goDofWar_skr
Herald
Sounds interesting..
could you please let me know what it does and how one would learn more about it..
Modify UPS with Li-ion/LiFePO4 battery and BMS
- Thread starter goDofWar_skr
- Start date
could you please let me know what it does and how one would learn more about it..
variablevector
Forerunner
There's a lot of videos on YT about doing this. Basically, you'll get much higher energy density and lifespan than a lead acid battery. With LiFePO4, it should be safer too.
The difficulty is matching up the charging/discharging voltages that the UPS uses, which are only designed for lead acid with what the LiFePO4 needs. But a 4S LiFePO4 pack will have a nominal voltage of 4*3.2=12.8V and a charging voltage of 4*3.6=14.4V. Afaik, that's pretty ideal and compatible with the UPS. The discharge cut-off voltage for a LiFePO4 is also lower than that of a lead acid battery, so you should technically be fine there too. Another problem is drawing too much current from it, because LiFePO4's are mostly limited to 0.5-1C where C is the capacity of the battery. A BMS will help manage some of these variables and protect the battery, just in case. But I've never done this, so I could be wrong about the specifics. If I do end up doing it, I'll let you know.
goDofWar_skr
Herald
Looks very interesting!
I wonder why the manufacturers don't do this by themselves in the first place... Maybe the cost of the LiFePO4 batteries?
I wonder why the manufacturers don't do this by themselves in the first place... Maybe the cost of the LiFePO4 batteries?
Anyone has any experience with these folks -
They sell LiFePO4 ups and battery packs.
They sell LiFePO4 ups and battery packs.
variablevector
Forerunner
BAK IFR 32140FS 3.2V 15000mAh LiFePO4 Rechargeable Battery 2C - Grade A
The BAK IFR 32140FS is a high-performance rechargeable lithium iron phosphate (LiFePO4) battery designed for applications such as electric vehicles, renewable energy systems, and industrial equipment. With a capacity of 15000mAh and a maximum continuous discharge rate of 2C (30 amps). LiFePO4...
quartzcomponents.com
These look kind of nice. 4 of them would cost 2.5k and when wired as 4S1P, you get a 12.8V 15Ah 192Wh battery that can do a continuous discharge of 2C i.e 30A. That's up to 384W. Not entirely bad for a UPS application, I think. By a conservative estimate, that should be able to run a 100W load for more than an hour or a 200W one for at least 30 minutes.
Combined with a 4S 30A balancing BMS like this one (not entirely sure about that website), that should make for a pretty good LiFePO4 battery pack under ~3k worth of parts. Although with the dimensions of these 32140 cells, it might be a tight fit into the UPS.
Hey @rsaeon as the resident battery expert here, is this a good idea?
rsaeon
Innovator
I really like those cells and they would be well suited for a low-load UPS with the BMS you found. I considered them for a powerwall project but this being India, the really good deal is in used LFP cells - they're usually priced at under 2000 for a 80Ah cell! (India has pretty great second-hand pricing for basically everything).
I've been hooking up my series-connected-cells-without-a-bms to the Inverter during extended power cuts and recharging each cell up manually afterward, LFP's efficiency is very impressive.
Most basic BMSs have a cut-off at 3.75V so those should be avoided, the one you found has the proper over-voltage and over-discharge thresholds.
Yes, this is the one drawback of using a LFP pack with any UPS/Inverter regardless of whether it was made for LA or Lithium batteries. 1C is usually the maximum, with some (older) packs allowing peaks at 3C to 5C.
However, some people on youtube have noticed some BMSs will block the UPS from powering on after the pack is drained completely. They got around this by installing a push button that'll bypass the BMS wiring to allow the UPS to sense that there's a battery so that it will power on. This is for UPS models that need a battery to turn on even if they're connected to the mains ac. Most UPSs will let you turn on without a battery with a long press of the power button.
I've been hooking up my series-connected-cells-without-a-bms to the Inverter during extended power cuts and recharging each cell up manually afterward, LFP's efficiency is very impressive.
Most basic BMSs have a cut-off at 3.75V so those should be avoided, the one you found has the proper over-voltage and over-discharge thresholds.
Yes, this is the one drawback of using a LFP pack with any UPS/Inverter regardless of whether it was made for LA or Lithium batteries. 1C is usually the maximum, with some (older) packs allowing peaks at 3C to 5C.
