Simple MPPT solar panel charge controllers

Hi pistnbroke,

All good points! And very helpful to me as I really don't know much about this part of the problem. Especially the exact values for floating etc.

I didn't discuss it above, but there are two potentiometers (roughly SE from the LM3914) that set the offset and span of the chip input, so you will be able to set it as you please when all is done. So, for now just take the values as tentative.

I did intentionally power the Flip Flop that is holding the memory of whether the device should be charging or floating from the battery so that the state memory would survive periods of no input. As described in the circuit description the system will come on and start charging (if there is an input) in either the Float or Charge mode until the first period when the enable goes low. Then the battery voltage will be sampled, and if it is in the range from 11 to 13 volts the Flip Flop will be set to the Charge state and the controller will try to achieve the full 15.6 volts. Of course the controller is trying to do the impossible for a low battery and will behave the same way, that is dump in all the power that is available, if the controller is set at either 15.6 volts or 13.2 volts. That's why I didn't bother putting circuitry in for lower than 11 volts.

Anyway, eventually the voltage will come up if the sun shines. From what I did read on this subject a small bit of gassing for a short time is a good thing as it stirs the electrolyte and keeps it from stratifying vertically with differing density layers. This is probably a non-issue on a boat where normal wave action will serve the same purpose, but then my understanding (or lack thereof) was that to get to a 100% charge the battery also apparently needs the brief high charging voltage as well. So, the higher charge voltage is set and held until reached (15.5 volts for now), then the next time the voltage is sampled the Flip Flop is re-set back to the float voltage and this voltage is held until the battery voltage drops below 13 volts again. The controller will not output any power at all during this period, because it "sees" a higher voltage on the battery than it is programmed to put out. Eventually the battery will self-discharge down to below 13.2 volts and the controller will turn on and commence to give it small kicks and maintain that voltage.

I was aware of the nasty tendency of an almost discharged battery to read a "good" 12.5 volts with no load no matter what, given enough time to settle out, and that is why I want to set the charge voltage so high. I'm assuming that that is sufficient voltage difference to get the battery to accept something near maximum current.

The design is limited by the fact that the comparators internal to the LM3904 step up by equal increments, but aside from that just about any offset and range could probably be designed or adjusted into the system. The points where the different states are picked off can also be chosen at will (the reset to charge points (diodes), the make no change points (no diodes) and the float point (diode). Also, the actual voltage that the controller attempts to put out can be anything, and if there was a compelling reason could also be more than two different voltages. I really haven't looked more than briefly at these issues and am completely open to suggestions! In fact I would love it if someone out there would do the research and find out how to optimize these variables. Unfortunately, I suspect they will be slightly different for different battery types, but that doesn't mean an individual couldn't program in what's best for his own system . . . if he knows what the parameters are.

BillyDoc

 

No it does not self dischage down(0.3%/day) to "below 13v"...slighty over 12.5 is the voltage of a fully charged battery that has stood for 24 hrs...
(battery test ..fully charge and stand the battery for 24hrs ..If the voltage is below 12.5v the battery is suspect )

Once you got it up to 15.5v then you dont want to let it drop all the way back to 13v you should keep it at 13.8 v ..thats floating ....alarm and stand by batteries sit at this for years..

Like you say battery types vary but some modern ones are so tightly packed with seperators ( to stop vibration destroying the plates) that there is no real exit between the plates for large volumes of gas so you should be sure your battery can go over 14.4v..

Its not bringing the battery to 15.5v thats the poblem but the time your 2A charger will take to do it and all the gassing involved ..maintanence issue .

If you restict yourself to 14.4 or 14.7 as a max and float at 13.8 with sun up cycles to 14.4 14.7 that would avoid many potential problems ...( jap cars using older battery technology go for 14.7v whilst european go for 14.4v)

 

Thanks pistnbroke, you sound like you know what you're about and I know I don't.

OK, so assuming a "marine" deep discharge flooded cell type of battery, instead of 15.5 volts, set the upper limit trip-into-float at 14.4 volts, and then float at 13.8 instead of 13.2. Is that about right? It's no problem if that's the best plan.

BillyDoc

 

go with the max the maker says 14.4 or 14.7 are common and 13.8 is good for floating . Something to bring it up to 14.4 at least once a week would be good as if its been sitting at 13.8 for 3 months it wont have the amount of stored charge that it might have and some of your savings in efficiency will be lost

 

Hi pistnbroke, sounds like a plan to me. The the voltages used are programmable (just pick a resistor) and could even be put on a switch if required. Same with the span and offset of the "voltmeter" chip. A seven day timer to force a charge cycle would be a bit more difficult to implement at this point, but would be easy if we want to go the extra step and add a microcontroller.

I hope to finish the last part of the circuit today, and then I will order the parts and build it. Then we'll see what we see.

Thanks for your help on this part!

BillyDoc

 

OK, a complete circuit, but not a final one. There were a couple of things in the buck/boost chip that I'm pretty sure I don't understand completely. At this point I really need to build it and put a scope on it to see what it does, so . . . that's what I'll do.

Meanwhile, here's what I have.

The MPPT and associated logic has already been discussed, so I'll leave that for now. I changed the voltage offset slightly based on pistnbroke's recommendations, and cleaned it up a bit but it is otherwise the same. The resistor values on the buck/boost feedback set 14.4 v and 13.8 v for charge and float.

