Solar Powered Low Power Embedded Device

[Kalcifer]

It’s been surprisingly difficult to find specific information on charging a NiMH. I want to SLOWLY (30uA or so, which is less than C/1000) trickle charge a NiMH to keep it charged. The idea is to keep it on a solar panel feeding it this approximate 30uA current indefinitely to power an embedded device.

What I don’t know is if this small current will damage the cell in some way. I’ve conducted some tests and at this current the cell does indeed charge, about 1mV/min increase in cell voltage. But what happens once the cell finally reaches its fully charged state?

When I’m trickle charging it, should I maintain a constant float voltage so that I don’t overcharge it? I’ve heard that NiMH don’t have a specific float voltage so maybe this isn’t possible? If not then how do you prevent the cell from overcharging? Will the cell be damaged even at this low charging current? I can’t imagine that 30uA will heat up the cell enough to damage it.

[bdunham7]

It's actually way too low and probably won't keep the battery at full charge due to self-discharge. 

https://www.vencon.com/charging-nicd-nimh-batteries/

[Kalcifer]

It's actually way too low and probably won't keep the battery at full charge due to self-discharge. 

https://www.vencon.com/charging-nicd-nimh-batteries/

Wouldn't that entirely depend on the quality of the battery though? I did a quick test like said with a NiMH battery connected to my solar charging circuit, and with a charge current of around 30uA, the voltage on the cell increased at a rate of roughly 1mV per minute which tells me that its charging just fine despite the self discharge rate of the battery.

[bdunham7]

It's actually way too low and probably won't keep the battery at full charge due to self-discharge. 

https://www.vencon.com/charging-nicd-nimh-batteries/

Wouldn't that entirely depend on the quality of the battery though? I did a quick test like said with a NiMH battery connected to my solar charging circuit, and with a charge current of around 30uA, the voltage on the cell increased at a rate of roughly 1mV per minute which tells me that its charging just fine despite the self discharge rate of the battery.

Yes, the self discharge will depend greatly on the characteristics of the battery as well as the state of charge.  Self-discharge is less at lower charge states, so you will charge the battery up to the point (eventually) where self-discharge equals your charge current.  What were the actual voltages involved?  And what type/size of battery do you have?

[Kalcifer]

It's actually way too low and probably won't keep the battery at full charge due to self-discharge. 

https://www.vencon.com/charging-nicd-nimh-batteries/

Wouldn't that entirely depend on the quality of the battery though? I did a quick test like said with a NiMH battery connected to my solar charging circuit, and with a charge current of around 30uA, the voltage on the cell increased at a rate of roughly 1mV per minute which tells me that its charging just fine despite the self discharge rate of the battery.

Yes, the self discharge will depend greatly on the characteristics of the battery as well as the state of charge.  Self-discharge is less at lower charge states, so you will charge the battery up to the point (eventually) where self-discharge equals your charge current.  What were the actual voltages involved?  And what type/size of battery do you have?

Yeah I figured as much. It'd go until it reaches an equilibrium point. but for me that's honestly no big deal. For my application, the density of charge doesn't necessarily matter. I just want to be confident that I will reliably have some charge stored in the battery that will be able to account for off times with the solar panel (night). All I'm concerned about is the faint possibility that I DO generate enough power to be able to overcharge the battery. I know that the current will still only AT MOST be a few hundred microamps, but that could still offset the self discharge rate.

As for the voltages involved, the VOC of the solar panel is somewhere around 6-7V or so; but when the solar panel is connected to the battery for charging through a series diode, the voltage of the panel drops to that of the battery plus the voltage drop of the diode. As for the size of the battery I'm currently testing, its a 1000mAh.

P.S. Would I be better off using a NiCd? I've heard that they develop a memory over time and are supposed to be completely discharged. In my project the battery would never be discharged, so I wonder if this would end up being a problem.

[bin_liu]

The internal will start to fever and consume energy, the phenomenon is that the voltage across the battery drops, this is -△V.
According to the temperature and -△V to determine whether the NiMH battery is fully charged.
However, these two points will not be obvious if the current is too small. SO, the electric razor we use supports long-term charging.

