge with enough current you can watch for temperature rise. That is, temperature rise above ambient. This makes it a little easier than trying to detect a delta temperature (rise per unit time). The plots below show a couple of bad cells being charged, but the temperature profile still matches a good cell being charged in that as it nears full charge the temperature reaches near 20 degrees above ambient for a AA size cell and 1 amp charge current.
This might not work though for a solar cell who's output varies a lot over time.

 

[ronv]

!00% agree. I'm not sure I know of a safe way to quick charge them from a solar panel. What do you think of the constant 1.5 volts? From your curves it looks like it would leave it less than fully charged. I've never tried it. Some intersting wiggles in the voltage curves.

 

[bountyhunter]

Hi,

If you charge with enough current you can watch for temperature rise. That is, temperature rise above ambient. .

You missed the point of the info I posted. NI-MH has an exothermic charge reaction so it gets hot by charging. It will show a rise above ambient long before it is fulluy charged.

 

[ronv]

Never said you were done when the temperature tripped the fast charge. Maybe I used a poor choice of words when I said back in the original post #7 that you could then switch to a safe trickle charge to equalize them. I should have said finish them up. The fact remains the picture you posted shows a normal temination on temperature followed by a .1C top off of 110 mah. (5%) Seems like if you were clever you could do the whole thing based only on temperature with a couple of comparators and transistors. Or you could just buy a chip.
Maybe I'll try it when I get some more time.

 

[bountyhunter]

"The fact remains the picture you posted shows a normal temination on temperature"

I guess I should have been clearer: the figure 2 that I posted above is a good example of why simple temp measurement is not good for NI-MH: the cell temp rises to 41C. It's from a Linear-Tech data sheet pimping one of their products. If you don't mind running the cells above 40C at end of charge it's great.... now ask a battery maker how long the cells last if you cook them that hot on every charge cycle.

The point is that even as far back as 1992 when gates Energy invented the first NI-MH cell, they were warning charger makers not to try to use simple temp rise to terminate fast charge (as had been industry standard practice for NI-CD) and the curves show why. The best method is to measure rate of temperature rise and then terminate on the inflection point wher the rate suddenly increases. That saves a lot of overcharging and temperature stress as it occurs at a cell temp of about 34C, not 40C. It also differentiates between a "smart charger" and a dumb one.

However, it's a lot harder to make a chip that can do that unless you have a uC that can datalog readings at fixed time intervals and calculate rate of change. Obviously, a linear IC house like Linear tech isn't building any such thing.

The Linear Tech article, which was incidentally written by somebody I shared an office with for four years, lumps NI-CD and NI-MH together to try to make you think the chargers for the former are fine to use with the latter. God help anybody who believes that, I hope their house is fireproof. Anyway, they think their part is the cat's pajamas but >40C is pretty hot to run a battery when you charge it.