[SimonRafferty]

I have a 2021 Polaris General which I originally bought with the intention of converting to electric. However, since it has the same drivetrain as a RZR, I plan to buy one of them as a test-bed and then transfer it to the general (which I use a lot) once the whole thing is proven.

There's not many examples of RZR conversions. Pretty much the only one I can find is EVWest who used an AC50 motor:

Unfortunately EVWest are not selling any of the useful bits, like the mounting frame nor CAN integration - but I can build them.

I wondered what other motor options might be out there? I'd like to repurpose the drive from some commercial EV if possible (mainly for cost reasons) . I wonder if there are any relatively small, light, low power motors worth considering?

The daft thing is, I have a Siemens Hybrid Bus motor & controller (from my Freelander) sitting under my bench - but it weighs about 100kg, and is too big to fit in a 600kg vehicle.

I was going to build a lous Elise using the 1PV5138-4W624-Z W12 motor & controller. I'd arranged to buy a Lotus Elise so dismantled the Freelander - then the seller changed his mind (while I was driving 300m to his house with my trailer)! Ever since, it's just been a big pile of bits under my bench! I never found another car I liked enough to convert - until the Polaris!

 

[Brown29brent]

Sorry I've been away for such a long time

I have a 2021 Polaris General which I originally bought with the intention of converting to electric. However, since it has the same drivetrain as a RZR, I plan to buy one of them as a test-bed and then transfer it to the general (which I use a lot) once the whole thing is proven.

There's not many examples of RZR conversions. Pretty much the only one I can find is EVWest who used an AC50 motor:

Unfortunately EVWest are not selling any of the useful bits, like the mounting frame nor CAN integration - but I can build them.

I wondered what other motor options might be out there? I'd like to repurpose the drive from some commercial EV if possible (mainly for cost reasons) . I wonder if there are any relatively small, light, low power motors worth considering?

The daft thing is, I have a Siemens Hybrid Bus motor & controller (from my Freelander) sitting under my bench - but it weighs about 100kg, and is too big to fit in a 600kg vehicle.

I was going to build a lous Elise using the 1PV5138-4W624-Z W12 motor & controller. I'd arranged to buy a Lotus Elise so dismantled the Freelander - then the seller changed his mind (while I was driving 300m to his house with my trailer)! Ever since, it's just been a big pile of bits under my bench! I never found another car I liked enough to convert - until the Polaris!

tks for sharing

 

[SimonRafferty]

Do any of you know if there's a decent BMS which is IP Rated / Waterproof?
All the components I've bought for this project are IP rated, except the Orion2 BMS, which is made of holes! I've put it in an IP68 enclosure - but it makes the BMS a lot bigger & adds difficulty getting the cell tap wires out of the box, keeping the IP rating.

Just wondered if there were any other options?

 

[SimonRafferty]

I've spent the last 6 months looking for a suitable RZR in the UK, but for some reason there is almost nothing for sale at the moment! Probably because the supply of all new cars has dried up due to Covid?

I've decided just to convert the General instead!

The first step is hacking the CANBUS. I've found very little info on line about how people approach modern vehicles full of CAN devices - so here's what I've decided to do!

I've bought one of these: Teensy 4.0 Triple CAN Board Include Teensy 4.0

And am in the process of writing a CANBUS bridge - something which will sit inbetween the vehicle ECU and the rest of the vehicle and pass messages in both directions, filtering some out and injecting some new ones.
The third port will connect to the Electric Drive CANBUS and act as a 2 way translator, reading things like RPM, packaging it up and sending to the dash, with a bit of translation. I'm going to use the rev counter to display something proportional to the power being delivered to the motor and the Fuel Level gauge as calculated range.

The first part of this was to decode everything I needed off the bus. If anyone else is considering converting a late model RZR, General or Ranger - this should be very useful!
GitHub - SimonRafferty/Polaris-General-2021---CANBUS-Hacking: Useful tools & Results from reverse engineering the canbus on my 2021 Polaris General XP1000

The photo is the Arduino driving the dash with the ECU completely disconnected.

