Homemade Battery Packs

Thanks so much DrkAngel! After some researching, I've decided to use Panasonic protected 3400's. Currently I only have a 25w soldering iron though I know I should get a better one to be safe. Is 40w enough of an increase to make a difference or should I go to 60w?
 
I would recommend an 80W soldering iron. You want fast heat. - In and out like Seal Team Six, then cool that cell down, baby.
 
drmacgyver said:
Thanks so much DrkAngel! After some researching, I've decided to use Panasonic protected 3400's. Currently I only have a 25w soldering iron though I know I should get a better one to be safe. Is 40w enough of an increase to make a difference or should I go to 60w?
Higher is better ... if, it has a temperature adjustment.

"Panasonic protected 3400's" ? If there is an added disc as protection on end of cell, it will likely be damaged by soldering (fragile-sensitve soldered components inside disk.)
Probably cheaper and better to get unprotected cells, but add a BMS.

If replacing SLA, Dakota LiFe 12V 10Ah packs are ≤$200 per pair delivered, and are plug and play replacements.
Can use SLA charger or get a fast, up to, 10A charger.
 
Ah ok, unprotected it is. I will definitely be including a bms (http://www.ebay.com/itm/1-PCB-for-18650-22-2V-6-Cells-Li-ion-Lipo-Battery-Pack-/400791608958?pt=LH_DefaultDomain_0&hash=item5d510a9e7e). I'm switching from a 6s5000mah lipo battery, so I already have this charger: https://www.hobbyking.com/hobbyking/store/__6478__IMAX_B6_AC_Charger_Discharger_1_6_Cells_GENUINE_.html
With this ESC: https://www.hobbyking.com/hobbyking/store/__10370__HobbyKing_150A_High_performance_Brushless_Car_ESC.html
I'm making a 6s pack so I don't have to replace the parts above.
And this is my motor: https://www.hobbyking.com/hobbyking/store/__18175__Turnigy_Aerodrive_SK3_5055_280kv_Brushless_Outrunner_Motor.html

I'm building this pack for an eboard, if I forgot to mention. There's a couple threads on 18650's in that section of ES if you're interested in seeing whats been going on there.
 
DA, Im confused. I intend (per say) to put together a pack with 208 cells 48 v. , 13s 16p (correct me if im wrong at anytime) It appears that most of your packs are wired (individual cells) in series then those strings are paralleled. I had never considered a BMS before but now im looking at the adappto unit and the onboard balancing seems enticing. My dilemma, 209 balance leads????? Now, it was suggested on an adappto thread that 16p 13s is the way to go ,which of coarse reduces the # of BL but wont moniter the individual cells. Can you give any input on this? I value your opinion. thnx
 
It would make much more sense to wire the 16P cells together first, and then series each group of 16P for 13S.
This way, you would only need 14 balance leads (One for each parallel group + the last lead). Just be sure that your parallel cells are roughly the same voltage before you wire them in parallel. (Charge to storage voltage of ~3.85V per cell)
 
douglashart said:
DA, Im confused. I intend (per say) to put together a pack with 208 cells 48 v. , 13s 16p (correct me if im wrong at anytime) It appears that most of your packs are wired (individual cells) in series then those strings are paralleled. I had never considered a BMS before but now im looking at the adappto unit and the onboard balancing seems enticing. My dilemma, 209 balance leads????? Now, it was suggested on an adappto thread that 16p 13s is the way to go ,which of coarse reduces the # of BL but wont moniter the individual cells. Can you give any input on this? I value your opinion. thnx
I build "in series" but then connect in parallel.
Then I connect balance leads to the parallel wires.
8s12p = 9 balance wires

file.php


As a final check.
I discharge at least once before adding parallel or balance wires.
Any bad-poor cell will show as much lower voltage.
 
douglashart said:
DA, Im confused. I intend (per say) to put together a pack with 208 cells 48 v. , 13s 16p (correct me if im wrong at anytime) It appears that most of your packs are wired (individual cells) in series then those strings are paralleled. I had never considered a BMS before but now im looking at the adappto unit and the onboard balancing seems enticing. My dilemma, 209 balance leads????? Now, it was suggested on an adappto thread that 16p 13s is the way to go ,which of coarse reduces the # of BL but wont moniter the individual cells. Can you give any input on this? I value your opinion. thnx
A BMS supposedly protects against over and under voltage on every cell.
Additionally, most provide an equalization feature.
This adjusts all cells to an identical voltage at full charge.

