Homemade Battery Packs

kje said:
Isn`t it dangerous to bulk charge all the cells simultaneously without balancing?

Why do you charge them to 4.25V and not 4.20V?

What copper wire and dimension will work for a battery pack like this?
4.2V into 1 bank of cells all connected together. They are all the exact same voltage-balanced. (3V to 3.5V has so little energy that ganging them together has a minimal equalization transfer rate.)

Charged to higher voltage so that self-discharge will be more immediate and readily evident. (4.25V no longer used, I tend to charge to .05V above final use voltage and allow a much longer period for "bleed down" - self-discharge.)
3.6V red Sanyo cells charged to 4.10V and confirmed above 4.05V retained as final use voltage, other voltages for 3.7V rated cells. See - Determining Optimal Charge Voltage

I like to maintain "rated" wire gauge for continuous use.
But also consider brief-surge use at ~200% of "rated".
14ga = 15A continuous = 30A brief or surge
12ga = 20A continuous = 40A brief or surge
10ga = 30A continuous = 60A brief or surge
My eZip bikes use a 40A fuse and 14ga wire ... I rebuild packs with 12ga or better - (very heavy tinned copper braid)

Do not forget to add a fuse to your pack build!!!
One little mistake-misstep can cause a whooole lot of grief!!!
 
DrkAngel said:
I like to maintain "rated" wire gauge for continuous use.
But also consider brief-surge use at ~200% of "rated".
14ga = 15A continuous = 30A brief or surge
12ga = 20A continuous = 40A brief or surge
10ga = 30A continuous = 60A brief or surge
My eZip bikes use a 40A fuse and 14ga wire ... I rebuild packs with 12ga or better - (very heavy tinned copper braid)

Do not forget to add a fuse to your pack build!!!
One little mistake-misstep can cause a whooole lot of grief!!!

My boat-motor have a max draw at 55 amps, does that mean I have to use 10ga copper wire? (wich I solder the batteries together with) Or should I use battery tabs?

Yes, I will add a 60A fuse to the pack.
 
My EZIP with 24V 20AH Ping pack has never blown the 30 amp fuse I installed. Of course, that's an unmodified drive setup, just with a bigger battery pack.
 
Gitten-er-done!

After draining my 33.3V 26Ah (9s10p) to 33.2V, ~3.7V per bank, I marked resultant voltages and added complementary 2s to each bank as deemed advisable.
Lowest bank got the best cells - best got the worst.
Lowest bank (3.65V) got 4500mAh, highest (3.71V) got 4100mAh, the remainder apportioned in between.
Since I will discharge to not below 3.8V, where cells were within 2/100th of each other and "good" ... now they are just "more good".

At 3.8V - 4.05V capacity is still >20Ah ...
I will measure Ah capacity after next deep discharge ... though I typically recharge immediately after any use and weather is too cold for any extended rides, so, full capacity test might wait till near Spring.

33V 31.2Ah.JPG

Last 2 banks were pulled from a different batch that had oddball tabs ... sorry, wanted it to look prettier!

I will monitor 1st few discharges ... mainly to check for bad-poor solder connections.
After that I will rely on bulk charging with the occasional balance check.

Edit - I did add 4 - 9mm, 1/2 cell height, blanks as support for the alternate banks.
And re-covered all cells with a wrap of boxing tape.
Both ends of cells were covered with closed cell polystyrene, (foam sheet, as is used for durable padding), to secure-stabilize cells tightly in pack.
 
What sizes of copper braid are you using for the connections?
 
Had to make sure this was still available.

Belden tinned copper tubular braid/ground strap 5/16" X 20' $12.99
10-9gauge Tinned Copper Braid - Great for connecting cells, solders nicely 30-50A continuous recommended

Used it for the "rails".
Seller describes as 9-10ga but weighs more than 4 stands of 12g, so I would rate at 100amp+ capable.

Intermediate braid was 1/8" flat, not round "18C"?, comparable to ~12ga.
Don't remember where I got it ... some eBay item?

The tinned copper braid is extremely easy to solder with and produces very durable solder connections.
But don't forget, you still need to use flux.