There's another Github repository by Milo Darling:
GitHub - milodarling/RZR_CAN_HACKS: Hacks to make an electric Polaris RZR work well!
But the CAN ID's he has discovered seem to be different to mine. Maybe he was working on an earlier vehicle & Polaris have since changed the ID's? His post was from 2017, so the model must have been earlier than that.

Tomorrow I'm going to cut the CAN wires going into the ECU and insert a plug & socket between the two halves. If it turns out not to work, I can just plug the two halves back together and it will be back to standard. The new CAN Bridge will just insert between the plug & socket.

 

[cricketo]

I've spent the last 6 months looking for a suitable RZR in the UK, but for some reason there is almost nothing for sale at the moment! Probably because the supply of all new cars has dried up due to Covid?

I've decided just to convert the General instead!
View attachment 131240

The first step is hacking the CANBUS. I've found very little info on line about how people approach modern vehicles full of CAN devices - so here's what I've decided to do!

I've bought one of these: Teensy 4.0 Triple CAN Board Include Teensy 4.0

And am in the process of writing a CANBUS bridge - something which will sit inbetween the vehicle ECU and the rest of the vehicle and pass messages in both directions, filtering some out and injecting some new ones.
The third port will connect to the Electric Drive CANBUS and act as a 2 way translator, reading things like RPM, packaging it up and sending to the dash, with a bit of translation. I'm going to use the rev counter to display something proportional to the power being delivered to the motor and the Fuel Level gauge as calculated range.

The first part of this was to decode everything I needed off the bus. If anyone else is considering converting a late model RZR, General or Ranger - this should be very useful!
GitHub - SimonRafferty/Polaris-General-2021---CANBUS-Hacking: Useful tools & Results from reverse engineering the canbus on my 2021 Polaris General XP1000
View attachment 131241
The photo is the Arduino driving the dash with the ECU completely disconnected.

There's another Github repository by Milo Darling:
GitHub - milodarling/RZR_CAN_HACKS: Hacks to make an electric Polaris RZR work well!
But the CAN ID's he has discovered seem to be different to mine. Maybe he was working on an earlier vehicle & Polaris have since changed the ID's? His post was from 2017, so the model must have been earlier than that.

Tomorrow I'm going to cut the CAN wires going into the ECU and insert a plug & socket between the two halves. If it turns out not to work, I can just plug the two halves back together and it will be back to standard. The new CAN Bridge will just insert between the plug & socket.

Have you got the electric drivetrain already setup ?

 

[SimonRafferty]

Have you got the electric drivetrain already setup ?

Nope, just planned. You have to do the first part of the CANBUS discovery with the original engine in place - so you can learn what you're trying to replicate.

I'm going to use a Hyper9, which I've already bought:

With a belt drive to this bit:

If you dismantle a secondary clutch from the original belt drive, it has a steel tube in the centre which engages with the splines on the gearbox input. Conveniently it's of a size for which you can buy a taper-lock bush, as pictured. The tube is through hardened and difficult to cut a keyway in (not impossible - I could get it ground or spark erroded). Initially, I'm going to try it with just the taperlock plus Loctite 638 to 'glue' it. I've had good results with this combo. The torque input is only 235Nm which is within the realms of possibility.

I've drawn this as the motor mount which picks up on the original engine mountings on the gearbox & frame. The dark grey cage around the motor body swings around the red hole at the top to allow belt tensioning / removal.

Originally I had intended to mount the batteries above the motor.

Something like this. However, having done a bit of measuring, I think there's enough space under the seats to mount the batteries once the fuel tank and the two vent tubes are removed. Hopefully, there's enough space under the seats for all the other electronics too. The seat box will keep everything better contained at the expense of easy access.

