ESC Cooling Concept- Theory and Practice

[questtek]

ESC Cooling Concept- Theory and Practice
I thought a new thread may be appropriate in response to the “Cooling Brain Fart” thread to discuss a bit of theory, few facts and present one potential workable concept for cooling ESC’s for FE boats.

The Theory….
ESC’s use conduction cooling. (so do electric motors but lets focus on ESC’s) The better the conductive path, the better the cooling. For aircraft ESC’s the high velocity wind from the prop across the finned heat sink is normally sufficient. In FE boats, the normal air cooling is not adequate which is why marine ESC’s are water cooled and/or forced air cooled by using a small brushless fan similar to that cooling a computer processing chip. Excessive heat is probably one of the major causes of ESC’s burning up in FE boats.

To improve cooling in an attempt to burn out fewer ESC’s we need to improve conduction cooling. This means using better “heat sink” materials to convect the heat away, increasing the surface area of the conductive “heat sink” and reducing the inlet cooling water temperature. Getting heat out of the FET’s is the goal.

In terms of better heat sink materials one needs to look at the thermal conductivity of the material used. Thermal conductivity is the quantity of heat transmitted through a unit thickness of material due to a unit temperature differential. (The term “unit” means that we care compare apples-to-apples so to speak). Maybe a simpler definition would be the rate at which heat is transferred by conduction when a temperature gradient exists. Simply put, the higher the thermal conductivity of a material the better it cools.

Comparing thermal conductivity of different materials………..
Silver has one of the highest thermal coefficients at 429 but copper is a close second at 400, (and a lot less expensive). Aluminum is 250 and Brass is 109. Since conductive heat transfer is directly proportional to the thermal coefficient, use of copper rather than aluminum for a heat sink would provide a 38% improvement. Can you imagine running aluminum wire from your ESC to LiPo’s and brushless motor? The building code for many areas now requires use of copper wiring and not the cheaper aluminum wiring (as in my 27 year old home).

Several great threads have been submitted on adding cooling to ESC’s such as the relatively inexpensive 200 amp Mystery models. Brass tubing is integrated into the existing aluminum heat sink. But the thermal coefficient of brass is 109 compared to 400 for copper. I think you can work the math here for the best tubing material.

A proposed open loop cooling concept ……………
In a previous thread I outlined using a 35 mm film can or canister with a cooling coil inside. The canister is filled with water than put into the freezer. The inlet water passes through this prior to going to the ESC and motor. Even this simple idea decreases the film canister outlet temperature but about 20 degrees. However I only use this on very small boats in the 12-16 inch range. It certainly would not help to any degree of most of the larger FE’s that are run by members of this forum. One comment I saw in the forum recommended just using water or alcohol in the canister (or really the cooling tank). Is this good idea? All one has to do is to compare the thermal coefficients…. remember the higher the coefficient the better the cooling. Air is 0.024, water is 0.58, alcohol is 0.17 and ice wins out at 2.18. Thus the cooling with all ice is 4 times as effective as water at the same temperature. The large ice value is a result of the latent heat of melting….if that makes any sense. (I am running some tests on the liquid used in ice packs and will post the results later.)

One Concept for “Canister Cooling” that I use on my Mean Machine with a 540 XL….
For lots of cooling we need lots of copper tubing. In the pictures you can see I purchased some small Tupperware-type containers: 4 for 99cents and the……….you guessed it, the local 99-cent store. I wound 20 inches of ¼ inch diameter copper tubing around a wooden plug so it fit perfectly in my 25-cent container. If you calculate the surface area of this length of tubing it is whopping 15.7 square inches. (Compare this to the aluminum heat sink that comes on many ESC’s) The entire cooling unit weighs 3.7 oz or 104 grams and when filled with water (ice) it weighs 6.2 oz or 176 grams. The dimensions are 3” x 2.3” x 2.4” high. (pics included)

Now for a canister cooling comparison:
1. My 35 mm film can with brass tubing has an effective total cooling number of 513. (You get this by taking into account material used as well as diameter and length of tubing)
3. My 25-cent Tupperware-type container with 20 inches of copper tubing has an effective total cooling number of 6,280. Enough said on this.

How well does this really work?…………….
Good question and that is what I am determining now. I use an Eagle Tree data logger that has 3 temperature probes (thermistors). One is on the ESC, one is the inlet water (for the fresh water lake in front of my house that is about 76 degrees F now), and the third is the water temperature on exit from the cooling canister. (pics included) Stay tune for the results.

