diy solar

diy solar

Alternative battery backup for 120V desktop computer

mcrn22

New Member
Joined
Sep 21, 2022
Messages
11
Hello! I have some APC BN1500M2 UPS I use for desktop PC and network equipment around my house.
The SLA batteries inside only last a couple years, and I have been using 12V 7Ah Mighty Max replacements.

I want to invest in something with greater longevity and more runtime.
It seems like a solar generator with LiFePO4 batteries could be a good option.

I am having trouble figuring out how a solar generator would communicate battery state and charge level with PC.

The APC UPS has a USB communication port that can be connected to the USB port on a PC. (Windows 10)
It tells the PC how much charge is remaining and what the power source is.
APC-UPS.png

When the remaining charge becomes too low, the PC can shut itself down automatically.
Power.png

Every UPS with more than 1000VA from APC, CyberPower, and Triplite I have ever seen includes such a communication port, but I have not been able to find any documentation about such a feature in any solar generator specs.

Does anyone know if there is a solar generator with such a communication port feature?
 
I would be tempted to replace the SLA with an LFP battery. Check the charging profile and it might be fine. The goal would be to use the existing UPS and software and just have a giant battery.
 
I would be tempted to replace the SLA with an LFP battery. Check the charging profile and it might be fine. The goal would be to use the existing UPS and software and just have a giant battery.
Thank you for that suggestion. Does this video describe what you mean?
 
I've thought about the same thing. Building a beefy UPS type power supply for my computer, however there would be no software control. Just a 100ah lifepo4 battery, 500w inverter and a mains charger big enough to replenish the battery in a few hours. I'd be monitoring the battery voltage manually or using a low voltage alarm.
 
Last edited:
Just be aware a big battery may make those units overheat and shut down. They are designed expecting the battery to die in X minutes so they don't bother with cooling too much.
 
maybe build a inverter/battery system and then plug a cheap UPS into the inverter, then plug the computer into the UPS.
 
I measured the APC BR1500MS2 charge voltage at 27.27V with the battery disconnected.

BR1500MS2_charge_voltage.jpg

The OEM replacement battery contains two 12V SLA batteries in series.
The third-party batteries I have been using say the float voltage is 13.6V - 13.8V.
So, I believe the OEM battery float voltage must be 27.2V - 27.6V.

I purchased a 24V 25Ah LiFePO4 battery that seemed like it might work.

ZeroFuel_24V_25Ah_voltage.jpg

The description says the nominal voltage is 25.6V, max voltage is 30.0V, and min voltage is 21.0V.
It says the continuous charge rate is 10A, continuous discharge rate is 35A, and peak discharge rate is 45A.

The APC BR1500MS2 max output is 900W. (7.5A at 120V)
If my understanding of all this is about right, then the draw on the battery could be 37.5A at 24V.
So, I used 8 AWG wire and crimp connectors to extend the battery cables.

BR1500MS2_standby_voltage.jpg

Then I left it to charge for 36 hours.
 
If you factor in the inverter efficiency, 85% is typical, then the dc current will be about 45A.
 
Hmm. I see. I guess I better not try to run it at full load then.

I tested it at 650-watts, and it ran for 6 minutes and 6 seconds.
 
These small UPS are not rated to run continuously for a long period of time without over heating, it's made to run just long enough to do proper computer shutdown until the small battery runs down.
 
UPS error or dead battery?
The APC UPS behaved as though the battery was depleted, but I am not convinced that is the case. The battery low voltage disconnect is 21.0v, and I never saw the voltage get below 24.9v. I do not know how to do the math to calculate the theoretical time that 25Ah 24v battery should be able to sustain a 900-watt load at 120v.

I plugged the APC UPS back into the wall, and the 650-watt space heater came back on, and the battery started recharging. I could not feel anything getting hot, or even slightly warm. It does have a cooling fan inside that I heard kick on when the inverter started up, but the exhaust air was not noticeably warm.

