[DRIVING] RX8 with Leaf EM57 motor and Gen3 Inverter

[Kelju]

My conversion project car is a 2005 high power version Mazda RX8.
Let me first start with a few (too many) words about how I came to start this project.
I used to own a second generation RX7 from the 80's. It was a fun car to drive especially in the winter time as it did not have enough power to go sideways on dry asphalt. as with most of these cars, this one also had the annoying feature of starting problems with a semi-hot engine. When these symptoms arise, it normally means a complete overhaul for the wankel engine so I decided to pass the car on to someone with the needed motivation.
A few years after in 2017, I found myself still thinking about the car and its great handling. At that time I also noted that RX8’s were quite affordable when importing from Germany to Finland and decided to go for it and bought a high power version. I was very much intrigued by the driving feel of the car once again. But there was a problem, though. I was constantly worried about the wankel engine. What if it fails on me? Quite fast, I decided to sell the car as a very enthusiastic buyer came along.
You know how it is? You don’t actually need another car – let alone another project car – and still in late 2019 I was once again checking if there is anything interesting available. Then I came across it. It was perfect for me. A bargain high power version RX8 with an already broken down wankel engine with endless possibilities to fit anything under the hood.
The initial idea was to shoehorn a BMW V8 in there as I have quite a bit of previous experience of those engines, but a few measurements later, I decided not to go that route. For a while I even thought of getting a refurbished wankel. One strong candidate was the Alfa Romeo V6 Busso, that I got lying around. That would have been a straight forward conversion…
I am an electronics designer by trade with several years of experience in embedded systems and power electronics so naturally the thought of going electric crossed my mind. I also had some previous experience with the open source Vesc project. Then I stumbled upon the Open Inverter forum and started “researching” the Leaf powertrain hacking. Finding an interesting deal for an EM57 motor and Gen3 inverter on ebay was the last nail in the coffin.
As it turns out, replacing the Nissan OEM brain with the Open inverter brain is not so straight forward. You can read up on the details here viewtopic.php?f=4&t=980
So at least for the time being, I am reverting to controlling the stock inverter via CAN. This obviously means that the initial goal of achieving similar or higher power performance will not be met. Not the biggest issue at this point as my first battery setup will not provide adequate power to reach that goal anyway.
My more realistic target is to get everything working and electrically safe so that I can get the car street legal, inspected and registered.

Here is the list of components I will be using for the build:

- EM57 motor from a 2018 Nissan Leaf
- 110kW Gen3 inverter from a 2018 Nissan Leaf
- Nissan Leaf PDM with 6kW AC-charger and Chademo fast charging
- 8pcs Enerdel NMC battery modules with total capacity of 12kWh from a Th!nk City EV
- Zombieverter VCU
- Th!nk City electric AC-compressor
- Th!nk City electric vacuum pump for brake booster
- What liquid heater should I use?

[arber333]

- What liquid heater should I use?

May i suggest outlander heater which is run via CAN? Outlander AC compressor is run via CAN as well...
https://openinverter.org/forum/viewtopi ... =19&t=1207
https://openinverter.org/forum/viewtopi ... =19&t=1997

[Kelju]

May i suggest outlander heater which is run via CAN?

Thanks for the tip. CAN control sure is the preferred way in my book.

[Kelju]

The plan is to have a beefy enough battery at some point to have good acceleration comparable to the original.
To get the best possible acceleration, I decided to keep the clutch in front of the 6-speed manual transmission. I am interested to see how it feels to change gears with an EV in a similar way as you would with an ICE powered car.
This meant keeping the flywheel also. As the Leaf motor output shaft bearings are not designed to experience any significant radial load, it was pretty obvious that the flywheel could not be just simply “bolted” on the output shaft. Some sort of extra bearing is needed to handle the radial load induced by the flywheel.
The Leaf motor and the RX8 transmission needs a spacer in between to house and support the flywheel. The spacer is also needed to prevent the Leaf inverter sitting on top of the motor from hitting the firewall.
The wankel engine is a modular design with cast iron end blocks and aluminum “spacers” with steel sleeves for housing the rotors.
As I had the defective wankel engine available, I decided to take that into use. I noticed that the eccentric shaft (equivalent to the crankshaft of a normal four-stroke engine) slip bearing housing is a separate part bolted on the cast engine end block. That part cannot be used as is, because of the slip bearings and lubrication relying on oil pressure which would not be available in my EV conversion. I took some measurements of the part and modeled a custom bearing housing. Also measured the eccentric shaft and modeled a custom shaft assembly with splined hole for the Leaf axle on one end, needed features to attached the RX8 flywheel to the other end and bearing inner race surface and retainer ring grooves along the axle.
The cast iron end block is used as is and bolted on the transmission. One of the aluminum spacers is machined to suitable thickness of 50mm. An 8mm aluminum plate with the correct bolt pattern is bolted on to the motor. The plate is machined with several threated M10 screw holes for attachment to the spacer assembly.

