Jake's Direct Drive build, NSW, Australia - Success!


#1

Ive achieved (pretty much) success with direct drive, so I thought i’d make a new thread going over the build, and to show the video of it working!

Heres the video:

and if that doesn’t work, the link to the video: https://drive.google.com/open?id=1SoZTYR4QNMF5R4qxD1gLNvYFFSyTU42I

Its surprising how efficient and powerful it is, in this video the current was limited to 50a battery and 90a motor for each VESC. I know from my pool testing that the battery current never exceeds roughly 35A because the ESC is limiting duty cycle because max motor current is already reached, so this means that it maxes out at just over 3KW (stationary) the RPM would increase slightly as it speeds up enabling more power throughput.

It uses 2 6364 190kv hobby king out outrunners they are only $60AUD each,

Here’s one of the motors epoxied with some 5 minute epoxy from the hardware store.

This is the motor system… About to design a duct, I thought I may as well test it first so i don’t make a duct for a bad prop…

In the electronics box I have 2x 10AH 6s 10c and 2x multistar 6.6AH 10c in parallel to make a 16.6AH 12S pack. The ESC’s are VESCS, they are struggling to stay cold, they have absolutely no cooling whatsoever, I’ll have to sort something out there. They can only handle a minute or so of riding before I have to wait for a few minutes to let them cool.

I did increase the max temperature to 110C which is as high as it lets you set it, and that extended the run time a bit before the power was dialled back.

in the less than 10 mins total of crashing and riding, the charger put back into the batteries under 190WH still above 4V per cell which was surprising.

I still need to address: (order of magnitude)

  1. VESC overheating
  2. Riding skills
  3. Slight water leak
  4. More Buoyant board? I find it very hard to stand up when the remote is in my hand. Probably my lack of skill though.

Links to the parts:
motors: https://hobbyking.com/en_us/kd-53-30-high-voltage-brushless-outrunner-190kv.html

ESC: VESC 4.10

Foil: https://www.aliexpress.com/item/top-quality-hydrofoil-carbon-foil-with-alumimum-mast/32857409005.html?spm=2114.search0104.3.30.59ebf533sSc11t&ws_ab_test=searchweb0_0,searchweb201602_3_10152_10151_10065_10344_10130_10068_10342_10547_10343_10340_10548_10341_10696_10084_10083_10618_10307_10131_10132_10133_10059_100031_10103_524_10624_10623_10622_10621_10620-normal#cfs,searchweb201603_43,ppcSwitch_5&algo_expid=27c94904-efee-492b-a7c2-656d8f94ac6b-4&algo_pvid=27c94904-efee-492b-a7c2-656d8f94ac6b&priceBeautifyAB=0

batteries: https://hobbyking.com/en_us/multistar-high-capacity-6s-10000mah-multi-rotor-lipo-pack.html and https://hobbyking.com/en_us/multistar-high-capacity-6s-6600mah-multi-rotor-lipo-pack.html

All other parts are 3d printed and I’ll post the fusion files here when I get my other computer.


#2

Great! Congratulations @jakebarnhill1!

Does it mean you run both motors/escs from the same 12s pack, so each ESC gets 48v when fully charged? In this case all two motors together consume 35x2=70a?


#3

As per @PowerGlider advice do you think to make a motor mount that overlaps the motor so the water won’t flow through the motor to minimize the centrifugal parasitical forces?


#4

Cool!! A larger board would have helped you a lot, but the set up is very cool!
Maybe, add a waterpump and watercooled heatsink block on each VESC, it’s more efficient than air cooling…
Also… get yor “bail out” routine dialed in better than now…for your safety!


#5

I would first start with some little stickable heatsinks like this:


(these are for raspberry pi, but just look up the size of the 4.X mosfets and get something suitable)

I had a similar problem with a VESC used as an ebike controller and that fixed it (since not much more cooling was needed).
And if it doesnt work you can still work out a little more complex watercooling design or similar…


#6

@Giga , Ill give that a shot. I have some aluminium and old heatsinks laying around so ill put that on, I don’t know how much of a difference it’ll make in a sealed container, but it should help, we will soon find out.

@Antonbit, yes both ESC are pulling from the battery together so the max current is around 70A. I have blocked the motor holes with a 3d printed cap at the prop end as you can soo in the photos, but I guess blocking the visible cooling holes could help improve efficiency, since they don’t heat up at all.

@MaB if the above doesn’t work which it probably won’t, ill install a water cooling system to the heatsinks. What is the best way to bail out?

Thanks everyone!