The above pack is 12.5kWh - and I've already bought the cells. Hopefully this will give me about 40 miles range - which is more than enough based on my usage since I bought the Polaris. The cells I bought have a max discharge current of 500A (32S 1P) - so will deliver around 50kW / 75Hp.

Initially I'm gearing it 1:1 which gives me about the same max torque at the wheels with a top speed of 60mph @ 8000rpm. If this needs increasing or decreasing, its easy to swap out the belt pulleys for different sizes.

 

[cricketo]

Nope, just planned. You have to do the first part of the CANBUS discovery with the original engine in place - so you can learn what you're trying to replicate.

Cool. I wouldn't waste my time and get the EV running first, then see what can be done about the blinky lights. Protocol decoding can be quite time consuming and unfruitful.

 

[SimonRafferty]

The protocol decoding is mostly done now. It only took a day.

It was necessary because I need both the ABS & Electric Power Steering to function, both are controlled over CAN. The 'blinky lights' are just a bonus.

I know most folks find it time consuming & unfruitful - which is why I posted the tools I used on GitHub. There are two simple tools which allow you to discover most of what you need.

[Feel free to drop me a line if any of you would like some help / pointing in a fruitful direction]

The mechanical part should be straightforward (that's my day job after all). That will begin next week, once the CAN Bridge is working.

 

[SimonRafferty]

This week I started dismantling the Polaris. At first, I was trying to minimise the amount of bodywork I removed - but it became increasingly obvious you pretty much have to strip it all off to get access to things like the fuel tank - which, incidentally was a complete pig to remove!

You can see in this one how the engine is just bolted to the side of the gearbox, connected by the (removed) CVT.
I'll make a frame to hold the electric motor in roughly the same place, by the same bolt mountings.

Removing the fuel tank leaves a fair bit of space under the passenger seat. I'm going to put 10 of the cells under there along with the Elcon Charger, next to the (previous) fuel filler.

The 2021 Polaris XP1000 engine weighs 61.65kg - surprisingly light for a 100Bhp engine.
I've not weighed everything yet, but I reckon all the other bits I've removed take the total to 110kg with an empty tank.
It takes 40l of Petrol which is about 31kg - so the total mass removed will be around 140kg.

I'm adding 3.35kg x 32 Batteries = 107.2kg. Hyper9 = 59kg + probably another 30kg of controllers - give or take 200kg.

A 60kg increase in the mass of the vehicle isn't too bad.

 

[SimonRafferty]

A tiny bit more progress. I removed the engine:

Then Plasma cut a bit of steel to check my measurements for the engine mountings.

Which looks good.
I found that my General is completely different to a supposedly accurate 3D Model I bought from Hum3D - so had to re-design the motor mountings a bit. The mount is a lot simpler and lighter. In 3mm Mild steel, it weighs 4.5kg.

Today I ordered the laser cutting for the motor, battery & electronics mounting hardware. Hopefully it's only a week or so away.

It turned out the gearbox input shaft i different to what I expected - different diameter. I picked up a different taper-lock bush 30mm ID which seems to fit OK.
I don't want to cut a keyway in the shaft. Instead, there's a splined collar which slides on to the shaft. I'll machine that down a bit & weld it to the taper-lock bush. That way, I can still change the final drive pulley easily enough.

Lastly, I 3D Printed a little panel for Regen braking & max speed / rpm controls.

 

[SimonRafferty]

I picked up the laser cutting for the motor cradle and battery trays. Welded the bits together yesterday and trial fitted the motor mount:

I was surprised how well it fitted!

Next I trial fitted the main battery tray:

The mounting position for this needed a bit of tweaking, moving forward 2" as the drive belt would have touched the bottom rear lip of the box.

Everything else seemed to go to plan. On Monday, I'll take the Aluminium bits to be Anodised.

