HUUUGE USB POWERBANK

The-J-Man

Member
Joined
May 2, 2017
Messages
263
Greetings,

I set out to build a powerbank for camping / Festivals etc and finished it last week.

I used the best cells I had to date that were similar, all from ebike batteries. Altogether I used 160 cells in a 4s40p config. It was 80x LG MG1's, 40x Panasonic NCR18650PF and 40x Panasonic NCR18650BE. They all had close to new capacities, testing in at 3000-3100mAh on the OPUS, the BE cells had 3250-3350 on the OPUS. I split the cells equally per group.

The steel case was originally a battery case all ready. I got these from a customer and the Li-ion puch cells configured in a 13s had all died. Upon pulling them out of the case, thesecond rubber internal case they were inexpanded to roughly double the size and broke open. These cells were really really really dead! and dangerous, but the voltage was super low so no problem there.
The case had some holes in it for the +- cables, charge port and switch but they were too small so I drilled out bigger hols for the 4cm wide USB chargers etc.

image_aiyvon.jpg

image_kdtoei.jpg


I fused them singly because I wanted to keep the option of easily dissasembling the pack again when I want to use the cells for something else.
A 10A fuse between blocks 3 and 4 acts as a mid pack fuse.

image_sxhxgl.jpg


I was able to fit 120 cells in the base of the case in a 10x12 and another 40 cells in 2x 2x10 on top of these. lying flat, essentially in the top part of the case.

image_hhkire.jpg


From the 4s BMS I used another fuse to connect to the 3 way position switch I had. Output from the swtich has another 10A fuse that then splits up into the two 12-24V dual USB chargers. Another positive goes to the magnetic Rosenberger charge port. BMS shares the same +- for discharge and charge. *** I have not yet tried to charge the battery yet because I still havnt managed to empty it...... ::D**


The current draw would be so low (max 8.4A at 5V so about 2 A at 16V (cell level) that I really did not have to worry about equal loads etc.
The one USB charger has a voltmeter built in, that way I know the capacity and I didnt have to bother with another device.

image_wjiptu.jpg

image_cejqjw.jpg


For some reason the packwere at 16.6V, but the BMS only ouputed 16.07V??? Anyone know why? cheap chinese BMS maybe?

I then added a lot of silicone around the switchto stop potential water ingress, and the terminals to glue them in place. The USB chargers and Rosenberger are waterproof all ready.


image_nrrbru.jpg


Padded the case with thick foam, hard plastic sheetand foam gives a safe layer between the 40cells on top and the 120 underneath.
Lots of duct tape covers all the contacts :)


Closed up the lid with a lot of silicone again, popped in the screws and let it dry.

Ended up using this for 5 people for 4 days all ready, charging phones and speakersand running 2 USB fans constantly.

remaining voltage, 15.6 from the 16.1 I started with.

Total weight, about 8kg.
Total capacity, about 1640Wh. :cool: (2850mAh average x 160, x 3.6V )

If I end up using it more I will add a proper 12V output socket as well so I can power some more stuff.
 
Which bms did you get? Possibly voltage drop from diodes
 
Yeah, I don't see any diodes, other than what's built in the FETs.

Check all the solder joints and make sure all the pads are making good contact. If not, hit it with the iron after applying flux to the area (this minimizes the chances of bridging)

I suppose it's also possible that there's a FET that isn't closing all the way, but I doubt that.
 
how are you measuring the voltage output? If you are using a multimeter then the voltage need to very close to each other. But if you are reading the voltage from the blue LED meters, those meters can be off in there readings. That might be your problem.
 
Agreed that is HUGE! Any plans for solar charging? Not that you really need it. 1.64kw is a lot of phone charge time. You could even keep a small fridge going for a day.
 
jonyjoe505 said:
how are you measuring the voltage output? If you are using a multimeter then the voltage need to very close to each other. But if you are reading the voltage from the blue LED meters, those meters can be off in there readings. That might be your problem.
When i first plugged in the usb charger with the voltmeter it said 16.1. so i was confused. I measured around with a voltmeter and at the output terminals of the bms it measured 16.07. so yeah...


Geek said:
Agreed that is HUGE! Any plans for solar charging? Not that you really need it. 1.64kw is a lot of phone charge time. You could even keep a small fridge going for a day.

I could certainly add it in, then i would have to get rid of the 40 cells on top tho to make room for thr cherge controller.
When i was deciding on how to put the cells in i noticed that i can put 120 in the base with using the stabdard cell holders but with cell holders that can package them of center to each other i can put 160 just in the base. So i can actually move all the cells into the base.
If i redo it id use the new cells i just recieved. NCR18650GA and LGMJ1. Then id have even more in it and have a solar charger etc.
maybe :)
A little fridge wouldve been great last weekend..... warm beer isnt the best ;)


Korishan said:
Yeah, I don't see any diodes, other than what's built in the FETs.

Check all the solder joints and make sure all the pads are making good contact. If not, hit it with the iron after applying flux to the area (this minimizes the chances of bridging)

I suppose it's also possible that there's a FET that isn't closing all the way, but I doubt that.

Interesting ok. Ill check it out when i open it up again... resolder etc.
in that case the FETs or the FET should get warm from the resistanceright?
 
If you're only drawing ~2A & the BMS is rated for 30A, the FETs should stay cold. If they're warming up, there's something going on.
 
