PROJECT: ENDGAME a karting inspired raceboard's development

[Dinnye]

Next up was the assembly, which actually went less smooth than expected.

Couple difficulties:

First of all rails to bellypan assembly. The source of this issue is that I absolutely need holes that are 100% impossible to make on the cnc, these holes are only possible manually. When I started the drilling process, I made jigs to hold the drill in the correct spot and perpendicular to the material, but these worked like shit. By the end I figured out that the best results I get by attaching the bellypan to the chassis, and using the bellypan itself as the guide to drill, and just sending it. But the first couple holes I did before figuring that out were a bit misaligned, so assembly was a bit more painful then it should be. At least I’ve got the process figured out now so next time it should go smoothly.

And then my arch enemy, which I suspected might be a problem, the slight misalignment of the shoulder bolt bores. Shitty drawing to explain:

The black rectangle is the guided diameter of the shoulder bolt (12mm), the red one is the M10 predrill hole (8.5mm); these two are made in the first side milling op so alignment is perfect. For the second side milling op however, I use a coordinate system reference that’s taken from the contours of the part, not from the bore itself - and this is where things go wrong. I get a slight misalignment, so the second op 12mm bore doesn’t perfectly line up with the first op - this makes screwing in the shoulder bolt abnormally difficult. On the first swingarm base with the worse order of operations this was a huuuuge pain in the butt, but I managed to save the part in the end. On the second base with the new CAM it wasn’t so bad but still got some improvements to do regarding the positioning accuracy between these setups. Good news is I know the next steps to improve this, so I expect it to go smoother next time.

I did successfully assemble the board though!!! And just in time to race it at the Budapest race

[Dinnye]

Next up: a bit of wheel talk in preparation for the Budapest race.

The track surface at Kistarcsa Gokart Farm is very worn down, it’s like asphalt but it lost all of it’s abrasiveness. Very low grip. The tires that work here are more or less what work at the slippery indoor concrete tracks. For such conditions, less contact patch (especially from a GOOD rubber compound) can often work better.

So I decided to give PMT scooter tires a try. These are the 250/65 version.

They are made for scooters, so the contact patch is quite little, and it’s a street/ race hybrid tire, so even though it isn’t a full on race tire, it is soft rubber and it was engineered to be very grippy. Furthermore, PMT is a very well respected manufacturer in the scooter and motorcycle racing scenes - so we are off to a promising start. In hand, the rubber looks and feels very grippy and great quality.

I did need to get a new set of open gears made with an extra set of holes, as my old ones only had MBS, and I wanted to be able to swap between these and the kart tires quickly anyways.

Lasercut from S355J2 steel, they lasted the race, look good still, but not expecting infinite life from them. I don’t need them to last long though, and it was quite cheap to get them made even on very short notice.

They do make the board look very long and narrow, so I prefer the looks of kart tires.

At this point, the mechanics are finally together. I also know that I am swapping from the 95+85 motor combo to 85+75, stealing the 75s from the mountainboard. The 75s are gonna go back on the mountainboard when I’ll finally have the drive for inrunners ready.

This is how far I got before having to start the drive to Hungary, Thursday afternoon. Race is Saturday afternoon, so cutting it verrrrry close. But at this point, I was fully commited to racing the new board.

[Dinnye]

After an insanely long 23h journey (sharing the drive with a friend), I got home to Hungary by Friday ~ 8pm. Quickly unpacked the car and got to work on a not so great workbench.

Given the tight timeframe, the only possible way was to reuse the pelican from the previous iteration - that made the electronics part very simple for the time being.

Wheels off, footpads off, mount motors, adjust pinion positions and loctite them… Apply harness tape to the phase wires and feed them into the bellypan. Connect the bullets to the old pelican case and secure the wires. All pretty easy work, but overall it added up in time. Finished around 2am, really wanted to finish this day to let the loctite cure.

Saturday morning, I started by setting up the vescs, then put on wheels for some photos. Took it for a short test ride, all good, initial suspension bushings were a little too soft. Put in some harder ones, one more short test ride, felt decet, then charged the board back up, had lunch, and started driving to the race. What a close finish!

[Dinnye]

After the vesc setup, I quickly threw on the kart tires for a couple pictures then immediately swapped to the PMTs afterwards.

[Dinnye]

Next up some pictures on the PMTs:

[Dinnye]

Next up is the actual race - it went really well, a very comfortable 1st place finish on a brand new board with zero experience on it, with tires I also had zero experience on.

I had a good fight with the EUC leader for a while, but I didn’t take the second direction very seriously so he gained some advantage by the end. I only did around 3 laps of practice on my board in the freeride time for the second direction.

At this point I was running the “stock angles”: front was 35+5=40, rear was 35-15=20. Turn radius was mindblowingly good. Too good even, but no stability issues. That said lower angles should be better.

The tires worked (comparatively speaking) very well on the slippery track. It was easy to catch a slide. My lines were very different with these than with the kart tires on previous practice sessions at this track.

[Dinnye]

Since returning home I put the kart wheels back on, and lowered truck angles from 40-20 (base angles) to 35-12.5. This is before the suspension articulation, actual angles during the ride of course vary a bit.

