jonescg's NEW electric racebike BUILD thread!

[Danny Mayes]

Aluminum is perfect for that, just a matter of working out what thickness...

The primary chain tension would probably be easiest from moving the motor rather than the jackshaft. moving the jackshaft would mess with your secondary chain tension...

These are good for this - http://www.smallparts.com.au/store/partslist/eccentricbushings/clampingandlocking/wide/1/ You wont get anything moving on you with these...

double row bearings will obviously handle more load for a given diameter. You can just go with bigger single row though. I guess it comes down to weight and rpm.

If you use a sleeve, it will need to be fixed to the shaft some how. would you spot weld it? I don't think it is an issue machining the shoulders onto the shaft. Otherwise just constrain the shaft with thrust bearings on either side...

Here is my build. A mini version of what you are doing http://www.endless-sphere.com/forums/viewtopic.php?f=28&t=31705

D

[jonescg]

Batteries.

Since I plan on using a 120 kW peak motor, which will probably be running an average of 35 kW throughout the course of a 15 minute race, 9 kWh would no doubt see a throttle-happy punter like me through. So with 7-10 kWh on board as a goal, and a 360 V nominal system (seems to be what the Rhineheart motor controllers are OK with) let's see what can be arranged.

360 V means 110 LiFePO4 cells in series, or 90 LiPo cells in series.

The only LiFePO4 worth using is A123's 20 Ah cells. A 1p110S arrangement would mean a 7.2 kWh pack capable of a maximum of ~500 A discharge. It would also weigh just 55 kg. So while I might be running out of juice on the last lap, I would have got there a bit quicker since the bike would have been pretty light. This would cost about $8.5k from Mavizen, or considerably less from Cell_Man.

LiPo gives a few more options. Using some of those Haiyin 6 Ah cells, I would need 90 in series. A 5p90s arrangement would give me my 10 kWh capable of a stupendous discharge rate. They are small format, so I can fit these into a Gixxer frame with a bit of clever planning and I'd be close to 77 kg worth of cells. At $13,000 that is just way too much money.

Turnigy cells work out much cheaper, and a 5p90s arrangement with these guys would make for a 8.3 kWh pack, weigh only 59 kg and cost about ~$6k after delivery. These cells would need some fancy arrangements too. Alternatively I can try to fit 6p90s, and have a 10 kWh pack weighing just over 70 kg, and cost about $7k.

OK, is there a problem with running a 1p110S arrangement? I'm expecting to be pulling ~500 A on occasion, which is 25C. If one of my cells is not up to the task, I have a whole system failure. A123 cells are pretty darn awesome; safe, easy to balance and monitor with a BMS, but would I be asking too much of smaller pack?

Are these Haiyin cells made from Unobtainium or is there a reason why they are so damn expensive?

Can I trust Turnigy cells to not fail/explode?

...

[Nuts&Volts]

Just wanted to let you know I found this source for a Rinehart here
http://vaxosystems.com.strawberry.arvix ... 100dx.aspx

They also sell Remy. Hopefully this gives you an idea of price comparisons

-Kyle

[jonescg]

Cheers Kyle, I was looking at their site earlier. Seems the 150DX is still 'in the pipeline'...

[Nuts&Volts]

Yea, I have only known of a few people using the 100kW version. I would assume thou that the new Lightning and the new Motoczysz use the 150kW model with the Remy Motors that each of them use. The 150kW controller is also listed on Rinehart Motions webpage and has been for awhile.

[jonescg]

On the subject of tensioning the primary chain...



42 links of 520 chain, rolling over two 17 tooth sprockets couldn't possibly need more tension than a few millimetres could it? The jackshaft unit itself, welded firmly to the triangular plates, would be pivoted on one of the motor mounts and a slot cut into the other side would enable you to tighten down once the correct tension was achieved. The other two motor attachment points would be utilised for frame mounting.

You could even put some kind of threaded rod on the side to push it up, but it's the sort of thing you would get right on the bench and then install it into the bike. Sure it puts the secondary chain out a bit, but that's my point of the primary not needing to be moved much, just enough to do the job.

[jonescg]

Should also point out that in the third dimension, there would be an identical plate on the other side with commensurate M10s holding the jackshaft assembly and plates to the motor. The primary drive chain would be on the RHS of the bike, with the jackshaft sprocket and secondary drive chain on the left. I was thinking a pair of M5 bolts threaded perpendicular to the M10 motor bolts could be used to adjust tension, and provide additional support to the friction being provided by the two M10s with big washers.



Or an alternative location for the adjuster bolts would be at the top since the bottom of the motor near the footpegs will be a bit tight for space:

[Arlo1]

I'm very excited to see how this goes.

[jonescg]

*Prof. Farnsworth voice*

"Good News Everybody!"

Neville from Evo got back to me. He said that the engineering team at Evo have agreed that a 360 V system would be ideal for for the AFM-130-3 motor, and encouraged me to bump the voltage up a bit higher. So a 400 V nominal system, delivering about 360 V under load should be perfect for this motor to deliver the goods.