I am sorry for this thread being so long but I wanted to create a common technical base so the different ideas and cooling gadgets from forum members can be numerically compared….apples-to-apples. As a teaser I think there is a MUCH better way to provide ESC cooling that does not require canister cooling, Peltier plates, fans, pumps or giant heat sinks. I have modified an idea used by the Russians in their MIR submersibles that have gone to the bottom of the North Pole and made over 100 trips to the Titanic. Again, stay tuned for the results……..IF I can make this concept work!

I look forward to your comments and suggestions

[Steven Vaccaro]

Until this is tried and tested, I'm moving it to the, questions and answers area. Since its theory and not a "how to" yet.

I like your idea, one thing to think about. The system needs to be blown out after each run, so you don't have frozen water inside the copper tubing.

[Jeff Wohlt]

Good info. I was surprised the # differences in alum to copper. I was surprised to see brass so low.

I do not see why it won't work. I saw those tuperware boxes at walmart but thought they were a little big. I am also looking at small tanks or vac forming something that would be thin and then put the coil in and seal it shut....the inlet/outlet would also be formed at the same time and then just epoxy it shut.

I would try the coil in the boat without anything else and run it just to see how flow is from the rudder. I would think it would be fine.

[Drag Boat Bob]

Great information! Keep up the testing.

This is a take on an old time car drag racing idea where dry ice was used in a canister to cool down the fuel prior to entering the engine.

With all of the ESC failures (many due to heat) out there, any new idea would be great IMO.

[DISAR]

Nice work!

Your theory is correct. Your water inlet temperature will be probably lower resulting in a lower ESC temp. How about the actual ESC cooling, what wil be the arrangement? Which ESC you will test? In order to have good results you should use a marginal ESC for your setup.
Are you going to measure the temperatures with and without the canister?

[questtek]

Nice work!

Your theory is correct. Your water inlet temperature will be probably lower resulting in a lower ESC temp. How about the actual ESC cooling, what wil be the arrangement? Which ESC you will test? In order to have good results you should use a marginal ESC for your setup.
Are you going to measure the temperatures with and without the canister?

I will post some pictures later of the test set-up and how I have tied the three temperature sensors into the cooling system and the Eagle Tree logger. I should also be posting some test results later today.

Totally agree on using marginal ESC's I have ordered 4 Mystery ESC's
1. 2 of the 100A units (2-6 LiPos)
2. 2 of the 200A units (2-4 LiPos)

The total cost for everything was $95.96, a great deal for experimenting.....IF they ever show up!

I will measure the temps with and without the canister. First on a bench set up then in the blue Mean Machine in the pics. It uses a 540 XL and a 4 cell, 5000mAh, 40 C pacs. Testing is easy since I live on a lake and have a 16 foot electric boat to "pick up the pieces".

[questtek]

Good info. I was surprised the # differences in alum to copper. I was surprised to see brass so low.

I do not see why it won't work. I saw those tuperware boxes at walmart but thought they were a little big. I am also looking at small tanks or vac forming something that would be thin and then put the coil in and seal it shut....the inlet/outlet would also be formed at the same time and then just epoxy it shut.

I would try the coil in the boat without anything else and run it just to see how flow is from the rudder. I would think it would be fine.

You might want to try your local 99-cent store...they have several different types of small "tupperware-type" containers. Remember, it does not have to be large. The small plastic box I am using is only 3x2x1.4 or about 8.4 cubic inches in volume. They have another box that is retangular 2.5x2.5x1.5 high or 9.3 cubic inches. The retangular box rould permit winding the copper cooling tube around a section of old PVC pipe. Remember, this copper tubing represents the same area as a 15.7 square inch heat sink made out of a material that is roughly 4 times better heat conductor than aluminum.

I have included a picture of the copper cooling tube in the rectangular "Tupperware" box. It is 24 inches long, about 17% larger than in the slightly smaller rectangular box. I must admit it is easier to wind a round coil. In this case I used a 1.75" OUTER diameter PVC pipe as shown in the picture.

I also included a pic of the EagleTree system Note the three thermistors; I built them into 'T" fittings and sealed the thermistors. This does not restrict flow since the thermister tips are just at the endg of the main flow. With these I measure inlet (black shrink tubing on end) and outlet (red shrink tubing on end) and the final thermister is on the ESC itself.

[Jeff Wohlt]

Excellent. Will you do cold water in the box or or the gel? I have an idea I will share later this evening...and something to show.

