Another Canon EOS 1100D peltier-cooling thread, and some noise tests

[elenhinan]

Hi

This is my first post on this forum, after lurking for half a year or so.

November or so last year, I bought a EOS 1100D for astrophotography use, after buying my first telescope, a Skywatcher 200pds on a HEQ5 mount. Since I like to tinker, and I'm kind of a perfectionist, I didn't like the idea of imaging with a less that ideal EOS camera for the job. So after doing the filter mod I started thinking and reading about cooling. Thanks to all the posts Gina has made about peltier cooling, I had lots of ideas on how to proceed. I've uploaded some pictures and a bit of information about my build on my webpage, as well as some tests of the bias noise as well as the thermal noise. Instead of reposting it all here, I'll just post the links, and then if there's any questions or suggestions please post them here

Canon EOS 1100D peltier cooling build log

Canon EOS 1100D peltier noise tests

Note that as I live in Norway, were it basically doesn't get dark during the summer months, only the fall/winter/spring is suitable for AP. Since the temperature in the months suitable for AP is more or less guaranteed to be in the range -10 to +10 degree celcius after sunset, I do not have the same need for extreme cooling as people living in warmer climate, so my estimates and musings are a product of that

I'll probably update my webpage with more tests as soon as the sky clears and I have some free time on my hands.

Picture of the camera below. Fan is 60x60mm, and size of the box is more or less the same as the original camera.

[Gina]

Excellent work   Thank you for posting the link   And thank you for acknowledging my work on the subject   I found that a very interesting read - particularly liked that your theoretical results matched my experimental ones.  Your control system response graphs were interesting too.  Your perfectly damped solution beats my relatively crude damping.  My temperature curve shows my system as under damped.  I used to know a fair bit about feedback and damping for control systems but that was a very long time ago and memories of that are no doubt among the many "little grey cells" that have since died off.

[TeBilePas]

Thanks for this exellent article ! I collected alot of info here and there and I'm already in the process of cooling and reboxing a Canon 1000d and I found this article very well written and very complete. I have to say that I've not that often seen  a so well documented and thought way to the solution. Congrats for that and thanks for the feedback to the community !

You gave me the idea of controlling the cooling vs the current in the peltier because I hadn't thought about the copper radiator being maybe less powerfull than the peltier... 

Clear skies

[elenhinan]

Thank you both for the feedback

Sometimes if I start at e.g. 25 C and set the controller to 5 C, the peltier power will reach 100% before the desired temperature is reached. In this case the system will overshoot the global minimum of the cooling curve (which can be seen in the plots for peltier current vs temp) and end up not reaching the set temperature. Meanwhile the heatsink is running hot. The fix is of course simple: find the % where you get the maximum cooling efficiency of the system, and limit the controller to this level instead of 100%. I've updated the source code some weeks ago with this fix, but I haven't uploaded it to the arduino yet as it hasn't really been a big problem. For my system the best performance is somewhere between 70 and 80% I recon. Ideally I would map this by setting the peltier duty % to a constant, let the temperature settle, log and increase the duty% and repeat. I should then get a experimental curve as well. I'll see if I find the time.

The results from simulating peltier and heatsink are not that accurate, since the insulation coefficient is difficult to guess. But it still helps in choosing components, and helps giving a intuitive understanding of the way it performs and were you should focus your work if the performance of the system is worse than expected.

My heatsink suffers from being low profile I think, air being sucked in will only pass the outer edges of the fins before going through the fan, leaving little air flow in the center. But it seems to be good enough, and is not hot to the touch in normal operation.

[Gina]

In testing I found the Peltier TEC gave no more cooling above 90-95% and was actually less at 100%.  I have my maximum set at 90% I think.  It's been quite some time since I last used the DSLR cooling system and forget now.  I have mono CCD astro cameras now for my main imaging but I am still interested in seeing whet can be achieved with a DSLR.  I'm a "dyed in the wool" research and development nut and love trying way out things

[TeBilePas]

I can't wait to be at that stage to make the tests ! ;-) I had ordered an additional temp sensor to take the temp of the heatsink because I thought it could be usefull  to regulate the fan. 

Do you think I could also use it to monitor that the fan is effective at dissipating the heat of the heatsink ? I mean, I suppose that in the case you mention, the problem is that the heatsink cannot dissipate the heat generated by the hot side of the peltier. Could I suppose that it does mean that the fan is not effective enough, thus that the heatsink becomes too hot ? It would mean that I can monitor that the heatsink doesn't reach a limit temperature taht I can determine by experimenting... Am I right ?