Then I went out for the first short T-race session on the setup and started out with a baseline of 12.87 with just a couple laps. Could’ve easily done better within a couple more laps. Even though I felt like I was really lacking power (lipos are not feeling great anymore) and of course smaller motors and less phase amps. And the track was also full of a bunch of small rocks.

T-race shakedown run on the new chassis

Even at these lower angles, turning is insane. The amount of lean the setup allows is even more insane.

The torsion box design fixed all the previous torsional flex issues, and even for the same angles I can feel the board turning noticeably better. Also made the bushings feel a bit softer.

I’m very happy with the ride feel of the new chassis. Still got some minor tuning to do, but I’m almost there. Couple more sessions (most likely next week) and then it’s mechanically perfect.

The next deadline within this project should be finishing the 20s6p and the proper ESC install before the Salzburg race which is 3rd of October.

In the meantime I also want to work on getting the V3 double stage drive design finished, as I’d like to start testing that around end of October.

[BenjaminF]

@Dinnye

Damn that’s a good looking wheel! What is it?

[Dinnye]

@Titoxd1000 milled them, but they aren’t normally offered for sale because even just the material cost is already quite insane.

[dimos15]

Your build looks awesome!!

@Dinnye

Lasercut from S355J2 steel,

Did you buy this from China? What is the price? I want to get some gears too . Xometry.eu has -200 coupon for September, 5 gears cnc are around 80 per gear

[Dinnye]

Thanks!

Got the 4 gears for ~30€ each including vat at the local laser cutting service, with same day pickup

[MichaelWA]

Can I pre-nomimate this for BOTY 2026?

[glyphiks]

Surely this beast is a contender for build of the decade!

[Dinnye]

Had a T-race session today to try out some adjustments I made. 12.48 toeside, so 0.05 away from my record from last year. Pretty good for a not very clean track and lower power than the version I made the record on, also lower ambient temperatures.

Definitely getting closer as far as adjustments go as well. But mostly i’m happy to enjoy some practice on the new setup while the weather still allows, didn’t have as many training sessions this year as I’d have liked.

[Dinnye]

12.09 on t-race today toeside, after seeing Tas get the world record I had to claim it back

Flipped front baseplate (-5 instead of +5) was definitely the right call.

I started pushing for 11.xx immediately after, but during the warmup lap for that a motor had slipped on the motor mount, disengaging a pinion so I ended the session. I’ll write this up to too loose backlash resulting in suboptimal effective contact angle between the gears.

[JeffyJ]

11s are going to happen! Crazy.

Congrats. New board is a beast!

[Dinnye]

Thanks! I think if I clean the surface, I might be able to do an 11.xx next toeside session already! I wonder what the hard limit actually is

[Dinnye]

11.xx here we come!

11.96 today

Can’t wait to put them into comparison tool I built since the 12.09 lap to compare against a numerically optimal lap to see where it’s still possible to gain time

Racing Line Optimization - compare your best lap with iteratively generated optimal racing lines

So I want to start off by saying that this project for the time being is still very much under development and I’m not even sure how advanced I want it to get yet, or exactly how much more work I want to put in it. I wrote some MATLAB code to calculate optimal lap times and optimize racing lines using the CasADi solver. It allows the import of RaceChrono telemetry data and comparing it directly against the simulated optimal line. [image] You can give it your vehicle parameters: Lateral acc…

[Dinnye]

Some more improvements done to the optimization tool under the hood. Here’s today’s 11.96 compared to the optimal lap generated by the tool.

Here’s what I can gather from the comparison features:

• I crossed the start line at the incorrect angle and coming in too hot - missed the optimal entry into turn 1. Lost 0.25 sec to this.
• I overdid the turn exiting the first corner, essentially sacrifing entry speed for a “straighter” straight. However, I failed to get on the throttle hard enough to make this tradeoff worthwhile, producing only ~0.4-0.5g longitudinal acceleration. 0.15-0.2 loss here.
• I nailed the braking, almost matching the optimal line’s speed from the start of the GPS logs braking zone till the turn in point. No further loss here.
• I was a little too slow throwing myself into the corner, building up that lateral G. However I did hit the 1.5G target. The slow buildup (jerk) cost me 0.05.
• Right after the apex I was trail braking a tad too long, lost ~2km/h and tightened up the radius a bit to compensate for that, but dropped to ~1.2-1.3G throughout the rest of the corner. That’s 0.1 lost between the apex and the exit.
• On the second straight I was lighter on the throttle than the simulation (or possibly just had too much battery sag by this point in the session) and therefore lost 0.3 before the turn in point compared to the simulation.
• Turned in slightly too strong at cone 4 and therefore lost another 0.1
• From the exit of cone 4 until the finish line I kept up with the pace.

Note: I have some ideas to further refine the simulation still. Ideally it’d match the vehicle dynamics as closely as possible, so that it’s more accurate about when I’m keeping pace and how far I’m falling off pace where I can’t keep up. But even at this current accuracy, I think this post goes to show how much knowledge can be gained from it. So glad I put in the work to make this tool.

[tuckjohn]

First time back to forum in a few weeks, this is batshit insane stuff dude. Incredible.

You’re the top of the sport​