yeah yeah yeah, I promised I’d post about field oriented control (FOC) and haven’t yet because that post is still awful.  The real motivation behind this project is to do Some Stupid Stuff with the inverter out of a gen 2 prius and a motor from another hybrid car.

Like a good MIT student about to graduate with a degree in mechanical engineering, I spend a number of hours every week browsing craigslist for junky motorcycles with search terms like “seized, rolling” and “title” – potential candidates for EV conversion.  In one such hour, I happened to spot this gem:


looks perfect.  I email the person, and hear nothing for a couple days, then, seemingly out of the blue I get an email from my buddy Mars saying:

Nick! Did you reply to a craigslist ad about a Yamaha XT350?

to which I replied:

yeah, I have a big electric motor I want to stick on it

Turns out I was around MITERS when Mars bought that bike a year or so ago.  Back then,  the engine was in bad shape but he was looking forward to getting it running.  He told me it had given him a real hassle and he just wanted to get rid of it.  Perfect.  Mars brought it over in the van and I traded him a brick of titanium I crufted a while back.  Even better, he had a friend who wanted the engine so I didn't have to deal with pulling it and getting rid of it.

And WHAM-O, a dirtbike looking thing appeared in MITERS.





The long-anticipated appearance of a donor vehicle turned out to be a great motivation and I spent that week fixing the brakes, stripping the frame, and fitting the motor and controller.

I cut out some new sprockets as the front sprocket was with the engine and the rear one was badly worn.  These were cut on the shiny new N51 waterjet which appeared after the Edgerton clubs got an ungodly amount of money dumped on them.  I screwed up and left out the nice holes on the big sprocket while making the toolpath.  dumb mistake.  we'll deal with it for now.


Mars drained the front brakes after a sticky master cylinder stopped the caliper from loosening.  Indeed the master cylinder was quite rusty.  I bought a $15 rebuild kit and cleaned everything out.  There were some sticky screws I couldn't get out of the cast aluminum brake fluid reservoir so I carefully TIG'd some heads to them.  Maybe there's some electrical or heat/cool stuff going on here too because I could loosen the (previously very stuck) screws with my fingers after this operation.






I gave the motor an open-up to install a shim to remove the couple millimeters of annoying axial slop.

IMG_0716 IMG_0718

with the back endplate off, I took the opportunity to make it round to provide a bit more room in the frame.  It was eyeball-centered on the rotary table.

IMG_0720 IMG_0721

I welded up a little subframe with some tabs to get a feel for where to put the motor.  The cardboard box motor model I made earlier was really handy here.  It is much easier to move and cut than the real motor.

somewhere in there, the idea of sticking the inverter heatsink out the front (radiator style) came up.  A quick template made of welding rod and some scrap 1/16″ wall square tube allowed this to come together very quickly.
IMG_0725  IMG_0727 IMG_0730 IMG_0732

at the beginning of that week, I preemptively ebayed the cheapest #428 chain I could find.  Two sections of 50cc dirtbike chain from china @ $15 shipped provided the necessary length and style at minimum cost.  If this thing ever ends up going anywhere then it might be able to justify itself a proper O-ring chain, but we're not there yet.


IMG_0733 IMG_0734

speaking of getting there, Bayley's new arduino shield thing came in.  he populated it and I figured out the connector situation.

IMG_0735 IMG_0736 IMG_0737

a bit of recalibrating sensors and swapping phase combinations resulted in the following bench supply fed test.

prius brick + ford fusion mg1 + XT350 – ICE =

naturally this was quite exciting and after hibernating for a number of hours we got back together and rigged up the necessary batteries to get the first proper ride in.  The past week has been an exercise in getting it done just enough to make it work so that we can move on to the next step faster.  Here, we've gotta prove that the thing can run stably at 80v with Real Current before wasting time making future stuff more legit.  The system uses a 12v battery to power the microcontroller, the logic side of the inverter, and the contactor.  The traction pack consists of two BremsPacks, 12s4p of LiFePO4 at a combined 80v.  The phase current limit was set at 80A but I don't think we got there due to some kind of control loop-related kthunk kthunking.  Likely the current loop crapping itself.

Still gotta figure out that throttle zeroing issue.  Zero throttle commands zero current which is all well and good, but the current loop kind of fails around zero because of sensor noise.  The result is that it “zeros” to an amp or two of current.  not enough to move anything but enough to whine as the inverter is trying to commutate.

huh, let's fix em.