The-J-Man said:
Geek said:
Agreed that is HUGE! Any plans for solar charging? Not that you really need it. 1.64kw is a lot of phone charge time. You could even keep a small fridge going for a day.

I could certainly add it in, then i would have to get rid of the 40 cells on top tho to make room for thr cherge controller.

...

Stick the charge controller to the back of the panel(s).
 
Redpacket said:
If you're only drawing ~2A & the BMS is rated for 30A, the FETs should stay cold. If they're warming up, there's something going on.

Oh yeah for sure. I was just thinking if one of them isnt closing right and it causes that voltage drop there should be enough resistance to get it to go warm.
I was going to test the whole thing ubder max load but ended up not having the time too..
Geek said:
Stick the charge controller to the back of the panel(s).

Good point ::)eep it external and plug it straight into the charge port. I like it
 
I could certainly add it in, then i would have to get rid of the 40 cells on top tho to make room for thr cherge controller.
When i was deciding on how to put the cells in i noticed that i can put 120 in the base with using the stabdard cell holders but with cell holders that can package them of center to each other i can put 160 just in the base. So i can actually move all the cells into the base.

Got a link to those cell holders?


I managed to find some: Cylindrical Cell Bracket
 
GerardAchternaam said:
I could certainly add it in, then i would have to get rid of the 40 cells on top tho to make room for thr cherge controller.
When i was deciding on how to put the cells in i noticed that i can put 120 in the base with using the stabdard cell holders but with cell holders that can package them of center to each other i can put 160 just in the base. So i can actually move all the cells into the base.

Got a link to those cell holders?


I managed to find some: Cylindrical Cell Bracket


Exactly thats the ones i woudlve gotten.
I had some cell holder from ebikes that i was playing with, wouldve had to chop some edges off and get rid of 2 slots for it to work, didnt want to do that :D
Im working in some other holders myself tho.
 
Korishan said:
Yeah, I don't see any diodes, other than what's built in the FETs.

Check all the solder joints and make sure all the pads are making good contact. If not, hit it with the iron after applying flux to the area (this minimizes the chances of bridging)

I suppose it's also possible that there's a FET that isn't closing all the way, but I doubt that.

The voltage dropis probably from the body diode of the charge transistors, in a configuration like an image below.


image_ldlowf.jpg


if the charge transistoris off and the discharge is on you will get a diode drop on the output.
 
Done.


image_cdphzp.jpg


Voltage drop @1 A (both directions) about 6,4 mV. With Rdson 5mOhm, 5 parallel 2 in series, plus 2mOhm from the resistors this is close to what can be expected.

On overvoltage from the cell at charging the Fets open, they are sitting in the Minus line.
I didnt check the 30 A shortcut.....

Its exaxctly as can be expected.
 
GregorScott said:
The voltage dropis probably from the body diode of the charge transistors, in a configuration like an image below.


image_ldlowf.jpg


if the charge transistoris off and the discharge is on you will get a diode drop on the output.

So a BMS with separate charge and discharge ports would solve the issue?

Cherry67 said:
Done.


image_cdphzp.jpg


Voltage drop @1 A (both directions) about 6,4 mV. With Rdson 5mOhm, 5 parallel 2 in series, plus 2mOhm from the resistors this is close to what can be expected.

On overvoltage from the cell at charging the Fets open, they are sitting in the Minus line.
I didnt check the 30 A shortcut.....

Its exaxctly as can be expected.
thanks you checking that out
No voltage drop when idle at all?
Makes me think I connected something wrong, then again thats not possible with such a small drop...
 
The-J-Man said:
Cherry67 said:
Done.


image_cdphzp.jpg


Voltage drop @1 A (both directions) about 6,4 mV. With Rdson 5mOhm, 5 parallel 2 in series, plus 2mOhm from the resistors this is close to what can be expected.

On overvoltage from the cell at charging the Fets open, they are sitting in the Minus line.
I didnt check the 30 A shortcut.....

Its exactly as can be expected.
thanks you checking that out
No voltage drop when idle at all?
Makes me think I connected something wrong, then again thats not possible with such a small drop...
No voltage drop (except the mentioned 6.4 mV) at 1 A charging or 1 A Load.
No "diode loss" in any direction.
Definitely.

I cant be sure that your BMS is ok, but mine shows as i described.
 
Cherry67 said:
The-J-Man said:
Cherry67 said:
Done.

Voltage drop @1 A (both directions) about 6,4 mV. With Rdson 5mOhm, 5 parallel 2 in series, plus 2mOhm from the resistors this is close to what can be expected.

On overvoltage from the cell at charging the Fets open, they are sitting in the Minus line.
I didnt check the 30 A shortcut.....

Its exactly as can be expected.
thanks you checking that out
No voltage drop when idle at all?
Makes me think I connected something wrong, then again thats not possible with such a small drop...
No voltage drop (except the mentioned 6.4 mV) at 1 A charging or 1 A Load.
No "diode loss" in any direction.
Definitely.

I cant be sure that your BMS is ok, but mine shows as i described.
Cheers for the info.
I contacted the seller and he said its normal. Of course he would say that but for 7.17 I cant be bothered taking it all apart again and sending it back. Maybe one day I pop a new BMS in.
For now the pack is standing on top of my powerwall :D
 
Back
Top