I did not know that batteries declined so much. Would you would replace personal UPS batteries every two years if they were at your residence?
Power for home lab/office
- Thread starter Adcadet
- Start date
I did not know that batteries declined so much. Would you would replace personal UPS batteries every two years if they were at your residence?
jtr1962
Storage? I am Storage!
In my personal experience lead acid batteries are sh*t after 2 or 3 years. That's one reason I stopped bothering with UPSes. Besides the cost, it's creating more electronic waste. And we almost never have power outages lasting more than a few seconds*. To me LiFePO4 are the only type of battery which makes any sense in a UPS, and even then they should be floated at no more than 3.4V per cell. That might require adjusting the float voltage if you're swapping them in place of a SLA battery.
* This got me thinking about a UPS based on using supercaps instead of batteries. The idea is enough capacity for perhaps a minute, basically to get you through the much more common very brief power interruptions, as opposing to a full-on blackout. Since supercaps are rated for at least 6 figures of cycles, and don't degrade with time, this would be a permanent solution. If you want to back up a system drawing 500 watts maximum, you might need supercaps storing perhaps 50,000 joules. For comparison purposes this is roughly the energy contained in one 18650 cell. To that end about 24 of these would be sufficient. Cost would be only $100 plus the cost of a balancing board. Maybe put them in a 6S4P arrangement which gives you a 333F, 16.2V supercap capable of storing about 43,700 joules. Assuming an inverter efficiency of 80%, you'll get about 35,000 joules at the AC outlet. This could run a 500 watt system for 70 seconds. It could run mine (average under 100 watts including the monitor) for almost 6 minutes.
It depends on where you live, but I experience a lot of short outages and then several that last a few hours. It's a PITA when your NASes are dirty from uncontrolled shutoffs and it is necessary to scrub them or whatever. I have a larger battery supply (~588Whr) in the main NAS simply because I want sufficient run time to operate a computer and accessories intermittently until power is restored. Generally I shut down after about five minutes. I'll look into a shorter replacement cycle.
In some of the lower load devices it may be less critical if I understand correctly that the performance under high loads will deteriorate sooner. I may not notice if a battery that usually runs for 12 hours only runs for 6 hours as compared to batteries in the main NAS that are draining at 800W.
In some of the lower load devices it may be less critical if I understand correctly that the performance under high loads will deteriorate sooner. I may not notice if a battery that usually runs for 12 hours only runs for 6 hours as compared to batteries in the main NAS that are draining at 800W.
Last edited:
Mercutio
Fatwah on Western Digital
FWIW, a lot of server systems I support need more than five minutes to shut down properly. 90 seconds for a desktop to shut down? Cool. VM or database host? Not so much.
The batteries I use are generic but I haven't found OEM batteries to work any better.
The batteries I use are generic but I haven't found OEM batteries to work any better.
jtr1962
Storage? I am Storage!
Easy answer is the marketing people would have a hard time selling a UPS which only lasts for a few minutes at best. Most people are idiots who look only at bigger numbers, even though in many situations a UPS which lasts a 1 to 5 minutes would cover 99% of power interruptions.
You could even do a hybrid with a battery and supercaps. You might be able to sell those based on their ability to cover brief power interruptions even if the battery fails.
Voltage isn't an issue with supercaps since you can simply put as many as you need in series to reach the required voltage. A 24V system might have a 12S2P arrangement of the caps I mentioned.
You could even do a hybrid with a battery and supercaps. You might be able to sell those based on their ability to cover brief power interruptions even if the battery fails.
Voltage isn't an issue with supercaps since you can simply put as many as you need in series to reach the required voltage. A 24V system might have a 12S2P arrangement of the caps I mentioned.
jtr1962
Storage? I am Storage!
The use case for what I describe wouldn't be to execute a shut down. Rather, it's to simply carry you through brief power interruptions lasting usually only a few seconds. Sort of like when your lights blink off, then come back on not long after. It's still annoying in that lots of devices have to reboot after that happens. They should have some sort of backup so that's not the case.
Aren't the servers supported at the datacenter where generators will kick in after some seconds? Or are they small companies that have old stuff in the backrooms?
jtr1962
Storage? I am Storage!
