Mazda RX-8 P2 build

wkerstens

Member
I have been reading along the forums for quite some time now, but as I received my P2 supercharger last week, it is probably time to start my build thread here as well.

The plan is to, just like Brettus his build, add a P2 supercharger to a Mazda RX-8. On paper, new, they claim do to 231 HP at the crank, but in reality this turns out to be a very optimistic number. The chassis is great, it's a joy to drive, but to me it has always been lacking that extra bit of power. It's great to climb through the revs all the way to 9000 rpm, but it's just not.. that quick.
I do not have easy access to a lift, so doing exhaust work for a normal turbocharger setup isn't that feasible, and adding a Turbo to a RX-8 isn't that straightforward. I've always had a big interest in batteries, BLDC motors and cars, so adding the P2 to a RX-8 seems like a great project. Especially being able to boost on demand sounds perfect.

The RX-8 with a 6-port engine is pretty much stock from the outside, but has some (reliability) upgrades:
  • IGN-1A ignition coils. Added this a long time ago before they became popular. Using a custom bracket and wiring harness
  • Using Versatuner to change fan temperatures, idle RPM and some other values. Will be used to tune the maps to allow boost
  • Cheap eBay decat midpipe, will eventually be replaced by something nicer
  • AEM intake as preparation for the P2
  • Lightweight pulleys because they look nice 😄
  • Bilstein B12 shocks which lower the car a bit while improving the suspension
  • Probably forgetting some stuff
So far it has been very reliable, but I'm only really using it for nice trips currently. Last year we did a trip from the Netherlands to Lake Como in Italy, the car was absolutely flawless. But since then we finally moved into a property with a garage for the RX-8 so it's time for some bigger modifications.

The goal is getting at least 50 HP out of an electric supercharger, and after seeing Brettus his build I'm confident this is feasible.

Supercharger:
  • P2
    • Huge, but allows low RPM which makes everything so much easier
    • Found a supplier in Germany, so no big import fees or shipping fees which makes it sort of affordable :).
Batterypack:
  • Strongly considering a 20S LTO pack, but getting good quality cells seems to be a bit of a gamble
    • Alternative would be LiFePO4, but that would require quite some cells in parallel to get to a decent current rating
  • JK BMS
  • Will be mounted in the trunk
  • 20 Ah or 40 Ah cells. Either should be more than enough current and the larger pack could even be charged at home
Motor and ESC:
  • HW MAX4 with 70125 seems to be the logical choice. Cheap and readily available. At €550 locally for the combo it's almost a no brainer
  • A TP power motor with MGM or VESC is probably more efficient and allows for more settings, but I want to get a feeling of what's required with the HW setup first. The nice thing about the HW setup is that it can easily be sold when it's time to upgrade.
Charger:
  • Victron Orion Tr Smart
    • Adjustable end voltage, can be set to any charge voltage and will run from the 12V system
    • Victron is used for many off-grid systems, boats and RVs, they are VERY reliable
    • Can be enabled/disabled using an external input
Control:
  • STM32 like board which should control the motor RPM based on a few inputs like RPM, throttle position, maybe load, etc. To be determined.
  • Versatuner to modify maps
I got a tiny CNC mill which should allow me to design / mill a nice motor bracket in the upcoming weeks/months. The only thing I'm not sure about is the belt.
If I'm not mistaken the P2 is using a T5 belt, and the more I look into this, the more I'm finding out that this is simply not suitable for the power we're putting into it.
What are the options here? I plan to just go with the T5 solution for now, but looking ahead, is this reliable enough?

The plan is to keep this thread updated, but it will be a slow build ;).
 

Attachments

  • IMG_5190.JPG
    IMG_5190.JPG
    210 KB · Views: 3
  • IMG_5157 Medium.jpeg
    IMG_5157 Medium.jpeg
    86.4 KB · Views: 3
  • 50ED33B9-A1F8-428C-8C5A-766C4CECD8F3 Medium.jpeg
    50ED33B9-A1F8-428C-8C5A-766C4CECD8F3 Medium.jpeg
    105.7 KB · Views: 4
One of the best chassis in the world (no bias here :LOL:) just begs for more power! Certainly looks like you are on the right path -good luck on the journey.
 
