Yet another 18650 battery build.

[llile]

YOW! A warning for all who work on batteries:

You are working on a (hopefully) low voltage high current power source. (Or if you are on a 300V Leaf battery, you are working on a high voltage high power source).

You probably have metal tools, a metal soldering iron, and work on a bench that is likely covered with little globs of solder and bits of wire. Any of these conductors (including the roll of solder you are using) can get across two live busbars when you are building a battery, and ka-POW! If your BMS is anywhere in this high current path, then Ka-POW means you can play taps for the BMS. Done it more than once.

Ways to mitigate this risk of sparks include: Insulate the other end of any loose wire. When I put the positive wire on the positive bus, it already has a bit of heat shrink on the other end, so that loose end can't wave around and touch something. Keep themetal tools off the bench unless they are in your hand an in use. Use insulated tools, like screwdrivers with insulation all the way down the shaft. You can make one with electrical tape. But this won't work for your soldering iron. Keep the bench clean of any stray bits of conductors. Use a nonconductive bench (duh!). Don't hook up the BMS sense wires when they are connected to the BMS, nor when the BMS is connected to anything else. Make the heavy BMS connections last. Double check the BMS sense wires visually, and once again with a voltmeter. They should test the correct voltage in sequence - 3.9, 7.8, etc etc.

I'm always wearing glasses or safety glasses in the shop, as required. Regular specs for soldering, wraparound safety glasses for almost anything else. Whenever I go in the shop without specs, I'll get something in my eye straightaway, it never fails. Should never solder without some kind of safety glasses, a hot solder ball in the eye is no fun. Don't ask me how I know this.

 

[Lurkin]

Sage advice. I built a (slightly wobbly) workbench out of some leftover wooden shelving and wood scraps specifically to spot weld on - to avoid putting things on a bench with bits of wire, grease etc which might conduct 8) I like the the tools warnings - the electrical tape tip ftw.

The plan is to spot weld nickel strip over premade spacers (not unlike yours but no dovetail joiners). The spot welding is very straightforward, I think its easier than trying to solder at speed. I think I will do multiple layers of nickel to to increase the ability to carry current etc.

I was more worrying about the cell layout but I think its about as good as it will get with the space/overall battery size and shape.

 

[llile]

Insulation!

Some time ago I found a cheap Ebay deal on 12"x12" sheets of Kydex.

Kydex is a heat-moldable PVC plastic sheet. You can bake it at 250F for 2 minutes in a toaster over, then shape it pretty much any shape you want. Once it cools it retains the shape. People often use it to make gun holsters and other durable molded objects. I was able to build some special shoe inserts to relieve my fallen arches, in some custom shoes, but that's a whole 'nother blog.

Shaping it involves either pressing a big piece of foam over the object, or using thick gloves to press it by hand. You can then make adjustments with a heat gun if the results aren't satisfactory. It is possible to weld two sheets together using a soldering iron, providing you can deal with the nasty fumes. I use a fume extractor fan to get them away from my face, probably should use more ventilation than that.

There are lots of other kinds of insulation that could serve. Fish paper is commonly used, but I didn't think that was the best idea here. Heat Shrink is commonly used as well. I wanted to have a little better heat transfer out of the center of my batt, so I wanted to leave that area open. Heat shrink would trap any heat.

I liked the idea of having the BMS heat sink aluminum sticking out the end, and set directly against a metal (preferably aluminum) case to get rid of heat.

Lots of other materials would serve, I wanted something that is a good insulator, would not absorb water if any got inside the case, not going to fail due to abrasion, Kydex seems to fill the bill, and I had a stack of it.

 

[llile]

One of my strictest safety protocols: Insulate ends of battery plus and minus conductors before they are even installed. Never have more than one uninsulated terminal at a time, except during initial construction when this is unavoidable. Initial construction is being done on an insulating mat, carefully swept of little bits of metal and any tools.

About the first thing I did with the first ebioke battery I got was to short the ends together while soldering on some andersons, thereby blowing up the BMS. Ooops.

 

[llile]

Here's a completed battery module, with BMS heatsink prominent so that it can rest against the aluminum case of my battery enclosure.