#7

Well, for the bike it was also a “water-resistant” case (so no airflow inside). I guess the problem is not the amount of heat/losses, but the transfer. These small heatsinks wont help you getting rid of the heat inside the box, but help a lot getting the heat spread away from the fets.


#8

Thats good to know, and I’m sure the heat shrink around the Fets would trap the heat so it should help a lot. cheers!


#9

The VESCS might enjoy themselves a bit more now :slight_smile:


#10

If you use such thick thermal pads be sure to put enough pressure on it, they are (only) used to compensate irregularities on the surface and compared the aluminium it self they are pretty bad at heat transfer (thermal conductivity coefficient of those pads is <20W/mK and aluminium is like 200W/mK).


#11

What do you mean by W/mk?


#12

That is the unit for heat conductivity. A higher value means better heat transfer.
You assume 15 Watts of losses in a single transistor on your Vesc for 90A motor current. This assumption is a little bit on the safe side but you have more losses the hotter your transistors get. Your thermal pad has (example values I found: http://www.sg-thermal.com/Thermal-Pad/) for example a thermal condutctivity of 6 W/mK and is 1mm thick. The surface area of the transistors you can use for cooling is about 1cm². So you can calculate the thermal conductivity of your thermal pad to be 0.6 W/K = (0,0001m²/0,001m)*6W/mK. This means with 15W of losses your temperature rises 15W/(0.6W/K)=25K (or 25°C) because of your thermal pads alone. If your thermal pads were pure aluminium with about 170 to 200 W/mK your heat would be removed much better with only a rise of <1°C under the same circumstances.

What this really should suggest is using only very thin thermal pads if you need electrical isolation to the heatsink. I have done many experiments with the Vesc 4 in my setup myself and can only suggest that for running any currents higher than 60A conitnously addidtional cooling beond a simple fan is necessary.
Removing the heat through the plastic case of the transistors (i.e. heatsink on plastic back of transistors) is not really worth the effort. Soldering some from of cooling to the remaining metal tabs on the transistors like PowerGlider did can help to acheive higher currents but the Vesc 4 is really hard to cool beyond 80A. You might just get there for your motor current with cooling fins soldered to each remaining tab on the transistors and a fan. I went a little bit beond that and have remounted all fets and attached a big watercooler to all fets on the metal cooling pad, the fets stay cool with over 100A now. But I have trouble with the current measurement shunts now. Those also make 10W of heat and are even harder to cool.
My suggestion would be to test your current setup with your current cooling. If you find you need just a little bit more power you can solder small copper cooling fins to the mosfet tabs. In any case add a fan to blow some air over the shunts and your cooling devices. Altough your heat is trapped in the waterproof box this should give you more time until your vesc throttles back.
In any case keep up the good work. I think I might have to add a second motor to my 6374 and go from my jet drive to your setup. Thanks for sharing your findings!


#13

Ah!ah! On your back, opposite side than foil! As soon as you feel getting close to the point on ‘non return’, jump out!


#14

Well, if you go beyond 60A cont. in dual setup you will have battery problems (discarge rate of multistar should not exceed 5C over longer period and runtime will decrease to like 10mins) before you have thermal problems…
So my guess is, that there is not very much missing…


#15

@Flo @Giga Ok I see, Aluminium alone is a lot better!

The heat sinks did help a bit I can ride continuously for over a minute now. Heres the youtube video of it now, I’m getting a bit better at riding, the heat sinks let me fly for much longer than before…

@MaB yes I see what you mean, it is very important to do it, otherwise its so easy to fall into the wing!

I might try water-cooling the heat sinks that are touching the top of the FETS for now, so i don’t risk stuffing the VESCS, if its not enough, Ill try something like what @PowerGlider did with his VESC.


#16

Has anybody tested other potential solutions such as:

  • to avoid salt contact (even if circuit is rinsed after each session), fresh water closed-circuit cooling as done on CNC router spindles, computer CPU. Water in a small bottle might require a separate safety compartment in/on the deck due to battery proximity though.
  • thermoelectric/peltier cooling ?

#17

Thermoelectric does not work. Power requirements too high and cooling is not very much.
Also, they only transfer heat a very short distance


#18

thermoelectric:peltier cooling does work (see Peter setup), i have tried 3 peltier in the row without sucess only because my heatsink on the hot side was to small and i need a fan when in static , the heat was going back to the cold side and the esc

i let go this type of cooling only because i have enough air cooling just by the heat sink on top of my case

cpu water cooling may work in closed loop, but this is starting to be heavy