[questtek]

Good info. I was surprised the # differences in alum to copper. I was surprised to see brass so low.

I do not see why it won't work. I saw those tuperware boxes at walmart but thought they were a little big. I am also looking at small tanks or vac forming something that would be thin and then put the coil in and seal it shut....the inlet/outlet would also be formed at the same time and then just epoxy it shut.

I would try the coil in the boat without anything else and run it just to see how flow is from the rudder. I would think it would be fine.

You might want to consider tight fitting rubber gromets for sealing. Add a bit of silicon if you want. In the picture shown I connected the white silicon cooling lines to the copper tube INSIDE the tupperware housing. The silicon tubes exit the housing through tight rubber gromets. I think this is easier and works much better becasue the flexible tubing gives you lots of options for the outlet and inlet.

You asked about just using a copper coil along. Actually this may HURT you rather than help you. The amount of heat transfer is proportional to both the thermal coefficient of the material as well as the temperature difference. With ice that difference is inlet lake water temperature minus the ice temp which is about 32 degrees F.

If you just use the copper tube inside the hull there is a chance that the internal heat from the ESC and motor..............and expecially if you warm your LiPos, may cause the copper to heat up. In this case you would be ADDING heat to the ESC and motor rather than revmoving it.

[questtek]

Excellent. Will you do cold water in the box or or the gel? I have an idea I will share later this evening...and something to show.

I look forward to hearing your idea. Good luck on the testing.
I am experimenting with both the ice and the gel. I could not find the thermal coefficient for the gel but I continue to search for it. I will just test both of them and see what happens. My guess is the gel will show a degree of improvement over plain water.

Remember, cold water has a thermal coefficient of 0.58 and ice is 2.18. Putting the box in the freezer to form ice is a tremendous improvement in cooling efficiency. Even if you use the gel, freez it!

[questtek]

I have just finished some preliminary tests. The pictures show the test set up.
The entire system includes the cooling system components including small 12 vdc pump to duplicate the water coming in from the boat's water intake. I have a large monitor to view the Eagle Tree data in the live mode plus also record it. I am measuring the three thermister temps, inlet, outlet and ambient. There is no ESC in the loop, I just wanted to see the temperature difference as ambient water cooled after passing through the copper tube in my tupperware ice both.

In the test set up, ambient water is in the red plastic cup, the small 12 vdc pump pulls the water out as it passes by the intake temp thermister.. The ambient water flows through the cooling container then into my High Fashon Drinking cup. The outlet tube has a thermistor attached.

The close up of the Eagle Tree temp guages, inlet and outlet pretty much tell the story. You can observe the temperature drop. I also noted the cooling started to decrease after a bit....the reason was the water was melting around the tubes previously covered by ice. and warming up) I just shook the tupperware cooling canister....to duplicate the vibration and movement on a FE and the temperature came down again. It appears you cannot have "stagnant" water next to the copper cooling tubes.

Now to try the gel then some real tests in my Mean Machine.

[Steven Vaccaro]

I dont see any of eagle tree pictures?

Also would it be better to have more space in between each coil layer? So there is more ice in between each coil?

[DISAR]

Pictures? How much is the temp.difference?

[Ken Haines]

You guys are doing great on this cooling R&D. I do not have anything to add, just wanted to let you know that I'm following this closely. My free time to field test this stuff may not be soon enough as you guys have really jumped all over this.
Maybe I can pledge that whatever is the best end result, I will race test at the 2010 FE Nats.
Keep up the great work!!!
Thanks, Ken

[questtek]

I dont see any of eagle tree pictures?

Also would it be better to have more space in between each coil layer? So there is more ice in between each coil?

Sorry, the pictures must have gone to electronic vapor-land but they should now show up below.

You are correct on coil spacing...you want the ice or gel surrounding each tube so a bit of open space is best.

While this method appears to work quite well in the test setup, I still need lots moe testing to refine it. I need to come with the size of the container and coils as a function of duration of run and size of ESC and motor.

It has been over two weeks since I ordered my four 200 and 100 amp mystery controllers but they have not arrived for the real serious testing.

[questtek]

Pictures? How much is the temp.difference?

