Clean Solar Project Suggestions Please

Thread Starter

AceTK

Joined May 15, 2012
13
Greetings all, this is my first post here.
I'm been learning electronics for years now and would like to work up towards being confident in power generation, regulation and storage. At present i'm working on what would seem to be a really simple project but i've encountered some niggles which I hope I can resolve after posting here lol. Basically the system is just a 10w 18v multicrystaline solar panel with a battery pack and a dc dc converter integrated so that I can have power on the move. At present its very basic and I simply have the output from the solar panel outputting directly to a 9.6v 1000mAh nicad battery pack via a 2N4001 diode. The dc dc output converter is strapped directly to the battery terminals after the diode.

Some of the devices I use are sensitive to undervoltage and would prefer no volts to low volts. Without resorting to LM IC modules what is a simple voltage trigger circuit? Ive tried using a 12v zener diode in reverse so that if the volts are over 12v the 2N3906 PNP transistor's base is saturated letting the main 12v pass to next stage but for some reason it just stays stuck on and the volts seemed to level out at 7v. This is on the output of the dc dc on the final stage. Ive also noticed that the dc dc outputs wrong volts (always lower not higher) when amps and volts jump around and down lots as clouds cover sun.

Really id like some better batteries to act as a larger buffer incase the sun goes in. Now at the moment ive been using nicad but to be honest they are not really the best solution and can suffer from erratic discharges and other issues. I wondered about lithium ion but then they can explode so then I thought perhaps lithium phosphate for safety. But if I buy protected cells can I trust the overdischarge and overcharge circuits? Should I have a separate charge protection circuit? If the batteries have protection PCB could I not just have the cells hooked up like the nicads? Or I wondered if perhaps I could use some of the LM324 quad op amps I have lying around to rig a comparator circuit to monitor the voltage and switch charger on and off. Im not sure how I can have the charger integrated on a system that is linked to the solar output and charging and discharging randomly as use fluxuates. Other systems ive seen seem to disconnect the source and drain to check battery voltage using a microcontroller. I have an arduino which im also learning to use if anyone thinks perhaps that is the way to go. As for the capacity of the cells im thinking 3.7v 2400mAh 18650 seems he most common place,I saw some real nice Headway 10000mAh 3.2v 38120 cells which would seem to have great capacity but I cant find anywhere other than alibaba to buy them in quantities less than 100x. I figure having four 18650's in series so that I get 14.4v with 18v solar panel to charge. Or I could have a dc dc buck converter to take any solar output and step it down to the exact charge voltage of four cells half series half parallel (8.4v) to give better overhead with solar input. I also wondered about setting the charge system to only charge batteries 80% based on max charge volts to avoid complicated circuitry and excessive battery wear. Anyhoo that's where im up to at the moment.

Thoughts appreciated, kind thanks. :)
 

Thread Starter

AceTK

Joined May 15, 2012
13
Erm that's not a helpful link I already know how to work out battery life, was my text too long to read? Is anyone else able to..... :p
 

wayneh

Joined Sep 9, 2010
17,496
...was my text too long to read?
A bit, yes. IMHO, you should stick to nicad chemistry because it's the most forgiving of non-ideal charging, in particular overcharging. Then, you should employ a charge regulator IC. They do a good job and are reliable as far as I know. I can't recommend any in particular, but I'm sure someone here can.

Under normal full-charging, there's no reason to buck down to battery voltage. The batteries will do that. You get ideal power transfer when the panel is at 80-90% of its open circuit voltage, and when the battery is half charged. What you DO need is to limit current so that you don't damage the battery.
 

#12

Joined Nov 30, 2010
18,224
was my text too long to read?
Uhh...yes.
My rule of thumb is: if you can't define the problem in less than 500 words, you haven't figured out what you want. Another point of view might be, "Why use up the time to read all that just to find out if I can do anything helpful?" Even helpful people can be lazy!
 
Use the Arduino a-d to check in on the battery voltage every so often. If high voltage for the LifePo4 cells of between 3.6-3.7V, then turn on a digital output to short the solar panel before the diode, with a NFet. If the battery is low, say 2.5V? then disconnect the ground to the dc-dc circuitry with another NFet. A comparator chip could do this, but is nice having the smarts of the micro to check in say every ten minutes and go back to sleep.