He also said the Rhineheart inverter (PM100DXR)would be ideal, as it operates at higher voltages and delivers the same peak currents as the PM150DX, and is smaller to boot.

As for the prices - I was quoted 5000 GBP for the motor and about 8500 GBP for the inverter. The inverter price is for the PM150DX, so the 100DXR could end up being about the same. That's a hell of a lot of money, but still a tiny bit cheaper than a UQM motor

AND I have just discovered Google Sketch-up, so in future you will stop seeing these MS Paint abominations and some better diagrams instead

Check it out!!!

[Nuts&Volts]

Wow that is super expensive for an inverter. I think a better option would be to get a Tritium 200 for half the price. Run a 450-460V battery pack and a AFM-140-4. I believe you still should be able to get 120kW with that setup. You only add about 1kg with the tritium, but save ~$9000 AUS which could be used for lighter wheels and better suspension, which would give you more of an advantage than any on-board weight additions. Just a thought. Tritium would also be local making it easier to setup your motor and controller.

Here is a curve I pulled from Ben Grant from the AUS electric car forum. It shows theoretical curves for a EVO 4 turn with a Tritium with 391V supply.


Only other controller option you really have is the Sevcon EVO5 which peaks at 400VDC and 400Arms for about 5 seconds. No idea on cost of that guy thou

Edit: Just to do more battery calculations. You could run 130s A123 cells which would give you about 8.7kWh, max 368Adc would be 18C max on the cells. Charge them to 3.65V and then drain them a little to rest at 3.5V or so. This puts you under the 475VDC cap on the Tritium. A Lipo pack could be more like 112s charged to 4.15V or 4.2V each and could be in varying Ah size.

[jonescg]

Kyle - yeah, I wouldn't buy anything from the UK if I don't need to. The very same inverter from VAXO is 8000 USD

Otherwise, Tritium's controller isn't looking too bad. Still haven't heard anything back from UniTek about their controllers.

The AFM140 won't fit in my bike. It's considerably heavier (13 kg heavier) and the 400 mm diameter means it will occupy the entire lower half of my bike. So there is even less room for batteries.

Still working things out... Do you think 120-130 cells in series and only one in parallel is a good idea? If one cell plays up it brings the whole lot down pretty fast. ?

[jonescg]

OK, more proof that Google Sketchup is awesome:
I downloaded a 3D model someone else was kind enough to spend days making, and positioned my scale model EVO motor in place:




It's reassuring to know that the motor fits inside the bike. And sits reasonably central. Also, it's good to know that there is at least 150 mm of space in front of the motor, as that's where part of the battery pack is going to have to slide in.

The only thing left to model now is how I'm going to finance this...

[AussieJester]

I like the "Fred Flinstone" legs/feet in the top pic, are these
for show or functional? i.e when the battery runs flat do
the feet yabba dabba doo fap fap fap you home

Good luck getting the dollars CHRiS..

KiM

EDIT if you want a Pro version of Sketchup lemme know, have latest AutoCaD if your interested also ?

[Jay64]

If I'm not mistaken, that google sketch up is from an '05 GSXR 600 model. Back in '05, there was more of a difference in frame size between the 600 and the 1000. Double check the frame dimensions to make sure. You might even have more room than you think. You can use the measuring tape tool in scetchup to see how wide they made the opening between the frame sides.

[Nuts&Volts]

jonescg,

Just reread your post about it being a AFM130. I didnt realize EVO had another model. What I said should still be true thou. You may be able to get a higher turn model of the AFM130 (ie 130-4) and run it at a higher V with the Tritium and still get the 120kW with less current (less losses as well). FYI i sent you a PM

About the series thing. I dont have direct experience. But I look at it this way. In a 2p setup with 40Ah, if one cell craps out you will only be able to get 20Ah or 50% of the entire pack before killing the other cell paralleled with the crap cell. Most likely you'll end up killing both cells. In a 1P setup if you have a cell crap out you will end up with a dead/weak pack, but you only have one dead cell. I guess my thought is that in either way you are kinda screwed.

So a low "p" pack has less points of failure, but a worse failure mode (only pack dead). A high "p" pack will have more points of failure, but a less hindering failure mode (only XX% of pack loss). Either way I wouldnt risk the pack and I would either remove/repair/bypass the bad cells before discharging more. Also the lower "p" means less cells to test for duds in the first place. This is all theoretical ideas and I havent really done any of this in the real world so maybe someone else can chime in a little

[Nuts&Volts]

Oh I wanted to add one of thing about the Rinehart. This controller does not have a self-learning tool for the motor, so you have to send in your motor and controller to Rinehart to have them set it all up on a Dyno and determine the ideal settings on the controller. I have heard this is quite expensive as well

Just thought I would pass this information along

[jonescg]

Thanks guys,

Jay - the 3D model I downloaded claims to be a Gixxer1000, but I was really just playing around. Measure twice cut once, but model a thousand times cause it's cheap

Kyle - Yeah, I reached the same conclusion about paralleled cells. Failure is more definitive when you have a 1P pack, which is good and bad. I can't really commit to anything until I have the frame in hand and I can measure the available space. If there is less room than I first thought, I will go for LiPo - no choice really. But if there is enough room to do what I want with LiFePO4 then I can go for a 1p120s A123 arrangement. Smaller pack, but easier to manage.