 

[llile]

Detail of case construction. I'm bending most of the case up out of aluminum sheet (Thank you, junk pile, all the skeptics who say it'll all be there at my estate auction can just lump it). Some of the bending is with a simple Sheet metal bending tool, the rest is done just by clamping the metal down to my bench with angle iron and pipe clamps, then tapping it with a hammer until it is properly bent. Could really use a proper sheet metal bending brake about now, but I can make do.

 

[llile]

Testing!

I'm running a test with a homemade dummy load. I made up some power resistors out of Nichrome and some ceramic insulators. (Thank you, junk pile, once again the naysayers who said I was a hoarder were wrong). Each resistor is about ten ohms, two in parallel are 5 ohms, then two 5 ohms in series are ten ohms, but 4X the power. For a nominal 36V (starts at almost 40V) this draws about 4 amps, 160 watts. The resistors get hot to the touch, but not dangerously so. I'm doing the experiment on a piece of cement board for fire resistance just in case, and being very careful about separation between battery terminal conductors. I have an old DC ammeter that was an antique the day I was born, and a decent digital voltmeter, so I can take data on the amps, volts and time.

I got sick of that after a half hour, and put the system on the charger. I've proved that the system can provide current, battery doesn't get warm at all at 4 amps, and the charging function of the BMS works properly. I'm going to build some more power resistors so I can get the system current up for 15 amps or more, so I don't run out of patience in a realistic test. I want to make sure the BMS actually kicks in on undervoltage protect. These are used BMS's, the seller said they were tested, but trust and verify with data is a good policy.

 

[garolittle]

So much progress, must catch this thread up with picvs soon.

Garolittle, you are right, spot welding is much better. I'm soldering cuz no spot welder. I also intend that this isn't my first build - get my chops on cheap batteries then upgrade with newer better faster spot-weldier ones. :lol:

That is *awfully small* wire and I am skeptical you can successfully and repeatedly spot weld it without making the weld into a tinier fuse. Also, is it not copper? Will that even weld properly to nickel steel? Or maybe I misread and you are using nickel plated steel wire. You know for sure, you've been practicing. What is your experience? I recommend using a thicker wire, I found 26 gauge to work pretty well, low resistance, yet the cell can still blow it if theres a bad problem. These fuses will only activate if something catastrohpic happens like cell rupture.

Sorry for the slow response. Your project looks great! I am learning many things by reading your posts. Regarding your questions above, I used the same tinned copper fuse wire as this gentleman used ...

https://youtu.be/H_b3OJuQpEE

The spot welding seems to work very well. However, you are correct in stating that this wire is really too small. My electric go kart battery pack will consists of two separate 10S8P modules connected in parallel and has to sustain short bursts of up to 200amps (12.5 amps per cell). The cells (Samsung INR-25) are rated at 20 amps continuous so I will need a fuse wire that blows at about 13-15 amps. I plan to purchase the same spot welder as seen in the video above but I welcome any feedback you or any of your fans have. Keep up the great work and keep us posted.

 

[garolittle]

I am guessing that this would be the ideal set up for individually fused cells. Any idea were I can find such bus bars?

 

[flippy]

I am guessing that this would be the ideal set up for individually fused cells. Any idea were I can find such bus bars?

buy and strip a tesla.

i would not recommend this solution for high current packs. it increases the IR way too much and is not consistent enough when done manually.

 

[garolittle]

I am guessing that this would be the ideal set up for individually fused cells. Any idea were I can find such bus bars?

buy and strip a tesla.

i would not recommend this solution for high current packs. it increases the IR way too much and is not consistent enough when done manually.

Thanks flippy. By “... done manually” you mean the way portrayed here? Ie. a single thin bus bar as opposed to a wider/flatter bus bar as seen in the picture I posted? Thanks for your input.

