What is the maximum voltage they are tolerant of? Will a CV/CC with a max of 14.4V and a reasonable current limit (1/3C) be good? Can you leave it like that or do you have to monitor it to prevent it from overcharging? I am kind of confused here. If the maximum voltage is 13.6V, how can you safely charge it with 1 volt more?

 

What is the maximum voltage they are tolerant of? Will a CV/CC with a max of 14.4V and a reasonable current limit (1/3C) be good? Can you leave it like that or do you have to monitor it to prevent it from overcharging? I am kind of confused here. If the maximum voltage is 13.6V, how can you safely charge it with 1 volt more?

As I said, there is no simple answer if you want to do it right. There is no magic voltage or current to charge it with that will be always correct. If you apply 13.6V it may take weeks to fully charge, if you apply 14.4V it may charge ok but you will need to drop the voltage when it is charged to stop the battery being destroyed. Some commercial chargers claim to do this automatically. Go have a look. Nome that I know of handle discharge properly though. They all have fixed cut off voltages or did last time I checked.

 

I will do something separate for discharge. So I am looking at this charger. 24V 4A means about 8-10 hours to charge. That is not ideal but is good enough. So do you think I can trust commercial chargers to not overcharge them, even when left for long periods of time connected to it? And the battery I am looking at says it is sealed lead acid. If that is the case, why does it say not to charge in a sealed container? Don't SLABs emit no gasses unless really abused?

 

I will do something separate for discharge. So I am looking at this charger. 24V 4A means about 8-10 hours to charge. That is not ideal but is good enough. So do you think I can trust commercial chargers to not overcharge them, even when left for long periods of time connected to it? And the battery I am looking at says it is sealed lead acid. If that is the case, why does it say not to charge in a sealed container? Don't SLABs emit no gasses unless really abused?

Ok. You are determined to make me do as much of the work as possible but there are limits ;-)
I've had a look at the link you provided but did not find a data sheet there. I did find in point form, claims about the charger, one of which was:
Full automatic, gushing, flat filling, water flow charge, full automatic shutoff
which I think means this: 'gushing' is boost charging, flat filling is equalise charging (if you don't know, equalise is elevated charge voltage applied as the cells become charged to force a charge into any weak cells and thereby equalise the charge in all the cells. You can see why they might call it flat filling) and 'water flow charge' is float charging.
If that is correct then that is very good except that there is no temperature compensation that I can see. If the battery is to be kept in a stable temperature environment then it may be just the thing. My only caveat other than the temperature compensation is that when it is done it should remain in float charge. It is unlikely that at the end of the equalise charge that the battery SoC will be any more than 90% at best.
If your charge time calculation is based on 4A versus battery AH capacity then I suggest you do a few trials as it will likely take a lot longer than you think. Just because you have 4A available does not mean the battery will accept every electron up to that 4A limit unless maybe it is a massive battery. I'm guessing you have a 30 to 40AH battery so you will probably see a decline in charge current to below 4A by the time the battery is at 50 or 60% SoC.
SLA batteries have an absorber in them for the gasses produced during charging and some, I think, from memory, have some kind of funky way of transforming the gasses into something else inert as far as the battery is concerned. In any event, these things can only cope with gas production up to a certain rate, beyond that the gas builds up pressure in the casing. In case any of this gas magic fails, the manufacturers cover their liabilities with the statement, do not charge in a sealed box (and the unspoken bit: in case our gas magic fails and the gas leaks out).
You might also notice that a regular (vented) wet cell battery (or VRLA etc etc) will boost at much higher voltages than an SLA. This is for two reasons; 1. they can, 2. they need the bubbling in the cell to stir up the electrolyte to stop stratification which will mean all the lead falls to the bottom of the cell where it is deposited back onto the plate which then builds up at the bottom (and gets thinner at the top) until eventually the plates touch at the bottom and the battery is dead.
Hope that makes sense to you. If you have a data sheet for the charger please share it if you can.

 

I will do something separate for discharge. So I am looking at this charger. 24V 4A means about 8-10 hours to charge. That is not ideal but is good enough. So do you think I can trust commercial chargers to not overcharge them, even when left for long periods of time connected to it? And the battery I am looking at says it is sealed lead acid. If that is the case, why does it say not to charge in a sealed container? Don't SLABs emit no gasses unless really abused?

I should also point out that not all batteries are equal. I answered another post a while ago on similar topic and the two batteries they were looking at were chalk and cheese with the cheaper one (seemed to me) to be the better one. Check the battery data sheet and pay close attention to cycle life expectancy graphs and figures. If you can't find any cycle life data then it is probably not a deep cycle battery or not a very good one.
If you do find the graphs then take note of the parameters specified. One of them will be depth of discharge (DoD) and you will probably find a maximum DoD of 70% (discharge to 30% SoC) will yield around 10 to 30 times as many cycles as a 100% DoD. They may also claim 100's of cycles with 100% DoD. If they do, read the fine print to find out how they determine when a battery is at the end of its cycle life. It may be at 50% original capacity, or 10%. Who knows. Best to check.