However, some people on youtube have noticed some BMSs will block the UPS from powering on after the pack is drained completely. They got around this by installing a push button that'll bypass the BMS wiring to allow the UPS to sense that there's a battery so that it will power on. This is for UPS models that need a battery to turn on even if they're connected to the mains ac. Most UPSs will let you turn on without a battery with a long press of the power button.
variablevector
Forerunner
Great! In that case, I'll pull the trigger on them and post an update here when I'm done with the project. It sounds fun.
I've been playing around with the idea of replacing my home's current power backup which has proven itself woefully inadequate for when the power goes out for more than 2-3 hours (1x150Ah Luminous LA and a 900VA Luminous inverter). I saw your post on the FE105 Narada cells from MDS Enterprises and I was already thinking about buying 8 of them and trying to build a 12V200Ah or 24V100Ah LFP battery first. Then eventually upgrading the inverter to a 24V 2-3 KVA Solar Inverter and then maybe only later getting some solar panels. Basically doing it in stages and cutting corners wherever possible so I don't have to spend all of that money upfront.
So the idea of used cells actually being cheap in India is very interesting. Where do you get them? OLX?
Ah, that's an easy/clever solution! I'll need to test the cheapo APC 900VA I bought to see if it does that or not. But based on what others were saying, it doesn't seem to power on with a dead battery. I assume that means l'll have to do this modification for it. Thanks for the heads up.
rsaeon
Innovator
@variablevector sorry, I forgot about your PM and your reply here.
The dealer is https://rojgarbox.com/ I last spoke to them in January:
The dealer is https://rojgarbox.com/ I last spoke to them in January:
Attachments
variablevector
Forerunner
No worries. I contacted the MDS guy and he quoted me 7.5k for a Hithium 280Ah "10,000" cycle cell and 3.5k for a Gotion 102Ah cell. I'm thinking of just getting 4x Hithiums for now and calling it a day. They're not that much more expensive than getting lead acid of the same capacity and probably more reliable in the longer term than these used cells you mentioned -- or so I would assume.
I'm sure with the recent drop in LFP prices that you should be able to get them for even cheaper from China, I see people throwing around prices of $50-$60/cell on diysolarforum but none of the Alibaba sellers I contacted were ready to ship to India either.
variablevector
Forerunner
I finally got the cells and wired this up. The battery pack looks like that because the 32mm cell spacers I got were too small for these "32140" cells. They're more like 33140 cells and need 33mm spacers. I taped them together with kapton tape for now. Yeah, I know it looks like shit, but it works. They're connected in parallel to the 9Ah lead acid inside the UPS. Connecting LA and LFP in parallel might seem insane but it's actually alright and may even be better for the longevity of the LA, provided the LFP has a BMS. There's a whole discussion about why this is (LFP's higher resting voltage keeps LA at float voltage constantly and LFP will always drain first due to lower IR) and there are a few people online who've tested it for years with much bigger battery systems.
Unfortunately, I found out that the APC UPS is trying to charge the batteries at 14.62V which is 0.02V above what the BMS on the LFP cuts off the battery at. The result is that the LFP only gets charged during the float charging mode, and only up to the float voltage of the LA i.e 13.5V. On the whole though, this actually isn't a big problem considering that is like 90%+ capacity for them. I wasn't sure how this would impact the charging time, but they happily charged from empty to full overnight.
I tested them with a load of about 150-200W (my gaming PC and 27" monitor at mostly idle, just playing Youtube in a chrome tab) and I got 1 hour 33 minutes of run time.
For reference, the brand new lead acid by itself only gets 33 minutes in the same test. And that was without the monitor connected to it!
Considering this is now a 9*12 + 15*12.8 = 300 Wh battery pack, I'd say that is exactly around the runtime you would expect. Neither too low nor violating the first law of thermodynamics.
After a complete discharge, the LFP was at a resting voltage of 11.21V and 11.69V for the VRLA lead acid. Both within spec. Both started charging after connecting to mains again.
So a 3x improvement in backup time with ~3.5k in parts. Easily scalable further by adding more cells. Not a bad deal, I think. @goDofWar_skr
rsaeon
Innovator
Sorry, I forgot to mention this, yes those cells need different spacers and robu.in usually has them in stock (out of stock now): https://robu.in/product/2-x-32140-33140-battery-holder-with-33mm-bore-diameter/
Some sellers differentiate these cells from the other 32650/32700 ones by calling them 32140FS. UTL actually uses these cells in their products, I remember seeing them in a youtube video about their testing lab.