You may be thinking the Aluminium bits look unneccecarily chunky? They definitely are from a strength point of view. They're made from 1050 Grade Aluminium (sometimes known as Pudding Grade ) which is soft, but has the best thermal conductivity. I have a liquid cooling plate on top of the battery pack and wanted to give it the best chance of cooling the whole of the battery when needed.

I have some self adhesive silicone heating pads (like you use on 3D printers) which will attach to the underside of the Aluminium casing to provide battery heating in cold weather. This is more because you shouldn't charge LiFePO4's below 5C, than to get more performance when running.

I plan to tie the heaters into the charging circuit - so they only come on while the charger is plugged in to the mains.

The liquid cooling just sends coolent through a labyrinth made from a sandwich of Aluminium with hose connectors on the top face.

Similar on the smaller battery pack. The Inverter drive has the supplied 'chill plate' but hopefully, the battery cooler will enhance the cooling of the inverter too a little bit.

 

[remy_martian]

The use of 3d printer heater pads is genius...as long as vehicle vibration doesn't wear through the insulation. Might want to do an isolation check now and then, if not built into the charge controller.

 

[SimonRafferty]

The use of 3d printer heater pads is genius...as long as vehicle vibration doesn't wear through the insulation. Might want to do an isolation check now and then, if not built into the charge controller.

The insulation should be OK as other than being stuck to the battery tray, they are not touching anything else. Silicone is pretty resilient. I guess, only time will tell!

I believe the BMS does check isolation - but it's a good call & I'll check it.

 

[electro wrks]

Initially, I'm going to try it with just the taperlock plus Loctite 638 to 'glue' it. I've had good results with this combo. The torque input is only 235Nm which is within the realms of possibility.

I don't want to cut a keyway in the shaft. Instead, there's a splined collar which slides on to the shaft. I'll machine that down a bit & weld it to the taper-lock bush. That way, I can still change the final drive pulley easily enough.

I've had good luck with Taper-lock bushings grabbing the shaft on installations without a key. Mark the shaft and the bushing, and check for movement. I would only weld the one to the collar as a last resort. Often, you will have to change the bushing size as the sprocket size is changed.

It looks like you're using HT sprockets and a belt. This will be less noisy than the Gates Poly Chain Carbon GT system, or its equivalents. The trade-off is, AIR, about half the power transmitting capacity for the HT system, every thing else being equal. Unfortunately, from what I've found, the belts and sprockets are not generally interchangeable.

You'll need a lot of adjusting slot length with these systems. The belt has to slip over the sprocket side flange(if both sprockets have them), and the selection of belt lengths can be limited.

 

[SimonRafferty]

It looks like you're using HT sprockets and a belt.

You're correct. I originally wanted to use Continental Silent Sync belts - but they seem impossible to obtain at the moment, in Europe at least. HT seemed the best alternative.

Both pulleys are flanged - but the alignment is pretty good so frictional losses against the flanges will be kept to a minimum.

I'm pleased we have similar experience of taperlocks without the keyway. The inner diameter is fairly small at 30mm on this - but might be OK. I'm going to try it with Loctite & the surface of the shaft roughened a little with a diamond burr. As you suggest, I'll mark both halves so I can see slippage.

I came up with a potentially better backup plan. The taperlock bush has a third hole (used to remove it from the pulley). I figured on welding a pin to the splined sleeve which inserts in this hole. Then weld a tube to the other side of the splined bush such that it can be retained by the original bolt in the end of the gearbox shaft. Then it's easy to remove.

I'll play it by ear - and adapt accordingly.

 

[bastens_com]

I wanted to chime in here and say what an awesome job you are doing and thanks for sharing. In addition I wanted to assure you that you are on the right track. I built an electric RZR but I used a Zero 75-7R (like the ME1507) motor. It works but is limited because anything faster than 40mph is greater than the constant power rating of the Zero 75-7R motor and eventually will overheat. I geared it down to overcome this but has has plenty of power and acceleration. It will do 0-40mph in 4 seconds! The Zero 75-7R is a great performer but if speeds faster than 40mph is desired then the hyper9 is the ultimate choice.. I wished I would have used a hyper9. I didn't go with the hyper9 because I was thinking the 40lb Zero 75-7R motor weight savings would be better than the hyper9 120lb motor which turned out to be not as big of a deal.

