Solar charge controller, what for?

[Sunking]

The reasons they do not work well in winter is I live in a forest so I don't get much sun till 9.30 am and lose it at 2.30pm. I of course get some trickle either side but it's only about 5 amps. Consequently they never get enough time on float to really work up a good full charge, add to that more drain with shorter days.......
This is the main reason I'm switching to LiFePO4's. They bulk charge upto about 90%, They will take a lot of amps in a short period of time. And lets face it, FLA are crap. Even when treated really well, they are mostly ornamental.

Well every battery has pro and cons. At the end of the day, FLA is still the default winners choice once you consider all the pros and cons. FLA's are less expensive and last longer.

The question at hand is, with hi + Lo V cut outs for protection, why do I need a charge controller rather than direct wire the panels. What benefits would it bestow? Main purpose on FLA is to hold voltage at float for ever, not a desirable function for LiFePO4's. My understanding is charge to voltage, shut it down.
You say I will still need a charge controller, but didn't state why.
I'm no expert, especially with LFP, so I'm really interested in the reasons so that I may understand, not just accept "the way it's usually done"

Thanks for your comments

OK you have a lot to learn. As of now you do not even know how solar works. You can as you say not use controller, and just use the old fashion antiquated relay to connect/disconnect panels. Do that and you change your 100 watt, 18 volt, 5.3 amp panel into a 67 watt panel. Solar Panels are current sources and Battery Current = Panel Current of 5.3 amps. Use a modern MPPT controller and you get a full 100 watts as it boost the 5.5 amps to 8.3 amps.

You can Float Charge LFP, and is the best way especially with a solar system. All commercial LFP chargers are simple Float Chargers. There are no stages like Bulk, Absorb, and Float. Trick is you set the voltage to less than 100% SOC around the 90% SOC range. For a 4S LFP we are talking 14.2 For example let's say around 2 in the afternoon, the LFP battery tops out at 14.2 volts so charging stops. However you can now use panel power instead of battery power assuming the load is equal to or less than what the panels can supply. That takes some of the load off the batteries and extend their life. You only use battery power at night or when power demands are more than the panel can supply.

As for using a BMS or HV or LV disconnects is not needed, especially in a solar system. You can use them if you wish, but there is no reason to use them. In fact us DIY EV builders would never use a BMS as they cause way more problems and battery destruction. The best and easiest way is to initially Bottom Balance your LFP pack, set charge controller to 14.2 volts, and let the Inverter LVD protect over discharging the batteries, then use your Inverters LVD to protect the batteries from over discharge. A 12 volt Inverter made for lead acid batteries has a LVD of 10.5 volts. LFP needs a cut-off of 10 volts so the Inverter LVD is more than adequate and will shut down when the LFP batteries reach about 10% SOC.

[bungawalbyn]

Was that Gold standard or Old standard?

If LiFePO4's live up to the claims, and the industry produces specific control gear to suit, then sooner or later, FLA will take it's place alongside the steam engine. Add to that the toxicity of FLA and lead in general, I say good riddance.
Of course the hoped for longevity of LiFePO4 may turn out to be of little consequence when the next big thing comes along, probably before the end of their lives. (See Lithium Titanate etc)

Rather than efficiency, which don't get me wrong, I'm all for, I've gone for capacity instead. When looking at the costs involved in obtaining high/optimal efficiency, it looks to me to be more cost effective to go big instead, and who doesn't want power in reserve for when the weather is not good, which happens here quite a bit. (High rainfall area) I don't want/need a Ferrari, I need a truck, that I can maintain myself. (Out in the sticks)
I've tried FLA, and I don't think it cuts the mustard, more like the cheese.


