Glad you’re interested! This is such a large project for me as I’m usually a pretty sloppy and cheap DIY guy but really looking forward to giving this a go. Hoping to have some updates about hardware I can share in the next few weeks
4 Likes
P112 Prius Gen2 Transaxle
I present to you the P112 transaxle, the latest in hybrid vehicle technology, if it were the mid 2000s. In this 100kg package we have a bell housing, which is the interface which connects to the ICE; 2 electric AC motors, one that is 50kw and another that is around 30kw; as well as the necessary gear systems, electrical and mechanical interfaces. With some clever welding of part of the power split device, we can cause both MG1 and MG2 to spin at the same speed.
Those wanting to learn more about this transaxle should watch this video: https://youtu.be/ia6jJO3cuOY
The current implementation for the Damien Maguire control board for the inverter only allows use of MG2 for propelling the vehicle, and uses MG1 for charging. An earlier board version did in fact control both motors but this function has since been removed. I may custom order my own board of the older version so that I can gain use of both motors as that will mean a combined power of ~80kw and 500nm+.
The rotational maximum speed of MG2 is rated at 6500rpm. The original vehicles maximum allowed speed was 103mph and at this speed MG2 is running at ~5700rpm. MG1 is rated at 10000rpm which means I have a bit of headroom to work with, not to mention I’ll never get to 103mph ever on the roads. This gives me piece of mind knowing this hardware can work as hard as I need it to.
I believe this transaxle has an oil pump built into the end of the outer housing for MG2, which is also a 3phase AC motor. This has a connection directly into the inverter which means that cooling for this can be easily done. To simplify things I will probably just use a 12v pump as that will make things cheaper. This transaxle, along with the inverter, require cooling as there is very little thermal mass given to cooling. I will be ensuring to build a robust thermal solution once installed in the vehicle.
In the coming weeks I do plan on opening this thing up and giving it a clean and making sure everything inside is as should be. There is also an AC compressor that I need to remove and decide if I want to use it. There is currently a small effort trying to reverse engineer this specific compressor for DIY use.
Here are a few photos, more coming in days to come:
If anyone has any questions ask away. Inverter is also on its way. That writeup should be more detailed regarding the underlying tech of running these motors
6 Likes
Bonus: here is Damian with a running configuration that uses the exact hardware I plan on using, minus a slightly different variant of inverter
3 Likes
Do you plan on using MG1 to get above that 6500rpm limit of MG2? Or are you just going to leave well enough alone with a 103mph top speed? 
I ask because as I recall, there have been a couple discussions here about using motors with different kV ratings (not sure if thats the case with these two motors, but the principal might still apply) where folks concluded that the back-emf from the lower-limit motor would outweigh the work that the higher-limit motor is putting out. Not to mention potentially confusing your motor controllers. Not sure how much or if any of that is applicable to your hardware though. 
Either way Im super excited to see this project moving forward!
Any progress in the battery department? Still planning on using the SPIM packs?
3 Likes
Awesome build man, don’t know how I missed it before, being that this is also a dream of mine one day, I will be following closely
2 Likes
So I don’t plan on going above the 103mph limit as this won’t be used at any drag strips or anything. This harware is interesting in that the controller simply gets given a torque command from the pedal then it sends that torque to the motors. Back EMF isn’t an issue I believe because these are 3 phase AC motors, correct me if wrong.
I am also still looking at the SPIM cells, @BluPenguin is getting a quote done on sending 72 cells to me. This hardware I’m using is cool because it will accept anywhere from 140-350+ volts so I can use a wide variety of different pack sizes and configurations. I will probably buy more cells in the future too as this project moves on and power requirements change
1 Like
Glad you are enjoying it! I hope that my particular DIY method is interesting as it’s easily one of the cheapest options for a DIY EV, its just super hacky in this early stage
2 Likes
Hacky is fine as long as it doesn’t catch fire
What I really wanted to do is to convert motorcycle one day, but make it have enough highway range to travel, and capable of fast charging, but given that I can count the number of charge stations for the whole country in the hands is a bit of a pipe dream for now
2 Likes
You’d be surprised how much range you can cram into a motorcycle if you’re creative. I’ve also seen people charging at like 5kw on a bike, but there are DIY solutions for faster charging (open source control boards for tesla chargers) if you have the space for it
1 Like
I am really looking forward to the documentation of how you actually mount all the parts and wire it up! The fabrication work is what has really held me back from looking into EV conversions more. (That and cost, of course).