[Conrad Hoffman]

Self-discharge depends on the battery. Eneloops, for example, are way under 1%/day. NiMH isn't a good technology for trickle/float charging. NiCads are better but I have a lot of test equipment that used them in that manner, and they didn't hold up well. Lead-acid is good, but I notice that gel-cells used for computer backup supplies seem to go bad on a regular basis. The gel-cells used in security systems get replaced on a regular schedule because they don't last forever. Can you get the job done with a super capacitor? How about a primary battery that you change every few years, because you'll probably be changing your rechargables that often anyway.

[bdunham7]

P.S. Would I be better off using a NiCd? I've heard that they develop a memory over time and are supposed to be completely discharged. In my project the battery would never be discharged, so I wonder if this would end up being a problem.

No, NiCad is toxic and should be banned, IMO.  They were commonly used in crude float charge systems like this because they can be float charged at C/10.  However, NiMH will usually tolerate C/30 if not C/20 and you can get 2X or more the capacity in the same size cell--thus they can usually be a perfectly good drop-in.  In your case, NiMH is clearly better. 

[magic]

There are NiMH cells designed for RTC backup applications and they are rated for durability under continuous trickle charge. I have seen a cell like that in some appliance and read the datasheet.
IIRC special electrochemical tricks went into them to reduce deterioration due to overcharging.

One thing about self discharge: I'm not entirely sure, but it may likely depend on temperature.

[Kalcifer]

However, these two points will not be obvious if the current is too small. SO, the electric razor we use supports long-term charging.

electric razor?

[Kalcifer]

Can you get the job done with a super capacitor? How about a primary battery that you change every few years, because you'll probably be changing your rechargables that often anyway.

I've been strongly considering and looking into just using a supercapacitor. It seems to check the box for reliability and safety. It's easy to predict, and simple to deal with. I'd definitely have to regulate the voltage somehow but that's no biggie. The only trickier part is finding a way to do it without large quiescent currents sucking up useful power.

As for the primary battery that i just recharge every few years, I was initially looking into using i lithium ion cell like an 18650-30q as the main power source. by my math, it'd last just over 2 years in an ideal scenario, but theres a few caveats for me.
1. I'm too lazy to charge a battery ;P /s   
2. Since I'm strapping this to the side of my house, I'm not super confident in the safety of a circuit using a lithium ion without an proper battery management system. I don't want to burn my house down.
3. I really love the idea of self sufficiency and autonomy of solar power. It just feels so simple and elegant having it all run on its own.

[NiHaoMike]

I've been strongly considering and looking into just using a supercapacitor. It seems to check the box for reliability and safety. It's easy to predict, and simple to deal with. I'd definitely have to regulate the voltage somehow but that's no biggie. The only trickier part is finding a way to do it without large quiescent currents sucking up useful power.

There are micropower voltage detectors that will work, or if the load includes a microcontroller, you can have it check the voltage every time it wakes up and alter its power usage in some way, for example by turning on some sort of load (e.g. a LED) or changing its sleep time.

[Kalcifer]

I've been strongly considering and looking into just using a supercapacitor. It seems to check the box for reliability and safety. It's easy to predict, and simple to deal with. I'd definitely have to regulate the voltage somehow but that's no biggie. The only trickier part is finding a way to do it without large quiescent currents sucking up useful power.

There are micropower voltage detectors that will work, or if the load includes a microcontroller, you can have it check the voltage every time it wakes up and alter its power usage in some way, for example by turning on some sort of load (e.g. a LED) or changing its sleep time.

The load is indeed a microcontroller, and that's a good suggestion. I will keep it in mind.
I was thinking of just putting a zener diode more or less in parallel with the capacitor that has a really low reverse leakage current. Something like the BZX55C4V7-TR. Passively accomplish the same idea.

Side note, what do you mean by a voltage detector? Is it able to do something when it detects a voltage? Could you give an example of said device?

[NiHaoMike]

Side note, what do you mean by a voltage detector? Is it able to do something when it detects a voltage? Could you give an example of said device?

Here's one example:
https://www.maximintegrated.com/en/products/power/supervisors-voltage-monitors-sequencers/MAX6375.html
It's a very low power comparator with its threshold set at the factory. When its supply voltage goes above or below that threshold, its output changes states accordingly. The biggest catch is that if you order some, review the part number very carefully to make sure you get the correct voltage!

Such chips are often used in simple battery balancing circuits, one per cell. When the voltage goes to the balancing point, it turns on a load to drain a little charge. In my 4S, 82Ah LiFePO4 pack, I added flashing RGB LEDs across the load resistors so they flash when the balancing circuit is active.