Here is a video about my build, maybe there are some ideas in there for your build ..

Cheers!

BASTENS LLC - unique 3D printed items - USB cables and chargers

 

[SimonRafferty]

I wanted to chime in here and say what an awesome job you are doing and thanks for sharing.

Hey Patrick - thanks for the comment! It was actually your video that made me decide to give it a go in the first place. So, thanks for giving me the idea & showing that it's possible!

 

[SimonRafferty]

A bit more progress.

This was just a test fit of all the batteries in the main, 22 cell pack. They were a bit tight sideways, so I had to slot the holes where the sides bolt on a little.

I bought some of this: SIL-X-600-FFF Thermally Conductive Silicone Sponge Sheet - Silex Ltd
Which is similar to the heat transfer pads people use on CPU's, but in a big sheet. The batteries will sit on top of one 1.6mm sheet, another one on top, sandwiched between the two Aluminium plates. These will both cushion the batteries a bit, grip them to hold them in place and most importantly, dramatically improve the thermal transfer to the Aluminium. You see sheets of it being used in commercial packs.

Bottom left of the photo, between the blue cooling hoses, you can see a SPAL Snail fan which will hopefully improve convective cooling of the motor. I've had the fan sitting under my bench for years - and thought "I wonder if?" and it fitted perfectly in the space! I think the Gods of EV conversions were smiling on me that day!

This is also a test fit & arrangement for some of the other components. The second, 10 Cell, pack in beneath the Aluminium plate.

I'm using Solid State Relays for both the 12V & 100V (low current) switching. I know not everyone likes them for a host of reasons, but I've found them ideal for switching DC Inductive loads as there are no contacts to weld closed and in every case I've experienced, they fail open circuit. They tend to fail about 10% above their rated current & they're cheap enough you can use them as a secondary fuse. The left-most SSR is 230V AC for the charger inhibit. Middle two are 200V DC rated, one to switch the cab heater & the other to switch the battey heaters. Right hand SSR will switch the engine & front radiator vans at 12V DC.

Type 2 Charge socket on a little 3D Printed stand-off as there wasn't much space behind it.

 

[SimonRafferty]

Main pack wired up with all the cell-taps. I've put rubber caps on all the bolts, sprayed it with conformal coating (almost like liquid insulating tape). There's a polycarbonate cover which slots in, in front of the battery terminals to make it a bit harder to get in there with fingers!

I pressure tested the cooling plates - and they seem OK.

Tomorrow, I'll wire up the 10 cell pack, put the lid on and can start on connecting up the hundreds of wires that come out of the drive components.

Initially, I'm putting them in to insulated screw terminal blocks - as they're easy to debug. Once I'm happy everything works, I'll solder them & cover with heatshrink.

Above is a bit of my connection spreadsheet. I find this a better approach than trying to draw a schematic for things like this - which always ends up looking like a spiders web! It makes debugging super-easy too.

 

[SimonRafferty]

Today, I connected up the Battery pack, BMS & Charger - and no magic smoke came out of anything! Always a relief!
At the moment all the cells are charging using a cheap Type 2 Adaptor bought on AliExpress. It all seems to work fine!

I've not quite finished the wiring - still a few (simple) things to connect up, like the reverse switch & cab heater.

The CAN network seems to work OK - the BMS can see both the charger & motor drive.

I had originally intended to make my own Info Dash, using a Nextion Display + ESP32, but decided in the short term just to use a Bluetooth OBD2 Adaptor and Torque running on an old phone.

This was the first time I've had the BMS & Charger wired up. It's sitting balancing the cells at the moment.

Next I'll get the Hyper9 up & running.