ESC Cooling Concept- Theory and Practice

[questtek]

How about putting a small generator on the drive shaft. Something thing that will generate enough power to run the cooler.

-t

Interesting thought but it may not be feasible for the following reasons:

1, The Peltier for example uses about 4 amps at 12 VDC which is approximately 50 watts of power (much larger models use about 400 watts) Assuming the generator is 70% efficient that means you would have to "pull off" the engine power about 71 watts.

2. The 540 XL motor has a nominal rating of 800 watts. Thus you must use about 10% of the motor power just to have a generator supply the voltage.

3. Net result is that it is MUCH easier to do with just another small 3 S LiPo pac.

BUT, most importantly the power requirement is a distant second to the amount of cooling required to get the Peltier device to work. It is this large cooling requirement that prevents their effective use to cool the ESC and not the voltage.

[questtek]

Closed-Loop ESC Cooling Test Results

My Mean Machine was set up with a closed-loop cooling system previously described in this forum. I blocked off the normal water pickup and relied on a small plastic container of ice and water instead. This was pumped through the ESC and motor by a small 12 VDC pump.

The attached pictures show the test set up in detail: ice and water in container, system mounted in the mean machine and the Eagle Tree logger with two temp probes, engine RPM and GPS.

The key thing I was interested in is to determine the ESC FET temperature (not just the heat sink) and the motor temperature (on the case in front of the cooling jacket).

Ambient or initial conditions were Ta/Tb or Temp of ESC/Temp of motor was 76.4/77.8 (all temperatures are in degrees F).

When I turned on the cooling pump these temps dropped to Ta/Tb 51.6 /65.8 (remember the ESC has a giant copper heat sink that I made) in the two to three minutes it took me to close the hatch and walk the boat from my shop to the dock in front of my house.

I put the Mean Machine in the water (pic attached) and decided to have about 4 minutes of run time that went from medium speed (30 amp draw on a 4S pac to 80 amp draw straight line speed runs)

After 4 minutes I opened the boat and downloaded the Eagle Tree data to my computer to look at it in detail. Here are the results after the 4 min run.
Hi Amps 80.01 amps
Hi watts 1113
Hi RPM 25,862
Hi Ta (FET Esc temp) 70.9 F
Hi Tb (motor temp) 85.5 F

Removing the red lid from the ice and water cooling container after the run I found only water but it was still cool, about 4-5 degrees that the ambient temp of the lake water that I would have normally been used for cooling.

I repeated this test a second time with a fresh set of LiPo’s and the results were almost identical. There are a few improvements I can make but the concept seems feasible at this time.

[Jeff Wohlt]

Cool to know that the water is the real deal.

I think you should go to a small nitro tank. It already has the tube hook ups! You would have a little more water and can drop more ice in. I do think you would have had ice left if you would have made a frozen bottom in it. Solid ice will melt far slower than pieces. The reason I wanted the nitro tank as I was planning about 1/2" in the bottom frozen. Overall I think it would be cooler longer.

What numbers! After that run and still fet temps are that low!

[questtek]

Cool to know that the water is the real deal.

I think you should go to a small nitro tank. It already has the tube hook ups! You would have a little more water and can drop more ice in. I do think you would have had ice left if you would have made a frozen bottom in it. Solid ice will melt far slower than pieces. The reason I wanted the nitro tank as I was planning about 1/2" in the bottom frozen. Overall I think it would be cooler longer.

What numbers! After that run and still fet temps are that low!

Picked up a small fuel tank for about $4 in the local hobby store. Looks perfects and fits like a glove in the MM. I can even use two if needed.

It is a good thing I have 4 sets of matched 4s, 40-C, 5000 mAhr LiPos so I can do continual testing..which I did today. I managed 8 runs in all different cooling conditions. I really pusted the MM and the 540 XL with max amp draws in the 85-90 amp range and max watts from about 1150 to 1270. That is really pushing this motor and it is doing just fine with acceptable motor temps thanks to the cooling schemes.

Details are way too much to publish but I am perfecting the system. I am really pleased that the 5 or 6 modifications I have made to the inexpensive 100 and 200 amp controllers are really working. Have not had a single hickup and they have performed better than any of the expensive high priced ESC's that I have used....at least for the few days of hard testing that I have done.

Will make more improvements and publish the results.....

[ray schrauwen]

Cool! literally!

Thanks for your hard work and experimenting. I will use this on a few boats.

Keep us posted

[Jeff Wohlt]

Excellent work. I am going to pick up a few tanks from the hobby shop...he has many sizes. Remember, ice floats so having a clunker in the tank on the intake hose would be good to keep it on the bottom of the tank where the coler water is..

Yes, the key is copper plate to draw the heat rapidly and exchange it.

Just think how this will do when it is 90 out and the water is also 90. That is when I think this unit will pay off very well.

I did find cpu coolers with pump and tank built in but they are waaaayyyy to pricey...like $60 or more.

Thank you for all your efforts! Very successful ideas.

[questtek]

Final ESC Cooling Tests using a Closed-Loop Cooling system with crushed ice and water in a 4 oz Nitro tank with Miniature 12 VDC pump.

