jonescg's NEW electric racebike BUILD thread!

[Gordo]

I nearly posted that video myself, but if I've understood right it means that all the clamping force has to come from a pair of threaded plastic rods. Plastics tend to creep under load, so the clamping force will gradually decrease with time. One possibility would be to use steel threaded rod with a series of insulating sleeves around it. The total length of the sleeves would have to be slightly less than the clamped length of the rod. The spacers would also have to be incompressible, so plastic is out again.

We have used that nylon rod for many years to prevent electrolysis on fishing gear, mounting 200 lb pullers with them. The gear never comes loose. It is relatively easy to maintain the clamping force, by using a 2 or 3 turn lock washer, essentially a spring. Marine aluminum blocks with a high copper content, used with a conducting paste would ensure good contact. There are lots of plastic compounds which do not compress. I expect to see Jack R. build a pack using this method soon.

 

[Arlo1]

Probably a little late for you jonescg, but i thought this may interest you..
http://vimeo.com/34702692
..Any idea who "mic ofOz" may be ?

Thats awesome!

 

[jonescg]

I owe you one! Next time you're on your way over to South Africa, drop in to Perth and I'll shout you a beer Nah seriously, name your price, that's awesome.

It's all good, but thanks for the offer! I'll try and come to one of the races in Victoria this year.

Unfortunately they are in Sydney, Goulburn and Ipswich (Qld) this year, and I can't afford to race without forgoing my new bike build. Unless someone wants to race it for me, of course.

 

[Hillhater]

..... I expect to see Jack R. build a pack using this method soon.

I believe he already has built "prototype" packs ... note the last sentence .
From the EVTV blog....(worth following.)

.....I've made this module THREE TIMES in prototype form. The last has some 48 cells in a 4P12S configuration for a 40 volt 75 Ah module. It is 19.75 inches long, about eight inches high and about nine inches wide and weighs about 75 lbs. It represents a 3 kWh pack.

The prototype has spaghetti wiring on it for the Cellog 8S modules to monitor individual cell voltages. We bottom balanced the entire set of cells at 2.75 volts during assembly. We'll be charging it and discharging it to see what happens to the individual cells. Under load, poor connections and so forth show up.

This is kind of a leap of faith. We used nylon threaded rod and some pvc nuts and by following Mic's instructions and including a nut with every insulated piece, we think we've managed the clamping forces as well as they can be managed. But we do not want flexing between the tabs and the spacers in the vibrating environment of a vehicle. So we've cast the ENTIRE thing in the urethane resin we've been playing with.

 

[jonescg]

It's interesting, but for now I'm sold on soldering them together, or at least spotwelding something solderable onto it, then soldering. If the cells are rock solid and pass QC, then there should be no need to undo them. A big ask for HobbyKing stuff :lol:

Also this method would only really work for cars, as it's too impractical a shape for a motorbike.

 

[Malcolm]

We have used that nylon rod for many years to prevent electrolysis on fishing gear, mounting 200 lb pullers with them. The gear never comes loose.

That's good to hear. I'm not trying to knock this method Gordo, I think it's very neat. Just pointing out that thermoplastics like nylon do creep with time: http://www.intechpower.com/material-information/creep-time-limits-and-yield-stress/

Nylon has better creep resistance than most, and using spring washers will help a lot, but the ambient temperature is likely to be 20 degrees higher in this application than on a fishing boat so the creep rate would be four times higher. I would still want to be able to check the tension regularly. You're right of course about the spacers, some plastics would be much better than others – something like Tufnol would be good.

 

[Gordo]

They are actually copper ones, just rather flimsy at the business end. The only time a termination got hot was when the connection was poor and the heat generated was sunk buy the 2 mm copper plate. Even then it wasn't that bad.
Provided every connection I do is a solid one, there shouldn't be any overheating worth mentioning under hard use.

I have had so much trouble with crappy jumper cables over the years, I've tried many solutions. I've never been able to find solid copper ones, but have found brass spring clamps for welding cables. All the "copper" automotive clamps I have found, are copper plated, only @ 0.005 of copper. It burns off on the first use. On the copper plated ones, the wire was poorly crimped to the clamp, again causing an arc and burn of the copper. Then after you leave them sit for a month, you have a rusty steel connection to the cable. I've silphos'd the clamp to the wire and then had the steel spring turn red and relax because of poor connection on one half of the clamp. I guess the worst were the aluminum cables, a few flexes and you were done

 

[jonescg]

Actually now that I look at it, Gordo, they are steel with piss-weak copper plating. Oh well, they seem to be holding out to my abuse.