 

[flippy]

yes, doing it manually will always yield slight differences. those small differences in wire length, solder, bending and so on seriously compound if you push 10+ amps on each wire making it load slighty uneven. doing that sonsitently over the entire pack compounds the issue more and more.

i am not arguing it does not work, its just not needed in DIY packs unless you have really shitty cells and want to protect the rest of the pack. but at that point you are already compounding problems already.

if you pull hard current you need to reduce heat production as much as practical, that means thick fat nickel strips in series and topped with 1 or 2 balancing parralel strips depending on the pack design.that gives a much stronger connection and one capacle of producing more power without heating up 200 potential fire hazards and failliure points.

 

[garolittle]

Thanks. This is very good information. At a risk of igniting a separate debate on a separate topic is the following summary of your statement correct? If you are using brand new cells from a reputable manufacturer there is no need to individually fuse each cell. In such cases you can simply use a spot welder and nickel straps. Correct?

 

[flippy]

you can do it like this:

or this:

depending on your design.

 

[garolittle]

very nice pack design. Tell me about that set up. What is your maximum amp draw? Do you experience any heat during your max amp draw? That is really nice. Thanks for posting that.

 

[flippy]

these are old pics of a few batteries i made last year or so.
the compact one is 1 block out of 3 of 16S18P made with panasonic PF's. it can do 180A continous at 60V and peaks probably into the 300A+. (10kW continous and 18kWish peak)
the smaller ones are for 2 bikes.

no, i dont expect much issues with heating as the peaks are very short (less then 2 seconds) and normal draw is around 30A, so less then 1.5A per cell.

 

[garolittle]

Good information. Thanks for sharing that.

 

[garolittle]

This is all great information I’m sorry to take up so much time on someone else’s thread. However it does emphasize an important point. The individually fused cells that you see on llile’s pack design are appealing since the chances of thermal runaway would be reduced by the fuses. However, I have to acknowledge that the increased resistance and heat that comes from using such fuses with high power applications is troublesome. There appears to be a dilemma between the safety offered by individually fused cells and the need to keep resistance (and heat) as low as possible by spot welding a nickel strips.

 

[flippy]

if you have good interconnects and spacing between the cells you get a lot less heating and if a cell pops you usually know it tue to the smell alone, the smoke might be a good second warning something is wrong or even the loss is capacity and a bms desperatly trying to balance the pack. if you use new/good cells there is zero reason to use wire fusing but plenty of reasons to not use it.

 

[garolittle]

....... if you use new/good cells there is zero reason to use wire fusing but plenty of reasons to not use it.

Very compelling points. I appreciate this information. I will be using brand new cells from a reputable dealer so your logic applies to my situation. Have you ever had a cell that was connected via spot-welding short-circuit?

 

[takyka]

Flippy,

You are wrong!

Appropirately dimensioned fuse wire does not generate significant losses.
My fuse wires are 0,3mm diameter and 10mm long. The measured resistance is 3,5mOhm. In comparison, the Pana PFs I use, has ~30mOhm internal resistance. ( by the way, how much is a spot welded nickel strip interconnection?)

For my entire pack (7S8P), the fuse wire loss is 7,7W@1,2kW peak power. You pay with 0,6% power loss for the safety! Is 0,6% matters?
Even if we trust, brand cells will never die, you have the added safety during the assembly process already. Examples are out there...

 

[llile]

Tesla uses fuses for a reason. Good cells shouldn't fail, but on the other hand, cars should not crash into each other either. The failure we are protecting against with fuse wires is a catastrophic failure that results from blunt trauma to the cells.

I also used fuses because they are way easier to solder. I can solder a fuse wire fast. Not so much a nickel strip.

As I said before, if I build another pack, I will buy or build a spot welder, and probably use nickel strips all the way. I might be able to figure out how to weld the fuses, as one guy says, but I'd experiment on this pile of dead cells until I was confident.

 

[llile]

I've been testing the packs, with a set of homemade power resistors, a digital voltmeter and an old analog DC ammeter.

I write down initial open circuit voltage, then take readings as 4 amps, 8 amps, 12 amps and 15 amps. I then run the cell at 15 amps, with a timer, and write down the measured pack voltage and amps every minute until the BMS cuts off.

The good news is, at 15 amps (about where I will be running) I feel no heat from the batteries, and just a very low heat from the BMS heat sink. I expect I will have no heat problems under real world conditions. I haven't measured temp on the BMS but it can't be more than 90°F by feel.