I guess I could try the same test with a SLA battery just for comparison. I wonder if the APC UPS could have some sort of predefined timer based on the load instead of just referencing the battery voltage. I find it hard to believe that the inverter is actually supposed to shutdown at 24.9v.

I should point out that I am not disappointed by these results. At this point I have a UPS that will shut down my computer automatically. I should not have to worry about replacing the battery for at least the next decade, and I should be confident that the capacity will not degrade so significantly during that time.

I was hoping for more runtime since I spent $225, and compared to a Bluetti EB3A it seems like I am not able to access the majority of the battery capacity. But whatever. This works.
 
Would be interesting if it went twice the time on 325w load. Or maybe there is a simple timer based on the original spec battery capacity.
Maybe plug it in 2 minutes and see if the timer resets to get an additional 6 minutes.

Could get a small inverter/charger to go with the battery to run until dead.
https://www.amazon.com/Tripp-Lite-Powerverter-Aps700hf-Inverter/dp/B00AM2HHF2
 
Last edited:
Battery: 25 amp-hours * 24 volts = 600 watt-hours
Test: 650 watts * 366 seconds ÷ 60 ÷ 60 = 66.1 watt-hours

Seems to suggest that the space heater used about 11% of the battery's described capacity.

If the battery really is 25Ah, and that math is right, then this UPS and this battery are a horrible match!
 
For anyone that ends up here seeking the same thing I am:

What I want does not currently exist under $400, and a DIY solution is not feasible.

I am unable to find a portable LiFePO4 power station that switches between utility power and battery power quickly enough to act as a UPS. Even if that were not an issue, I am unable to find one that supports RS485 communication that desktop computers use to communicate with battery devices.

Nor have I been able to find an automatic transfer switch with sufficient speed to assemble a DIY solution. Relay based options are even slower.

The only solution under $400 is to continue using traditional UPS and regularly change the SLA batteries. The UPS could be connected to a portable LiFePO4 power station for more runtime.

The only UPS I have been able to find that do not use SLA batteries are the APC Smart-UPS Li-Ion products which start at $460, use lithium-ion battery chemistry, and are intended for rack-mount applications.

Rather than invest in those, I would be better off saving up for a low-frequency inverter/charger and a few hundred amp-hours worth of LiFePO4 batteries.

Thanks to everyone for your helpful input.
 
Would be interesting if it went twice the time on 325w load. Or maybe there is a simple timer based on the original spec battery capacity.
Maybe plug it in 2 minutes and see if the timer resets to get an additional 6 minutes.

Could get a small inverter/charger to go with the battery to run until dead.
https://www.amazon.com/Tripp-Lite-Powerverter-Aps700hf-Inverter/dp/B00AM2HHF2

IMG_20230318_123109539_HDR.jpg

I got a 24v 1000w inverter and a 24v 10A charger.
I charged the battery up to 27.2v and then connected a 2A load to the inverter. It sustained power for 2 hours and 44 minutes.

240 watts * 164 minutes ÷ 60 = 656 watt-hours
656 watt-hours ÷ 24 volts = 27 amp-hours

Oh I also got a 24v fuse block and used a 15A fuse for the charger.

So at least the battery seems fine.
 
View attachment 140173

I got a 24v 1000w inverter and a 24v 10A charger.
I charged the battery up to 27.2v and then connected a 2A load to the inverter. It sustained power for 2 hours and 44 minutes.

240 watts * 164 minutes ÷ 60 = 656 watt-hours
656 watt-hours ÷ 24 volts = 27 amp-hours

Oh I also got a 24v fuse block and used a 15A fuse for the charger.

So at least the battery seems fine.
BTW, what is the power draw of that 24V 1000W inverter with no load connected?
 
BTW, what is the power draw of that 24V 1000W inverter with no load connected?
I do not have an amp meter, so I am not sure how to measure that. I checked the user manual that came with it, but I cannot find that information.
 
Back
Top