Sadly this is only reasonably effective if all cells are of good condition and all banks are near identical capacity.
I forgot to mention ... I tend to view a BMS as a Band-Aid for a defective pack.

I feel it essential to thoroughly test every cell for:
1. self-discharge
Then test & rate every cells
2. capacity
I do also, especially with mixed brands of cells, test for
3. comparative IR
Then I build banks of,
4. equal capacity and IR

Rather than constantly fighting to patch a defective pack with a BMS ...
I strive to build to a one time "proper" build that is monitored and alarmed at a bank level.

See - Bulk Capacity Testing - for specific details and methods.
 
DrkAngel said:
I build "in series" but then connect in parallel.
Then I connect balance leads to the parallel wires.
8s12p = 9 balance wires
DA, can you elaborate on this? Perhaps a better picture showing all of your connections clearly?
 
Picture has 6 batteries of 8s2p.
(Battery" actually refers to identical components arrayed and working in concert, eg artillery battery)
These 6 batteries are then paralleled using the bare "parallel" wires.
The balance connector is connected to the ends of the parallel wires, 5 on top (shown) and 4 underneath (hidden)

file.php

Cells with white discs are Positive - bare w/red ring Negative.
Pairs of cells were pulled from battery packs, left connected with the handy extension "tab" for connecting in series.

Many want to separate every cell, test, then reassemble.
I test in pairs, sometimes in 3's and most recently in 4's with the 10,400mAh packs for latest project.
If pairs etc. are substandard I will separate, hoping for 1, or some, good cells.

I strive for banks of equal capacity, rather than cells of equal capacity.

After these ... hmmm ... 7 years of Lithium pack building ...
One of the surprising things I noticed is that Lithium cells tend to deteriorate at a reliable percentage rate rather than at a capacity rate.
eg Same brand 80% capacity cell paralleled with a 40% capacity cell, tend to diminish in capacity at the same rate. 80% to 40% and 40% to 20%, rather than 80% to 40% and 40% to 0%.
This is reasonable, when you take into account that the weaker cell is charged and discharged at the same residual C rate as the better.
In series, by comparison, the weaker cell would be discharging twice as fast at double the residual C rate, and might deteriorate at 4x the rate as the stronger cell.
Oh! C rate should be calculated at present, rather than original, capacity!

So, after removing self-discharging cells (the bane of any pack builder) I build banks of equal capacity, resting assured of an stable and equal deterioration.
While I do not! use a BMS I do employ a bank level low voltage meter-alarm. <$5.
I bulk charge ... rather than balance charge = no need!
After removing self-discharging cells and having banks of equal capacity, (= equal future deterioration), banks remain balanced. (other than due to physical damage eg broken solder connections!)
I do monitor bank voltage at "full" and "empty".
Any variance indicates a problem and is hunted down!

Sorry! I do not use a BMS ... corrected above.
 
Thnx DrkAngel, Ive studied your threads and consider you the guru of grassroots battery building, no shortcuts. I plan to use the bike im building for traveling, months out on the road so my interest in the adappto with bms is, firstly, that I might moniter(dashboard readout) the cells without having to open the box, and secondly, to have the 'bandaid', till I get home for repairs. That bms can be disabled (until needed), but without going thru it , im not sure I can moniter the pack with the adaptto readout(convenience). I intend to buy new Panasonics and capacity test the whole batch before building. Anyway in that picture it appears to me that each cell has a BL and the (say) #2 cells in every S string , the B leads are connected in parallel with each other . Is that correct?
 
Each "Balance" lead is connected to the corresponding "Parallel."
 
Thanks a lot Dark, I think im beginning to grasp some of this. I do understand your meaning of a 'proper' and thank you for the education. Im all set up for testing and I wont go on a trip without a proven battery. Can you tell me what you set you LV buzzer at?
 
douglashart said:
Thanks a lot Dark, I think im beginning to grasp some of this. I do understand your meaning of a 'proper' and thank you for the education. Im all set up for testing and I wont go on a trip without a proven battery. Can you tell me what you set you LV buzzer at?
My most common cells are "empty" at about 3.7V.
I set my LV buzzer to 3.7V ...
Now, with voltage sag, this gives me a bit of limp home juice without dropping actual cell voltage below 3.7V.

Every cell type is different tho!
3.7V seems to be at the high side, but typical with many of the brands I've tested.

See - Capacity Mapping
 
Typically tho...
As the setting for low voltage alarm.