3 Likes
Honestly I’m looking forward to that too because I’ve got no idea right now 
I need to start making measurements of the current transmission and figure out things like whether I need to get new shafts and whether the current mounts will be even remotely useful
1 Like
I’ve done a few sketches on some common motorcycle frames
Ideally I would want to cram at least 20 kWh, not that easy given that the only one with something near are the Enégicas, giving almost 200 km range on highway
Here’s some measurements for those interested. The last two photos shows the outer distance between the two shaft holes for the half shafts. There is a decent difference here so that means I will need at least 1 different half shaft depending on where the transmission will be mounted. I need to get the old one out so I can see how the mounts will work. It’s looking like there will be a decent amount of welding though unfortunately
4 Likes
Before you put that monster in, will you verify that it works ok?
I want some videos 
2 Likes
Absolutely! Once I’ve got the main parts there will be plenty of videos showing how it all works. Its gonna take me a couple months to get everything so bear with me, but the end result should be cool
2 Likes
This looks like an interesting project for sure! I like the idea of a manual transmission coupled with the electric motor quite a bit.
Bookmarking, since prius motors are cheap and might be a fun microcar propulsion device…
A couple points of interest:
I’m not familiar with the laser, and it’s been a while since I’ve played with something of this vintage, but I suspect the prius’s 6.5k rpm likely meshes well with the original tach, and that the tach/(speedo?) output is a convenient cable coming out from the top of the trans to your cluster.
From what a coworker that formerly worked in PHEV batteries tells me, most OEMS opted for cell blocks of 14-16S (iirc, tesla excepted) which were then strung together, have you given much thought into your BMS setup?
From the discussion on the SPIM08HP cells, it seems that they sag pretty quick, so you may need a larger initial pack for testing, particularly as the Wh/km consumption is going to be orders of magnitude greater than our standards.
I would have mentioned Chevrolet Volt batteries as a good price/performance option, but those modules have the packaging issue you’re trying to work around
The transmission plate adapter looks like a good candidate for a waterjet, possibly secondary processing with a mill to clean up the mount holes if they’re on shoulder bolts/dowels. This should be cheaper than finding a place to EDM your profile, but that may be worth checking for a comparison, particularly as you’d lose a good deal of profile tolerance on the back side at thicker dimensions with the waterjet.
I don’t have access to one anymore, but when I had needed unobtanium (2d) parts, I just scanned them on the xerox, loaded into corel/illustrator and auto traced the outlines. a bit of minimal cleanup if I needed something with a dimension better than +/- .5mm and I had my new gasket or w/e to laser/waterjet.
Your local EV group will probably have far more valuable input, so I’m interested to see what they have to say about some of the roadblocks.
2 Likes
Lots of useful info here, and I love having discussions about this project so keep it coming.
I’m no longer using the original transmission as the Prius transaxle contains everything I need. The old tranny will be coming out this week depending on the weather.
This is true of the old transmission, not sure if the new one uses a manual speedo cable still or if it’s electric. I am scouring the Prius service manual to determine this. The document is over 4000 pages long and not in order so taking me some time to get through it. If the Prius transaxle does not have a manual cable, then I will use a laser to figure out the final drive RPM then drive a small motor in the instrument cluster that runs the speedo at the correct speed.
Honestly BMS is pretty low on my list right now. When I start, as the build is being tested and stuff, I was planning on getting a few of those charge only BMSes from Ali that let you daisy chain the modules for much larger series packs. Once I have made final decisions regarding batteru s I can start to figure this out. As long as I have a way to check each series voltage I’ll be happy. I don’t plan on adding extra gauges and stuff like DIY EV builds usually do. I want to keep the original look and feel of the stock cluster.
My issue for batteries is that I’m in Australia, where we don’t really sell EVs besides Tesla (just seen my second ever Leaf on the road today). I have been looking at the SPIM08HP for a couple parallel groups (at around 50S for 200v) which would be located close to the battery, and then I would have a bunch of high capacity lower amperage cells. This method is currently employed by @MoeStooge in his race boards successfully. If my testing shows that this won’t work then I’ll either just use SPIM08HPs. Those cells can be bought and shipped to Australia, which I’m currently having sorted. BatteryHookup sometimes has some Leaf (22kwh) packs for only $1600US but due to weight (650lbs) it would be nearly impossible to get it shipped here without going bankrupt.