The picture shows the inlet temp on the right and the cooled temp after passing thru the coils on the right. ambient air temp was about 74 degrees F

[questtek]

[QUOTE=Ken Haines;131750]You guys are doing great on this cooling R&D. I do not have anything to add, just wanted to let you know that I'm following this closely. My free time to field test this stuff may not be soon enough as you guys have really jumped all over this.
Maybe I can pledge that whatever is the best end result, I will race test at the 2010 FE Nats.
Keep up the great work!!!
Thanks, Ken [/QUOTE


Thanks a bunch, that would be great. When I get closer to developing a much more refined coolong unit, backed with lots of test data, I will send a system for you to expertment with. If you have a newer V-3 Eaggle Tree I can send it along with three temp probes. This will allow you to measure inlet water temp, post cooling water temp and actual ESC temp.

[MarkF]

This is great stuff. I have burned up more speed controlls at SAW events than I care to admit. What about using a small piece of dry ice so the cooling effect last longer. We had a guy who use to race with us a long time ago who made a little wood box to put his controller in and then he put a small piece of dry ice in the box with it. It worked great untill the dry ice melted and got condensation in the controller.

Mark

[Ken Haines]

[/QUOTE


Thanks a bunch, that would be great. When I get closer to developing a much more refined cooling unit, backed with lots of test data, I will send a system for you to experiment with. If you have a newer V-3 Eagle Tree I can send it along with three temp probes. This will allow you to measure inlet water temp, post cooling water temp and actual ESC temp.

Yes I do have Eagle Tree Systems, 1Pro, 1 Standard. the Rev I do not know. They are the older big data loggers with the remote readers, there is plenty of room in both boats that I'm considering. The class at nats that this most applies to is LSO. Hopefully they will still be running it as a 4 minute timed event. This is where the heat came into play at the 2009 nats, LSH would also be a consideration to maintain a very aggressive set-up by better cooling.
Thanks, Ken

[Jeff Wohlt]

Just tested my little ice pack (see below) I used a coke bottle full of 90 degree water but gravity feed...no pump. Water flow seemed about what the rudder would do on my boats...slow steady stream

Water was 75 degrees coming out. I measure it using my digital meat thermo. Pretty accurate for what it is. Actually you will probably be slowing flow as it goes around everything which is even better.

So now I wonder what a couple more tubes will do and using copper tubing...which I did not have.

BUT...the good news is it does drop water temps considerable for only 15" of brass being in the ice.

I would like to pump more thru it and see how long it lasts.

[properchopper]

Joe, this is great info and I commend you for addressing this issue, one which is of considerable importance to FE. One concept which has been previously discussed, without complete consensus, is the flow rate between the cooling medium and the heat source [e.g. motor and/or speed control]. I'm sure you'll agree that the thermal transfer rate, or cooldown effect, is intimately related to this factor. I'd like to hear your take on this variable. Thanks for your efforts & keep us posted

[Jeff Wohlt]

hmmm...just added 3' of silcone to the exit and same test...now water is 70 degrees coming out of tube. Slowing flow is the key. 5 degree drop.

so about 20 degree drop in water temps...

My goal is to create temps from running in summer time....I doubt any of us need cooling when the water temps are 60 degrees anyway. 90 degrees is probably more than needed.

Going to try 75 degree water now.

[Jeff Wohlt]

Interesting....74 degree water only came down to 70 degrees.

So hotter water is cooled much better.

[Jeff Wohlt]

So obviously Joe has more tube in the ice and using copper. His temps went down more. Mine has little in it and was actually surprising how well it did for so little and easy to make. to drop 20 dergrees is pretty good.

Yes, pulling heat from the ESC is the key...it has to have good contact with fets. I think plates should have more than 2 outlets on a plate on each side when we need to cool all rows of fets...another reason the Pletier could be excellent...full contact on fets.

Maybe the Peltier is key or the water circ method but I worry about more motor interference from anything powered. Peltier might be better as there is not spinning motor. Then water cool the top heating part of it. But doe this take away from the cooling side?

They pull a bunch of power is the problem. A small 3S pack should be able to pull 5-6 amps to keep it running. I know it will drain a deep cycle battery over a few hours. I have done it with my cooler.

[questtek]

Miniature pump information…

Years ago I bought a bunch of small 12 vdc pumps at surplus. I used most of them for the “blood injection system” for the remote control great white shark fins that I built. Going thru my spare hardware I did run across two of the pumps that remain. I will provide you with some details so you can consider what may be available if in your cooling situation you opt to go with adding an additional pump to improve ESC and motor cooling.

I have included pictures of the two miniature pumps. One I covered in black shrink tubing, the other is the bare case. The pumps were made by Oken Seiko Co. Ltd, Tokoyo and designed as a rolling pump. (This type of pump no impellor but a unique settle of wobble diaphragms to push the water). If it would do you any good, the Model Type of the pump is P0532197Ce27-3379851-Rev 1 (Good Luck!)