I'm having trouble though. Three of the packs have works OK, but cut off a bit early, should measure 7 amp-hours but only measuring about 5 amp hours. One of the packs will work at 4, 8, and 12 amps, but cuts out if I put the 15 amp load on it, voltage drops to zero. It will restart at lower amps, so this looks like low voltage cutout or overcurrent cutout.

At first I thought I had a defective BMS. However on closer inspection, I found that when the packs cut out, most cells are at 3.6-3.7 volts but one cell is at 3.0 volts, which is the proper low voltage cutout. I am pretty sure that all of my packs are suffering from balance problems and one layer of cells is too low on voltage.

I'm only charging at 40.5 volts for 10S. This should help the battery achieve high cycle life. Full charge voltage on these cells is 42 volts for 10S, or 4.2 volts per cell. It is possible that I'm not charging at a high enough voltage to achieve balance. Also it is possible I have a bad cell in one of the stacks. All of the cells were tested before assembly, about half tested full charge discharge charge, the rest tested just to charge, and all ended at 4.2 volts. All were disconnected and tested after a few hours for leakage, still held more than 4.1 volts.

I've left two tested packs on the charger all night (CV/CC+BMS, it's safe) to see if they cells balance better. I'll check each cell layer's voltage and test them again to see if I get better capacity and consistent cell voltages.

Whaddya think, am I charging at too low a voltage to achieve balancing? Or bad cell theory?

 

[flippy]

sorry but i think you have the wrong impression about why wires are used in industrial setups.

the reason tesla uses wires is extremely simple: it is the fastest and cheapest way to make a huge amount of connections with a robot. it has nothing to do with inherent safety of the cells, its just a free bonus to keep the uninformed masses from complaining. the risk of cells popping is too low to take into consideration for construction of the pack.

i have done a LOT of testing with wires as i build custom batteries almost full time now and i needed to find the best and fastest (not cheapest) way to make the connections, especially on big packs this is a huge issue as time really adds up on 30.000+ welds on a single pack.
my results were simply not consistent. i have probably much more sensitive equipment (scopes, 500.000 count multimeters and so on) and the differences were simply too great over time. after a few months it would mean the bms would be balancing non stop during charging to keep the pack balanced. that is not sustainable and a bad product. tesla uses robots to keep things perfectly the same and basically the best BMS system in the world today capable of pumping power between cells. also the most expensive. we mortals dont have that.


more ontopic with your issue, i think you have a crap bms. most bms (certainly the cheap ones) dont start balancing until they get ABOVE 4.2V. below that its doesnt do turd.
you need to get a better bms that can handle the current better, is programmble and you can read data from. better specs means less wasted heat, programmble means you can set the balance voltages much lower and reading data means you dont have to guess anymore. check the bluetooth bms topic.

 

[llile]

I'm beginning to believe you about the crap BMS. I think these were like $10, and I'm paying for that mistake now. I put one battery through a couple of charge-discharge cycles, and used a 44V CC-CV charger so I'm at the right voltage for balancing. This one battery still has 6 cells out of 10 that are 3.86- 4.05V after the charge current stops, the rest of the cells are at 4.2V. Voltage on the low cells hasn't changed in 2 charge-discharge cycles. Another battery has four cells that are low, six that are 4.2V. I'm really beginning to believe these are balance problems, and the BMS may be the culprit. These batteries are never going to hold full power if banks of cells are not fully balanced.

Let's quit banging on about the fuses on this thread, thanks for educating me, It's pretty clear I wouldn't build that way again, and wouldn't solder fuses again if I did the project over. Folks have made some great observations about fuses here, I've learned a lot from y'all. Tesla can pull it off with their cloned robot armies. Us mortals should be welding nickel strips. The rest of that discussion pros and cons should go in another thread.

Besides getting a battery out of this project, I also set out to get a lot of knowledge about building batteries from scratch, which has certainly been the case. Thanks guys!

Got any recommendations for a BMS rated >30A, 10S Li-Ion, with an aluminum heat sink, not crap?