 

This is the battery:
https://www.amazon.com/UPG-UB12350-...rd_wg=ULkYR&psc=1&refRID=9MPQSX6QJA1YDNTBP659
It says up to 14.9V in cyclic use, so the 14.7V (per battery) charger should be okay. I may get the 7A version to charge it faster.

I googled UB12350 and this looks like the datasheet for it.
https://a89b8e4143ca50438f09-7c1706.../003/518/original/ubD5722-spec.pdf?1440174717
It looks like not discharging it to 100% DoD will significantly extend the lifetime.

 

This is the battery:
https://www.amazon.com/UPG-UB12350-...rd_wg=ULkYR&psc=1&refRID=9MPQSX6QJA1YDNTBP659
It says up to 14.9V in cyclic use, so the 14.7V (per battery) charger should be okay. I may get the 7A version to charge it faster.

I googled UB12350 and this looks like the datasheet for it.
https://a89b8e4143ca50438f09-7c1706.../003/518/original/ubD5722-spec.pdf?1440174717
It looks like not discharging it to 100% DoD will significantly extend the lifetime.

If you have a bench supply that can deliver the current I would suggest a few trial runs before you make the investment in a more powerful charger.
I'd also be skeptical about those life cycle curves. After 200 cycles of 100% DoD battery capacity is between 90% and 50% of the rated capacity ( so between 32AH and 18AH). So the 100% DoD curve is meaningless. The 50% is much better showing 90% to 95% capacity remaining after 400 cycles. Only problem is, who's going to limit the discharge to just 50%!!!??? That's crazy talk! You could try an extrapolation for 70%, hassle the manufacturer for better data or do what I did and research the thing until you feel your life force giving up.
At 30% it looks great with 1000 cycles, except that it should be more in the range of 3000 or more cycles at that shallow a DoD. See if you can find a sonnenschein (spelling questionable) battery of similar capacity just to compare the specs. Don't look at the price of the sonnenschein, special preparations are required before doing such a fool hardy thing.

 

Here's a hysteretic switching circuit I designed and simulated (but not built) to charge a 12V lead-acid battery.
It has a three-stage charge protocol of constant-current initial charge, constant voltage topoff, and then a lower voltage trickle/maintenance charge.
It gives you an idea of how it can be done without a microprocessor.

And here's a simple, low-voltage cutoff circuit that may also be of interest to you.

Both use a low-cost TL431 voltage reference to accurately determine the charge and cutoff voltages.

 

That reference looks like a viable option here. I am considering this one.
https://www.mouser.com/ProductDetail/863-TL431BVLPG
So what is a typical circuit for it? How do you get a variable voltage from it, based on a potentiometer? When it says .4%, does that really mean its output is accurate to .4% of the desired voltage, or are there other things that would make it less accurate in reality?

I will need to use a different circuit because it must be adjustable and it must use a latching relay. I was considering something like this. There would be a pull-down resistor for the mosfet, snubbers, and some other additional things. This is just a rough schematic.

So can you recommend a suitable op-amp here? And the goal is that if the battery is below the reference, it turns on a mosfet, turning off the relay. It can only be turned back on by the pushbuttons. Is the op amp set up correctly for that?

 

That reference looks like a viable option here. I am considering this one.
https://www.mouser.com/ProductDetail/863-TL431BVLPG
So what is a typical circuit for it?

Below is a typical circuit for the low-voltage cutoff as discussed here.

You can make it adjustable by changing R3 to a 20k pot.

If you want it to latch off, then connect R1, R2, and Q1 to the output of the relay instead of the battery.
A push-button connection from the battery to the relay coil will turn it on (to protect the transistor also add a 1N4148 diode in series with its collector, anode towards the collector).

When it says .4%, does that really mean its output is accurate to .4% of the desired voltage, or are there other things that would make it less accurate in reality?

The main additional error is due to the 50 PPM (.005%) / °C change in voltage with temperature.

 

How do you set up just the reference part? It is hard to tell what does what, and what is needed for the reference there. And is there anything major wrong with the circuit I designed? I know that there are some other things to add.

 

How do you set up just the reference part?

The TL431 has a reference voltage of 2.5V at its reference pin (the one coming out the side of the symbol).
If you want to use it as a voltage reference (rather like an adjustable Zener) then you use the circuit below from the data sheet:
If you connect the reference pin directly to the cathode without R1 and R2, then the output voltage will be 2.5V.

And is there anything major wrong with the circuit I designed?

Nothing apparent expect it requires more parts, but I haven't really looked at it in detail.

 

So Vka is the output voltage? And the DoD curves for that UB12350 battery I am looking at appear to be kind of standard. Looking at other similar batteries, the datasheets show very similar curves.

