Direct Drive 6384 eFoil Build From Queensland Australia

[Joez]

This is my first post so first of all, big thanks to this awesome forum/community. I won’t be able to pull off the build without it. So I thought I’d contribute to the community by sharing my experience building the project.
I have mechanical engineering and software engineering background and I consider myself a somewhat handy person but I didn’t have any RC building or surfing/boarding experience. I used ChatGPT for high level planning since I didn’t even know where to start. I also did lots of googling and that’s when I found this forum. After reading posts for a couple of days, I got a plan for the main components:

• 6384 motor direct drive controlled by 75100 VESC and Flipsky VX3 remote
• Second-hand windsurfing board to cut off both ends
• Battery and electronics will be in a waterproof case strapped to the board (I’ve seen experienced DIYers struggling with waterproofing in-board compartment so I didn’t want to try my chance)
• Commercial hydrofoil wings (I don’t have the skill to make my own)
• 3d print the motor mount and other small parts

I decided to source all the materials/parts on Taobao (the “Amazon” in China) and ship to Australia by a shipping company. Here is the part list (price approximated in USD):

• Flipsky 6384 waterproof motor - $111
• Flipsky 75100 pro VESC - $70
• Hydrofoil wings set and mount box - $278
• 48V Li-ion battery 28.6AH (13s I guess?) - $180
• Fiberglass, epoxy, themal epoxy/paste - $28
• Water alert - $4
• Flipsky VX3 remote - $68
• Battery charger - $17
• Hall effect switch - $5
• Aluminium plate/heatsink - $5
• XT90 connectors - $8
• Bolts/nuts/inserts - $12
• Waterproof case - $12
• Waterproof cable connectors - $5
• Strap/mushroom lock tape - $6
• Cable/rubber duct - $8
• Foot strap and magnet - $6
• Shipping - $77
• Second-hand windsurfing board - $39

Total cost: $939

The parts arrived after 2 months and the build began.
First, to cut the board and fit the foil.
This is the original board:

After cutting both ends:

Then I fiberglassed both end, cut the hole for the mast mount and fiberglass on it:

Finally, putting the foil on:

With the foil and board done, it’s time to work on the “e” part:

I designed and 3d printed the motor mount (after a couple of failures):

and attached the motor to the mast:

Propeller is using @V_S 's fixed 2 blade prop 3D design 156 0.39 1.09 15 4 Loch - Tinkercad
All 3d printed parts were epoxied for better strength and waterproofing (since I used plain PLA).

I wanted to have a kill switch that kills the motor if I fall off the board (and forget to release the trigger). Originally I thought of using a high current relay that’s controlled by a magnetic proximity switch but then I found out the new VESC firmware version supports killing the motor on ADC2 signal.This made things a lot easier. I just used a hall effect switch connected to 5V and ADC2. In operating condition, a magnet will be placed close to the hall switch (inside the case) so the switch is closed and motor can run normally. If I fall off the board, the magnet (attached to the foot strap) will be pulled away from the hall switch and 5V will be sent to ADC2 and kill the motor. I tested the setup and it worked perfectly:

(the yellow thing on the battery is the hall switch)

The foot strap with magnet:

On to the waterproof case. The phase cables go into the case through 3 waterproof connectors:

When the nut is tightened, the rubber ring in the connector will create a seal with the phase cable insulator.

For the ESC cooling, I drilled 4 big holes on the case and epoxied an aluminium plate on the outside. To compensate the strength affected by the holes, I use themal epoxy to glue another aluminium plate on the inside that the ESC will attach to. This turned out to be unnecessary and problematic later.

I then themal epoxied a heatsink to the aluminium plate outside:

For added safety, I wrapped the battery with heat shrink bags (so minor leak won’t get into the battery). I also put a water alert in the case (the white little box on the right hand side):

The hall switch and the remote antenna are taped (for testing) to the top along with some battery guard so it doesn’t move.

Then I did a leak test (with grease applied to the rubber seal of the case) by submerging the case for 15 mintues:

The result was good - no water got in.

The last part was to attach the case to the board. I used mushroom lock tape to prevent the case from moving and 2 straps to hold it against the board:

A big metal washer was taped on the case for the magnet.
Now it was finally ready for the first water test:

[Joez]

It went relatively well for the first water test. No water leak. All parts works until the prop broke after half an hour.
I made a video of the build process and the first water test:

DIY efoil build (first water test)

The prop broke because I didn’t pay attention to the seam position (it was right at the root of one of the blades) and didn’t have enough wall thickness and infill.
I fixed the propeller issue and did 2nd water test. The prop held up for the whole session without problems. However, I did notice the ESC temperature (showing on the remote) was a bit high (up to 70c). It was probably because I could stay on the board longer (I was too busy falling off the board the first time). I thought it could be due to the themal paste not fully cover the back of the ESC where it attached to the aluminium plate:

I then applied more themal paste and did the 3rd session the next day:

