No announcement yet.

A project cooling a camera that was not intended to be.

  • Filter
  • Time
  • Show
Clear All
new posts

  • A project cooling a camera that was not intended to be.

    I did some horsetrading with a friend and swapped my Princeton Instruments 11MP color camera for a Lumenera LT16059HM which has USB3 with a KAI-16070 interline CCD from OnSemi. It seems to be an early production version without a mount, which works just fine for me. The biggest issue with this camera is it runs hot, like 50C hot. The hotter the sensor runs the more noise you have. So I went about to cool it, why? Because that camera is about $15k and there is no way I could afford a real astro camera with this kind of sensor.

    Lumenera lt16059hm by Jerry Biehler, on Flickr

    I got a Canon EF mount added to it with an adapter I made from PPS. It is a pretty good thermal insulator.

    I machined up some parts to do the water cooling with peltiers and got it running. First try I got the CCD down as low as 4C That was good enough for me to know this has a decent chance of working out.

    The 1e/S point for this sensor is 7C so if I can get below that I should be good.

    Here is the mess of testing out the cooling scheme.

  • #2

    I finished up the cooling system last week and everything seems to work pretty well, where the sensor was running at around 50c without cooling I can now cool it to -15c without too much issue and I am not running the pelters at full output, maybe 80%.

    To keep the camera dry I picked up a sealed enclosure off ebay, I put colder CPC style bulkheads on it for water, a sealed MS style connector for the TEC power and thermistor feedback, a sealed USB3 connection, and a fancy milspec multipin connector that provides power to the camera and monitor board as well as USB2 and interlock for the monitor. The monitor board is built around an old Teensy 2.0 I had lying around with a custom board I designed and had made that mounts it to a red 2x16 OLED display. It ties in to a i2c current sensor for the TEC, a combo pressure/humidity sensor, and also two panasonic EX-F71-PN leak sensors. These all have set points that shut down the cooling system and power in case of a water leak, humidity too high? OFF. Water leak, OFF. The board displays internal humidity, temp, TEC current, and dew point. It also displays any problem that shows up that will shut things down.

    I ended up making a new housing for the camera to allow air flow though it to ensure I got the internals as dry as possible. I also installed a little Hirose 6 pin connector to connect the thermistor to the sensor heat spreader for feedback to the TEC controller.

    You can see most of that stuff in this pic: Lumenera lt16059 Cooling Project by Jerry Biehler, on Flickr

    The monitor board:Lumenera lt16059 Cooling Project by Jerry Biehler, on Flickr

    Lumenera lt16059 Cooling Project by Jerry Biehler, on Flickr

    Once it was all together I popped in a clay based desiccant pack and purged the case with dry nitrogen. This dropped the humidity down to the useful accuracy of the sensor, which I did not find out till later. From the sensor tests I have found it is probably dry as a bone in there now.


    • #3

      The cooling unit is built in an old communications box from some military equipment. It holds a two fan PC cooling radiator, a water reservoir, brushes mag drive gear pump, interlock relay and switches, fan controllers, and a DC/DC converter that provides nice, clean 12v to the electronics no matter the input voltage, from 9 to 32v DC. Anderson connectors provide the power connection. There is a temp sensor in the return after the radiator that will increase the fan speed if the water warms up. So far I have seen nothing. A toggle switch and pushbutton allows me to set the pump speed and the fan temp profile. Two more CPC connections connect silicone hoses to the camera.

      It might be a little over kill.

      Inside the unit. It was a tight fit getting it together.

      Lumenera lt16059 Cooling Project by Jerry Biehler, on Flickr

      Finished cooler. Connector on upper left goes to the camera supplying 12v and interlock return. TEC controller is installed in the lid, all the cable can wrap up and stow inside the lid.

      Lumenera lt16059 Cooling Project by Jerry Biehler, on Flickr


      • #4
        That is very sophisticated, but in the end it is still air cooled, why not use a Peltier?


        • #5
          Originally posted by old mart View Post
          That is very sophisticated, but in the end it is still air cooled, why not use a Peltier?
          OP's post #1: "I machined up some parts to do the water cooling with peltiers and got it running."