Made some mockups today before making the actual brackets. Not entirely happy yet, bolt pattern is a bit weird, it isn’t using a circular pattern for the M10 (?) holes. And the pattern doesnt align with the belt cover pattern 😅.
Tried to add a belt tensioner but needs some more tweaking. The bracket has some mounting holes to add a plastic cover at a later time, for “safety” 😁

IMG_5320.jpeg
IMG_5321.jpeg

Need to buy some additional aluminium stock and some shorter M6 bolts and M10 bolts as well.
 
Last edited:
I have much the same components as you and have just found a weakness. The M5 grubscrews holding the toothed pulley are insufficient for the job. The one that tightens onto the flat part of the motor shaft basically twisted in it's thread and allowed the shaft to rotate past it. My plan is to fit longer grubscrews and increase from two to four. Fingers xd that will be sufficient. Never had this issue with the 2028 Castle motor but that spins faster so .... less torque on the shaft.
 
What do yuo mean by twisted in its thread? Did it back itself off and come lose or did the shaft twist under the screw like it wasn't there?
 
If I'm not mistaken the P2 is using a T5 belt, and the more I look into this, the more I'm finding out that this is simply not suitable for the power we're putting into it.
What are the options here? I plan to just go with the T5 solution for now, but looking ahead, is this reliable enough?
I'm not confident about this arrangement either given the belt temps I've measured over short test durations of no more than 30 seconds at a time. My belt is on the narrow side at half an inch and it is not a "T" series belt but 240 deg F free wheeling without pressure resistance in place is destined to end up some where in the 300 deg range and no doubt failure at typical service loads so a 1" wide belt is mandatory.

I would not recommend the VESC ESC to anyone involved in this project, I have not been able to get it programmed right to apply and hold full power on demand and it is also doing some abnormal things like applying low level voltage after power input has been reduced to zero resulting in unusually high motor temps.
 
I was afraid this was going to happen eventually, thanks for letting me know.
I'm currently looking into taper lock pulleys to prevent this, but that might not work perfectly since the shaft has a flat spot / d-shape. Also, the minimum tooth count would be 24T.
@brettus8 what would you say is a good starting tooth count with the 70125? I read in your thread that you went up from 22T. Would 24T be a good starter?

@maticulus I intend to use the T5 belt drive system to figure out the ideal impeller speed, to then start working on a direct drive system using a TP5870 or similar with the right KV rating, but that will also have its own downsides having to make a custom shaft and get the tolerances right.

Update on the build:
ESC + Motor is on the way. Went with a slightly different setup:
Mad Components MAX20. More or less the same footprint as a HW Max4 ESC. Should be good for 400 A continuous, has built in BT and slightly better cooling, two (larger) fans instead of one. Also the app should have more settings available.
Motor is the same 70125, 560kv. Should arrive in the upcoming weeks :).
 
Last edited:
I was afraid this was going to happen eventually, thanks for letting me know.
I'm currently looking into taper lock pulleys to prevent this, but that might not work perfectly since the shaft has a flat spot / d-shape. Also, the minimum tooth count would be 24T.
@brettus8 what would you say is a good starting tooth count with the 70125? I read in your thread that you went up from 22T. Would 24T be a good starter?

@maticulus I intend to use the T5 belt drive system to figure out the ideal impeller speed, to then start working on a direct drive system using a TP5870 or similar with the right KV rating, but that will also have its own downsides having to make a custom shaft and get the tolerances right.

Update on the build:
ESC + Motor is on the way. Went with a slightly different setup:
Mad Components MAX20. More or less the same footprint as a HW Max4 ESC. Should be good for 400 A continuous, has built in BT and slightly better cooling, two (larger) fans instead of one. Also the app should have more settings available.
Motor is the same 70125, 560kv. Should arrive in the upcoming weeks :).
I'm looking at the TP5860 as a potential direct drive with its 50k rpm, the TP5670 with its 85k rpm gear reduced to 2:1, or slightly lower ratio to torque multiply it up to the 15kW level of the 5860 and the LMT 3080 and 30100 which is what I should go for to cover the full range of power level I may eventually strive for. The problem so far is the ESC (looking at the Max21), I just don't want to spend $1k on an ESC and am purposely staying in the 16s range battery power to help keep current loads below 400 amps at max compressor speed. I'm not sure that the LMT motors will be capable of that while delivering 7psi according to the BorgWarner matchbot computations.