Kind of what I was thinking. If you have backup generators, a UPS like I described, which has a few minutes run time, is perfect. That's more than enough to tide you over until the generators kick in.
Mercutio
Fatwah on Western Digital
Not always. Some businesses still want hardware on premises. Data centers generally have power sorted but sometimes they charge extra for access to the giant batteries or fancy real time power monitoring.
The last nasty long term outage where my servers are, the pumps that feed diesel into the on site generators failed, so one of the Ops had to spend a couple days pouring fuel in by hand until they could replace it.
If you're supposed to keep things running always, there's no such thing as being too careful. Shit happens.
jtr1962
Storage? I am Storage!
Speaking of things running always, I took this today:
It's apparently been at least 206 days since our last power blip, probably longer. Last time I may have rebooted after installing software, not because of any power outages. The electricity here is pretty reliable. Only two causes of outages lasting more than a few seconds. Sometimes the wind causes branches to break off and take down local power lines. That's usually fixed within a few hours. It only happened 3 or 4 times in the 45 years we've been here. The other cause of course is major blackouts, usually weather related. Last one was during Hurricane Sandy when we lost power for about 4 days. I'm sure the Internet was out also, so even if my PC had power it wouldn't have been terribly useful.
If they underground the lines, probably the only time blackouts will occur is when the entire grid goes down.
It's apparently been at least 206 days since our last power blip, probably longer. Last time I may have rebooted after installing software, not because of any power outages. The electricity here is pretty reliable. Only two causes of outages lasting more than a few seconds. Sometimes the wind causes branches to break off and take down local power lines. That's usually fixed within a few hours. It only happened 3 or 4 times in the 45 years we've been here. The other cause of course is major blackouts, usually weather related. Last one was during Hurricane Sandy when we lost power for about 4 days. I'm sure the Internet was out also, so even if my PC had power it wouldn't have been terribly useful.
If they underground the lines, probably the only time blackouts will occur is when the entire grid goes down.
ddrueding
Fixture
Power here is so reliable that even when battery systems are installed they don't normally install a standalone isolator switch. No spikes either. Everyone just plugs direct to the grid; apparently it has been years. Amazing what happens when you put all your power infrastructure underground in serviceable tunnels and have a distributed generation system.
Mercutio
Fatwah on Western Digital
Where I live, it's all overhead lines. We have a lot of tall trees. This means we can have power outages just from WIND.
The people are squeezed into a relatively small area in Denmarken so power distribution is less of a concern over distances than in many parts of the states. In live in an underground area yet there are still power outages, perhaps from farther away. Sometimes transformers fail in various locations and sometimes there are trees. I've lived in many places and the immediate surroundings don't seem to correlate to the outages.
A few years ago one branch of the company had an on-site infrastructure failure and destruction of hardware so nobody could access systems. Then they removed most of IT staff and all equipment from the region. Now everything (both computing and storage) is on the global Amazon which supports all Regions. I have no idea how it works, but supposedly it can handle local disruptions and natural disasters with full availability, almost anything but thermonuclear events. If it conks out, it's not my problem anyways.
Now a completely different UPS shut down and took out everything with it, even the clock. 
The TV and YAMHAH are not connecting and energy outlet is not triggering itself. At least the main STB is on Lithium 12V.
The TV and YAMHAH are not connecting and energy outlet is not triggering itself. At least the main STB is on Lithium 12V.
Everything is on again after removing all power and batteries for a while, but it's not clear why a YAMAHA was not communicating with the box on Lithium. I guess everything is supposed to be in a certain order or something. Maybe I should just remove that UPS from service for now.
How do you test the UPS batteries? My battery thingamajig is suggesting C/20 and somebody suggested as much as 4C (576W for 12Ah), which is way beyond the power I can handle. Besides, isn't that wearing out the battery just by testing it?
jtr1962
Storage? I am Storage!
Basically, the UPS tells me the battery is crap when the red "no battery" light starts flashing. At that point when I test the battery it's pretty much an open circuit. Sadly, this seemed to happen within 3 years almost all the time.
Doing a capacity test on a battery every few months won't wear it out. For lead acid, it's probably a good thing actually as they degrade just being floating without discharging. You need a really hefty electronic load to discharge a big SLA battery at 4C. For someone who only does this occasionally, it's not worth it.