The pump specs are as follows:
Size: 2 1/4 “ long x 1” dia.
Weight 2.1 oz or 60 grams
Power required is 12 vdc and draws a max of 200 mA when pumping water
Pumping rate as I just tested them was 12 oz of water per minute.

So, for a 4 minute run, as many members have discussed, the total volume of water used if the system is open loop (water pumps out) is 48 oz. Since there are 128 oz per gal and water weighs 8.34 lbs/gal this means the weight of water to provide 4 minutes of continuous cooling with this pump would be 3.12 lbs or 1419 grams of water. This does not seem practical to me which why we need coolant materials that have a much higher heat transfer than just water.

You could control this pump by a spare remote switch on your transmitter.

Remember, the rate of heat transfer from your ESC is a function of the thermal coef of the conduction material, temperature difference AND flow rate of the cooling liquid to remove the heat. Keeping this flow rate high is imparative

Ideas, suggestions, comments?

[D. Newland]

Joe-great work on this!

Where are you in Ca?

[questtek]

Blue Ice - It has been discussed by others in this forum so now it's time for a bit of detail on this magic blue gel.

This blue gel substance, normally called Blue Ice, is a re-freezable ice substitute. It is currently marketed by RubberMaid and a few other vendors. It needs at least 3 hours to completely freeze.

I just picked up two differnt pack of Blue Ice, a small hard plastic container and a soft plastic bag. You can see them in the attached pic relative to my hand.

The Blue Ice in the hard plastic container is 2.8”x1.3”x4.9” and weighs 6.2 ox or176 grams. It costs 94 cents and is found in the sporting good section of WalMart. The soft patic bag of Blue Ice is the same price.

The Blue Ice mystery liquid seems to be a trademark secret as far as I can tell. I could not even find an MSDS safety sheet for the product. I was just curious. I really do not know what is in it JUST how well does it work! To find this out I filled two of my first generation 25 cent tupperware containers with Blue Ice from the soft plastic pac and put them in the freezer. There was more than enough to fill both test cooling tanks with their internal coils of copper tubing.

Tomorrow I will continue testing but tonigt it's Blue Ice chill time.

The upside of this project is how little money needed to experiment. I probably have less than $10 invested and that includes 4 tupperware containers, 1/4" dia copper tubing, gromets and even the magic Blue Ice mystery fluid! That's less that the postage to have another ESC delivered! But, add the three Eagle Tree data Loggers, 12 thermocouples probes and other sensors, large LCD data display monitors, tablet computers.................Opps, my wife is coming in!

[questtek]

Joe-great work on this!

Where are you in Ca?

Thanks, this is a practical and fun project. I keep a bag of burnt our ESC's on my desk for motivation!

I live in Lake Forest, Ca....near Newport Beach - about half way between LA and SD. I am vey fortunate that I actually live on a man made lake which is absolutely perfect for FE experimenting. No driving, just walk 20 feet from my workshop to the dock. Even have a 16 foot electric boat as a recovery vessel.... (and cocktail cruising. I welcome anyone in the area to stop by.....and have include a few pics in case anyone takes me up on it!

[Jeff Wohlt]

I like those pumps. I have found some close to that but am looking for cheap.

I still think a pump with closed loop back into the tank is going to be faster cooling but more headache to build.

I was going to slit the jell hard pack and put the coil in it but wanted to try some tubes thru it first. It does work as it came from 90 to 75...so if you can keep that water in more tube you certainly should be cooling.

Good stuff

[questtek]

I like those pumps. I have found some close to that but am looking for cheap.

I still think a pump with closed loop back into the tank is going to be faster cooling but more headache to build.

I was going to slit the jell hard pack and put the coil in it but wanted to try some tubes thru it first. It does work as it came from 90 to 75...so if you can keep that water in more tube you certainly should be cooling.

Good stuff

Jeff, The surplus pump I showed was only a few dollars each. They or something similiar should be out there for low cost.

Great work with the Blue Ice gel pac. I will use my Eagle Tree set up to test the units I made up with the copper coil from cooling. I will use one the pumps I have but calibrate the flow rate and see how it changes the temp. At 12 VDc the output is 12 oz per minute but at 6 or 9 volts I a sure it will be considerably less. I am anxious to hear your results.

I am not quite convinced yet the use of an additional pump is necessary. The amount of flow even from a rudder intake is quite large when the boat is moving at a reasonable speed. Only testing will tell!