So the Prius Gen2 transaxle has a final drive ratio of 4.11, and the KB Laser final drive ratio is 3.77. I’m unsure if this will be problematic but considering MG2 is rated for 400nm and the Laser’s ICE is rated for 92nm I shouldn’t have any torque issues.
Due to the DIY nature of the controller I’ll be using, I don’t think it has any sort of output designed for reading RPM, but in saying that I know the main connector for the inverter has some spare pins and the project is open source so I could always just create the reading myself. Servo on the speedo might be easier honestly
For the fuel gauge the plan will be to use this: https://www.zeva.com.au/index.php?product=126
Its a simple product and has an analog voltage output which means its perfect for my IP. For voltages I do plan on having some way to read that but hopefully will either be hidden or very discreet, haven’t decided on this yet
I’m not sure on the exact number, but I should have at least a couple hundred pounds spare for batteries, but I will be weighing the empty shell at a weighbridge before the new transaxle goes in so I know how much headroom I have
Regarding the welding, I’ve always wanted to learn and think it would be a great skill to have. This project is the perfect opportunity but I will definitely need some practice. I’ll have to go out and buy a welder and some tube steel and get going!
I wanna make sure that the transaxle is properly supported, so I’ll probably use the original transmission mount with some extra support structures coming off another structural source. I’m gonna have to have a harder think regarding batteries in the rear. I have considered cutting out the floor of the car where the fuel tank is and essentially extending the area at the back of car that is lower and using that for batteries. That will come in the future I think
(Discourse has a bug that fucks the rotation)
Here we have a package about the size of a large toaster, which is capable of outputting potentially hundreds of kilowatts of power. Not only does it contain two inverters, there is also a boost converter which takes the 203V battery and boosts it up to the 500V that the motors are capable of using. In the bottom there is also a DC-DC converter that is capable of 100A+ at 14.2V. Inside there is lots of complicated circuitry that goes over my head but I do know that it uses IGBTs to achieve its power ratings.
Here you can see the main DC battery connection, as well as the multipin Ampseal connectors which are really cool because they are waterproof. I am unsure of where these connectors will be going yet as the current setup for using this hardware doesn’t have defined things like connectors yet. This will come in the future.
This unit is watercooled as it is very high power but due to its size and very little thermal mass. Even when doing bench testing it is recommended that you use even a simple 12v pump and water because otherwise it’ll burn up very quickly.
Toyota’s engineers have really hit a jackpot here as this is a super powerful little unit and its basically indestructible. It supports battery voltages of ~120V-~400V which means a broad variety of batteries can be used. When trying to pump too many amps into this thing, it won’t blow up, it’ll just shut down. When trying to overheat this thing, it won’t blow up, it’ll just shut down. It handles faults gracefully which means that this should be perfect for me as someone who blows stuff up regularly.
I’m about to do a teardown of this thing so if anyone wants photos of anything in particular just let me know
Damien also released a new board revision yesterday for the Prius control board, and he specifically added the required components for me to be able to run both MG1 and MG2 in parallel, hooray!
That seems pretty clever, kind of a reimagination of a clamp style DC ammeter. My knee jerk was that it seemed expensive, but I haven’t seen anything else that handles DC current, particularly with the convenient outputs.
I definitely support the the impetus and learning welding, so show off your test pieces and skill development here too.
You might be able to reinforce/replace whatever subframe was present, since I’m guessing that was stamped steel, so a tube structure should give you some more strength down below.
The rear will be more difficult for sure. Besides the strap for the fuel tank and hangers for the exhaust, you’ll have to look around for hard points. I’m not sure what the suspension is back there, a torsion beam (the era makes me uncertain, but being a mazda, possibly…)? Tying the two sides together would make for a humorous driving experience, but maybe not what you’re looking for
I was impatient watching the referenced youtube videos, so I’m interested in the teardown for sure. the 500v step up seems particularly interesting.
So I just got the old transmission out of the car. Honestly wasn’t that hard, jiggling the axles put was the hardest part.
Seeing the engine bay without the transmission allows me to visualise better how I want to do things:
In this image I show how the transmission is currently mounted, horizontally with the axles at the back, in green I have shown possibly how the new transmission could be mounted, vertically with the axles at the bottom:
The current fuel tank simply bolts into the body of the vehicle, so I think I will just reuse the tank and bolt it back in to keep things easy for me, at least for now.
That video of the transaxle I have referenced like at least a dozen times in the last few weeks and I’m sick of watching but old mate does such a good job of explaining things. I will be starting the teardown tomorrow as it’s raining so I’ll be stick inside all day. I tried today but I have up quickly due to the heap of black silicone holding the two halves together
One funny thing I noticed was when I was removing the speedo cable, on the end was a plastic gear. I swear I remember a heated discussion about not using plastic gears in our esk8 gear drives. This vehicle has 140000kms on it with no broken gear…
A speedometer drive is not really transmitting any power.
A spiral gear like that has very little load on it and will wear very slowly unless something jams up in which case you want it to fail to save the transmission.
Just finished the teardown of the inverter. Not a full teardown, but I needed to figure out how to get into the enclosure because I will need to replace a board eventually
Here under one of the top covers is access to one set of 3 phase connections as well as an interlock circuit that will stop the inverter from being able to run without the cover on
Underneath the top lid is the circuitry for the current sensors. Here you can see a section that is covered by silicone of some sort. Good way to waterproof
And here is the star of the show. The top PCB is the control and interface circuitry for everything and the one underneath is the power stage PCB. Under that is the IGBTs and then cooling fins and a separate cooling channel for the water cooling. This top board will be the one that I have to replace with Damian’s one
Just a small update, I have been able to confirm that the electric motors in the Prius transaxle do work.
I’m gonna play around with encoders in the motors tomorrow see if I can get smooth operation out of them. In these videos you can see that there is lots of stuttering from the transmission, which if you watch Damian’s video he did on the first spin he noticed the same thing. I think once the planetary gear set has been welded in the way it’s supposed to be I should be good
@Gamer43 any chance you’d know what the letters actually correspond to? I’m gonna guess the ones ending in G are the encoder signals but that’s simply a guess
Looks like each signal has its own ground reference (the G suffix), doesn’t look like MRF is needed (looks like the sin cos carrier signal)
Those waveforms… bleh, connecting them is not going to be straight forward. I don’t really have experience with such resolves, so I don’t think I can help much :(, sorry.
Basically, it looks like the MCU will have to do a bit of math to use MRF to demodulate the sin and cos signals to get the the raw sin and cos values.
Well in that case, I’ll connect MCS and MSN to H1, H2 on vesc, MRFG, MCSG and MSNG to ground and see what happens. I’ll try MRF as well. Thanks for the help mate