This time I was able to plane on my knees (the remote showed 17A current draw) and the themal paste didn’t seem to fix the cooling issue. The temperature went up to 80c within a minute and current draw slowly went down, resulting a slower speed (I guess it was the ESC’s temperature protection kicking in?). I had to let it cool down for a while but it was kind of OK because I still fell off a lot. I tried to keep the temperature below 80c and practiced for about 3 hours. The remote flashed “battery votagle below cutoff value” a couple of times during the period but recovered immediately so I didn’t pay too much attention. The remote always showed 100% battery.
When I felt like going back, something dramatic happened. The motor cut off in the middle of a ride and the remote showed 0% battery. I thought it was because I didn’t set up the battery information correctly in the VESC tool so the percentage didn’t display and now the battery was dying out. I thought I could lower the cutoff voltage and squeeze some juice to get me back to the launch point at least (I was 2km away). However, the bluetooth never connected. I had to swim with the board attached to my foot strap for about 1 km and a kind gentleman let me hold on to his kayak and dragged me back…not a fun experience
I opened the case when I got home and immediately smelled something burnt. Remembering the battery warning on the remote during the session, I suspected something went wrong with the ESC so I disassembled it and found one of the MOSFETs appeared to be burnt out…

[Joez]

This is where I’m at now. I’m going to order a new ESC and implement water cooling. I thought I could get away with air cool but apparently not. I’m thinking of either:

• Flipsky 75100 water cooling version + external bluetooth module
  or
• Makerbase 75200 + water cooling block + external bluetooth module

I’m leaning toward the 75200 option since it’s cheaper and 75200 is supposed to be able to handle more current. For the cooling water, I’ll try ram feed first but keep the space for a water pump in case the ram feed water is not enough to cool the ESC (I’m in Brisbane so the water temperature is relatively high).
I’ll also add a fuse between the battery and the ESC. I was going to use an anti-spark switch as the power switch but the Makerbase 75200 seems to have a power switch feature (I’ll need to confirm this with the seller).

Appreciate any comments/suggestions

BTW, I’m still curious to know why the MOSFET burnt. I thought 80c wouldn’t really destroy it and the ESC has temperature protection (I guess?). So was it really caused by my poor cooling design or could it be a QC issue?

[S_Roger]

Well you had a real baptism-by-fire having your battery fully spent out on the water like that! Don’t those big salt crocs live in Queensland?
Did you have the remote connected via UART or PPM? I’m guessing UART if you had the temp read out. Did you adjust the battery setting on VX3 itself? One of the options it has is the number of series cells in the battery. Generally I think the best practice is to plan out where/how far you ride to try and match the battery capacity. Then when the first cutoff warning comes stay close to shore. Also when you get low in the pack there can be voltage sag on some packs and it can be harder to get back up on foil if you’ve fallen.

I think 75200 is a better choice for eFoils over the 75100. It can work with passive cooling (search the forum for “Fliteboard adapter plate” and you can find a few examples).
I think with your 75100 keeping in it’s housing did it no favors. It may have lasted longer and cooled more effectively if you tossed the housing (it’s not waterproof, just some impact protection and extra weight) and placed the esc’s bottom plate directly on your aluminum panel (with a small bit of thermal paste). This would give much better heat conduction and is the strategy that works quite well in the fliteboard based builds. You could try this with a 75200 and skip the water hose and I bet would be fine

[sharrow]

If you want good cooling you have to remove the aluminium case and put the esc without the black case directly to the cooling block that goes outside your case with a very thin layer of cooling paste. The more paste you use the more worse it gets. Go for the 75200 I used it with a 70131 motor without any additional cooling only with the original Black cooling case inside a pelican case. No water cooling was needed and no extra cooling block. Depends also on your settings in the vesc how much heat you will get in the esc.

[Joez]

Good to know the 75200 would work better. Thanks for the suggestion. It’s interesting that there’s such a big difference between 75100 and 75200 - one would overheat within 1 minute (with poor cooling though) and the other doesn’t even need any extra cooling. This gives me confidence that I may not need water cooling after all.

[Joez]

Don’t those big salt crocs live in Queensland?

I was testing in a reservoir so there wasn’t dangerous wild lifes in the water. If I were in the river or sea, I’d surely be worried.

My remote is connected via UART and I did set the number of cells for the remote. I think I also set the cells in VESC tool but I can’t confirm it now since it doesn’t connect anymore. I’ll surely make sure the battery indicator works before heading out next time.

I think I’ll try 75200 with the hoursing removed. If I can skip the water cooling, it will save me a lot of headache. Thanks for the suggestion.

I may also test how ineffecient my current cooling setup is then improve it and re-test. I suspect the 4 holes filled with themal epoxy is the bottle neck of the whole head transfer path. I’m planning to merge them into a bigger cut and put a 3mm aluminium bar in place instead of themal epoxy. I can’t attach the ESC to the outer plate directly because the plate (and the side of the waterproof case) is not big enough to fit the whole PCB. I’ll also add another headsink to the outer plate then hopefully it’ll be enough to keep the ESC cool.

[JonathanC]

Using 75100 removed from factory case and directly in contact with external heatsink works fine even in efoil constant use. We can run for 40 minutes or longer with 10S6P and 6384 motor. That’s 30 plus degrees Melbourne summer.

[Joez]

That’s good to know. I think my issue was caused by the case didn’t transfer the heat properly to the inner plate and the inner plate didn’t transfer the heat quick enough to the outer plate and heatsink. I did some test that I blowed hot air to the ESC case to simulate the heat generated and measured the temperature of the heatsink. It only increased 0.9c after 2 minutes and the ESC increased 10+c. I’ve replaced the thermal epoxy between the inner and outer plates with a solid aluminium bar. Hopefully that will increase the heat transfer rate. I’ve also ordered a Makerbase 75200 v2 so let’s see how the new setup holds up .