Enough ranting


Here's a pic (difficult lighting conditions) of the base board and cradle. I didn't bead the sides as there are cross channels in the bases that I guess are to provide air to the vertical channels for ventilation. I used a bamboo/plastic composite board for the compression panels, non conductive both electrically and thermally, just the right size, and cheap from the offcuts heap at the hardware store. As these batteries are effected by temperature, to use metal side plates that would create a temperature difference on the end cells doesn't seem like a good idea. (Sit on aluminium)

P1060717.jpg

[Living Large]

The reasons they do not work well in winter is I live in a forest so I don't get much sun till 9.30 am and lose it at 2.30pm. I of course get some trickle either side but it's only about 5 amps. Consequently they never get enough time on float to really work up a good full charge, add to that more drain with shorter days.......
This is the main reason I'm switching to LiFePO4's. They bulk charge upto about 90%, They will take a lot of amps in a short period of time. And lets face it, FLA are crap. Even when treated really well, they are mostly ornamental.
Which is all beside the point.

I don't have a system, but I designed one for the lower Adirondacks - low sun hours and I was planning for the same window as you in winter - 9:30AM to 2:30PM. I just wanted to say that I chose LFP for much the same reasons as you - faster charge rate in low insolation conditions. I can't agree that "FLA are crap" or "ornamental" since I believe they are the gold standard for solar systems and hey, I don't even have a system. However, the outgassing and watering and equalizing I didn't want to deal with. I was going to go with about the same SOC window as you, with BMS (non shunting), LVD and HVD like you. So on paper, it sounds similar. This and $2 US will buy a cup of coffee.

Regarding whether to use a CC - from what I have read, you will get substantially better efficiency under most conditions - you may appreciate/need this in the environment that you describe. Good luck.

[bungawalbyn]

Thanks PNJunction for your concern and encouragement. I wear glasses and will be doing a suitable size fuse. I'm in the middle of making a base board on castors and compression cradle for the batteries, being careful not to block the ventilation channels formed in the battery casings. I'll post a pic later when finished

And yes Karrak my charge disconnect will be a separate circuit from the BMS. I realise MPPT is supposed to give a bit more in suboptimal conditions, but here in Oz any charge controller big enough to handle 100amps is horrendously expensive, whereas another panel isn't. + we come back to not being designed for LiFePO4. I've designed a triple layer protection (sort of) with the BMS as last resort. I closely monitor my present set up, I have meters set in the wall by the door than I can see from all over my house. A little obsessive maybe but I know what's happening with my system at all times. (When I'm home)
Thanks for the link to your project.

[karrak]

Having run a solar setup with LFP batteries for just over two years I agree that your setup should work OK with the following observations.

I would make sure that the circuitry to disconnect your solar panels from your battery at the end of charge is totally separate from the BMS disconnect circuitry to stop any single point failure damaging your expensive battery.

Make sure you don't scrimp on the quality or specs of the disconnect relays. You could substitute transistors for the end of charge disconnect relay?

As PNjunction has stated, the final SOC of the battery is very dependent on the charge current which of course can be vary variable when using solar energy as the charge source. I have found that charging to an end voltage of 3.38-3.4 volts/cell with solar, can result in an SOC in the range of around 95%-80%. I would look at using an end voltage of around 3.4 volts and letting the current taper off to around .02C before stopping the charge.

This Open Source Battery Monitoring and control project that I designed may be of use or give you some ideas. https://github.com/simat/BatteryMonitor

Using an MPPT controller can give you an extra 20-25% of power on those cloudy days when you need it the most.

Simon

[PNjunction]

So far, I think you'll do fine.

Keeping a 4S / 12v battery happy isn't too hard with lifepo4 for an interested DIY'er who knows the importance of monitoring. If your application is in the "Sub-C" or 0.5C or less charge / discharge, the cells will be pampered if not taken to extremes.

Look into banding those cells together somehow too.

This is my first foray into LiFePO4's and despite having read everything that I've been able to find, there seems to be no consensus on the "Right" way to do all this.

There never will be. The right way is the way that works for you best and just keeps the cells healthy and you safe (fusing etc).

I've based most of my thinking on what the guy from Compass Marine has to say. He's put forward detailed and well reasoned arguments backed up by testing. Even so he's application is more slanted to mains and generator charging, there seems little out there that is specific to solar on large 4s packs.

He's a major player with a lot of experience and good advice. Take special note of his cautions on fusing, ie don't use cheapo automotive stereo fuses. So far though, I think only you and I are in the 12v arena. You'll learn a lot, and while your capacity is about 20 or more times mine, we will only lose up to 4 cells if we make a big mistake.

