ESC Cooling Concept- Theory and Practice

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • questtek
    Senior Member
    • May 2009
    • 556

    #106
    Originally posted by Jeff Wohlt
    Dang I feel like the poster child getting anything I want!

    You are doing a big service to us all...just so you know that. Anyone not following this thread would be silly!

    Yes, A small tank would be a better testing ground...just a simple tupperware container where the water is going back in the tank. Then prop up and pull some power! :) Are you finding that the tank is at a continuous level with water coming back in? Of course priming the system then filling with water would be good in a boat if using a very small tank and lots of tubing.

    Also, a chance at geting a motor temp would be great....even if just by your hand!

    Thanks man!
    ESC Testing with a Small Container of Iced Cooling Water

    Inspired by the degree of ESC cooling using a large, quart picture of iced water I wanted to know just “How Low Could You Go?”

    Let me describe the test set-up, again an apples-to-apples comparison done at the same power levels, same motor, same ESC, same ambient temperature, same music playing, etc.

    First test is where we left off from the previous post. I am using the small copper plate with two edge-cooling tubes that I built. First 6 minute test is without water, i.e.: NO WATER FLOW. You can see how the ESC FET temperature increases aver the 6-minute test interval.

    Immediately at the end of the 6-minute test I turned by small pump that now sucked ice water from a very small 25-cent Tupperware container. I included a picture to show you the tiny size of this ice water container. The amount and degree of ESC FET cooling was amazing! Here are the results:

    First with NO WATER flow and the small plate heat sink with two copper edge tubes.
    Time 0 30 sec 1 min 2 min 3 min 4 min 5 min 6 min
    78.9F 81.9F 84.7F 90.1F 85.7F 100.6F 104.6F 109.1F

    NOW, let me immediately turn on the 12 VDC micro pump and supply cooling from the very small container of ice water.

    Time 0 30 sec 1 min 2 min 3 min 4 min 5 min 6 min

    109.1F 88.3F 79.6F 70.3F 66.2F 64.9F 64.9F 63.9F

    Is there any doubt now that cold water-cooling REALLY works?

    Are we now in the Golden Bubble of ESC cooling? Unfortunately NOT. During the testing I put a thermistor in the caps. The temperature of the Caps continued to rise throughout the two tests, even with water cooling (remember the water cooling is designed for FET cooling). The temp of the caps went from 77F at the start of the tests to 107F at the end of the second test. Something to think about.

    Motor is cool, but remember the test is only at 75 watts!
    Attached Files

    Comment

    • Drag Boat Bob
      Senior Member
      • Apr 2009
      • 304

      #107
      I have to say that I am simply blown away by all of this great testing. Thank You!

      I cannot wait to see if the results are as rewarding when running at 800+ watts.

      Comment

      • Jeff Wohlt
        Fast Electric Addict!
        • Jan 2008
        • 2716

        #108
        So was that with ice in the water or simply ice cold water? That is a bit smaller than nitro tank I am looking at a. It will have to maintain colder water going out than coming back in so I think the solid block of ice in the bottom of the tank can cool water fast simply because the ice has more surface cooling the water coming back in after that just ad ice cubes to it. Water will continue to cool as it sits. then we go out for a 5 minute run/heat. I am not looking for 52 degrees but I see this as much better than 85 degree water coming in.

        I bet going back your block things would be diffeent as well.

        Man, does this tell us how hot fets get? yow! And we wonder why they blow. Good work!

        It certainly leads me to spend a few bucks on a tank and pump and see what I can do with a chunk of copper. I can't see a downside to it. And for the big dog boat owners running 3 boards and 6 up to 12S and CC 240s it can only have positive effects.
        www.rcraceboat.com

        [email protected]

        Comment

        • questtek
          Senior Member
          • May 2009
          • 556

          #109
          Originally posted by Drag Boat Bob
          I have to say that I am simply blown away by all of this great testing. Thank You!

          I cannot wait to see if the results are as rewarding when running at 800+ watts.
          I did the baseline tests with my Mean Machine. This baseline configuration used my modified Mystery 100 amp controller with large copper heat sink and two tube cooling. The boat was set up with a 540 XL, x8 that has a 2084 KV. I used 4S, 40 C, 5000mAhr Lipos. The running gear was stock and not tweaked in the least bit. The prop was an old beat up plastic pile of crap I just found on my workshop floor...good enough for the baseline since I am not looking for performance, (at least not now). I used the latest V3 Eagle Tree data logger with GPS sensor, brushless RPM sensor and thermistors on the ESC FET's and on the motor just in front of the water jacket but in direct contact with the case.