I cut of the end of the eccentric shaft that connects to the flywheel to use it for visualization and figuring out the correct dimensions for the flywheel bearing assembly:

Here is an overview picture of the planned assembly:

The first part I had made was the Leaf axle spline adapter which bolts on to the right side of the above assembly:

Second part was the adapter axle featuring the correct angle, locking pin and threads for the big nut fixing the flywheel:

Once there was an axle to attach the flywheel onto, it was time to work on the flywheel. I am guessing the clutch had slipped at some point as the surface of the flywheel was not perfect. it was straightened out on the lathe. Also, it turns out that the flywheel of a wankel engine is not balanced on its own. There is a considerable lump of iron on the other side of the flywheel and the combination of the eccentric shaft and the flywheel is balanced as a set. In order to make use of the flywheel, it needs to be balanced on its own, so the iron lump needed to go. While on the lathe, all the excess material (at least as much as we dared) was removed along with the lump:

Here is a picture of the part set:

The eagle eyed viewer might have noticed from the previous picture that I made a mistake with the fixing holes of the bearing housing. Somehow I managed to model eight holes instead of six. Luckily I was able to fix this by some additional holes. I decided to thread the leftover holes so that they can be used for pulling the part out of its seat as it is a quite tight fit:

Here is a picture of the complete assembly:

[muehlpower]

I'm a bit confused about the attachment of the Leaf axis spline adapter. 4 screws from below, 4 from above and different sizes?

[Kelju]

I'm a bit confused about the attachment of the Leaf axis spline adapter. 4 screws from below, 4 from above and different sizes?

That is a very good question indeed.
I was quite worried about the tensile strength of the M6 bolts under the stress induced by the motor torque.
This is why the spline adapter fitment to the bearing axle was made in such a way that in order to attach, prior to the assembly, the spline adapter was kept several hours in the freezer and the bearing axle was heated up in an oven. The actual assembly was done in a hydraulic press.
The threads of four of the spline adapter fixing screw holes where actually drilled out and the holes where used for alignment of the other four fixing holes with guiding pins, because after pressing, there would not be any possibility to fix misalignment of the fixing holes.
The guiding pins were screwed into the threaded holes on the bearing axle.
After pressing, the guiding pins were removed and the two pieces were finally fixed together with 8 screws. Four on either side due to the availability of the threads.

Here is a view prior to pressing the parts together. The guiding pins are also visible:

[Kelju]

Continuing on the walk through of the motor assembly.

Adapter plate attached to the motor. Centered with a specifically made centering tool:

The reused wankel engine block parts attached with the custom bearing housing:

Flywheel in place. The sliced-up wankel cylinder used as a spacer, can be seen from this angle:

Clutch disc and pressure plate assembled on the flywheel:

And here the motor is finally in the engine bay attached to the gearbox and ready for dimensioning of the motor mounts:

[johu]

Cool, you did as arber suggested to me back when I did the Touran conversion: use part of the engine as a spacer.

I haven't given up on the control board yet. Maybe you can try to convert your design files to a more common file format? Or redo the design in KiCAD and using the mini mainboard (https://openinverter.org/shop/index.php ... duct_id=70)

I have a very motivated tester who'd stick the inverter into his conversion, provided we find the silly round connector somewhere.

EDIT: I'm sure Arber will join in when it comes to using a clutch in an EV conversion

[arber333]

Argh! No.... clutch....no.....☠ RIP.

Well i think by now you all should know how i feel about slipping surfaces and electric torque request. Why o why would you feel the need to use a flywheel. It pulls down your leaf motor and restricts its torque development.

Seriously, that is really sound mechanical engineering. Commendable. I think if we would focus more on using the front part of engine shaft and bearing there would be less problems with assembly.

Also how do you think 3rd gear is doing? Is there lots of vibration and noise? We found out just about every RX8 transmission had worn out 3rd gear.

[Kelju]

I did not dare to leave out the clutch as I am worried about the endurance of the synchros. I would be happy to hear about succesfull experiences with changing gears on the fly without the clutch.

How do you think 3rd gear is doing? Is there lots of vibration and noise? We found out just about every RX8 transmission had worn out 3rd gear.

This remains to be seen. I bought the car with a busted wankel which had already been taken out by the PO.

[arber333]

I did not dare to leave out the clutch as I am worried about the endurance of the synchros. I would be happy to hear about succesfull experiences with changing gears on the fly without the clutch.

You still have an option not to clutch while shifting. Try how it goes.
I have tried many things and i found out synchros are not the issue if your MTX is good.
The main thing for good clutchless shifting is to release the rotor. No regen or torque command with shiftimg.
I do this using a BMS pin on openinverter with a button i built in the shift lever. It works really well omy Mazda MX3. But i see a lot of play in the gearbox with 2nd gear...
https://mazdamx3ev.wordpress.com/2018/0 ... lopka-ven/

[johu]

Yes, same here. I hardly ever shift at all and if I do I just temporarily disable regen (via cruise control button) and just shift

[Kelju]

Does anybody have experience of running the cooling circuit backwards? To be more specific, it seems that for simplicity of hosing, it would be beneficial for me to run the coolant backwards at least in the PDM. Any thoughts?

