Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Lithium Ion UPS
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Lithium Ion UPS

randvegetarandvegeta Member, Host Rep

Hello all,

I just bought 150kwh worth of lithium ion batteries. 28 Tesla modules, 5.3kwh and 25v each.

Max voltage = 25.2
Min voltage = 18

I also bought a bunch of inverter/chargers that can handle Tesla's particular voltages, as well as have adjustable cut offs so I can limit max SoC to 80 or 90% min SoC to 10 or 20% to improve battery longevity.

I will end up having a bunch of old lead acid UPS, which of course operate on different voltage range. Im wondering if anyone here knows how to go about repurposing these older UPS to work with Lithium Ion voltages. All my lead acid UPS operate with 8, 16 or 20 12v batteries (96v, 192v, 240v).

If the old UPSs cant be repurposed, i guess they will need to be sold or tossed.

Same goes for our diesel generators. Better to add more battery capacity than maintaining a generator indefinitely.

«1

Comments

  • vimalwarevimalware Member
    edited October 2020

    Oh , this is going to be one interesting thread.

    Livestream the battery room. The potential pyrotechnics will break youtube.
    JK. I wish you all the best.

    Read up. Stay safe and prosper Mad Max.
    Edit: The Tesla modules should have decent safeties. BMS is baked in?

    Thanked by 1coolice
  • MrRadicMrRadic Patron Provider, Veteran

    You'll need to setup a separate inverter for the different types of batteries. With proper MPPT, they should be able to feed a single source based on demand. You can have them going into a single breaker box (with their own breakers).

    Thanked by 1vimalware
  • letboxletbox Member, Patron Provider

    If this what I’m understanding you shouldn’t mix those batteries together please keep noted that very dangerous don’t try it if you don’t know what you are doing.

    I’d recommended you to bought 2 Parallels inverter with builtin MPPT charger and make sure setup enough safety fuses/ breakers

    Regards

    Thanked by 1vimalware
  • jsgjsg Member, Resident Benchmarker

    @randvegeta said:
    Hello all,

    I just bought 150kwh worth of lithium ion batteries. 28 Tesla modules, 5.3kwh and 25v each.

    Max voltage = 25.2
    Min voltage = 18

    I also bought a bunch of inverter/chargers that can handle Tesla's particular voltages, as well as have adjustable cut offs so I can limit max SoC to 80 or 90% min SoC to 10 or 20% to improve battery longevity.

    I will end up having a bunch of old lead acid UPS, which of course operate on different voltage range. Im wondering if anyone here knows how to go about repurposing these older UPS to work with Lithium Ion voltages. All my lead acid UPS operate with 8, 16 or 20 12v batteries (96v, 192v, 240v).

    If the old UPSs cant be repurposed, i guess they will need to be sold or tossed.

    Same goes for our diesel generators. Better to add more battery capacity than maintaining a generator indefinitely.

    I doubt that your old UPS can be re-purposed except for some parts like display, interfaces, etc., a major reason being that (at least the) large brands optimize for low cost production. Another major reason being that the highest efficiency ones are resonant ones where the tank is closely specified to a given case and if that was designed for 12V then it either won't work at all or work very poorly with 18V.
    But there are good news too because nowadays we have e.g. SiC transistors (and GaN but they are too expensive for most cases) and can enhance efficiency quite a bit both in PFC and on the secondary side (actively switched rectification).
    So, if you happen to know a good electronics engineer with solid experience in that field he could design and build you a couple of UPSs with efficiency north of 98% total.

  • raindog308raindog308 Administrator, Veteran

    @randvegeta said: Same goes for our diesel generators. Better to add more battery capacity than maintaining a generator indefinitely.

    Generators can be refueled while they're operating. I once had a situation where we operated a DC for a month on diesel power without any interruption. Just kept adding fuel.

    You can't do that with batteries. They have their place but so do generators. I think you're confusing UPS and alternate power. UPS is there to clean power and keep power flowing until alternate power can take over in the case the mains fail.