 

So Vka is the output voltage? And the DoD curves for that UB12350 battery I am looking at appear to be kind of standard. Looking at other similar batteries, the datasheets show very similar curves.

It is possible that some of the batteries you are looking at are actually made at the same factory and badged differently. I did try to find curves for some of the better battery manufacturers but didn't have much luck with it. I know, even though they won't quote a curve, some models of sonnenschein batteries have cycle life expectancy of 3000 cycles or more at 70% DoD and more like 300 cycles at > 90% DoD.

 

So Vka is the output voltage? And the DoD curves for that UB12350 battery I am looking at appear to be kind of standard. Looking at other similar batteries, the datasheets show very similar curves.

The TL431 is an industry standard and a great part to know about. I've been using these since about 1988. Yes, in the 'before time' but after the dinosaurs. Like the 555 timer, there are hundreds of circuits out there based on the TL431 that do all kinds of useful functions like voltage triggered switches, timers, level detectors etc etc.
If you are using it as a voltage reference, beware the stability is only certain with a parallel shunt capacitance of either less than a few nF (I use 1nF generally) or greater than some 10's of uF. Check the data sheet, there will be a graph showing the stable range of operation against capacitance. There was a time when every major silicon manufacturer had a 431 in their line up.

 

So it is reccomended to use a 100-200uf capacitor with them? Can you have too large a capacitance for it to function properly? And why can you not use just a few uF or a few hundred nF? I know about inrush, etc but that is not too big a concern.

 

So it is reccomended to use a 100-200uf capacitor with them? Can you have too large a capacitance for it to function properly? And why can you not use just a few uF or a few hundred nF?

Here's the graph from the data sheet showing where it's stable.
As can be seen, anything below about 6nF or above 3μF is stable for all conditions.

 

How do you set up just the reference part? It is hard to tell what does what, and what is needed for the reference there. And is there anything major wrong with the circuit I designed? I know that there are some other things to add.

In the circuit for a low voltage cut-off, the reference is internal to the TL431 and is 2.45V. There are one or two things I'd say about the circuit as presented; the values of the resistors include two or three that are in the E96 range which always makes me suspicious about whether the circuit has ever actually been built or only simulated. The 453k resistor is in series with the coil resistance which may be significant enough to be a relevant variable against resistances chosen from the E96 range. The 431 requires a minimum current to operate which is typically about 1mA and the 499R resistor is supposed to allow that current without turning the transistor on but it may still develop 0.5V which means the transistor will be close at room temperature but may already be on anyway with 0.5V at elevated temperature (false triggered). The 806R resistor is a unnecessarily fussy value. It could be more like 1k and the 499R could be 330R and the circuit would be ok, the resistors easier to get and it would be less temperature sensitive.
A time constant somewhere would also be wise to avoid a glitch from false triggering the cut off. Ideally the time constant would not interfere with the positive feedback that gives the circuit hysteresis so a cap across the 10k would not be the best place to do that.
A fixed cut off voltage like this will also not protect the battery from excessive discharge, it can only protect from a discharge at the point of 100% DoD with any accuracy. Attempting to set a voltage for 70% DoD for example will need temperature and load current compensation of the trigger voltage at the very least and it still won't be particularly accurate in DoD terms.

 

In the circuit for a low voltage cut-off, the reference is internal to the TL431 and is 2.45V. There are one or two things I'd say about the circuit as presented; the values of the resistors include two or three that are in the E96 range which always makes me suspicious about whether the circuit has ever actually been built or only simulated. The 453k resistor is in series with the coil resistance which may be significant enough to be a relevant variable against resistances chosen from the E96 range. The 431 requires a minimum current to operate which is typically about 1mA and the 499R resistor is supposed to allow that current without turning the transistor on but it may still develop 0.5V which means the transistor will be close at room temperature but may already be on anyway with 0.5V at elevated temperature (false triggered). The 806R resistor is a unnecessarily fussy value. It could be more like 1k and the 499R could be 330R and the circuit would be ok, the resistors easier to get and it would be less temperature sensitive.
A time constant somewhere would also be wise to avoid a glitch from false triggering the cut off. Ideally the time constant would not interfere with the positive feedback that gives the circuit hysteresis so a cap across the 10k would not be the best place to do that.
A fixed cut off voltage like this will also not protect the battery from excessive discharge, it can only protect from a discharge at the point of 100% DoD with any accuracy. Attempting to set a voltage for 70% DoD for example will need temperature and load current compensation of the trigger voltage at the very least and it still won't be particularly accurate in DoD terms.

I know voltage is temperature and load dependent. But you could set it to something conservative like 12V and it would probably discharge to less than 100% DoD. It may not be perfect but would probably be good enough.

 

Here's the graph from the data sheet showing where it's stable.
As can be seen, anything below about 6nF or above 3μF is stable for all conditions.

View attachment 154857

I'll use a 1.25F supercapacitor just to be on the safe side. I will probably go with a common 27uF capacitor, to eliminate any issues.