NOW, I got my PCBs in the mail!!! PCBFast are fast, good to deal with and the product is perfect... but man it would be expensive to get 40 of these made up

This is how they stack up:

Obviously I would remove the protective tape, balance charge all of them in one big paralleled pack and then slide them into position. Then I would fold the tabs down to parallel each set of 3 cells, trimming with side cutters if need be. This will be important on the cells where the tabs are really close together; maybe cut the tabs in half and fold them over like two interlocked Ls? Then I'd carefully buff each exposed tab with emery and tin them with a big generous blob of solder.

Then I'd take my tinned copper bus bars, place them across the appropriate tabs and stick a big hot mofo of a soldering iron on the top and melt it down. I'd hold the copper down with a bit of dowel so it got a good solid connection. Then when it's set, dab the top of the copper with a sponge in ice-water. Probably a good idea to place the balance wire in place while I do this, or maybe tap each block so I can screw a ring crimp into it? Any ideas are welcome.

It might also be worth while devising some kind of protective wall for the adjacent copper bus. If the solder spreads too far it could potentially short...

At the very least, I'll have a 36 V, 15 Ah e-bike pack with a plug for balance charging

 

[Nuts&Volts]

Hey that's much better than this

and probably won't kill nearly as many cells. I can guarantee you that the number of failed cells we have would more than pay for the PCBs needed to build your pack

PS - made from 25C 4s/6s packs taken apart (desoldered, un-stickied together) and then resoldered/welded together in parallel/series groups

 

[jonescg]

ROFL! :lol:
Kyle that's terrible! I feel pretty good about my ham-fisted duckshit soldering now

I think it will be important to use a big iron on the bus, and do it as quickly as practical. Big copper terminations means no unnecessary heating, and that's heat which would ordinarily go straight into the cell. Any ideas on the balance wires?

 

[Nuts&Volts]

Haha yea I know. Hey when you dont have a lot of money you can't beat that approach. Please note I didnt build or design that pack, I only inherited and am trying to get the most out of it. It delivers 90kW+! but at 6.5C a 430V pack dropped all the way to 403V on 25C cells Seems like a lot to me, a 35Ah battery drops 0.24V on average at only 6.5C when its rated to 25C cont.

Anyways, solder on balance wires wasn't too bad from what I heard and none of them are coming loose. I actually think it would be cool to utilize the PCB for the balance wires. Route them all to one point and through hole solder on a JST connector or something that could be dual purposed for monitoring (Cell-log) and balancing (RC charger). I got the idea from this pack http://www.metricmind.com/audi/14-battery.htm and simplified it.

Also the most failed cells are the ones that were soldered to the big copper busbars on the ends, because they needed a lot of heat to melt the solder and actually get a connection. Basically the outside cells in the picture. I've learned this from having to disassemble them.

 

[jonescg]

I thought about that, but it would make things kind of difficult. If the wires are free to be moved, it makes assembling and disassembling the whole pack that little bit less of a PITA. It gives me options if the wires are accessible. I'm wondering whether or not to include higher current taps in the pack so I can balance charge 14 cells at a time fairly quickly...

I still need to solder and unsolder any dud cells if (nay, when) I encounter them, so having a solid fixture mounted on the copper bus would help with removal of key interconnectors, such as between each 10s sub-pack.

Wow, that Metric Mind pack is EPIC!

 

[jonescg]

OK while hiding inside from the heat, I managed to assemble 90% of my first 10s, 3p lipo pack. The cells were all about 3.83 V, so that's a good start. I unsoldered my previous monstrosity and cleaned it up. A bit of heat goes into the top of the cells when you do this, which should be avoided if you can help it. Do it once, do it right

I racked them up three by three, and bent the tabs over in the correct order. Seems you have to do very little tab trimming if you fold them in the right order. I didn't want any tabs overlapping, as this makes the surface uneven, and might mean a poor tab collection after soldering.

This process would be greatly furthered through the use of a jig, designed to compress the cells a little from all sides. Then they can be taped up with one round of cellotape. I have nothing between the cells, so hopefully this is not a problem in terms of rubbing.