AND due to the massive size of your bank, you'll want wiring that can take the momentary surge of an accidental short *while* the fuse is actually blowing. You don't want both your fuse blowing and your wiring going up in smoke at the same time, even though that can take only milliseconds. Note that he has on hand a very large pair of insulated cable-cutters as a last resort in case things actually fuse together when the fuse / infrastructure is supposed to open!

Look into heavy duty contactors, such as Blue-Sea or other heavy duty dc switching, even if your application is relatively low current. Ie, in case of accidental short, you don't want your dc-disconnect switch going bang in your hands. What about GOGGLES? Are you wearing them while making your connections etc? I don't want to sound like a paranoid shop-teacher, but you are dealing with cell capacity in the very big leagues.

Keep it safe. If you smell sweet perfume, that is electrolyte heating, so don't stop and smell the roses, but disconnect all loads / sources and inspect.

Just being cautious. I want your project to be SAFE and successful.

[bungawalbyn]

What surprises me is why this post was even started if the answer is already known by the OP.

Take it easy guys, I'm no expert, I'm looking for flaws in my understanding and proposed design. I'm happy to receive every ones advice, especially if backed up by good argument from experience and/or technical expertise. This is my first foray into LiFePO4's and despite having read everything that I've been able to find, there seems to be no consensus on the "Right" way to do all this.
I've based most of my thinking on what the guy from Compass Marine has to say. He's put forward detailed and well reasoned arguments backed up by testing. Even so he's application is more slanted to mains and generator charging, there seems little out there that is specific to solar on large 4s packs.
As has been pointed out, "direct connect" is something of a misnomer, though accurate in that the panels will be directly connected rather than through a proprietary charge controller, they will not be unregulated in that there will be hi and lo cut offs etc

I thought given the sparse 1st hand information for this specific application that others may be interested in the journey from ignorance to implementation and adjustment to achieve a workable cost effective set up that doesn't rely on makeshift disabling of equipment designed for other battery types.

Most of all though, I guess I fall into the category of compulsive DIYer, I not only want the power, I also want the fun associated with the learning and nutting out of the how's and why's.

I do read and think about your comments and do further research if necessary, but will also argue the point if I am uncertain or don't agree.
The wheel is constantly being reinvented.

[PNjunction]

I thought that might spin your dials. I had them top charged before delivery to what the factory and distributor recommend. I take your point about leaving them like that so have been bleeding off charge all afternoon.

Ok, but that is another bit of a red-flag. I'm no hazmat expert, but I'm pretty sure that the regs don't allow for shipping lifepo4 fully charged or topped off to the end-user. While you may have been accomodated by the seller, just be careful in your dealings.

Right - you have a diy controller setup - it will be interesting to see how that goes with proper monitoring. Consider that MUCH smaller 12v lifepo4 such as motorcycle start batteries in the 2.5 to 12ah range like Shorai, Antigravity and so forth have no controller other than what the bikes regulator's hvc is set to. Most smart guys drop the voltage down a bit for adjustable regulators since they are no longer charging Pb batts. The manufacturers rely on very closely matching the cells for both capacity and internal resistance. Off road starter batteries are another application so we'll leave that there.

I think you can do it, but you are going about it on your own terms. Not recommended for those with plug-n-play mindset.

[inetdog]

You might as well stop beating your head against a wall. Apparently bungawalbyn has made up his/her mind and is going to wire the pv panels directly to the batteries. Sounds as thought he/she has enough knowledge about battery technology but in the long run it's his/hers money to spend as they want to.

What surprises me is why this post was even started if the answer is already known by the OP.

Yes, it seems clear that when he says he will not use a CC, he is depending on this BMS to limit the upper voltage and cut off the charging to the batteries (by shunting, I think).

That is NOT the same as connecting the panels directly to an unmanaged battery bank.

But he still should avoid charging into the top 20% of SOC if he wants optimal battery life. I hope his BMS is adjustable.