I might try LiFePO4 in one of my UPSes. I have a bunch of surplus A123 26650 cells. I'll need to put them in a 4 series arrangement, with as many in parallel as will fit in my UPS, then add a BMS. Also if possible adjust the UPS float voltage to something like 13.6V.
jtr1962
Storage? I am Storage!
It's already at 83% capacity. Generally 80% is when it's time to consider taking a battery out of service. I guess ~3 years is pretty much the lifetime for SLA batteries in UPS service.
BTW, LiFePO4 would probably hold 12.8 to 13.2 volts for most of the discharge. The nice thing about LiFePO4 is the very flat discharge curve. Basically, it would stay above 12V until you're about 95% discharged.
BTW, LiFePO4 would probably hold 12.8 to 13.2 volts for most of the discharge. The nice thing about LiFePO4 is the very flat discharge curve. Basically, it would stay above 12V until you're about 95% discharged.
I'm pretty sure that I've never seen the capacity at near 100% at C/4 or C/6. VRLA batteries are typically rated at C/20 and of course they are never quite 100% even then. I think that one above is about 90% of new, so still good for the relatively low load purpose.
I bought some weird LiFePO4 battery thing a few years ago. It was like this when new.
IIRC the max curent was rather low, so it was not very useful for a full power UPS. I was planning to use it directly for the 12V telecom devices. I'm not sure where it is now, probably stone dead from disuse.
IIRC the max curent was rather low, so it was not very useful for a full power UPS. I was planning to use it directly for the 12V telecom devices. I'm not sure where it is now, probably stone dead from disuse.
jtr1962
Storage? I am Storage!
Correct that lead-acid is rated at C/20. Most likely that battery is still good for another year, at least.
That's where the A123 26650 cells come in. 50A continuous discharge, 120A pulse discharge. Looks like 2 of these can fit in a 2.56" width. Put them in a 4S2P arrangement and you get 100A continuous, 240A pulse.
Could be that the UPS they're in doesn't cook them as much.Now I'm getting 89.2% at C/4 for some batteries that I'm almost sure are from 2019. For some reason they are doing better than the 2021 batteries at C/6.
jtr1962
Storage? I am Storage!
Definitely not A123 cells. Probably China cells. That said, the battery might still be OK. I just revived 2 old Craftsman power screwdrivers which had dead NiCd batteries using A123 cells I got in 2006. I had to add an LiFePO4 charging board and USB C board, but now they work great, and are USB rechargeable to boot. I remember the NiCd took overnight to recharge. These recharge from dead to full in under 1.5 hours.
The 2019 batteries are external in a 21°C room. Maybe that is part of it?
jtr1962
Storage? I am Storage!
Definitely. The high temperatures in a UPS cook the heck out of batteries.
No, they are cheap stuff. I discovered there is no good way to use it as a UPS
So where does one buy a good UPS battery that is safe and legal? From what I read the A123 are Chinese, maybe good Chinese.
ddrueding
Fixture
I have a couple CATL batteries in my workshop for an EV conversion project. They have very impressive capacity, but wouldn't fit in a PC rack
So far I just have the BMS wired in and managed over CAN bus, charging circuitry is next.
So far I just have the BMS wired in and managed over CAN bus, charging circuitry is next.
So I tested the internal 12V 14Ah battery and capacity results were great. The OEM was 12Ah IIRC, but after 17 hours of charging the 2021 battery it is at 13.17V. That seems like forever to charge.
I have a couple more smaller batteries to test assuming another UPS charges them enough. At 13.56V will that fully charge a 12V battery in human times?
Is there a preferred brand of SLA batteries people use in UPSes? I don't want to mess with retrofitting a LFA into my CyberPower. My CyberPower OR1500PFCRT2U OEM battery pack is a 4 x 12v 7AH pack that costs ~$200 (on Amazon) but I can find individual cells for around $40/each locally. I'm debating trying to rebuild the pack versus just getting the OEM version. The OEM battery does come with 18month warranty vs 12month for individual cells.
I sent you a conversation. I prefer local since you can check the battery voltages in the store. Lead batteries that go too low are permanently damaged and if you get an old battery online it is a PITA or impossible to return. Local stores also take the old batteries for recycling.