diy solar

diy solar

UPC Uninteruptiable Power Supply Battery Replacement

thepartyhound

Christian Prepper
Joined
Sep 14, 2022
Messages
146
Is there any reason that I could not remove the worn out SLA battery from a small APC battery backup and replace it with a LFP battery? Obviously the UPS was designed for use with a SLA battery, but the BMS in the LFP replacement should keep everything happy...or at least safe, right?
 
Complex question.

The UPS wasn't designed to charge LFP to the right voltage exactly. it is close, but not perfect.

Some LFP vendors sell batteries which claim to be "drop in" SLA replacements, so they either have added circuit changes to adapt the voltages or they are just boldly claiming compatibility without those changes.

For example:


Is it really compatible? Hard to know.

A 4S LFP pack wants to be charged precisely to 3.6 volts/cell or 14.4 volts.

A 6S SLA battery (12 volts) wants to be bulk charged to about 2.4 volts per cell bulk and 2.3 volts per cell standby float use. This is 13.8 volts to 14.4 volts. So it seems okay on that basis. But UPS often just do constant current charging, and the LFP battery really doesn't like that, they want to be constant voltage at the end of charge, and then stop.

I've always wondered why you can't buy proper LFP UPS right now. The lead acid ones are such garbage. But I guess the only thing the market cares about is being cheap to buy and then just throwing away the UPS in a few years when the lead acid battery dies.

If you go for it, let me know how it works out and which UPS you used it in. I have a pile of lead acid UPS units with no battery that I could use.

Mike C.
 
I've always wondered why you can't buy proper LFP UPS right now. The lead acid ones are such garbage. But I guess the only thing the market cares about is being cheap to buy and then just throwing away the UPS in a few years when the lead acid battery dies.
Easy to replace a UPS battery.

Many UPS have enough SLA battery to last 10-15 minutes. That implies a 4.0 to 6.0 C discharge rate. Much too high for a LFP battery. At 1+ hour of run time, LFP batteries start making sense.
 
 
Easy to replace a UPS battery.

Many UPS have enough SLA battery to last 10-15 minutes. That implies a 4.0 to 6.0 C discharge rate. Much too high for a LFP battery. At 1+ hour of run time, LFP batteries start making sense.
I did the math on my UPS and posted in a previous UPS thread:

My Cyberpower 1500VA UPS uses a pair of these small form factor batteries that are very commonly used. 1500VA at 24V is 62.5 Amps. For a 7Ah battery that's a 9C discharge rate. The cute little LiTime 6Ah battery with that same physical form factor can do 30 Amps for 5 seconds.
 
But UPS often just do constant current charging, and the LFP battery really doesn't like that,

I don't think that is really an issue. As a LifePo4 battery reaches full charge, current inherently drops. If the UPS is putting out lets say 14v, once the battery reaches that, the current will drop to practically zero. Now, it may take a lot longer to recharge, but with a UPS, that is not normally an issue as it's usually only used once in a while.
 
Chargers are simpler than people think. They are just power suppliers.
You set the desired end voltage and let it run.
The current is limited by two things how much the charger can output (usually settable) and how much the battery will accept. (Reduces as desired voltage is reached)
 
Have you tested it? That battery is rated for 10A (128 Watts) or double that for 5 seconds. The larger home UPS units often demand large currents from their battery at rated output.
Good question. At the time I bought it, I wasn't really well versed in battery capabilities and just went for what the mass said worked. I tested it enough, I suppose? The equipment it's on doesn't take more than 100w or so at idle.

I wonder if that's a major reason why they still use SLA batteries for UPS solutions - because they can withstand high-amperage draws for a short period of time.
 