[SunEagle]

The main problem with leaving out the charge controller is that you are then depending on getting a precise, consistent voltage output from your panels.
That will not happen, since Voc and Vmp both depend weakly on the insolation at a given time and strongly on panel temperature.
If you want to be able to drive full current from the panels during the Bulk phase, you will have to be able to size you panels so that Vmp is close to the low SOC battery voltage while Voc is not any higher than your target charge cutoff voltage.
Since the difference between Vmp and Voc for that typical crystaline silicon panel is about 20%, you may not be able to hit the values you need. And the variation of both voltages with temperature can be as much as 25% in a very cold winter climate and a high summer panel temperature.

And you would be completely ruling out the opportunity to select your panels for best power per watt, which an MPPT CC would make possible.

You might as well stop beating your head against a wall. Apparently bungawalbyn has made up his/her mind and is going to wire the pv panels directly to the batteries. Sounds as thought he/she has enough knowledge about battery technology but in the long run it's his/hers money to spend as they want to.

What surprises me is why this post was even started if the answer is already known by the OP.

[bungawalbyn]

The main problem with leaving out the charge controller is that you are then depending on getting a precise, consistent voltage output from your panels..

Am I? How so? If panel V is less than batteries, no charge, if greater, some charge. LifePO4's especially of large cell size will take all my panels will put out and MANY times more amps, with minimal resistance, doesn't this keep the operating voltage at about SOC of the battery? Doesn't this mean that the voltage will only rise to cut out when the battery reaches the SOC of this voltage?
Remember these are not FLA and operate quite differently under charge due to their very low internal resistance.

Why do I want the Voc to be no higher than my target cut off voltage? I'm not relying on the charge just fizzling out due to voltage equalisation, I've got sensors and cutouts for that.

The business I bought the batteries from turns out to be doing something similar to what I propose, having previously done comparisons between "direct connect" and using MPPT controller and found little difference. These guys, one would hope, have greater knowledge and experience in using LiFePO4 batteries. They do EV cars, and run their workshop on a 'direct connect" set up.




The batteries after draw down are now sitting at 3.31v all cells after an overnight rest. so the balance is very good. Sweet.

[inetdog]

The main problem with leaving out the charge controller is that you are then depending on getting a precise, consistent voltage output from your panels.
That will not happen, since Voc and Vmp both depend weakly on the insolation at a given time and strongly on panel temperature.
If you want to be able to drive full current from the panels during the Bulk phase, you will have to be able to size you panels so that Vmp is close to the low SOC battery voltage while Voc is not any higher than your target charge cutoff voltage.
Since the difference between Vmp and Voc for that typical crystaline silicon panel is about 20%, you may not be able to hit the values you need. And the variation of both voltages with temperature can be as much as 25% in a very cold winter climate and a high summer panel temperature.

And you would be completely ruling out the opportunity to select your panels for best power per watt, which an MPPT CC would make possible.

[bungawalbyn]

[QUOTE=PNjunction;163084]OUCH! From this distance I have to throw up some red-flags unless you are making typos. A fully charged cell after at least 12 hours rest should measure about 3.38 to 3.42v or so. Granted, you had them charged beforehand, and some huge capacity cells, but if they are *resting* at 3.6v or more, that is NOT good and means they were overcharged to begin with. Normally cells that have not been charged come sitting somewhere in the flat part of the curve, at about 3.2v, which of course does not mean balance. In your case, I surely hope that those big cells were not overcharged.

I thought that might spin your dials. I had them top charged before delivery to what the factory and distributor recommend. I take your point about leaving them like that so have been bleeding off charge all afternoon.
Ive got the bank sitting at 13.24, and the voltage drop has slowed right up, I think I'll take them down to abot 13.2 or a touch lower. Interesting to note virtually no voltage bounce back when loads switch off, Also individual cell voltages stay within 0.02 but each time I measure they alternate ranking. Curioser.
Hopefully this will put them on the "Flat" cause they will have to sit for a week or two

My system gets a regular daily workout, when I go away it's good for the FLA's they get some good float time, but with these I think it would be best to turn everything off.

So I'm not overly concerned with them sitting on high states of charge for long

P1060715.jpg

This was the start of drain down. I since hooked up an inverter and wired in a voltmeter I can see through the window. I've been going out to do the cells regularly.