    Thanked by 2risharde Pwner
  • @randvegeta said:
    Im wondering if anyone here knows how to go about repurposing these older UPS to work with Lithium Ion voltages.

    You don't want to do that. You don't know how the UPS electronics were designed. They could be dependant on the characteristics of lead-acid batteries, which are far more durable and tolerant of abuse than li-ion. In short even if you adjusted for the voltage difference the UPS could be unstable on battery under load or could shorten the life of your batteries.

    Thanked by 1vimalware
  • @rcxb said: You don't want to do that.

    He will do this. Do not try to convince otherwise. Just provide knowledge how to do that safely.

    In my POW, all equipment needed for adaptation will cost more than to buy normal acid batteries.

  • @LTniger said:

    @rcxb said: You don't want to do that.

    He will do this. Do not try to convince otherwise. Just provide knowledge how to do that safely.

    In my POW, all equipment needed for adaptation will cost more than to buy normal acid batteries.

    You got your own POW? Damn son.

  • @serv_ee said: You got your own POW

    Actually few of them. Volume 1 and 2.

  • @randvegeta said:
    Hello all,

    I just bought 150kwh worth of lithium ion batteries. 28 Tesla modules, 5.3kwh and 25v each.

    Max voltage = 25.2
    Min voltage = 18

    I also bought a bunch of inverter/chargers that can handle Tesla's particular voltages, as well as have adjustable cut offs so I can limit max SoC to 80 or 90% min SoC to 10 or 20% to improve battery longevity.

    I will end up having a bunch of old lead acid UPS, which of course operate on different voltage range. Im wondering if anyone here knows how to go about repurposing these older UPS to work with Lithium Ion voltages. All my lead acid UPS operate with 8, 16 or 20 12v batteries (96v, 192v, 240v).

    If the old UPSs cant be repurposed, i guess they will need to be sold or tossed.

    Same goes for our diesel generators. Better to add more battery capacity than maintaining a generator indefinitely.

    Hi,
    I suppose the modules were not taken from one car? Wikipedia says the Tesla S has the biggest capacity with 100 kWh.

    If so, my advice is to check the voltages (fully charged to discharged) and capacity of each cell. You want them well balanced.

  • randvegetarandvegeta Member, Host Rep

    @Baris said: I suppose the modules were not taken from one car?

    I purchased 2 packs. They should be fairly balanced straight out of the pack.

    @rcxb said: You don't want to do that.

    All that stuff about tolerances for Lead Acid not being suitable for Lithium Ion... yes I know that. That's why I asked this question to see if it was feasible. The UPS are fairly old anyway though, so it makes sense to retire them regardless. As I have said, I'm going to be using some brand-new inverters to replace the old UPS. But if I can also modify the UPS (safely) then I can have them as a backup.

    @raindog308 said: Generators can be refueled while they're operating. I once had a situation where we operated a DC for a month on diesel power without any interruption. Just kept adding fuel.

    We dont own the building. Generator is inside our unit. It occupies a huge amount of space, and if the power goes out, the lifts wont work and getting fuel would be tricky. Not impossible, but tricky. It would probably be easier/cheaper to rent a generator on a truck and throw some very thick cables out the window to supply power. But HK has extremely reliable power. So although I know a UPS and generator serve different purposes, the reality is that the generator has cost many times more the value it's provided. Adding batteries, I think, is cheaper, long term. If our UPS can last long enough to call the rental generator, then this is ideal.

    @MrRadic said: You'll need to setup a separate inverter for the different types of batteries.

    I dont think so. We will have multiple inverters any way, and I dont plan to mix different capacity modules. But even if the modules had different capacities, as long as the voltages for given SoC is the same, if they are in parallel, it should not be a problem (AFAIK).

    @vimalware said: Edit: The Tesla modules should have decent safeties. BMS is baked in?