As far as I can tell our local group no longer exists. Updates haven’t been posted to their website for like 18 months. The statewide EV group also was unable to run their yearly EV Expo this year due to lack of interest for t group. I’ve found as I delve further into this niche that there really isn’t anyone interested in it, which is disappointing. We live in the most exciting time for electric vehicles now is a great time to go out and build your own.
I’ve noticed you know a lot about materials technologies, I was wondering what you think would be the best way to design mounts for the new transmission? It’s bigger and heavier than the old one so obviously I wanna make it as strong as possible. I’ve never welded before but am willing to learn if it can be done fairly easily
My reading comprehension is a bit better in the light of day
Removing the old trans will certainly save a good amount of weight and package space. My only check in that case is that you’ve verified the final output rpm of the new transaxle to the laser’s original first and fourth/fifth gear to make sure you have enough starting torque along with your top speed.
otherwise, it may be prudent to keep mg1 and mg2 independent to take advantage of the plantery system.
You might have already seen this, but it’s a site that I found informative a decade(!) ago: http://eahart.com/prius/psd/
though I’ll need to dig to get the flash player functional to play with the sliders…
by 2000, I’m guessing that toyota moved to the wheel hub mounted speed encoders, the rpm of the motors should be something that you’d be able to read from the motor controller itself like how we read out erpm on the vesc (how open and accessible it is for you to do that is another matter (I still need to delve into the controllers you mentioned)) in the end, I’m guessing the easier option will be to use the laser (or maybe a simple hall encoder) and a servo like you mentioned.
I just went back to your picture of the IP and noticed that the cluster omits the tach, so only one output to deal with in any case.
I wasn’t aware of the ali chainable BMS, but that sounds like it has promise. I’d still want a way to read out the min voltage of any of the P-groups though, just because the sheer number of cells compounds failure incidence. Maybe using the fuel gauge for overall voltage/coulomb count with the option of a toggle to flip to read out minimum cell voltage? Or recessing a voltage readout behind the existing battery emergency light? Some sort of pack health monitor in any case…
The modularity is definitely what makes it a cost effective setup. A sort of charge and ride on a larger scale, I guess.
…that’s unfortunate. as good as the internet, video overviews and whatnot are, there’s still a lot of value in the tactile interface that you can get hands on with someone else’s project. Nothing to be done about it, just a lament on my part.
Looking at the wikipedia entry for the laser, the E series’s iron block, plus the transmission must be heavier than the prius’s aluminum housing+steel stator(s)?
I think as long as you can secure the new motor transaxle assembly, countering the hanging moment arms, you should be ok.
I’d definitely try to use as many of the existing motor and transmission mount points as I could. that way you can keep/source oem dampers to avoid stress risers and fatigue on the sheet metal in non structural areas (I think the weight in the fuel tank will be a greater concern there…)
Look at how the mounts are setup to see if it’s likely that the engine itself was a stressed member that formed part of the structure to inspire your spaceframe.
Beyond the existing motor mounts, possibly a “K-brace” between the strut towers and firewall?
In any case, welding is a great skill, I highly recommend picking it up for the doors it opens up for you.
If you ask around, you’ll probably have most people pushing you towards TIG for the control you get.
Since I don’t have that finesse, I’m more fond of MIG since it’s a few less things to coordinate about while you’re casting lightning and slinging molten steel. Both are fairly quick and learnable to the point that your work is structural, if not aesthetic, though, so that’s ultimately up to you. You’ll get more respect for knowing TIG . 
Look into tube stock that racers around you would use for their roll cages. CAD would naturally be best, but again, as a seat of the pants person, I think you could get away with a lot of hand measurements and test fits. (besides, getting CAD of a 36 year old vehicle sounds implausible without paying far too much money for a cloud scan)
Make sure you form triangles with your joints, and your biggest challenge will be in cutting the tubes so they mate cleanly and you’re not trying to stitch them together with weld.
…this is where you’ll curse not having CAD as you go back and forth grinding fishmouths into tubes to guess and check fit
3 Likes