[elenhinan]

I run my fan at 5V even though it's originally a 12V fan. This reduce noise, number of wires needed (temp and humidity sensors also run at 5V) and possible vibrations (not that I think those are a problem). If you use ds18b20 temp sensors, you could add as many as you want. For research that would be useful, for day-to-day use you probably only need one or two.

With a better heatsink you'll move the global minimum of the temperature curve to the right, and if the global minimum is at or past the maximum voltage/current for the peltier, then you can run it at 100% for maximum cooling. Remember that the heat generated by the peltier is proportional to the current squared, while the temperature difference has a linear relation to the current.

I've uploaded the peltier-efficency python script to my download page, so if you're into python you can try it. It's not well commented (none of my code is), but it should be well structured and you should easily see how to add new peltier elements or heatsinks. Basically, you only need the thermal conductivity of the heatsink (which isn't always specified) in units of W/C. Most will probably give it in C/W, so just invert it like the ones already described in the source.

[TeBilePas]

Great ! Thx, I'll check this one ! I think I've got the spec of the heatsink, I just have to find it back (the project was "in the fridge" for a few months because of some transformations in my house).

[russellhq]

Excellent post, thanks for sharing. This might help me kick-start my 1100D cooling mod that's taken a back seat recently:

http://stargazerslounge.com/topic/209720-canon-1100d-cold-finger-peltier-cooling-wip/

I really like the fact that below 0C, there's very little to be gained in reducing noise.

Can I ask how you intend to keep dew away from the front of the sensor? As it looks like you have the circuit boards protected but I don't see how have protected the exposed sensor.

[elenhinan]

Can I ask how you intend to keep dew away from the front of the sensor? As it looks like you have the circuit boards protected but I don't see how have protected the exposed sensor.

By keeping the sensor above the dew-point. Adding the dried out silica-gel pack to the camera-box gets the humidity down <10%, which keeps the dew point a lot lower than the sensor temperature. I haven't protected anything, really. If you check the dew point here, you'll see that it shouldn't be a problem with dew on the sensor as long as the humidity inside the box is low. As a precaution, the peltier-controller won't cool the sensor below the dew-point by design.

Right now I'm using it with a coma corrector, so unless that starts to dew up on the front I'm in the clear. I guess with an apocromat I would use a field-flattener, so I don't see any problems there either.

[elenhinan]

When deciding upon the solution I wanted, I just went for broke and dismantling the camera totally and putting it in a box. I lost the screen, manual control etc, but in the end I figured I wouldn't use these anyway. It was kind of hard to take the leap of totally modifying the camera, making it useless for anything but AP. But once the decision was made I think it made everything afterwards a lot easier. Having all the electronics in a sealed box makes life easier for controlling dew, light leakage, robustness etc. Also, with the mirror causing vignetting with my F5 newt, there was no downside to removing it when the prism and autofocus were already removed.

I guess the only downside might be no live view on the camera body when focusing. But I've added a stepper motor to the focuser, so the next project is to auto-focus from the laptop so it doesn't really matter. I have live view in BYEOS

[russellhq]

By keeping the sensor above the dew-point. Adding the dried out silica-gel pack to the camera-box gets the humidity down <10%, which keeps the dew point a lot lower than the sensor temperature. I haven't protected anything, really. If you check the dew point here, you'll see that it shouldn't be a problem with dew on the sensor as long as the humidity inside the box is low. As a precaution, the peltier-controller won't cool the sensor below the dew-point by design.

Right now I'm using it with a coma corrector, so unless that starts to dew up on the front I'm in the clear. I guess with an apocromat I would use a field-flattener, so I don't see any problems there either.

I'm still not sure how you are keeping the sensor away from the external atmosphere which should have a higher humidity than your camera-box. When you remove the coma corrector, does it let humidity into the camera-box?

[elenhinan]

I'm still not sure how you are keeping the sensor away from the external atmosphere which should have a higher humidity than your camera-box. When you remove the coma corrector, does it let humidity into the camera-box?

Yes. But I don't, so it doesn't

I have the option of adding a filter to the t-ring adapter as well, so inserting a clear filter there would isolate it from the external atmosphere if I want to use it without coma-corrector. If I were to do it again, I might have added a IR/UV filter to the t-ring, and then removed both original filters from the front of the sensor. Then I would've gotten a better seal.

[Gina]

Yes, I'm using an IR/UV filter as optical window on my camera.

[TeBilePas]

Hi, I'm just doing the same : a T clear filter is sealing the optical path of the camera box I made, so that I can disassemble the parts or change my camera (for example when doing some planetary captures) without having to re-dry my cooled box every time.