    Each module has a BMB (mini BMS for each module). I thought I could use those, but that requires some advanced controller, which seems to cost quite a lot. I'm looking into other BMSs but they also seem expensive, and if I understand the requirements correctly, would end up costing me more due to the volume of modules. I found a controller that works with the native BMBs, upto 64 modules. About $2.5K. If I get another kind of BMS that seems popular with Tesla modules, the cost looks to be roughly $500 per BMS, which either works with just 1 module, or only a few modules. So either way, if my plan works out, the large controller for 64 modules would be the cheaper option.

    I don't yet have a BMS to install, so my plan is to set the max charge (voltage cut off) to about 60%. If the cells become unblanced, it shouldn't be too bad with a relatively low max SoC. I will of course manually check individual cell voltages regularly, and continue to do so until such time I have a BMS that can handle it for me.

  • If you are actually asking how to repurpose these UPSes to work with this, I think you need to go back to the drawing board. Monitoring cell balance by hand, really? I wouldn't want to be in the same building as this janky contraption you are planning.

    Frustrating that there are all of those youtube videos encouraging people to experiment with big strings of lipos. Wonder how many fires need to occur before it stops

    Thanked by 1vimalware
  • randvegetarandvegeta Member, Host Rep

    @JahAGR said: I wouldn't want to be in the same building as this janky contraption you are planning.

    Don't worry. You're not invited.

    But in all seriousness, the length of time the modules will be without a BMS will be at most a couple of weeks. Plan is to test the setup with the new inverters, and then to test with different kinds of BMS. When we know which we like best, we'll order enough for the foreseeable future.

    The voltage variance between 0% and 100% is actually kind of large. Keeping SoC at around 50-60% should give a fair bit of margin for any imbalances. Several volts across the module.

    The charger does not supply constant power to the batteries so the chance of over charging once the desired voltage is reach is nil. I can check the few modules being tested by hand every day.

    I have some experience reviving dead modules (from <1v for a whole module, back to 25v. Even these 'problematic' modules seemed to charge / discharge quite evenly. If I detect any imbalance at all within a short period of time, then it would actually indicate a problem with the module and so should not be used, or at least not used unattended.

  • How much did you spend per kWh and in which country?

  • randvegetarandvegeta Member, Host Rep

    @Baris said:
    How much did you spend per kWh and in which country?

    This is in Hong Kong. I got a good price, but I also fix and export these cars on the side so I get better access and prices than most.

    ~$50/kwh

    Thanked by 2vimalware Baris
  • Gamma17Gamma17 Member
    edited October 2020

    One note on "repurpose old ups" thing...
    If you look at lifepo4 voltages they match surprisingly well with lead-acid batteries. To the point when some companies actually produce and sell lifepo4 battery packs designed to be compatible with old lead-acid ups and other equipment.
    It is also possible to connect 4 cell lifepo4 battery without any extra circuitry directly to 12v lead-acid ups (more cells for more voltage as needed) and it will work, but charge voltage will be slightly lower than nominal (which will mean battery will stay slightly undercharged) and you still need some sort of balancing circuit.
    It is also much safer that liion/lipo batteries, does not catch fire when damaged etc.

    Might be a better way, since using whole bunch of liion batteries without proper protection/management circuitry sounds like recipe for disaster (at least do some thermal monitoring/cutoff, otherwise it will be far too dangerous to even consider)...
    Lower voltage also does not guarantee safety. I've seen old 18650 cells from laptop batteries which were bad enough to start heating up to dangerous temperatures before reaching nominal charge voltage (while charged with really small current, like 1/10C), luckily i was using charger which had thermal monitoring...

    Thanked by 2randvegeta vimalware
  • randvegetarandvegeta Member, Host Rep

    @Gamma17 said: Lower voltage also does not guarantee safety. I've seen old 18650 cells from laptop batteries which were bad enough to start heating up to dangerous temperatures before reaching nominal charge voltage (while charged with really small current, like 1/10C), luckily i was using charger which had thermal monitoring..