I am very pleased with these results. They represent running the pump and closed loop cooling system for two minutes, (the time it takes to close the hatch and tape it to insure no water enters), and a very hard, nearly continuously full-throttle run in my Eagle Tree instrumented Mean Machine with a 540 XL and my highly modified 100 AMP controller.

Initial Ta/Tb or T ESC FET temp/T Motor Temp in front of cooling jacket in degrees F.
Ta/Tb initial = 68.9/69.0
Ta/Tb at the end of the 2 min. pump-on and closed loop ice and water cooling = 47.6/60.3

After a hard 6 minute run the Eagle Tree showed:
Lo Pac V = 13.9
Hi Amps = 83.77
Hi Watts = 1201
Total mA = 2795
Hi GPS Speed = 29 MPH
Hi Ta or EXC actual FET temperature = 79.7 F
Hi Tb or Motor Temp = 112.3 F

At the end of the run the ice was all melted in the tiny 4 oz “Nitro” tank but the water was still a few degrees lower than the ambient pond water.

Just to complete the documentation, here are how Ta/Tb increased during the hard 6 minute run of my Mean Machine:
after 1 minute Ta/Tb, (ESC/Motor Temps)= 54/63 F
after 2 min = 59.2/73.4
after 3 min = 60.1/78.1
after 4 min = 65.4/81.3
after 5 min = 71.9/93.2
after 6 min = 79.7/112.3

Even with the best 540 XL cooling system, the motor actually got a lot warmer, (by about 33 degrees F), but was still well within an acceptable operating range. The effectiveness of the large copper heat sink and copper pipes as well as other modifications to the ESC I made have kept the actual FET temperatures down to a comfort zone that probably never dreamed possible!

[ray schrauwen]

Fantastic!

I like how you changed the wires so that they come out the same end for tight spots!!!

[Jeff Wohlt]

Nice temps for the ESC!!!...I really thought motor would be lower. Still not a problem on a hard 6 minute run. I can only imagine what normal water would do from the lake with 80 degree temps.

4oz is still very small...I am thinking double that and still would be minimal weight.

It would also be a good device to help with CG if needed.

I congratulate you on all your findings! Thanks for taking the time and effort to do this.

The nice thing is water temps do not affect this system!

[ray schrauwen]

I think I'll pick up one of those pumps at our hobby show next weekend, actice surplus always has some there.

[Jeff Wohlt]

You can also get a fuel pump for RC nitro and either revoce the cover or leave it on. They are small and can run 6 or 12v. I would go for a 6v and use a small 2 cell pack or even a 4 cell AA pack which would run it plenty fine. I did find some other 6/12 pumps for $8 and dim are 78 x 33mm Long but thin.

[ray schrauwen]

Thanks Jeff, I like the lower power consumption idea or lower voltage small pack.

[Jeff Wohlt]

Questtek, Are oyu running stick radios with those ESCs or pistol?

Here is a link to small pumps I found

http://www.lightobject.com/3LMin-min...-HHO-P122.aspx

[ray schrauwen]

I don't think those esc's need to run on stick radios, most don't now.

[haidiacr]

Sorry for pulling up and old post but after reading this thread i was wondering why you havent focused more on the transfer of heat between the fets and the water. From what i have read most of you are using a copper plate with either copper tubes true them or soldered on them, why not use a design closer to that of a water cpu cooler. Or preferably put the water in direct contact with the fets.

[m4a1usr]

Bringing back an old thread is a good thing. Always helps to remind folks of whats been covered and dont bother rehashing something clearly previously defined.

There is no way any design could put the media we run in, water, in direct contact with the FET's. The reason is simple. Corrosion. When electricity is passed thru conductors it forms what is known as galvanic reaction. The reason is because water is composed of many properties. While in its simple form (h20), there will be less reaction, the problem arises when contaminates form bonds and allow precipitation of chemical reactions due to those individual chemicals. Pure water is not our media. We run in water with many contaminants. And the chemical make up depends on what part of our nation or global location you are situated in.

John

[AndyKunz]

Oh man! I ran at a SAW event on Long Island a while back (setting a P-Hydro record at 39 MPH - stop laughing, it was fast at the time!). I had buffed and polished props. Problem was, after a single run they would be dingy brown! The steel shafts had rust spots on them where before they were shiny. This is after maybe 10 minutes in the water!

The shore had needles and assorted garbage all around it. Pretty trashy place, imho. I did get to see a Concorde landing while I was there. That was pretty neat.

Andy

[haidiacr]

Good to know you can bring up old threads here. I see your point about corrosion, the reason i brought it up is because this reminds me of the time i was in to oc'ing computers. And with watercooling the efficenciy of the heat exchange between the cpu and the water was the key factor. But forgot that cpus had heat spreaders and the liquid used was designed to prevent corrosion. But still most of the theory should apply here too, you could build something similar to the copper block at the bottom of this pichttp://www.build-your-own-computer.n...cooling-03.jpg And get even lower temperatures, but the cooling boxes will heat up faster

[Jeff Wohlt]

Looks like more intesest in this...Dr. Wayne. Smaller nitro tank and then fill system alll the way thru so it is full and purged. Add ice chunks and start pump. Very easy idea and works. Read this thread.