UniTek got back to me about their BAMOCAR-D3-700-400-x controllers!

Dear Mr Jones,
Our drive will work with the EVO motor. We already have an application running with this motor. There are a few limitation up to now. We can use the motor only in a star-delta connection. There is no fieldweakening.
best regards
Thomas Labod

UniTek Industrie Elektronik GmbH
Hans-Paul-Kaysser-Str.1
71397 Nellmersbach
Germany

He attached the same AFM130 datasheet which I emailed you. I'm assuming star-delta format means I'd have to run the very high voltage system?

[jonescg]

So I visited my local Ye Olde Bearing Shoppe and they were rather helpful. Assuming 25 mm is a standard size for most output shafts, and if splined appropriately, the right size for a Gixxer front sprocket. They suggested a pair of 6305 bearings which take a 25 mm ID and 62 mm OD. They are 17 mm wide and are good for 11,000 rpm and take a high load. And at $17 each, they are the least of my expenses!

I have minimised any unnecessary turning of the shaft, so hopefully I can get it made up with Z-13 splines at either end easily enough. Sure, I could probably slide tube over it to give me inner support for the inner races, but then it wouldn't be able to take any kind of radial force without it being fixed to the shaft somehow.





The whole unit is now only 80 mm in diameter, which leaves enough room (9 mm worth) for welding the plates to the sides. Would this affect the tube diameter so much that I couldn't get the bearings pressed in and out? Should I have them welded then get the steps turned out? I can't work out how to draw a circlip groove in Sketchup... Anyway, hopefully I don't need to get this turned out of a solid rod of ally, but I'm sure heavy duty ally tube is available.

I also stopped in at the wreckers to see what they had. Not much, but there was a Megelli there in satisfactory condition. You could have a lot of fun with it with a small motor or couple of motors. I also measured up a '00 GSXR600 frame and it's not much smaller than the thousand. I still think the thou will be better- more room and all.

[jonescg]

Just a bit more playing with Sketchup:





This is a scale model of a 10 kWh battery pack positioned inside the model of a GSXR1000 (Again, much credit to whoever went to the effort of making this model!). It uses 540 of those individual LiPo cells from Turnigy, 5p108S. To my surprise, they appear to fit relatively well, even with an extra 12 mm worth of acetal plastic holding them together. I also did a model where my motor goes in there too and while it appeared to fit everything, this is purely a simulation.

Pretty cool huh?

Anyway, better just focus on Voltron's next race, and saving enough dough to pay for this little exercice

[jonescg]

Well for my birthday, the Lovely Katherine gave me a copy of TurboCAD 18! Holy crap is that one complicated bit of software! Not nearly as easy to use as Sketchup. Good thing is, I can draw stuff in Sketchup and open them in Turbo, then save them in a format that suits most CNC/ laser cutters.

[SplinterOz]

Well first off... Happy Birthday.

I also think this Katherine must be a keeper

[voicecoils]

Happy birthday Chris!

What a sweet gift

It's super convenient that a CAD model already exists for the GSXR1000 and enough lipo can fit in a sensible looking vertical stack.

Over at elmoto.net, teddillard is scheming up a battery mount idea where a group of lipo bricks can slide in and out of the big pack blade server style. I wonder if something like that could be incorporated into your stack?

If you could slide out a row of 6 bricks (or 12 for the top of the stack) as one tray, they could easily be balanced, serviced, bad packs removed etc. With some clever use of bullet connectors and busbars, you might be able to minmise wiring make each tray hot-swappable.

[jonescg]

It is so easy to Ctrl-C and Ctrl-V $6000 worth of LiPo and stuff it into a bike

Yeah, there are pros and cons to having a pack that consists of smaller sub-packs. Pulling it apart to swap out dud cells would be a lot easier with smaller sub-units, but at the same time, that is a lot of connectors and tabs/pins to contend with. It also makes waterproofing a bit harder to maintain too.

I haven't drawn the acetal box for the cells yet, but I'd be getting ahead of myself if I did since I don't really know how much room I will have. An L shape is ideal because it makes use of all of the tank space and still allows a fair bit of weight to slide down into the bike, keeping the CoG a bit lower too. I can make a 120 cell A123 pack in the same manner and it would still fit, but not quite so easily as the LiPos.

Of course a pack of this size is going to weigh 65-75 kg, so you would need a winch to hoist it out of the bike. If I used A123s, I'd virtually never need to do that But if it was Hobbyking LiPo, then yeah, I'll probably be lumping it out quite a lot of the time. Even if the one-third of the pack was separate, it's still a 40-50 kg mass to haul out.

Gotta hand it to Google - they really made a sweet program with Sketch-up. TurboCAD is bloody clunky and counter-intuitive in comparison.

[voicecoils]

Yeah, lots of fun things to think about!

How about one of these (or fabricating your own) for next year:



At least you & Ripperton could both make use of it

[Nuts&Volts]

Happy Birthday Sir! and Go Fast