    The modules I'm using are in good condition to begin with. They are not 'revived' cells, so to speak.

    Again, I fully intend to use a BMS to monitor cell voltages and temperatures.

    I will likely run 4 modules per 6kw inverter. So even at full load, the max power draw will be only ~0.3C. The charger is rated for 85amps, so ~0.1C. I may even run more modules on a single inverter, so the C rating will be even lower.

    If need be, I may also consider hooking up a reservoir of ethylene glycol and pump through the coolant tubes on the modules to regulate temperatures better.

    But a BMS to balance the cell voltages is of course a given. Just need to get the right ones.

    @Gamma17 said: It is also possible to connect 4 cell lifepo4 battery without any extra circuitry directly to 12v lead-acid ups (more cells for more voltage as needed) and it will work, but charge voltage will be slightly lower than nominal (which will mean battery will stay slightly undercharged) and you still need some sort of balancing circuit.

    It is also much safer that liion/lipo batteries, does not catch fire when damaged etc

    Really? Don't you need to cut off the power once full (or near full)? Lead acid can seemingly be connected to a permanent power source, but lithium cells apparantly keep on charging, even if the voltage is below their rating (weird...). Does lifepo4 really act like Lead Acid?

  • Gamma17Gamma17 Member
    edited October 2020

    @randvegeta said:
    Really? Don't you need to cut off the power once full (or near full)? Lead acid can seemingly be connected to a permanent power source, but lithium cells apparantly keep on charging, even if the voltage is below their rating (weird...). Does lifepo4 really act like Lead Acid?

    Charging algorithm for most chargers is constant current, then once target voltage is reached - constant voltage, then very low current "trickle charge" for some time, then complete disconnect and periodic "top-off" charge to compensate for self-discharge. This is exactly what most UPS do with lead-acid too. And since 12v lead-acid is usually charged to ~13.7 while 4-cell lifepo4 will have max charge voltage of 14.6v this will result in slightly undercharged battery and should be completely safe. However my experience with this does not go much further than seeing people use lifepo4 in a car without proper BMS (just balancing circuit) and doing the same with small (2kva) UPS (though i actually used BMS with thermal monitoring for extra safety), so proper research is obviously needed before committing to something...

    In all the stuff i've read (specifically about lifepo4) before using it i've never seen the issue about "charging even at lower that target voltage" even mentioned so it is either not an issue, or i just used bad sources. I know that keeping li-ion at max charge voltage constantly is bad (old laptops did this, killing batteries fast), but it seems to be different issue alltogether...

    Thanked by 2vimalware randvegeta
  • randvegetarandvegeta Member, Host Rep

    @Gamma17 said: (though i actually used BMS with thermal monitoring for extra safety)

    This is a must if the batteries are going to be left connected to a charger 24/7.

    I have a bunch of dead modules (sub 1V) and of course these will have suffered severe degradation/damage, and special care must be taken. But, it is interesting to see how they charge up. Forcing a charge into the module does indeed charge up the cells, and they are surprisingly well balanced off the bat. But I would not leave these things unattended charging for long without a BMS. With a BMS, I think it should be fine.

    Despite the modules being dead and likely having lost a great deal of it's original capacity, they still seem to hold a significant amount of charge. It takes forever to charge them up and discharge them.

  • raindog308raindog308 Administrator, Veteran

    @randvegeta said: the length of time the modules will be without a BMS will be at most a couple of weeks.

    Murphy loves the IT industry.

    Thanked by 1randvegeta
  • jsgjsg Member, Resident Benchmarker

    Funny things I read here. Do yourself a favor and look closely at the charge curves for lead acid and LiPo4 plus look at the details. Example: A LiPo4 is ca. 90% loaded when the absorb voltage of ca. 14.5V is reached. From there the voltage is kept constant and the current decreases. So, charging LiPo4 with a lead charger will not work properly. Also note other differences like e.g. timing of stages, natural loss compensation, etc. which seem to be similar but are not.