Since there is no turbine blocking the exhaust that should equate to a lower turbine shaft power requirement than the bot calculator indicates I believe.
 
@brettus8 what would you say is a good starting tooth count with the 70125? I read in your thread that you went up from 22T. Would 24T be a good starter?
Depends on your goals ......... 22 gave me 4psi in midrange and 3psi up top at 250A but I want to max out at around 300A so went with a 25T. Should give me another 1 psi.
 
Started to work on the motor mounting bracket. More or less the first “real” thing I made on my desktop CNC, really happy with the result:IMG_5744.jpeg
IMG_5742.jpeg
IMG_5746.jpeg


Also:

- ESC + motor should be in somewhere in the next two weeks
- BMS is also on the way, went with a JK BMS 300A variant
- 24T pulley is in
- Need to order a belt so I can start testing

Also need to order some additional stock to fabricate the bracket that holds the motor.
 
Last edited:
I too have been experimenting with FreeCad and a small cheap and nasty CNC machine that really isn't up to the task. They claimed it could handle mild steel so I figured that like all sales pitches that mild steel would be just beyond it's abilities, so aluminum should be fine. 🤣

But it is really right on it's limit with Aluminum and a 500w 12,000rpm spindle

 
This one has got a 200W spindle @ 12krpm and doesnt seem to have any problems with aluminium. It’s not super fast, but I’m removing about 2500 mm3/min while roughing, which isn’t too bad.
The community firmware hacked a 3D probe into the workflow, that has been great, makes everything so much easier.
 
i am getting bit chatter, its too flimsy, i am experimenting with slower feed speeds and less depth per pass. It won't hurt me if it takes 3 times as long to cut.

But i might look at getting some t-slot stuff to try and stiffen up the gantry on it
 
What worked for me was dropping the “high efficiency machining” and just doing 60-80% stepover with 0.5 mm DoC and high feedrates. Yes it will wear down (but so far they're fine) the cheap aliexpress endmill at the tip, but it’s fast and easy on the machine. Having great results with that.
 
Last edited:
I don't know if you have seen this thread? https://www.electrifiedboost.com/threads/raspberry-p-s-i.163/

Towards the bottom of the page they provide some STL files.
I didn't, thanks, that was a very interesting thread. Looks similar to what I plan to do.

Got the motor controller, motor and BMS in. Still waiting for my belt, but made a quick test setup using a 3D printed TPU belt and 6S lipos. That was very very sketchy, but allowed me to get to about 3000 motor rpm. Already insanely scary and impressive, and that's only like 15% of the final RPM...
Should get my final belt tomorrow, which allows me to test fit the belt to then start working on the aluminium bracket to hold the motor in place.

Also ordered my LTO cells. Went with 24 x 40 Ah cells, it's probably a bit overkill but they can handle the current comfortably and with over 2 kWh of power storage I can probably even get away with just charging the system at home. I'm currently thinking about putting 20 in series, and then just charging them to a maximum voltage of ~50V, which should be within the 12S ESC limits.
 
Last edited:
Belt is in, got to about 5000 motor rpm but then got scared of the 3D printed motor bracket vibrating apart. Working on the aluminium version now:

photo_2025-07-10 09.25.57.jpeg

Finishing the other side tonight, I hope.
My current 6S lipos are also starting to show their age, so getting to max rpm on 6S might not be possible with that setup. Will just need to start working on my final battery pack.
photo_2025-07-10 09.27.41.jpeg

I'm also afraid I will need to take apart the P2 supercharger and clearance all bearings properly... As I previously noticed it does not seem to spin freely, and getting to 5000 rpm the bearings were incredibly noisy. Probably will replace the bearings with ceramic hybrids as well, I want to hear the compressed air, not the bearings ;).
 
Top