Haven't messed with lico batteries, and would probably go with a charger IC if I did.
 

THE_RB

Joined Feb 11, 2008
5,438
Actually I think the worst posts of people asking for help is when they post very little info on their "secret" invention and then want someone else to design it all for them. Thank you for taking the time to post a comprehensive description of your project. :)

I don't think lithium batteries with inbuilt smart controllers will be a good idea, they tend to charge in a "cycle" which may not work well with your portable solar periods of being in the sun.

I think you are on the right track with a NiCd battery, and just put whatever solar power you can into it. As NiCds can reduce capacity with age and bad usage you might want to invest in new good quality cells, maybe larger ones too. Then keep them charged as they degrade if left in a discharged state.

NiCds are simple enough, and you are using a solar panel that is basically a constant current source, so the only protection you need is something like your zener diode to limit peak charge voltage. Maybe something a bit more sophisticated like a comparator or opamp IC and a resistor load, to limit the max volts the battery will charge to.

However if you increase the NiCd size you may just be able to hook up the panel and rely on the fact the panel's max current of about 650mA will be low enough to be tolerated, as it is usually a rare case foe a solar powered battery device to have an absolutely full battery anyway.
 

Thread Starter

AceTK

Joined May 15, 2012
13
Actually I think the worst posts of people asking for help is when they post very little info on their "secret" invention and then want someone else to design it all for them. Thank you for taking the time to post a comprehensive description of your project. :)

I don't think lithium batteries with inbuilt smart controllers will be a good idea, they tend to charge in a "cycle" which may not work well with your portable solar periods of being in the sun.

I think you are on the right track with a NiCd battery, and just put whatever solar power you can into it. As NiCds can reduce capacity with age and bad usage you might want to invest in new good quality cells, maybe larger ones too. Then keep them charged as they degrade if left in a discharged state.

NiCds are simple enough, and you are using a solar panel that is basically a constant current source, so the only protection you need is something like your zener diode to limit peak charge voltage. Maybe something a bit more sophisticated like a comparator or opamp IC and a resistor load, to limit the max volts the battery will charge to.

However if you increase the NiCd size you may just be able to hook up the panel and rely on the fact the panel's max current of about 650mA will be low enough to be tolerated, as it is usually a rare case foe a solar powered battery device to have an absolutely full battery anyway.

Thank you, to be honest I did read the forum a bit first and noticed lots of short posts where other users asked for more info so I figured I should just go for it and explain where I'm up to fully.

With regards to the project it remains unchanged at this point and I have to agree with you on increasing the NiCd capacity so that charge current is not an issue and even perhaps adding extra cells in series so that the charging max voltage is within tolerance of solar output. Still though I can't help but wonder if id be better off with lithium cells simply from a no charge memory perspective but il roll with the NiCd's for the time being until I can make a functioning circuit that protects from overcharge and overdischarge. Do you think the LM324 quad op-amp is suitable for this project? The basic functionality would appear so but I'm not sure if I'm overlooking something like reliability or nuances in this application.


A bit, yes. IMHO, you should stick to nicad chemistry because it's the most forgiving of non-ideal charging, in particular overcharging. Then, you should employ a charge regulator IC. They do a good job and are reliable as far as I know. I can't recommend any in particular, but I'm sure someone here can.

Under normal full-charging, there's no reason to buck down to battery voltage. The batteries will do that. You get ideal power transfer when the panel is at 80-90% of its open circuit voltage, and when the battery is half charged. What you DO need is to limit current so that you don't damage the battery.

I have noticed the battery buck effect when charging and see the potential to save me a buck converter on the solar output stage for sure. As with the reply above I'm hoping adding extra cells will allow me a large enough overhead so that overcurrent does not become a problem although it would be nice to know I have that piece of mind should I put it outside and forget about it. Perhaps though if I just had the solar output set to disconnect once the battery voltage reaches max set charge voltage. Thanks.