[Kalcifer]

Here's one example:
https://www.maximintegrated.com/en/products/power/supervisors-voltage-monitors-sequencers/MAX6375.html

Very cool! That will come in handy. For this case, what would be the pros and cons of that when you compare it to say just using a zener diode with its reverse leakage current being similar to the quiescent current of that device? The main pro I could see of using that chip is that it offers more customizability/functionality with how you want to handle the overvoltage event.

[NiHaoMike]

Zener diodes tend to be "soft" in their response, while a voltage detector "snaps" on and off with a bit of hysteresis to prevent excessive cycling.

All in all, for your application, having the microcontroller adapt its power usage would be the easiest and gives the possibility of adding functionality. For example, if it's reading the sensor, it could give readings more frequently

[magic]

LM385-1.2 will work as a 1.2V zener with <3µA leakage when it isn't clamping. Put a few in series for more voltage or use the 2.5V version with 10µA leakage (due to internal voltage divider).

I think it could even work to protect a NiMH cell from overcharge, while still allowing it to charge to a usable fraction of its full capacity.

[Kalcifer]

Can you get the job done with a super capacitor? How about a primary battery that you change every few years, because you'll probably be changing your rechargables that often anyway.

I've been strongly considering and looking into just using a supercapacitor. It seems to check the box for reliability and safety. It's easy to predict, and simple to deal with.

The one problem I'm seeing with this after looking into it more is that Supercapacitors appear to have a pretty large leakage current (relatively speaking).
If you look on page 10 of this datasheet, you see that for the 15F capacitor the self discharge current is around 60uA. Well my solar panel will typically only produce around 30uA-40uA, so that's a slight problem.

I need a supercapacitor on the order of 5 or so volts with a capacity >10F with a super small leakage current. Any suggestions?

EDIT: Upon further thought, I may just have to deal with a lower Capcitance. There’s a few capacitors out there with a lower leakage current, they just have a substantially lower capacitance, like 1-5F instead of 10F+.

[bdunham7]

Well my solar panel will typically only produce around 30uA-40uA, so that's a slight problem.

I didn't realize that was your power limit.  How do you have a solar panel that produces so little current? Even tiny LED-sized ones produce hundreds of times that amount.

[Kalcifer]

Well my solar panel will typically only produce around 30uA-40uA, so that's a slight problem.

I didn't realize that was your power limit.  How do you have a solar panel that produces so little current? Even tiny LED-sized ones produce hundreds of times that amount.

Its an amorphous silicon panel. The AM-5904. here is the datasheet.
It was my understanding that these amorphous silicon panels were optimal for low light conditions because they produce their open circuit voltage even at very low light levels, and it was just the current that changed.
If you are able to link a better panel I am all for it! I'm not attached to this specific panel at all, I just thought it was ideal for the scenario.

[bdunham7]

That datasheet shows a lot more current for that part number than what you are saying. 

[Kalcifer]

That datasheet shows a lot more current for that part number than what you are saying. 

As you can see here, the power and current values that you are noting for the AM-5904 are for outdoor use with a lux of 50k. But if you look at this, you see that for indoor (or low light) use with a lux of ~10, the current drops to what I expect.

[bdunham7]

OK, I didn't glean that you were using this indoors at such a low light level.  10 lux is pretty dim. However, the part number you gave is not one of the indoor models.  You might want to do some actual testing because I think your results might be significantly different than you expect.

[Kalcifer]

OK, I didn't glean that you were using this indoors at such a low light level.  10 lux is pretty dim. However, the part number you gave is not one of the indoor models.  You might want to do some actual testing because I think your results might be significantly different than you expect.

I am not using it indoors; however, the light levels that I will be expecting will be similar to that of an indoor device. This device will permanently be in some form of shade. I can do my best to orient the panel to get more light, but generally it will be in a shady area.

[bdunham7]

An experiment is certainly in order, as we're all guessing, but my guess is you will have a lot more current than you think with that panell.  Are you using more than one cell in series?  How many?  I suspect that the simplest solution is also going to be the best, and that is just your panel, up to 3 cells of at least 600mAh and a 1N4148 blocking diode.  If you don't have parts yet, a cheap experiment could be done with parts from those solar pathway lights the sell at the dollar store.