    Also look at what an IPS is. It's basically a down-converter whose secondary voltage "just happens" to fit what's needed for charging, a battery manager (charging, discharging, temp. and other feedback loops) and an up-converter which is fed either by the down-converters output or by the batteries. Obviously UPS manufacturers design their products around the battery system trying to squeeze out the maximum possible which again usually translates to optimizing around the battery system.

    If you go the "Oh well, they are similar enough" route you'll end up with inefficiency (in the most expensive part at that) and probably not insignificantly reduced battery life-time.

    Good luck anyway.

  • Gamma17Gamma17 Member
    edited October 2020

    @jsg said:
    Funny things I read here. Do yourself a favor and look closely at the charge curves for lead acid and LiPo4 plus look at the details. Example: A LiPo4 is ca. 90% loaded when the absorb voltage of ca. 14.5V is reached. From there the voltage is kept constant and the current decreases. So, charging LiPo4 with a lead charger will not work properly. Also note other differences like e.g. timing of stages, natural loss compensation, etc. which seem to be similar but are not.

    Thing is, for lithium batteries (in general, be it lifepo4 or something else) charging to lower voltage actually works and provides meaningful trade-offs. There is no rocket science in charging either, it is as simple as it gets - constant current up to target voltage, constant voltage down to some current (commonly ~0.01-0.05C), top-off charge for some time (~1 hour). With multi-cell batteries - simple balancing circuit. And it will work fine at lower voltage. The reason why it is often said to be complex is to discourage people from playing with stuff which can blow up in their face. For lead-acid the same algorithm is used (regardless of the fact that the curve might look differently, which happens because the battery itself works differently), except instead of limited top-off charge in standby systems like UPS this top-off charge is often applied constantly. This is definitely something to consider, but is not "completely different" at all.

    As for trade-offs - it reduces capacity slightly (for example ~15-25% for li-ion charged to 4.0V), increases longevity exponentially (number of cycles, for example from ~300 at 4.25V to ~1500 at 4.0V) and makes battery safer (much less likely to "vent flames"). It also greatly reduces requirements to charger complexity/precision, because if you are charging li-ion to 4.0V instead 4.25V you suddenly have whole 0.2-0.3V safety margin within which battery will not "vent flames" or even get damaged. While if you are charging to 4.25V (as many mobile devices do) you are already slightly above maximum safe voltage and going just 0.1V above it will definitely result in fire.

    Why is not it done commonly then? Capacity and energy density are very important in some cases. Like would you buy a phone which claims to have battery which is guaranteed to last 10 years with daily charge but has 30% less capacity? 95% of people will not...
    For the stuff discussed here though such trade-offs seem not only acceptable, but beneficial...

    Thanked by 1vimalware
  • jsgjsg Member, Resident Benchmarker

    @Gamma17

    I would largely agree if this was about someones private lab/research/makeshift solution - but if I'm not mistaken this is about a multiple rack or room UPS in a DC and there is a simple and strict guideline: Do not waste el. power because that is one of the very major cost factors.

    One just doesn't risk losses of "a couple of percent". Quite the contrary, one tries hard to not reach "only" 97% efficiency but to squeeze out yet another half percent.
    loading batteries with about 1V lower voltage (in a good device) means a different resonance tank, different protection triggers/circuitry, probably a different main inductor and different charging algorithms likely mean different software and potentially a different MCU.
    Plus the quite different chemistry which is a can of worms in itself in a tightly controlled, insured, inspected, etc. setting with millions and millions of $ worth of at least in part critical equipment.
    Plus of the temperature issues; keep in mind that a car moves in free air while DC racks do not ...

    In short, I agree with you that the era of lead acid UPSs is coming to an end sooner or later and lithium based solutions are among the top candidates to replace it - but to become established and trusted in DCs (as opposed to cars) that technology will require more research and lots of engineering.

  • randvegetarandvegeta Member, Host Rep

    I am amazed people think that the demands of a DC are greater than that of a car.