Ya gotta laugh!!

Not long after I first hooked it all up I went to check the voltages and was feeling the temp of the cells when white plasticy smoke wafted over the batteries. I had contractions. When the moment of panic subsided I searched for the source. I'd stood on the soldering iron, I was making a burnt offering of my sole.
Phew!

[PNjunction]

I invested in a DC power supply the other day so I can bench test everything before going live.

Forgot to mention that this is great for your needs.

Just know, that if for some reason your bms fails or for whatever reason you go beyond say 85% DOD down into the deep part of the discharge knee, which happens relatively FAST, the OPPOSITE is true for coming out of it. That is, below about 85% DOD one should apply no more than about .01C until the cells reach 3.2v again, and THEN you can apply your normal charge current. Kind of a gentle absorb only in reverse!

While going below 85% DOD doesn't do the cell any favors, applying too much current coming out of the deep discharge does secondary damage. I hope the op in the video above with extreme capacity testing took it easy recharging initially. In other words, don't go into a panic and hammer the bank with your genny if you have gone too far. Use your controlled power supply instead to recover back to a normal state. Once again, an 80% DOD max is the right choice overall as even a moderate solar system capable of .05C is too much when the bank has gone too far down the slope to recover nicely.

Of course if you flatten your batteries, this extreme low current revival applies too, BUT you MUST get to them ASAP. Staying too long in a deep discharge from a safety / recovery standpoint has been discussed elsewhere, but this is a gentle reminder when your adrenaline might be flowing.

I'd love to see pics of your project and keep us updated...

[PNjunction]

I had them top balanced and they arrived at 1 cell at 3.62 1 @ 3.63 and 2 @3.65 so not far apart.

OUCH! From this distance I have to throw up some red-flags unless you are making typos. A fully charged cell after at least 12 hours rest should measure about 3.38 to 3.42v or so. Granted, you had them charged beforehand, and some huge capacity cells, but if they are *resting* at 3.6v or more, that is NOT good and means they were overcharged to begin with. Normally cells that have not been charged come sitting somewhere in the flat part of the curve, at about 3.2v, which of course does not mean balance. In your case, I surely hope that those big cells were not overcharged.

Either way, DISCHARGE those cells NOW to about 3.2v each. This simple video (going too low but for extreme capacity testing) might help showing some simple loads. At the very least, an automotive bulb can be applied to each cell do drop that voltage now... :

https://www.youtube.com/watch?v=oJ8AfJsOhsM

Of course, I have to ask - after spending all that on the batteries, I hope you aren't using a shirt-pocket $20 multimeter. Invest in a good quality Fluke or similar.

Keep in mind that as a non-professional speaker I get tongue tied, so I'll try to simplify and reiterate the above when using low-current charge rates, like .05 - .10C ...

No matter what voltage you choose as your HVC, be it 3.47 to 3.6v, given enough time your cells WILL fully charge to 100% SOC, so once again voltage is not a TRUE indicator of full charge. Keep an eye on absorb current. Basically, no matter the voltage, once your cells just start to dip into / below .05C absorb, STOP no matter what. Those that try to charge well below .05C, or spend waaaay too much time at .05C will eventually witness the change from battery charging to electrolyte heating.

That is one thing a commercial controller has the advantage with - safety timeouts as compared to just a simple voltage / current triggering bms. Ie, even my simple Morningstar pwm will only spend 1 hour at absorb before falling back to float. Yes, we're talking a Pb charger, but at least it does something which I accommodate and shoehorn into the wrong application.

Example - on a cloudy day, I could possibly supply ridiculously low currents, say .0025C which isn't enough to drive my cells to the 3.5v cutoff. BUT, if I don't use the batteries and they get this treatment day after day, they *could* sit at 3.45v for a week, eventually fully charge, but still never reach the 3.5v cutoff. So it spends a whole week basically in absorb - and eventually on day 5 they go into electrolyte heating mode which does finally trigger the hvc... I think you get the idea. Sometimes too little current can be damaging according to the environment / application at hand. Hard to account for all the variables in a forum ...