The Buck Boost controller is a ***** to understand (or I'm too stupid) and there are a couple of things about it I am not at all clear about, namely the current feedback effect on the PWM, and the compensation circuit for the feedback signal. I hope to know more when I play with it.

There really isn't much for me to say about the Buck Boost circuit, because I mostly just copied it from the front page of the LM5118 data sheet. My only contribution was to add a means of digitally changing the feedback set point (charge or float) which I think is self explanatory . . . well it is when you remember that the transistors used have a very low ON resistance less than 0.1 ohm. Oh, oh, speaking of transistors I see I forgot to put labels on the Schottky diodes and the power MOSFETS! They are IRFR3607 and STPS20120D respectively (sorry, I'll correct with the other corrections yet unknown). Anyway, you should be able to see the basic logic of the entire system now. PLEASE let me know if you spot some dumb mistake, or even a smart one!

BillyDoc

 

Billy, you drive me crazy with these diagrams and explanation. But thanks....
Anyway, how much you think something like this cost in parts. And how would it compare with a more expensive unit like Outback. Of course it doesn't have the fancy displays but does it do the same you figure?

 

Hi mydauphin,

They drive me crazy too! But sometimes it's worth it.

I don't think this circuit will ever be as good as something like the Outback products, because they use a microprocessor to optimize the various variables continuously. I DO think it will get close, however, probably within a few percentage points in terms of efficiency. As for costs, the only way to find out is to go shopping and see. Most of the parts are fairly cheap, for example the LM5118 is $4.16, and I think it is the most expensive IC used. Here's a quick sampling of the major parts from digikey.com:

LM5118 4.16
STPS20120D 1.28
IRFR3607 1.91
IRLD024 1.40
AD8214 1.98
ICM7556 1.33
LM3914 2.60
LT3010-5 3.13
74AC74 0.55
1140-101K-RC 100 uH 9.56

Plus a board to solder it all on, a bunch of capacitors, resistors, etc., then a box to put it into . . . it all adds up! As a very rough guess I would say about $75 to build one.

Other major places to try are newark.com and mouser.com.

THIS IS JUST TO GIVE YOU AN IDEA OF THE COSTS INVOLVED WITH THIS APPROACH. PLEASE DON'T BUY ANYTHING YET!!!! GIVE ME A COUPLE OF WEEKS TO GET THE PARTS AND MAKE SURE THIS THING ACTUALLY WORKS!

BillyDoc

 

Billy, Do you have software to test circuit? I have a cousin in UCF studying EE and he has software to test circuits. Also how are you going to make circuit board. Are you going to breadboard prototype?

 

Hi mydauphin,

I will actually be using this controller to charge a battery in a cabin my wife and I own in the middle of 160 acres of forest. We go there every weekend, so the usual battery load is only for lights and fans in the summer and an electric furnace blower in winter, which doesn't take much.

So I will be building a "permanent" device by hand-wiring on a circuit board with soldered connections. I should get it done in a couple of weeks and I'll report back then (I just got a bunch of other work I have to do as well, however). I'll take some pictures.

Thanks for the software offer, but I've already tried to use that approach on the National Semi web site (lots of manufacturers have modeling capability on site) and the model they have for their LM5118 is limited to roughly a six amp output. I want to go to higher current pulses into the battery, but make them brief and intermittent, and this is not really how this chip was designed to be used. So, the modeling doesn't really provide the answers I need. I think my old-fashioned oscilloscope will though. The only reason I can see for the six amp limit is the gate charge into the MOSFET transistors being a limiting factor, but I have used much higher current transistors in my design with almost the same gate charge characteristic . . . so we'll see.

It also wouldn't be the first time I did something stupid, so we really aren't going to know until I build the thing and make any necessary corrections.

BillyDoc

 

Hey Billy,
What happen? I miss your totally above my head electronic conversations...

 

Hi mydauphin,

I am so sorry! I know I promised to get on with this stuff, and I'm seriously late. I hate to do that too. The problem is that I suddenly got a bunch of work . . . and with the economy being what it is I can't afford to turn any work away . . . so it's been on a back burner. I'm in the office working today and will be tomorrow too. Hopefully I can get to it next week. I did order the parts, and they're here. And the actual panel has been put together, so I won't be starting totally from scratch.

Again, my apologies!

BillyDoc

 

Just pulling your chain. I know for the first time in my life I look forward to getting work. Have a good one 4th of July.

 

Solar powered recharger

Hello BillyDoc,

Thanks for all that information on solar recharging. A friend of mine and I are working on a school project on designing a solar powered recharger to recharge lithium ion batteries efficiently. We are looking for a low cost solution and were wondering if the voltage regulation is necessary. Can't we simply use a diode in series with the panel and the secondary battery to prevent current flow from battery to panel? The battery voltage can be regulated by a comparator as you suggested, and whenever it is low, the panel can be charging the battery up. Since the battery voltage cuts off the power supply beyond a certain voltage, why do we need a voltage regulator at all?

Thanks

 

Welcome to the fourum ..We like all types of quetions here and you will find a lot of clever people with a vast knowledge who started like you on projects ......You need to do a lot of research on the specifc requirement of charging lithium Iron batteries which are unlike any other in there charging requirements ...If you wanted to charge a car battery with a low powered say 20w panel yes you could just put in a diode ....but a lithium Iron is not a car battery ..
Like you said its your project so do the research on the internet ..getting someone to tell you how to do it is not research or a project ... Tell us all what you find out and what problems that is giving you and many will help.

What size panel do you have and what is the voltage and capacity of your lithium batteries ?