          The lake water was about 77F. The initial EXC and motor temps were 76.9 and 77.4F

          For the 4 minute run which just had bursts of speed and only modestly fast turns...
          Max amps 82
          Max watts 1156
          High RPM 29,880
          Max GPS SPeed in MPH 29
          Final ESC temp 101.3
          Final Motor temp 102.9

          With this base line complete, I have now modified the Mean Machine and moved from open-loop lake cooling to closed loop ice and water cooling with a small 12 VDC pump. I attached pictures to show this set up.

          Testing is later so stay tuned............
          Attached Files
          Last edited by questtek; 10-24-2009, 12:35 PM.

          Comment

          • questtek
            Senior Member
            • May 2009
            • 556

            #110
            Originally posted by Jeff Wohlt
            So was that with ice in the water or simply ice cold water? That is a bit smaller than nitro tank I am looking at a. It will have to maintain colder water going out than coming back in so I think the solid block of ice in the bottom of the tank can cool water fast simply because the ice has more surface cooling the water coming back in after that just ad ice cubes to it. Water will continue to cool as it sits. then we go out for a 5 minute run/heat. I am not looking for 52 degrees but I see this as much better than 85 degree water coming in.

            I bet going back your block things would be diffeent as well.

            Man, does this tell us how hot fets get? yow! And we wonder why they blow. Good work!

            It certainly leads me to spend a few bucks on a tank and pump and see what I can do with a chunk of copper. I can't see a downside to it. And for the big dog boat owners running 3 boards and 6 up to 12S and CC 240s it can only have positive effects.
            It was small ice cubes and water. The crushed ice will actually give you the largest surface area for cooling. The smaller the pieces (within reason) the larger the suface area so my testing will be with crushed ice and water.

            I have a couple of carbon fiber Starships. They go like crazy BUT they keep blowing speed controllers. I put my Eagle Tree inside one (with a shoe horn) and found out the ESC temps were over 200 F.....and that was not even measured on the FET's but just a thermistor stuck inside the shrink wrap. These ESC's have a tiny aluminum plate single tube cooling system which is certainly not adequate. I will most more on this later. I may do a separate post on what I have learned from these increbile but tiny carbon fiber Starship FE's!
            Attached Files

            Comment

            • ray schrauwen
              Fast Electric Addict!
              • Apr 2007
              • 9472

              #111
              I have a Pelliter plate. Just the right size for a Mystery esc.

              Have been hesitant to try it since it will ice up fast but, it would work well if you used it on Ch.3 Aux. and switch it on for the last 1/2 of a race or run. You could have a servo switch the battery power on and off with an external power battery, (used for ballast).
              Nortavlag Bulc

              Comment

              • questtek
                Senior Member
                • May 2009
                • 556

                #112
                Originally posted by ray schrauwen
                I have a Pelliter plate. Just the right size for a Mystery esc.

                Have been hesitant to try it since it will ice up fast but, it would work well if you used it on Ch.3 Aux. and switch it on for the last 1/2 of a race or run. You could have a servo switch the battery power on and off with an external power battery, (used for ballast).
                Yes, I have done a bit of experimenting with them. You can see that there are some just about the perfect size. I mount the cold side to a copper plate that is in direct contact to the FET's. However the big problem is getting rid of all the heat on the hot side! There have been good suggestions involving keel cooling, ice baths, etc. but you really need to heat the heat out to make the cold side operate properly. It is worth some more thought and I hope that upi do some experiments with it.

                For those not familiar with Peltier Plates I have included a pic of the top and side view of one of the units I have been testing. Note the Deans connector since it takes lots of 12 VDC to make it work. The current supplies is directly proportional to the amount of cooling.........and, unfortunately the amount of heat on the other side you have ot get rid of!
                Attached Files

                Comment

                • domwilson
                  Moderator
                  • Apr 2007
                  • 4408

                  #113
                  Why not just put a copper plate between the Peltier and the FET's and cool the heated side with a conventional watercooling plate?
                  Government Moto:
                  "Why fix it? Blame someone else for breaking it."

                  Comment

                  • ray schrauwen
                    Fast Electric Addict!
                    • Apr 2007
                    • 9472

                    #114
                    Thats about it.

