Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Big Badass UPS Build, any help is appreciated :D
Hey everybody, 
after watching hours and hours of jehu garcia, hbpowerwall and averagejoe i decided to get involved as well. I have a connection at a recycling center with a virtually unlimited laptop battery supply. Right now i have about 500 cells tested with another 400 to go, i just bought some extra OPUS testers to help speed things up and i am very curious about other peoples experiences in starting a project. 

I used 40 tp4056 on a 2 pcb's to charge the cells(which saves me alot of time). I have a 10 amp power supply and a 60 amp powersupply, both from aliexpress. Unfortunately the 60 amp blew while i only used 40 tp's, which should be around 40 amps(or maybe 45 if you have a lot of losses, i use thick wires so that shouldn't be a problem). So yeah, kinda sucks that it blew and now i am buying 4 smaller 10 amp power supplies to charge 8 cells at a time. The one i still have seems to work fine, although gets a bit hot. I noticed that during charging with the tp the cells all tend to get very warm, the tp gives about 1 amp if i am not mistaken so for me that is a nice benchmark to toss bad cells, although the line between warm(which is sort of okay) and hot(which is a toss) is not always very clear. Do you guys have any experience with that, maybe a rule of thumb that you use? It does scare me bit to leave them charging alone so i built a tiny battery bunker with bricks in case they catch fire. I have tried really hard with shorting them out to get them to pop but i havent managed so far. How do you guys handle the safety aspect and am i a bit too careful? I also bought 2 second hand fire proof cabinets to store all my cells just in case. 

The opus tester i like a lot, i bought one for 40 euro's from a legit company in the Netherland(yes, im dutch  Big Grin ), and i ordered a couple more from They work okay but i think they are fake, i have had a lot of weird problems(display not giving amps etc), it works but it feels like its not the real stuff. Anyone have some experience with that?

Most of my laptop cells are around 2000 too 2200, most cells perform pretty well except for the red ones, they get really hot. Since there are so many brands and types i havent yet bothered with looking up every type of cell, maybe i should. Right now i am at the point that when i have a batterypack in my vice, i crack it open and if i see dark blue or red i just toss them out. 

My main project will be to build a ups from li-ion, i am thinking about 7s50p to start with a 24V system. The power comes from the net and the batterypack is put in series with a group so that when there is a powerout you still have power. 350 cells x 8.1 Wh(3.7x2.2Ah avg) is about 2.8 kWh. Given that there are some losses and you dont want to discharge your cells to hard you might end up with about 2kWh of usable energy, with 50 cells in parallel and 1 amp max per cell you get a max load of 50A x 24V = 1.2kW. 

The one thing i am not sure about is the logic that switches between the powerwall and the net. What i have in mind right now is to connect the net to the charger, that charges the bms, and from there the battery is connecting with the inverter to whatever it is used for. The downfall is that you always run on your battery, the upside is that you dont need any complicated switches that instantly switch when the power goes out. Does anyone have any experience with that? Im not sure which way is the best.. 

Is my math a bit legit here or am i overlooking stuff? The purpose of this powerwall is to run a refridgerator, lights, computers etc on it(obviously no ovens or anything above 700 watts(a microwave might still be okay)). 

I have found an inverter on banggood that does 24V to 1500 watts for about 120 euro's which is fine(see link below). Now what i still need is an inverter to get 220V down to 24V(29.4V is what the BMS wants) but i cant really find it anywhere. Does anyone have recommendations for a proper BMS that is not too expensive and is fit for this configuration. I have an 24V 8 amp 7s ebike bms but i am having trouble with how it works exactly. For example if i take an inverter that steps 220 down to 29.4V, what is regulating the current? Is there currentclamping in the bms? Or does it happen in the charger, i'm not really sure. I can image that if you put 29.4 volts on an empty batterypack without any current clamping you blow the powersupply because you only have the internal resistance of the cells and since that is really low you have almost infinite amperage, thus blowing something in the circuit. Is my reasoning sound?, or am i just uttering nonsense?

Really looking forward to all the feedback i can get. I added some pictures of my setup, i'm also working on my own testing rig, where i can test 20 cells at a time with computer integration and temperature sensors. The pcb is use now for charging is the first iteration of my testing rig but there were some faults in it so i just used them for charging. Would anyone be interested in 20 cell tester which usb and ethernet connection? My idea is to hook it up to the pc so you can get nice discharge curves of the voltage as well as temperature under load, although when reusing batteries im not sure if all of that is necessary, thought it be cool though  Tongue . In raw parts it will be around 15 to 20 euro's, i'm not looking to make a profit, would be cool if other people want it since the opus is expensive and only does 4 cells. 