As the first reply stated it's not that simple.
It will have issues charging the battery fully and it will have issues with runtime alarms and battery level will be incorrect.
Basically your alarm will go off at the standard discharge setting that it has for an SLA battery. The LFP battery can of course run a lot longer but none the less you will keep hearing the beeping.
If you are using it with a PC and have the cable attached it will shutdown the PC long before the battery is exhausted.
Another point seems to be drifting of the batteries, assuming your using two.
Overall it seems like it can work but a BMS in each battery is really needed and from time to time you might need to re-balance the batteries. This last part was what made me abort the project. There is software to get rid of some of the issue like re-calibrating the battery meter and people have used Putty to get into the UPS software and modify the charging voltage.

This guy seemed to have the Best video on the subject.

 
So, what I think I'm understanding as a whole is that because these units charging circuits are designed with SLA batteries in mind, LFP will not be fully charged. It may work, and if a small load is all that is ever needed for just enough time to properly shut aomething down, it may even seem to work well. I'll need to do some looking to find out exactly how much power my UPS would need to supply before making a final decision, but it is looking less and less plug and play.

Heres another option: LFP UPS Has anyone played around with something like this. I would assume it was designed for use with LFP from the start...but I have been wrong before.
 
You can just build your own if you want.

Inverter large enough for the load.

Battery charger that is sufficient to meet the load + charging.

Battery pack that is as large as you want / chemistry that you like. It just needs to be capable of charging and discharging at the required rates and not be higher than the recommended rates.

For example, if you want 1000 watts of actual discharge rating, but usually 300 - 500 watts actual use.

- Assuming 24 volt setup, that is ~ ( 1000 watts ) / ( 24 volts ) ~ 50 amps max, 15 amps commonly.

So a possible solution would be 2 each battle born 100s wires in series to make a 24 volt / 100 amp-hr pack. max recommended discharge rate of 100 amps, so you are well within the spec at 50 amps discharge max.

Buy a 2000 watt / 24 volt inverter that you like.

120 vac - 24 volt charger for 30 - 50 amps charge rate.

You can even add solar if you like.

That is what I did - rolled my own.

Runs a long time so I have plenty of time to do stuff.
 
So, what I think I'm understanding as a whole is that because these units charging circuits are designed with SLA batteries in mind, LFP will not be fully charged.
Maybe, maybe not.

It all depends on how the UPS does charging of the SLA. SLA is very tolerant of various charging schemes that can vary in sophistication.

If the UPS does a proper bulk plus float charging scheme using constant voltage, the LFP is probably going to be fine and last a good long time.

If the UPS charges at bulk voltages, and then does a constant current float, the LFP is going to be unhappy. The constant current float will slowly overcharge the LFP. This either means the LFP BMS clicks out, or it has to have some special current shunt (simulating SLA float) to keep the voltage within tolerance.

The fundamental difference between SLA and LFP is that SLA will tolerate charge current above 100% SOC and LFP won't.

Heres another option: LFP UPS Has anyone played around with something like this. I would assume it was designed for use with LFP from the start...but I have been wrong before.
I hope we see more units like this that are purposefully designed to be LFP.

Mike C.
 
The constant current float will slowly overcharge the LFP.

How? The term constant current is being misused in this example. Unless the UPS is putting out more than 14.4v or so, it can't overcharge as the LFP will reduce the current draw to zero. The UPS can't "force" current into the battery unless it raises it's voltage about 14.4v or so.

The fundamental difference between SLA and LFP is that SLA will tolerate charge current above 100% SOC and LFP won't.

Again, you can't force current into a LFP without raising the charge voltage.
 
I built a logger for my Cyberpower 1500 UPS when I expanded the small lead-acid to bigger lead-acid a few years ago. The charging 'algorithm' is the simplest possible: charge to 27.2 volts (400mA max) forever. If you run it down to the low voltage cutoff (21.9V) it takes over a day to recharge fully. Seems like a good candidate for a LFP drop-in.
 
I thought floating for LFP isn't recommended...

Most of these UPSs will just float the battery, either at 14v or maybe a bit less.
Will this not degrade the LFP battery fairly quickly?
 
I thought floating for LFP isn't recommended...

Most of these UPSs will just float the battery, either at 14v or maybe a bit less.
Will this not degrade the LFP battery fairly quickly?
I thought the same thing.
 