With battery at optimal DOD, apply standard throttle, note battery voltage.
Set Low voltage alarm to this voltage.

Or slightly above ... to allow a bit of extended travel.
 
douglashart said:
Thnx DrkAngel, Ive studied your threads and consider you the guru of grassroots battery building, no shortcuts. I plan to use the bike im building for traveling, months out on the road so my interest in the adappto with bms is, firstly, that I might moniter(dashboard readout) the cells without having to open the box, and secondly, to have the 'bandaid', till I get home for repairs. That bms can be disabled (until needed), but without going thru it , im not sure I can moniter the pack with the adaptto readout(convenience). I intend to buy new Panasonics and capacity test the whole batch before building. Anyway in that picture it appears to me that each cell has a BL and the (say) #2 cells in every S string , the B leads are connected in parallel with each other . Is that correct?
"Traveling, months out on the road."
For emergencies, I recommend a few spare cells with leads attached..
If there is a problem with a bank of cells, eg voltage dips lower after or during discharge, tack or jumper additional cell(s) to that bank.
 
I recently reviewed a video that indicated that one type of battery deterioration was directly related to time spent charging.
It seems that the graphite becomes plated with lithium during the charge cycle.
Reducing the time spent charging, increasing amps, reduces the degree of plating per Ah of charge.

So ...
I will upgrade my 25.9V 5A charger to 10A. (25.9V 25.92Ah battery)
and
I will upgrade my 36V 4.5A charger to 10-15A. (33.3V 31.2Ah battery)
Li-ion (Laptop LiCo) are variously recommended at .5C-.8C charge rate
 
Got 108 excellently capacity matched cells ready for my 33.3V 31.2Ah, (1kWh), eZip pack build.
33.3V x 31.2Ah = 1038.96Wh (1kWh+)
Compared to the OEM
24V x 10Ah 240Wh/2(SLA) = 120Wh
So 800% the usable capacity ... at about 70% the weight!

Will be using 6 12pks of Sony and 3 12pks of Sanyo

file.php


Recharging cells now and should have pack built this weekend.

Looking to prolong life as long as possible, so will, typically, be charging to 4.05V per cell, possibly lower.
Charging to 4.10V instead of 4.20V is reputed to double cycle life, 4.05 likely to triple it!
Will determine optimal voltages after capacity mapping Sony cells.

Not built as a replacement for my re-re-recycled 33.3V 31.2Ah eZip pack.
file.php

Now at less than 50% of original capacity ...
But all cells have diminished at a nearly identical rate and banks are still remarkably equal in capacity.
(These cells have been built and rebuilt into multiple of my pack builds)
I will retire old pack to Winter duty ... don't want to take any extended trips in the cold and snow ...
Pushing 60 and the cold reminds me of every ol' "poor decision" I've ever made.
 
With multiple batteries and certain batteries dedication for certain bikes or seasons, Snow Beast etc., I thought I'd better investigate battery storage procedures.

Turns out most batteries store best at 0ºC or 32ºF.

Found this chart which examples Life saved vs lost, due to storage temperature.
Also, the preferred voltage level for storage.

file.php


See - Batteries vs Cold
 
33.3V battery seems to sag to the ~31V LVC (Low Voltage Cutoff) of my 36V controller.
Rather than a distinct disconnect, controller seems to progressively reduce current-amps!
Especially in the cold of Winter, this reduces performance and usability of my 33.3V batteries.

Fortunately I have found a 24-36V compatible controller ... no HVC (High Voltage Cutoff).
Yi-Yun YK31C - Now - $13.29 delivered!

Connectors must be adapted to fit the eZips, but this allows 24V, 25.9V, 29.6V, 33.3V, 37V ... up to the 50V limit of the caps.
Great.
I even purchased the 36V version of the controller as a means of using a 36V battery as well as the easy to construct 44.4V pack.
Sadly, the 24V and the 36V version seem identical with the 50V caps ...
Too close for extended use at the 50.4V rated voltage of a 44.4V battery.
So I purchased and will install 63V caps.
This should allow 24V - 48V battery packs on the same bike.
A "switch" could even re-path a modest 24V 40Ah battery to 48V 20Ah for high performance "Turbo" mode!

Seeming identical ... I might be able to directly compare 24V and 36V version to determine LVC circuit and adjust for a better 33.3V LVC, Same for adapting 24V controller for 22.2V pack! ...?
 
Back
Top