De-soldering note:
I've always had mixed success using de-soldering wick, until I tried coating it with flux.
Logical, but I just seemed to miss that point ... initially.
 
FWIW I've used braid saved off the outer shield of old RF coax cables from TV stuff. (since it was already tinned it solders real easy, and was free with a bit of effort stripping the insulation off) for bridging cells in pack repairs.

I've also used that RF braid shield crrimped/soldered into lugs for the interconnects between Thundersky prismatic cells.
 
After disassembling laptop batteries I leave the cells connected in pair (parallel) and charge them. Should I disassemble them individually? When I discover a bad cell-pair, is there a possibility that it`s just one of the cells that`s bad?
 
9s 33.3V battery looks to be my new voltage of choice.

I have 2 capable eBikes, my Winter (Snow Buster) and my Casual Cruiser.

Snow Buster is a 2008 eZip Mountain Trailz with oem gearing, 33.3V pushes it to 20mph+ ... even with 2 studded tires. Maximum motor output is right near 12mph, so it mushes through the slush and snow nicely.

Casual Cruiser is geared up with the 16T mod. While this does provide assist noticeably beyond the "legal" 20mph, PAS mode provides a moderate 18mph, with the option 25mph++ in TAG mode, with moderate pedal assist.

24V vs 33.3V.jpg

33.3V Chargers
Not the easiest, there seems to be a very limited, or costly, options available.
I did mod a MeanWell S-350-48 as a 36.45V 9A charger ... but that seems wasteful.

I am working on stretching the voltage of the MeanWell S150-24 to 36.45V as a 4A charger but am still waiting for the 50V caps.

Alternately, I can simply run a 12V computer power supply in series with a a 24V MeanWell for a 6A+ charger. Computer power supply fan could be directed to cool MeanWell also, or it's 12V to power fan for MeanWell.

Even Better ... I just got a pile of external power supplies for some Dell mini computers. These are nice compact 12V 10A units.

Dell 12V 10A.JPG
2 in series with a S-150-24V (modded to 12.45V) will provide a healthy 36.45 10A charger, with plenty of 12V options for cooling fan(s).
 
DrkAngel said:
Bonus! Capacity testing, for large packs is not "necessary".

Thank you. :) Why is this not "necessary"?

DrkAngel said:
Build pack in 3 equal banks.
Discharge ... till one bank dips below ~3.7V.

Should I discharge them simultaneously? Or can I discharge one bank at the time, using a stop watch and compare the time results till one bank dips below ~3.7V?

DrkAngel said:
All 3 banks should be of various voltages, mark as H M L (High Medium Low).
Recharge to equal voltages.
Remove cell (cells) from H bank and add to L bank, or add cell to M and more cells to L bank.
Repeated monitored cycles will allow you to create nicely balanced capacity pack.

So the more cells in parallel (more capacity), the longer the voltage stays (less voltage sag). I think I understand :) but correct me if I`m wrong.
 
Sorry ... capacity testing individual cell, not necessary - bulk pack.

1st eliminate any obviously bad-poor cells ... bleed down-self discharge test.

Just build equal quantity cell banks,
partially discharge entire pack in normal use,
add cell(s) to any bank that discharges faster.
Recharge-discharge repeat.

Of note is that voltage sag is more a factor of usable capacity ... than number of cells.

Another important observation is that discharge and charge rates should be determined by actual capacity, not by original rated capacity!
 
DrkAngel said:
Of note is that voltage sag is more a factor of usable capacity ... than number of cells.

Another important observation is that discharge and charge rates should be determined by actual capacity, not by original rated capacity!

Battery Deterioration
18650 laptop type cells exampled

Most battery users notice that when a battery begins to deteriorate, it's deterioration accelerates.
Basic understanding and simple mathematics explain this.

High charge-discharge rates are damaging to batteries.
Batteries are typically rated with maximum and recommended discharge and charge rates.
The recommended noted as for best-maximum life.
Now, if we factor actual capacity rather than the new rated capacity ...

40A actual capacity battery
A 40Ah pack, with 40A controller, might surge 1C and sustain .5C. = near to recommended
A 10A charger charges at .25C, (safely below the .5C charge limit for common 18650 cells).