This sounds like it should be fixable, either with a longer cable or by perhaps adding/adjusting a trimpot inside the UPS? This has been possible on other models: http://www.jjoseph.org/notes/apc_smartups_battery_float_voltage
That's an excellent upgrade, thanks for sharing!
goDofWar_skr
Herald
Nice! and it will last longer than the lead acid batteries too I suppose..
Hoping that someday I get the time and motivation to try something like this myself
variablevector
Forerunner
Yeah. But most things on robu seem perpetually out of stock. Seems the only way to get these spacers right now is directly from China.
That's what I was thinking too. I was going to add some XT90 connectors and make the battery detachable anyway, so adding some longer cabling seemed reasonable. But that is the same as adding a low resistance resistor in series, right? And in order to get a voltage drop of 0.03V at 2A charging, you would need a resistance of 0.015 Ohms. But while discharging at 20A, that would create a voltage drop of 0.3V. That seems not ideal. It probably depends on whether the gain in the top end would justify the loss on the bottom end.
That article is great. I didn't see any trimpots inside the UPS, and this being their most basic model, there definitely is no serial port for communications. But I was just going to keep them at 13.5V because according to this paper, cycling LFPs at a lower SoC range causes less capacity fade than cycling them at the top SoC range. They're claiming that the longer you keep them at a higher voltage/SoC, the more lithium inventory loss occurs, which is a degradation mode for LFPs.
Making this post to bookmark this thread as I'm going to need to refer to it later, as I bought an APC BR1000G-IN yesterday only because it had the option to add an external battery pack, and I will be doing so in the near future as I have to leave home for a fortnight with family and need a robust power backup solution to keep the CCTV-DVR and WiFi router up and running in case of a power outage.
idestinytechlab
Contributor
Check myinnovation.in - I got LiFePO4 12V Battery for my aquarium backup power supply and using it from past 2+ years and backup is still great. Not feeling any difference.
rsaeon
Innovator
Had an extended power cut the other day, so I needed to hookup some LiFePO4 cells in parallel with the Amaron Quanta batteries to make it through.
I used a set of car battery jumper cables but they warmed up really quick so I added in a second set of cables in parallel.
It wasn't ideal (the LFP cells were charging the LA batteries and providing power to the inverter) and it looked awful but it worked:
I used a set of car battery jumper cables but they warmed up really quick so I added in a second set of cables in parallel.
It wasn't ideal (the LFP cells were charging the LA batteries and providing power to the inverter) and it looked awful but it worked:
Attachments
idestinytechlab
Contributor
For how much you did the Prismatic cells? is it from the conversation screenshot from RojgarBox?
rsaeon
Innovator
Those cells are from my long-running project:
RojgarBox would never reply on time whenever I have funds so I haven't purchased from them. Just MuscleGrid, MDS Enterprises and Myinnovation.in
Taking a detour from:
techenclave.com
Led me to exploring this option. Most 48V server rack style batteries available abroad are in a 16S configuration and having learned my inverter charges at 14.2V per 12V battery, for a total of 56.8V, divided by 16 resulting in a per cell voltage of 3.55V which is pretty good for a LiFePO4. Bumping it up to 14.6V gives us 3.65V which is pretty much the recommended charging voltage for this chemistry. Float at 13.9V works out to be 3.475V which is a little over the 3.4V...
Working with Inverter/UPS Lead Acid Batteries
I'm going to consolidate my experiments and adventures with inverter batteries here. Previous posts on the topic: https://techenclave.com/threads/diy-wi-fi-battery-monitor-for-inverters-ups-iot-lead-acid-battery.217210/...
techenclave.com
Led me to exploring this option. Most 48V server rack style batteries available abroad are in a 16S configuration and having learned my inverter charges at 14.2V per 12V battery, for a total of 56.8V, divided by 16 resulting in a per cell voltage of 3.55V which is pretty good for a LiFePO4. Bumping it up to 14.6V gives us 3.65V which is pretty much the recommended charging voltage for this chemistry. Float at 13.9V works out to be 3.475V which is a little over the 3.4V...
RojgarBox would never reply on time whenever I have funds so I haven't purchased from them. Just MuscleGrid, MDS Enterprises and Myinnovation.in