                    I figured if you use a 3rd ch. then you could turn it on and off at will. It produces so much cold that its not needed to be on all the time. A thermal sensor/switch would be nice. If the plate reaches a certain heat, it breaks the cct until it resets.
                    Nortavlag Bulc

                    Comment

                    • domwilson
                      Moderator
                      • Apr 2007
                      • 4408

                      #115
                      Originally posted by ray schrauwen
                      Thats about it.

                      I figured if you use a 3rd ch. then you could turn it on and off at will. It produces so much cold that its not needed to be on all the time. A thermal sensor/switch would be nice. If the plate reaches a certain heat, it breaks the cct until it resets.
                      One could use a circuit similar to this. Of course you will need a higher current PNP transistor (Q1) and a pot for R4.

                      Government Moto:
                      "Why fix it? Blame someone else for breaking it."

                      Comment

                      • questtek
                        Senior Member
                        • May 2009
                        • 556

                        #116
                        Originally posted by domwilson
                        Why not just put a copper plate between the Peltier and the FET's and cool the heated side with a conventional watercooling plate?
                        I would recommend that you try running some experiments with the Peltier device. My experience has shown that:

                        1. You need 3 to 6 amps of 12 vdc to power the device....not bad in itself.
                        2. When you plug the device into the power source the "cold side" gets very cold for about 1 or 2 seconds. After that it is pretty much at room temp.
                        3. The "hot side" gets very hot, almost too hot to touch.
                        4. To cool the "hot side" down you need a tremendous amount of cooling, MUCH more than just a copper plate and water bath. You need a heat sink that has an area of about 30-50 square inches along with a blower or CPU fan. Without this massive cooling the cold side remains pretty much at room temperature, or maybe just a few degrees below it.
                        5. I have the perfect copper finned cooler with 4 heat pipes and fan that cools the peltier device perfectly. Unfortunately the device is 6 inches in diameter and about 5 inches long! Not practical for any FE application I have.

                        This is only my personal experience using several different devices. I highly encourage you to try it yourself and see if you can come up with a more effective solution to make the Peltier device work. Large copper heat sinks mounted to the bottom on the boat would be one (impractical) idea.

                        Good luck and please keep me posted.

                        Comment

                        • domwilson
                          Moderator
                          • Apr 2007
                          • 4408

                          #117
                          Found an interesting article describing Peltier cooling. After reading it, and considering the relatively high wattage that we are dealing with as far as ESC's. I will put this one on the back burner till I do some further investigation and weighing in some various options. http://www.heatsink-guide.com/peltier.htm
                          I have used peltier's in the past with electronics and CPU cooling, but ESC's produce relatively more heat in relation to these lower power units.
                          Government Moto:
                          "Why fix it? Blame someone else for breaking it."

                          Comment

                          • questtek
                            Senior Member
                            • May 2009
                            • 556

                            #118
                            Yes, I have been down the Peltier Path........but consider the following comparing ESC cooling to CPU cooling:
                            1. In a desktop computer you have a 350 to 1000 watt power sup[ply giving you all the "free" 5 and 12 VDC you could possiblly want for powering th Peltier device) (In a FE using this required more LiPos or dedicated batteries, money, charging, space and weight)
                            2. The inside of a conventional computer case has LOTS of available room so it is easy to add more heat sinks, pipes, etc. If you do not have the room your can always add it to the exterior. (Not the case inside a FE)
                            3. Not sure what type of heat a CPU generates but it is pale compared to the 1000 watts put our by a 540 XL.......or the 6600 watts we are using with some of the coke-sized can brushless motors.

                            Maybe it would be like trying to add Peltier cooling to a laptop? After all, some of the larger Peltier devices are rated at 400 watts.......but take almost 30 amps of 12 VDC.

                            Comment

                            • domwilson
                              Moderator
                              • Apr 2007
                              • 4408

                              #119
                              Yep. After reading the article and seeing that under some conditions, the peltier can actually create more problems than it resolves. I will put this one off for awhile. Although I do have some ideas that may resolve some of the issues.
                              Government Moto:
                              "Why fix it? Blame someone else for breaking it."

                              Comment

                              • tharmer
                                Senior Member
                                • Mar 2009
                                • 342

                                #120
                                How about putting a small generator on the drive shaft. Something thing that will generate enough power to run the cooler.

                                -t

                                Comment

                                Working...