When capacity degrades to 30A
A 40Ah pack@30A actual, with 40A controller, might surge 1.33C and sustain .75C. = beyond optimal - damage accelerates
A 10A charger charges at .33C, (safely below the .5C charge limit for common 18650 cells).

When capacity degrades to 20A
A 40Ah pack@20A actual, with 40A controller, might surge 2C and sustain 1C. = pushing maximum safety limits - rapid damage
A 10A charger charges at .5C, (right a the .5C charge limit for common 18650 cells).

When capacity degrades to 10A
A 40Ah pack@10A actual, with 40A controller, might surge 4C and sustain 2C. = dangerous heat and catastrophic damage
A 10A charger charges at 1C, (dangerously beyond the .5C charge limit for common 18650 cells).

As noted battery abuse-damage is geometrically progressive.
So ... It might be a good idea to build larger than "needed" packs or "retire" packs to less demanding use, while still of reasonable condition!
 
So ... It might be a good idea to build larger than "needed" packs or "retire" packs to less demanding use, while still of reasonable condition!

I do that with electronics and appliances as well as batteries. As a result, I have about 6 working video recorders, including a Super BetaMax. ;)
 
One easy step I am doing after removing cell pairs from the laptop packs is a quick voltage sag test. I test cell voltage, then voltage sag under a constant load (flashlight bulb - about 2a). I then sorted all my cells into 5 groups. From 0.1v to 0.5v sag. The .4 and .5s will probably not be built into a pack, and there were very few.

This makes bulk capacity testing easier and more efficient since they will be tested in these 5 separate groups. Perhaps I will find that voltage sag will be correlated to capacity.

I wanted to tab weld my cells, but my homemade tab welder is not working good enough. Looks like I will use the braided wire and solder. Right now I have 400 cells, or an estimated 2.2kwh.

Tip for opening laptop battery case - put in vise vertically, just enough to hold it. Use large adjustable wrench on the other end and twist both directions. That gets a lot of the hard work done, usually. Also, a rounded plastic tire lever is good for 'scooping' the cells out if they are really glued in well. I also use a screwdriver along the seem and twist. I've gotten a lot better at opening them and usually spend an average of 1-2 minutes for each.
 
DrkAngel, I recently saw a you tube about taking a transformer from a scrapped microwave oven, and using it as a spot-welder with a foot-switch for the on-off. There are several different videos, and another experimenter added an adjustable timing circuit so that each weld lasted precisely the same length of time.I just thought you might be interested...I'm not familiar enough with electronics to know what I'm doing, but I might still try this.
 
Even Better ... I just got a pile of external power supplies for some Dell mini computers. These are nice compact 12V 10A units.

BTW Those dell supplies work great, you can vary the voltage from 3.6 to 12 volts easy. I have worked out a circuit that make those units do CC,CV. AND EOC
I run 4 in series and use comparators to drive a string of optos to single the sense wires.

Black and white are negative, Red and brown are plus, and green and yellow are the sense.
 
spinningmagnets said:
DrkAngel, I recently saw a you tube about taking a transformer from a scrapped microwave oven, and using it as a spot-welder with a foot-switch for the on-off. There are several different videos, and another experimenter added an adjustable timing circuit so that each weld lasted precisely the same length of time.I just thought you might be interested...I'm not familiar enough with electronics to know what I'm doing, but I might still try this.
I am still toying with the idea of

AC > momentary contact switch > bridge rectifier > large caps > HD contacts or contact switch > welding contacts on a sprung drop arm

In normal position contact switch is engaged filling the caps to capacity with rectified DC.
Pulling the arm down disengages switch and positions contacts, pivoting handle on swing arm engages HD contact switch, discharging precise cap current through welding contacts.

I picked up a little pile of coke can sized 360V caps ...

105V DC is not ideal for spot welding, so placing a transformer between the HD contact switch and the welding contacts, lower voltage-higher amps, is likely advisable ... and safer!
 
torqueon said:
Even Better ... I just got a pile of external power supplies for some Dell mini computers. These are nice compact 12V 10A units.

BTW Those dell supplies work great, you can vary the voltage from 3.6 to 12 volts easy. I have worked out a circuit that make those units do CC,CV. AND EOC
I run 4 in series and use comparators to drive a string of optos to single the sense wires.