    Cars will charge and discharge at >1C, and are exposed to very harsh environments and fluctuating temperaturess. They also tend to get cycled more and higher.

    A DC is typically in a controlled environment with pretty favourable temperatures. For longevity, batteries may be charged to only 80%, and given the denisty improvements over Lead Acid, should still mean plenty of capacity for the same space. The batteries need not be cycled at all even.

    Sure, lithium packs a punch and can be hazardous, potentially, but Im replacing a giant generator operating IN DOORS with these batteries. Which is more hazardous? Im gonna go with the one that actually combusts in order to operate as designed.

    Thanked by 1vimalware
  • jh_aurologicjh_aurologic Member, Patron Provider

    I'd be careful, DC at high amps is worse when it goes wrong. Probably you should include a electrican in your planning. You're running at the risk of severe burns if you make a mistake, at least you dont need to be scared to get electrocuted ;)

    Thanked by 1vimalware
  • I suggest proper local electrical and engineering consultations - not low end engineering

    Thanked by 2vimalware randvegeta
  • randvegetarandvegeta Member, Host Rep

    @Unbelievable said:
    I suggest proper local electrical and engineering consultations - not low end engineering

    Not use low end engineering on low end talk? Outrageous!

  • randvegetarandvegeta Member, Host Rep

    So, I wont be leaving the chargers connected unattended under any circumstances. :D

  • randvegetarandvegeta Member, Host Rep

    In case anyone is interested in an update, as I shall probably be documenting this further as I think it's an interesting project/subject, batteries were received!

    So I bought 2 packs from 2 salvage Tesla Model Ss. I thought both were 75KWH packs. Turns out 1 was an 85KWH pack. Yay, an extra 10KWH (not really), so an extra 2 modules 'free'.

    That gives me a total of 30 modules.

    I found what appears to be a decent BMS / Control Module, but they are quite pricey, and this one only supports 20 modules. So I decided to sell 10 modules. But apparently the buyer wants an entire pack, so I sold the 85 pack (16 modules) and I'll make do with the 14 modules. The profit made on the 85 pack covers the cost of the BMS and even an inverter.

    As the BMS manages all the modules as a single cluster, I will now build the 1st bank of batteries into a single parallel bank. So that upto 20 modules in parallel. Checked with the inverter/charger manufacturer and it's perfectly fine to hook up multiple inverters/chargers to the same bank. Just need to make sure I don't overload the bank. With upto 20 modules, that's >100KWH, so hooking up 4x6KW inverters will only get me to a max discharge rating of 0.24C. Likewise, max charge rating would still be just 0.08C. Heat should not be a problem, and as such, will not need to hook up the cooling tubes.

    Another interesting 'discovery' is that I can also use the lead acid UPS we already have.

    Our UPS uses 16 x 12v batteries and charges up to 219v. Incidentally, if you hook up 9 modules, that's 24.3v each. Well below the 25.2 max. The only problem with this setup would be the discharge voltages. Lead acid has a higher range than lithium, so although it won't overcharge with 9 modules, it WILL over discharge. But that may or may not be a problem. With a 6KW UPS and 45KWH of batteries (charged up with ~37KWH of energy), that UPS should last >6 hours at full load before the battery voltage drops below critical. But we wont be doing this as using the new inverters should be a superior/more efficient solution.

    Now I need to design a shelf / rack mounts for the batteries.

  • SplitIceSplitIce Member, Host Rep

    @randvegeta Please at a minimum ensure that you have a well tested
    a) Fire Detection system located in close vicinity to the Batteries; and
    b) Temperature Monitoring on your batteries.
    c) Dry Chemical fire extinguishers at various locations (i.e at entries to room)

    I've been involved in projects involving Li-Po batteries for (in-device) battery backup purposes. I've seen what can happen with small cells under stress testing without protection circuits. Multiple KWH? eek.

    Thanked by 1vimalware
Sign In or Register to comment.