Actually, many with 2 SLD batteries operate at 24 VDC with both batteries in series, many of APC's UPS's use 2 X 12 volts batteries in series
 
I can tell you from experience lfp @ 29.2v all day everyday kills them within 18 months.

(From when i first started out and used a single stage pwm reg to charge everything)
 
How? The term constant current is being misused in this example. Unless the UPS is putting out more than 14.4v or so, it can't overcharge as the LFP will reduce the current draw to zero. The UPS can't "force" current into the battery unless it raises it's voltage about 14.4v or so.
Many SLA float charging stages are simple resistors to higher voltages like 15 or 16 volts.

Again, you can't force current into a LFP without raising the charge voltage.
That is correct but SLA is so tolerant, that a crude float charger could end up exceeding the voltage on an LFP. A properly designed charger shouldn't do that, but SLA is tolerant of bad charger design, so they are out there.

Mike C.
 
I thought floating for LFP isn't recommended...
It isn't.

Lithium batteries (including LFP) suffer the most stress at 100% SOC (SLA is at 0% SOC, they like to be constantly float charged). This is why folks with EVs often terminate charge at 80% or 90% to extend the service life of their batteries.

Lithium also wants very precise termination voltages, SLA is very tolerant of sloppy voltages.

I worry that with the wide variance in SLA charging designs and tolerances, an LFP pack will be abused.

Mike C.
 
Do you have any real world observations to prove that it doesn't hurt the battery?
Only my own personal experience.
I haven't overcharged (above 3.65v per cell) , or over discharged (below 2.5v per cell) . And I don't push them hard (I stay at or below 0.2c rate). Or let them get too hot/cold.
But I do charge to the top of the knee (3.56v percell) , and float there (3.5v per cell) daily.
The only degradation I have seen is calendar aging (1% per year).
I also have one battery that I used to charge to 3.625v per cell, daily.
Which is now part of a portable system that doesn't see much use. And just floats 24/7 at 3.5v per cell.
I have only been running LFP for a little over 2 years. But I see no reason for the results to suddenly deviate from the expected.
Is the same true if floating at lower voltages than 100% SOC ?
Yes
But you will limit capacity if you don't balance the cells regularly, in the upper knee (Above 3.45v per cell).
 
Only my own personal experience.
I haven't overcharged (above 3.65v per cell) , or over discharged (below 2.5v per cell) . And I don't push them hard (I stay at or below 0.2c rate). Or let them get too hot/cold.
But I do charge to the top of the knee (3.56v percell) , and float there (3.5v per cell) daily.
The only degradation I have seen is calendar aging (1% per year).
I also have one battery that I used to charge to 3.625v per cell, daily.
Which is now part of a portable system that doesn't see much use. And just floats 24/7 at 3.5v per cell.
I have only been running LFP for a little over 2 years. But I see no reason for the results to suddenly deviate from the expected.

Yes
But you will limit capacity if you don't balance the cells regularly, in the upper knee (Above 3.45v per cell).
Interesting...
Did you notice any bloating of the cells?
What cells are you using (brand, Ah, etc...) ?
 
I thought floating for LFP isn't recommended...

It depends on what voltage it is floating. I would float a 12v LFP at 13.6 forever and not give it a second thought. And I'm pretty sure Bluettis and similar float too while left plugged in. What good is a power station if it's not ready to go?
 
And I'm pretty sure Bluettis and similar float too while left plugged in. What good is a power station if it's not ready to go?
I'd say they don't really care as long as it outlasts the warranty. On the other tentacle, you _should_ be able to fully charge it and leave it for a year, no?
 
The idea to use LPF as UPS batteries has appealed to me too. Wondered about it.
I am glad to have found this thread.

I guess SLA, AGM and LPF each has its own place.
 
Last edited:
The idea to use LPF as UPS batteries has appealed to me too. Wondered about it.
I am glad to have found this thread.