Use the Arduino a-d to check in on the battery voltage every so often. If high voltage for the LifePo4 cells of between 3.6-3.7V, then turn on a digital output to short the solar panel before the diode, with a NFet. If the battery is low, say 2.5V? then disconnect the ground to the dc-dc circuitry with another NFet. A comparator chip could do this, but is nice having the smarts of the micro to check in say every ten minutes and go back to sleep.

Haven't messed with lico batteries, and would probably go with a charger IC if I did.

Yeah that seems about what I imagined. Is it ok to short the solar panel out? Would simply disconnecting the positive/negative be better? Or perhaps I could have it switch to a dump load? Also when I have the arduino checking the voltage should I disconnect the solar panel and output while taking readings or is it ok on the fly? Also is every 10 minutes too long between arduino readings? Would something like every 60 seconds be better? Thanks
 

wayneh

Joined Sep 9, 2010
17,496
Perhaps though if I just had the solar output set to disconnect once the battery voltage reaches max set charge voltage.
A solar light I took apart used a dump circuit to begin diverting current thru a resistor and transistor once the SLA battery voltage hit a limit. It was real simple (and posted here), using just a zener and resistors to set the threshold voltage. So, if your panel is capable of say, double the current the battery can tolerate at full charge, you could use the dump circuit to divert the excess. It would not work without heat sinks and such if the panel was huge, but it wouldn't be too tough to dump an amp or two.
 
Yeah that seems about what I imagined. Is it ok to short the solar panel out? Would simply disconnecting the positive/negative be better? Or perhaps I could have it switch to a dump load? Also when I have the arduino checking the voltage should I disconnect the solar panel and output while taking readings or is it ok on the fly? Also is every 10 minutes too long between arduino readings? Would something like every 60 seconds be better? Thanks
Shorting the panel I believe is O.K., but you could use a Pfet switch, still need the blocking diode though because of the usual internal reverse bias body diode. A dump load would be fine too, like some leds or something else useful?

Ten minutes might be a little long for your size panel. Don't like zeners so much, because it is higher bias current than either the comparator or the microcontroller method. I would take the readings on the fly. There is always going to be a bounce back when a load is disconnected, account for it.
 

Thread Starter

AceTK

Joined May 15, 2012
13
A solar light I took apart used a dump circuit to begin diverting current thru a resistor and transistor once the SLA battery voltage hit a limit. It was real simple (and posted here), using just a zener and resistors to set the threshold voltage. So, if your panel is capable of say, double the current the battery can tolerate at full charge, you could use the dump circuit to divert the excess. It would not work without heat sinks and such if the panel was huge, but it wouldn't be too tough to dump an amp or two.
Ah interesting. So I'm guessing you just have the zener diode rated at the max charge voltage bridging the battery charge input connections in reverse so that any excess voltage overflows into a series of resistors or any resistive load like say an 18v halogen bulb to waste the excess? Where does the transistor come into play though perhaps I missed something. Do you have the link to that circuit also? Thanks
 

wayneh

Joined Sep 9, 2010
17,496
The search function seems to be down at the moment. Just do an advanced search of this forum, on my screen name and "solar light". I'm sure you'll find it. It's on the right-hand side of the schematic, probably the last schematic in the thread.

Here it is. Be sure to read the post following with a correction.

The zener starts to conduct at 5V, I think it was, thus starting to open the dump transistor.
 
Last edited:

Thread Starter

AceTK

Joined May 15, 2012
13
Thanks that's much better I see now. The zener biases the transistor to allow the dump. The resistive divide is there to make the volts equal zero by calculating the ohms back to ground and/or perhaps draw some current to make that n-channel bipolar transistor's base conduct. Would perhaps a n-channel MOSFET would be better in place of the bipolar in order to have base/gate isolation? Also could I then scrap the resistive divide and plug the output from the zener direct to the n-channel MOSFET gate?
 

wayneh

Joined Sep 9, 2010
17,496
No, you'd need a much higher voltage to turn on a MOSFET. The transistor starts to open up around 0.6V. I believe that is reached at about 1mA across the zener. Base current isn't a problem since current is exactly what you're trying to get rid of.
 
Top