Thanks in advance, awesome forum, awesome people here! Hoping to hear from you guys! 

Greetings from the Netherlands, 
Yesterday i finished sorting the packs. I had all the batteries tested and sorted by capacities. I thought about exactly sorting them so the capacity of each pack would be identical but then i realized that the standard deviation of measurement on the OPUS is rather large. I've had cell double tested with 2 different tester of the same make and there sometimes was a difference of 100 mAh between them. Seeing how large that difference is, there is really no point in exactly sorting them because the uncertainty is so large. I have sorted my cells at intervalls of 100, ranging from 1500-1600 to >2500. My trays with 2100-2200 up till >2500 where already sufficient for the 350 so i started with the >2500 tray and just evenly distributed the cells over 7 packs. Then i worked my way down to 2100. So on average the cells i my pack are around 2300 and if we can trust on the math, then if we make the sample size large enough there should be convergence to identical sized packs Big Grin. My packs are 350 cells total, i cannot imagine the hassle it takes to individually sort 1400 cells, calculate the perfect pack etc. etc., probably takes a day.

I started a thread where is was looking for the right inverter. I dont mind paying a bit for a decent model, i got some different advice. I even contacted the manufacturer because the inverter i want is only for lead acid. I think the conclusion is that as long as you can run constant current from the inverter to the pack at a voltage rang that is similar to the li-ion, you should be fine. The inverter im looking for is 2000W 24V Iconica hybrid pure sine inverter and this one is made for 24V lead acid. The constant current cycle of this one goes to maximum 28.4V. I have 7S so in theory i would need 29.4V to fully charge them. But i want to be nice to the cells so 4V per cell is fine for me(28V). My plan is buy the inverter and set it to lead acid cycle, there you can set the constant current it charges too(a limit of 20 amps i think) so i want to set it to around that, this means that at 40p there is maximum half an amp going into the cells. Now in the inverter you can also set the maximum battery charge. If i set this to 28V then in theory the inverter should stop charging when the pack is at 28V and still in the constant current cycle, so the constant voltage cycle will never be turned on. Im afraid of that constant voltage because either the inverter will not like it or there will be too much current going to my cells. I dont want the current going into my pack to be limited by the physical limitations of the inverter, thats just asking for problems. Anyway, anyone have some experience with this? Is this a legit, safe and responsible approach?

My next phase is to connect all the packs. I build a spotwelder at home from an old microwave that works perfect and i am fusing all the cells. My problem is that i cannot find any fuse wire that is spotwelded. Either the spotweld doesnt take or it gives a spark. Spotwelding busbar to the negative side works like a charm and i can recommend it to everyone. Anyone have a suggestion about a specific type of thin wire to use as a fuse? I would love to buy a whole roll of it. I tried resistor legs but they didnt take so well with spotwelding. I can regulate my spotwelder between 0-50 ms and 0-100% power. My best result i achieved with 0.75 mm^2 zinc plated wire(used for gardening Tongue ).

I have done 2 out of 7 packs now with a combination of spotwelding and soldering. The next challenge will be to find a nice way to connect connectors to it so its safe, reliable and neat. Probably going to check averageJoe out, he had some nice methods with anderson connectors.
Your second novel is better than the first.
"only for lead acid" - typically means the voltage range selection is pre-set but there are a lot of different lead-acid types that need different voltage ranges - decalcification voltage is different. Plus, lead usually needs a higher "boost" voltage before finishing the charge at a lower "float" voltage. Lithium does not need separate boost/float voltage.

"2000W 24V Iconica hybrid pure sine inverter" - the battery voltage range can be programmed.

"My problem is that i cannot find any fuse wire that is spotwelded" - propper fuse wire has a low melting point (far lower than the steel can of a cell) so will never spot weld. Solder. Either fuse wire or axial glass fuses.
Korishan likes this post
(08-30-2018, 03:44 PM)completelycharged Wrote: propper fuse wire has a low melting point (far lower than the steel can of a cell) so will never spot weld. Solder. Either fuse wire or axial glass fuses.

I have seen a few videos of spot welding fuse wire / glass fuses.. never tried it so I have no idea how it would work in reality

Korishan likes this post
You could spot weld axial fuses as the legs are standard steel
completelycharged likes this post
Proceed with caution. Knowledge is Power! Literally! Cool 
Knowledge is Power; Absolute Knowledge is Absolutely Shocking!
Certified 18650 Cell Reclamation Technician

Please come join in general chit-chat and randomness at Discord Chat (channels: general, 3d-printing, linux&coding, 18650, humor, ...)
(this chat is not directly affiliated with SecondLifeStorage; VALID email req'd)

Forum Jump:

Users browsing this thread: 1 Guest(s)