[arber333]

Does anybody have experience of running the cooling circuit backwards? To be more specific, it seems that for simplicity of hosing, it would be beneficial for me to run the coolant backwards at least in the PDM. Any thoughts?

Whatever... motor doesnt mind. Its the order for electronics that you need to keep. I dont think there is a check valve inside PDM.
So mind this loop order:
0. Radiator
1. Coolant pump
2. Inverter
3. Charger
4. DCDC
5. Other electronics
6. Motor
0. Radiator

Coolant pump should be the lowest part in your loop because for it to work propperly it needs to be fully submerged.
Motor should be last in line because it can take much more heat than electronics and dump it to the radiator. Radiator in turn does not need to be the same as OEM. I replaced my Peugeot 406 huge radiator with a simple thin Fiat Uno radiator. I kept the OEM fans and AC condenser because they were upfront in the bumper and i wanted a functioning AC system.
Also you can use the same loop for inverter, DCDC and charger since you use either one or the other and not both at the same time (usually). You just need to setup the coolant pump to operate on demand though. Also you want to have a bleed valve on the highest point in your system accessible so you can bleed air if you need to refill the system because of any maintenance.

[Kelju]

I assume it does not come as huge surprise to anybody that the PDM does not fit on top of the Leaf stack as the Nissan engineers originally planned it. The unit is very bulky, but I managed to find a place for it under the hood without sacrificing the original radiator and AC condenser assembly.
Once the PDM is in place, there is hardly any space left to run cables and hoses. Hardly any space to fit the liquid heater, let alone the AC compressor. The PDM is creating such hurdle that I am beginning to question the meaning of the abbreviation. At least I will be calling it the "Painfully Dimensioned to Mount"!

Anyway, I found a place for it. Sorry for not posting a picture of it mounted. I did not take a picture and already took it out since I noticed that I need to figure out and assemble almost all the cabling and plumbing beforehand.

Then I started looking at a suitable mounting location for the Th!nk City AC-compressor. That thing is not small either.
Does not fit on either side of the motor nor the front.
Might be possible to fit it assembled vertically beside(ish) the motor on the driver side.
The most promising location seems to be on top of the stack, but upside down due to the location of the refrigerant line connectors and shape of the unit. If placed in its natural orientation, part of it will hit the hood.
Does anyone no if these things a picky of their mounting orientation e.g., to secure adequate lubrication?

Also interested to know what is the most compact HV-electric AC-compressor one can find.

[m.art.y]

Also interested to know what is the most compact HV-electric AC-compressor one can find.

What about the original 2019 Leaf AC compressor? It has a little bracket and mounts to the side of the stack.

[johu]

What about the original 2019 Leaf AC compressor? It has a little bracket and mounts to the side of the stack.

And I will be using it as well AND it has its own hack-thread: viewtopic.php?f=4&t=2136

I wonder if anyone ever fit the PDM on top. Even in the tall Touran engine bay it didn't fit.

[wamort]

I am looking to make a spline adaptor and can't find any pressure angle info on the leaf spline. Can you provide the spline callout you used on your adaptor?

[Kelju]

I realise I have a ton of updates in between to catch up with this build thread, but as it stands - I made my first test drive with the car today!

I will have to post some pictures later, about the mechanics and electrics side of things.
On the SW side I have the following things sorted:
Modified the SW:
- to get the DC-link voltage from the Leaf INV as I do not have the ISA-shunt
- to get the current reading from a LEM CAB300 (because on ISA-shunt)
- to send out a huge amount of CAN messages to get the Think City battery modules to spill their guts on the CAN
- to receive and decode a huge amount of CAN messages to have the Pack voltage, min+max+avg cell voltage, min+max+avg cell temperature
- to calculate SOC and maximum allowed charge and discharge currents based on the battery data
- to update idcmin and max based on battery limits regulate the torque request
- to adapt main contactor actuation at the end of pre-charge based actual battery voltage rather than fixed value

[j2digital]

Hey man! I'm not sure how often you check the forum but I thought I'd ask how you like the Enerdel batteries you have in your build. I found an ad for some "new" ones on eBay that seem to fit the bill but I don't see many projects with people using them. I am sure it could very well just be me not searching hard enough or the right way for more projects with them but wanted to ask you anyway as our builds may be similar in spirit. Either way would be cool to see an update on your build!

[Kelju]

I opted for those batteries in the first place as I had some modules readily available. I was able to package half of a Think City battery in to the trunk, so 11,5kWh. However, the maximum continuous power output from this small pack is only around 20kW.
As I wanted to preserve the sporty performance of the RX8, I changed the battery. At the moment I have Mercedes EQC modules in the car.

[Kelju]

Around a year ago I "finished" switching to the EQC modules and managed to shoehorn 40kWh of energy to the trunk of the RX8 (half an EQC pack, as it conveniently has two parallel 400V strings). Then I got the conversion registered.

Today the car passed it's first normal MOT, so once again another year ahead to drive.

The plan is to update to adjustable suspension. Already got the kit.