I guess SLA, AGM and LPF each has its own place.
We use UPS at customer's houses to keep their phone service working in the event of a power outage. Back in the days of copper we only had to fire up a generator at the CO and everyone's dial tone would continue uninterrupted. Now, in the days of fiber, the backup power has to go at the CO as well as at the customer's premisis. Lead acid varieties of UPS do not seem to last more than a few years on average. It would be great to find a good LFP solution.
 
Is there any reason that I could not remove the worn out SLA battery from a small APC battery backup and replace it with a LFP battery? Obviously the UPS was designed for use with a SLA battery, but the BMS in the LFP replacement should keep everything happy...or at least safe, right?
Here we go again. First investigate the specs of the UPS before throwing money at it. Most have a garbage stepped-wave output that your protected electronics will hate. Additionally the H-bridge is designed to overheat just as the standard battery runs out of power. Adding expanded batteries will not extend the runtime. Even if you run at a fraction of the rated power, the internal timer may ignore the battery voltage and shutdown the UPS anyway.

If you are going to spend money, I would instead look into an “all-in-one” unit that already has the solar input. You can grow the system into a proper prepper. Watch this:
 
At a slightly higher cost, you could go the route of building a diy solar generator, good for daily protection of your equipment (computer, tv system), and good for camping or such.

Basically, just need LiFePO4 battery, inverter, battery charger (pulls from your grid) ...

Advantage is more functionality, good diy project, we aren't reverse engineering UPS internals, etc. Just replacing older tech with newer tech.

Hope this helps ...
 
Here we go again. First investigate the specs of the UPS before throwing money at it. Most have a garbage stepped-wave output that your protected electronics will hate. Additionally the H-bridge is designed to overheat just as the standard battery runs out of power. Adding expanded batteries will not extend the runtime. Even if you run at a fraction of the rated power, the internal timer may ignore the battery voltage and shutdown the UPS anyway.
Yep these things are all possible. In the case of my Cyberpower 1500, it's a pure sine wave model. I saw it had a temp sensor on the output transformer and a fan that ran when inverting. Then I tested the bigger battery while keeping a close eye on it. At a mere 10% load it didn't generate much heat. The estimate of time remaining was comically bad for the whole 4 hour runtime but in the end it only cared about going until it hit the low voltage cutoff.
 
Yep these things are all possible. In the case of my Cyberpower 1500, it's a pure sine wave model. I saw it had a temp sensor on the output transformer and a fan that ran when inverting. Then I tested the bigger battery while keeping a close eye on it. At a mere 10% load it didn't generate much heat. The estimate of time remaining was comically bad for the whole 4 hour runtime but in the end it only cared about going until it hit the low voltage cutoff.
Most but not all are bad. Love my APC “XL” models that are pure sine, low-frequency for starting motors, rated for continuous operation and have external battery connectors. Getting difficult to find nowadays. Downside they are heavy, loud and not very efficient at low loading.
 

Attachments

  • ups.jpg
    ups.jpg
    97.8 KB · Views: 8
  • 35AH battery.jpg
    35AH battery.jpg
    88.8 KB · Views: 7
Most but not all are bad. Love my APC “XL” models that are pure sine, low-frequency for starting motors, rated for continuous operation and have external battery connectors. Getting difficult to find nowadays. Downside they are heavy, loud and not very efficient at low loading.

I do love my APC unit - the external battery config units are getting harder to find it seems. I picked up one of those "twin" sets where there's a whole addon battery designed for it a few years ago. Luckily I have another year or two before I need to rotate the batteries out of it.

I have two cyberpower 1500va units in my living room running my synology and PoE/security gear. And (feels good to be able to say it now) whole home is running on UPS backup from two EG4 Indoor units, so I'd need to burn through 28kWh worth of battery before the smaller UPS's even get touched.
 

Attachments

  • 2024-05-20 21.27.38.jpg
    2024-05-20 21.27.38.jpg
    115.5 KB · Views: 5
Back
Top