I'm still waiting on 5 more cells (Hong Kong, Singapore, Switzerland (WTF?) and then Perth). I might get to cutting up my bits of copper. I reckon provided there is a small hole in the bus bar, I can leave a bit of 14 or 16 AWG wire in the hole for a balance tap. Then just hot-glue these down the middle, or around the sides might be easier? As they will form a maximum stack of 17 or 19S, it might be easy to make the 10s pack with wires out one side, and the 7 or 9 S pack have them coming out the other. I'll see.

Makes for a nice, compact pack though!

 

[SplinterOz]

That is not very big at all. Looks like it will be small light and powerful pack.

 

[jonescg]

I know! I was surprised to see how compact it is when you terminate this way. I will feel a whole lot better about a pack with a PCB separating my solder efforts from the cells below

They shouldn't exceed 140 mm tall, and are exactly 50 mm wide and as thick as the number of cells you want.

I made a bit of a pictorial for how I'm putting it together:

First, cut your copper to size, and drill any fixtures you want. I have drilled a 3 mm hole so when soldering the pack together, I can stick a balance tap wire in there. I will try to run these down the side. Then, buff the crap out of it. If it's not shiny, the solder won't stick.

It takes a while... I will have to do about 170 of these!

Then with a big soldering iron, tin them on one side. No real need to tin both sides, but when it comes to soldering the balance wire in, it will help a lot.

This is approximately what it will look like when complete:

Now, I stopped here for a couple of good reasons. I want to wait until all of my cells have arrived so I can cello-tape them in a tight bunch. Why cellotape? Cause it's transparent and you can see an impending cell failure before it happens

Even better if I had a compact jig to hold them in place while I tape it. Then I will go along and buff each tab. Maybe not with the dremmel, but at least some emery paper. Then I need to tin the tabs. This is tough cause if it doesn't stick quickly, you end up heating the cell below. I'm tempted to put them in the freezer first or something... Then when a nice healthy blob of solder is on each cell to contact the bus, I can solder the lot on. But in order to do this well I need a source of pressure to hold the bus down while I do it. I want to make sure every cell has good surface contact with the bus while I push down.

I better see what I can knock up in the shed

 

[adrian_sm]

Then, buff the crap out of it. If it's not shiny, the solder won't stick.

It takes a while... I will have to do about 170 of these!

Could you chemically etch them instead? I think that is how the home brew PCB guys get their traces (mask and etch), so the supplies shouldn't be too hard to find.

- Adrian

 

[Bluefang]

you could always clean then up by putting them in some acid for 10min. Just buy some acid from the local pool shop, that should be strong enough

 

[jonescg]

Err, no thanks. I might just buff them with a wire wheel, very carefully

I'm devising a system for applying pressure on the middle of a busbar, and heating with two soldering irons either side. Provided about 60 N of force is applied they should sit down very nicely.

Better pack my bags - off to the nation's capitol tomorrow! In amongst hanging out with fellow e-bike-nerds, I will be working at ANU all week, so PM me for a catch up!

 

[SplinterOz]

wow time flies... will see you over here.

 

[jonescg]

I am back from Canberra where I spent daylight hours doing plant molecular biology on a microfluidics chip, and night hours drinking beer with good friends

I finally got my last 5 cells in the mail! So I got myself a second soldering iron and hopped to it!

I scuffed up the tabs using a grinder wheel on my Dremel. I tried the wire-wheel but it just seemed to polish them. Emery paper was tedious so the little grit wheel was fine. Although I would probably grind a bit more for the next pack. I want to see nothing but shiny copper - no grey.

That was all of them tinned.

I set up a (rather flimsy) rig where I could put good pressure down on the middle of the copper while dumping the heat of two soldering irons into the bus. The giant wrench was just the right amount of pressure. I used a small bit of FR4 to insulate the steel rod from the copper, just to make heating that bit more efficient.

I then put two 80 W irons either side of the copper and heated that sucker till it meted and sat down snugly on the tabs.
Here's a video of me doing just that, but I noticed a large gap in one of the tabs, so I added a bit more solder. Probably took far too long, but all of the other pieces of copper went on in a matter of seconds. The chilled water was perfect for sucking the remaining heat out of the bus.

http://www.youtube.com/watch?v=PbKCw6r5wZQ&feature=youtu.be

And so the final product will be a bit more refined than this.

I didn't get a chance to solder the balance wires on cause I don't have three hands. When I build a better rig for soldering tabs this way, I will try to incorporate some means of holding the balance wire down too.