Black and white are negative, Red and brown are plus, and green and yellow are the sense.
Got 14 DV-1.
So I would be interested in any mod info.
Getting a pile of DV-2 also ... maybe ...
 
spinningmagnets said:
......I recently saw a you tube about taking a transformer from a scrapped microwave oven, .......

If you are going to try the microwave oven transformer mod keep an eye out for very-old microwave ovens that have that type of transformer. The new-old microwaves are much more light-weight then the old-old ones, and have more light weight transformers then the old-old style. I have a couple of the lightweight transformers and I am not sure they would work for spot welding. I have not tried them yet, but they look light duty.
 

Attachments

  • New-old Microwave Transformer.jpg
    New-old Microwave Transformer.jpg
    61.3 KB · Views: 5,634
DrkAngel said:
torqueon said:
Even Better ... I just got a pile of external power supplies for some Dell mini computers. These are nice compact 12V 10A units.

BTW Those dell supplies work great, you can vary the voltage from 3.6 to 12 volts easy. I have worked out a circuit that make those units do CC,CV. AND EOC
I run 4 in series and use comparators to drive a string of optos to single the sense wires.

Black and white are negative, Red and brown are plus, and green and yellow are the sense.
Got 14 DV-1.
So I would be interested in any mod info.
Getting a pile of DV-2 also ... maybe ...

Did you mean DA-1 and DA-2 ?? I have found the DA-2 doesn't do well at lower voltages could be defective. Its the DA-1 am am working with,
I am using variations of the Fetcher current limiting circuit. Overall the circuit works well there are issue to be addressed, the CV needs have a reference off the battery side of the shunt. EOC is twitchy do to small voltages difference across the shunt. Performance has improved greatly from solder less to soldered board. That said I am running with the plan.

I will try to post my Microsoft schematics to give you an idea of what I am doing .
bear with me
 
Here are the schematics for controlling those dell power supply's, Mind you, that this is Microsoft paint which Sucks.
 

Attachments

  • ES.jpg
    ES.jpg
    42.3 KB · Views: 2,425
I have two EZIP packs left that I don't need, and aside from selling them I'm thinking about putting together an inexpensive 28 volt (is that right for maximum voltage for the late model OEM controllers?) rack pack for my mountain bike for Summer. I was wondering if any of the pack builders here have a set of "grade B" cells from their testing that would perform well together without a BMS? I'm hoping for about 15AH at 28 volts, so I can ride my usual 12-15 miles quickly. I'd be willing to trade one of the packs I have for the cells, possibly with some cash as well. I think I need tabbed cells, because while I can solder I have zero experience with soldering on tabs or buses...
 
OEM eZips come with 24V leds on the throttle.
With my 25.9V Li-ion pack upgrades, (29.4V fully charged), the leds are semi usable hitting red at empty. - (Voltage limit for newer eZips)

Unfortunately, with my 33.3V upgrades, these leds are worthless, always green, even at (my) empty - 33.3V.
2009 eZip with 24V throttle and no HVC controller.

Almost as bad is my 2008 with 36V controller and throttle.
2nd green led doesn't even light at my 36.4V full charge point.
Red led lights with any full throttle start, and have a yellow-red flicker at cruising speed (16-18mph).
2008 eZip with 36V controller and throttle.

Gotta rig a 37V capable volt meter as my fuel gauge.
Analog with led light might be preferable to digital?
Throttle disengage should give prompt and reasonable "fuel" level.
 
Those throttle power lights are just comparing the voltage input to some preset voltage dividers or voltage references with external resistor, usually. So you could open up the throttle assembly, pull out the voltage lights boards, and trace the circuits out, note down the resistances, and experiment with different ones in their places (or if the chips are marked you might find their spec sheets and see what it says about how to do this).

In my DayGlo Avenger thread, there's a set of posts where I have closeup pics of the board in a typical ebike throttle, and I had planned to go do the above stuff--but I never got around to it due ot other things coming up. :(
http://www.endless-sphere.com/forums/viewtopic.php?p=314579#p314579
the chips on mine are TL431C. Dunno if my pics are clear enough to make a schematic from.
 
Back
Top