          • #6
            Impressive, I’d be tinkering with a pid controller and getting approximately nowhere, a place I seem to visit a lot!, well fabricobbled (learning ave)


            • #7
              Beautiful job Jerry.


              • #8
                There are plenty of CPU air coolers which will handle 150 watts, how much heat does a camera make?


                • #9
                  Originally posted by old mart View Post
                  There are plenty of CPU air coolers which will handle 150 watts, how much heat does a camera make?
                  CCDs work by generating an electric current when exposed to photons. Unfortunately, as they get hot, they also generate current through the production of thermal electrons. This is known as "dark current" and can have a significan impact on image quality, especially in astrophotography where you are taking long exposure images in a low light environment. The obvious solution is to keep the sensor very cold. In some cases, this may even mean going down to cryogenic temperatures.

                  So, while an air cooler can handle large wattage, it cannot cool below ambient temperature. That's where the TEC comes into play. And you could put an air cooler on the hot side of the TEC, but that's not very efficient and could be quite cumbersome if Macona is planning on mounting the camera to a telescope. Using liquid as the thermal fluid allows him to better control the system and maintain better thermal isolation between the camera (cold) and the environment (hot).

                  Fun project, Macona! And nice score with the camera. You'll have to post some images once you start using it.


                  • #10
                    The camera uses about 48 watts. The amount of heat a TEC can move is a function of the temperature differential, I am currently running these off a 12v source and they are designed to run up to 15.2v @ 10a. So I am currently running them at about 2/3rd power. These are also double stack TECs to get a higher temperature differential and at full operating power they can each handle 58 watts, so figuring in the derating I am closer to 70 watts for both TECs combined and then you have to figure losses.

                    CPU coolers are not designed to handles nearly the temp differential. Also since I need to get way below freezing dew and frost is an issue which is why I went with water cooling, it is the easiest way to keep the camera dry while getting the heat out.

                    Boslab, I looked at making my own TEC controller but I found these on ebay and its just not worth rolling my own for how little I paid for it. Plus it even has a little tabview program to handle tuning.


                    • #11
                      I guess that watercooling also removes a potential source of vibration (fan rotation) from the camera compared with air cooled peltiers.

                      Also air cooled peltiers are seriously NOISY, which I'm guessing would totally screw with the ambience. 2 big fans on a radiator are going to be able to move the same (or more) amount of air but at a much lower noise level.

                      Cool build Macona, I really like it. I've used peltier cooled cameras before for fluorescent imaging, but none as large as that one.


                      • #12
                        I used

                        The pump is probably the noisiest part of the system since I used a positive displacement gear pump. They just have an inherent whine when they run, I can slow the pump down though, the brushless motors has a PWM in option that I found works with PC fan controllers.

                        Were you doing something with a fluorescence microscope? I have a photometrics quantix camera that has a smaller 6.3mp mono sensor in a vacuum chamber that was originally used for that kind of setup.


                        • #13
                          I might have missed it, but did you measure the input and output water temperature? That might give you some info on how much water flow you need for a given heat load.

                          Alot of my work is with fluorescent proteins, typically green and red, expressed in a small roundworm called C. elegans. Most of the time a plain jane uncooled monochrome camera works just fine (epifluorescence on a compound scope) but for long exposure times to capture low signals a cooled camera is better. I'm pretty sure the cameras on confocal laser microscopes are cooled too, but I never looked at one closely enough.


                          • #14
                            I did check the water tank temp which is right after the radiator and even with the fans at the lowest speed it never gets more than a couple degrees above ambient. Im running it off a big 12v supply which a built years ago, I will be curious to see if there is much of a difference running of battery power. I have a 147AH gel cel in the back of my car where the spare would go to run stuff like this.

                            Yeah, usually those are all cooled to some degree.

                            As you can see, the lens mount gets quite cold. I will probably have to add a ring of resistors around the perimeter to keep it a little warm,

                            Untitled by Jerry Biehler, on Flickr