Next task it to attach these leads, solder some decent lugs on for testing, and loading it to determining capacity. I need to work out how good the tab-to-copper junction was for all of them. A continuity test should work I suppose.

 

[Hugues]

Hi Chris,

congratulations, saw your video, you make it look simple ! I like the part with the 2 soldering irons, looks like you're about to give these little bastards an electro choc !

Good idea for the rig too !

That pack seems pretty compact, what's the voltage, aH and max amps ? and then dimensions, while you're at it.

I'm still struggling to fit enough batteries on my chopper project, i think i'll have to scale down a bit on range i'm afraid, because the GBS batteries from Elite take quite some space. Or I go with your idea, seems much more compact per wh.

But one question: are you planning to register your bike at your local DOT ? or race only on private tracks ? I need to register mine in order to commute, and reading all the regulations i need to comply with, i'm not sure i'll be able to build the pack myself and be compliant. Maybe i should make the "jump" now with an Elite pack, register the bike, then later on build my own denser pack for the fun rides. 8)

Keep us posted !

 

[jonescg]

Hi Hugues,

I will be building a much better rig for doing this! Also, one of my irons has a chisel tip, so it doesn't dump the heat as quickly as I'd like.

The pack is made from individual Turnigy 40C LiPo cells. The arrangement is 10s3p, so 15 Ah capacity, and about 40 volts when ready to roll. I don't plan on drawing any more than about 18C out of these cells, at an absolute upper maximum. 250 A should be about right. It's destined for a 700 V pack. This little sub-pack stands 270 mm high, 50 mm wide and 140 mm deep. I don't know where to run the balance leads as they don't really fit down the side. Total weight is just over 4 kg.

I might register the new bike but that's not on my radar for another couple of years. My current bike is also unregistered, but I think I should at least try to get it sorted this year. It would make an awesome commute, but it is definitely a dry weather bike.

 

[jonescg]

So when it comes to balance charging, I have a single 1S-6S balance charger from Jaycar. So I made two 5S balance taps with the second one sharing a negative with the end of the first one.

Soldering the leads on was no big deal, again I had to heat the copper up, stick it and chill it again. If I had a third hand I'd hold the balance wire on at the same time as soldering the bars.

 

[jonescg]

Today I soldered a couple of longer copper plates to the + and - terminals so I have something to clamp onto when doing my load testing. One thing that was abundantly obvious was the need for a third or forth hand. I wish I was Elzar, but alas, I need to build that jig for compressing the cells and soldering the bus bars. Might get onto that tomorrow.

But I started thinking of a decent battery crate to put all these cells into. Its size really depends on what format I go with or whether I make it one rectangular box or a stepped, or L shaped box. Since an AFM140 will take up the entire lower half of the bike, I might as well make it a near rectangular box.

What to make it out of? Well acetal is the most easily machined material, but I don't know about it's UV resistance or weight, and whether something better might work. In any case it will be containing >70 kg worth of LiPo, so it needs to be pretty tough. I have drawn the basic shape using 10 mm acetal sides and base, while the top can be something a bit lighter. It would screw together using long, countersunk M4 screws. The slot at the bottom is to connect the front half with the back half. A piece of 2 mm thick and 20 mm wide copper would join them, soldered at both ends. The front and back plates would then screw into the sides with a bit of that yoga-mat type foam to keep things snug.

I really want to include the contactor, shunt, fuse, balance taps and CA inside the battery enclosure, but it's already 460 mm high. I think I'd be struggling to fit 6 columns of cells across the box, as the frame isn't quite big enough to fit such a cube.

I was shopping around on the FX website and saw a couple of mid 2000s GSXR1000 trackbikes for sale for $7000. I hope they are track-ready (i.e. suspension is sorted) for these prices. Still, I reckon I can flog the motor and exhaust for a small sum.

 

[jonescg]

After realising that soldering balance leads and HVC/LVC wires to the copper blocks was quite a challenge, I have decided to get a proper PCB made up with traces out to holes. The only thing I'd need to do is to make sure the adjacent tabs have plenty of solder to make a good connection. No big deal. It also makes getting it wrong when loading the cells a bit more memorable

So without any idea how to use PCB software, good old MS Paint comes to the rescue.

At the end of a turn, I would simply flip the board and populate it with cells going the other way. Rather than make it hard to comprehend with holes everywhere, I reckon this way I'd have to try pretty hard to get it wrong. PCBs can have traces on both sides right? It costs as much whether you have one trace or ten I suppose?