Canon 1100D Cold Finger Peltier Cooling - WIP

[russellhq]

Finally decided to give this a go after months of poor weather driving me to distraction.

I've built the first prototype, the purpose being just to see if I could build a working cooler and it turns out I can!

First thing was the cold finger, this proved pretty tricky as I had to do it all by hand as I couldn't find my clamp. It's made from 18 gauge (1.2mm) copper sheet.

Next is the heatsink; I decided to go for one out a PC as theses are designed pretty much for this task: http://www.evercool.com.tw/products/HPL-815EP.html

It's not too heavy, and has good thermal performance. Here it is mounted in a project box (the box will change in the next design).

Behind the heatsink is the Peltier. It measures 30mm x 30mm and spec is: Max Watts 37,Max Temperature Differential 75 degrees C,Max Amps 3.9,Max Volts 15vd

Next, the inside of the camera (this is a broken camera I'm using to try out the mod on). I just added some foam to cut down on drafts:

Now the assembly process was quite tricky, as I had to fit the sensor onto the cold finger before attaching to the camera:

A bit more of the assembly process:

And finally a shot of the thermometer. This is show the temp difference between ambient and the cold finger beside the sensor:

Now that I've gotten things to work, the next task on the list is condensation prevention/protection and trial it in the working camera under load and see how things perform.

[libraryman]

Great work Russ & great photo's, well done

Ray

[russellhq]

Thanks Ray. It took a lot of work to get to here, and I've still got a long way to go yet. But I think this is the hardest part over now.

[russellhq]

To protect the electronics from condensation I was thinking about either a silicone grease or a conformal coating:

Grease

http://uk.farnell.com/dow-corning/2793695/compound-silicone-dc4-tube-100g/dp/537019

Coating

http://uk.farnell.com/electrolube/dca200ml/coating-conformal-aerosol-200ml/dp/298293

I've used the grease before for protecting PC electronics and it worked well. The only issue was it's quite messy. The conformal coating is not something I've had much experience with. I would welcome any discussion on the subject.

[Luis Campos]

I use slow curing Epoxy to encase all the electronics, it's easy to apply and ensures the thing to work everytime even covered with dew

Cheers,

[russellhq]

Thanks Luis. Have you any pictures of your boards with the epoxy on them? Also, what make did you use?

[Gina]

My latest incarnation of a cooled DSLR will include all the camera electronics within the dry chamber.  The camera will have three zones - the inner zone will be cooled by Peltier TEC band  will include the sensor and image processing PCB plus the shutter (since I can't put insulation between sensor and shutter ) and out as far as the optical window.  The cold zone will be thermally insulated as much as possible from the rest of the camera.  Then there's the sealed dry zone bounded mainly by a copper box that sits against the hot side of the TEC and the optical window.  This will contain the camera electronics, shutter etc.  I'm hoping the heat from the hot side will stop the optical window from steaming up.  The outer zone will contain the Arduino control system and filter wheel.  This outer casing will be on the front of the camera only, the back having a finned heatsink and fan designed foe PC CPU cooling.  This is described in greater detail in my 450D thread.

[russellhq]

Thanks Gina, I think I've read almost all your build threads, very inspirational!

I was considering beefing up the cover glass on the sensor with a thicker piece of glass to slow down condensation and I can remember someone recommending an online supplier but I cannot find any reference to it any more. I think it even might have been mentioned on one of your threads but I'm not certain.

[Gina]

Thanks Russel - glad you like my threads   I'm hoping to be getting back to the 450D with cooling and filter wheel project in a few days time. 

[russellhq]

Been doing a bit more research (can't work on the camera at the moment), and this is the sort of end result I would like: Tidy, light, efficient

http://astrophoto-sv.com/index.php?p=1_84

[russellhq]

I've been making slow progress with this, but am getting there slowly.

After reading elenhinan's thread where he stripped down the camera and sealed it in a box with desiccant to prevent dew, I decided this was the way to go for me after my struggles with dew.

http://stargazerslounge.com/topic/213990-another-canon-eos-1100d-peltier-cooling-thread-and-some-noise-tests/

So far I've been getting to grips with the arduino and have managed to get PWM, Humidity, Temperature, PID control and LCD all working. Here's a picure of the prototype board:

Here's some pics of the prototype case an camera (I'm using a broken camera and test case to try everything out on first)

Here's a close up of the stripped down camera:

And one of the cold finger:

Hopefully I'll get it assembled and running in the next week or so.

[russellhq]

Been doing some more testing but think I might have done something to the mosfet. It no longer responds to the pwm output from the arduino and seems not to be switching off the 12v supply to the TEC.

Is there any way to test a mosfet to check if it's not been damaged?

I may have wired the gate and sink the wrong way round but I don't know if that would have been enough to kill it?

[Gina]

I think it would   At least they're cheap enough from ebay   I keep a stock of them and maintain my stock from Hong Kong etc.  Such as these.

[russellhq]

Nice price. The one I have is about 10 times the cost of those! Think I'll have to buy some.

Here's the link to the one I'm using:

http://www.bitsbox.co.uk/index.php?main_page=product_info&cPath=140_151_154&products_id=1034

I've got a spare, so I'll try it out and hopefully it should work. Can't understand how I got the first one wired the wrong way round though :-(

[Gina]

Yours have a higher current rating than mine so that's probably why they cost more.

[elenhinan]

There are some drawbacks of using just a mosfet (without filter) with the PWM output from the arduino to drive the TEC.

If you run it directly from the mosfet to the TEC, you will at 50% duty cycle for a 2A TEC at 12V generate 12W of heat.

With a filtered output (using an inductor) you'll have 6W of heat, and the same efficency for heat transfer.

If you use analogwrite of the arduino, the PWM frequency is very low and you need a very high inductance to get a DC level out, and to keep EMI low. Using the built-in timer and PWM of the arduino, you'll get a much higher switching frequency.

That's why I used a DC/DC converter with a 300 khz PWM frequency and built-in filter instead. Bit more complicated to solder, however. Price-wise more or less the same.

Whether of not this will affect the performance a lot in the end is difficult to say. The closer you run to 100% duty cycle, the less of a difference it will make.

[Gina]

I use an inductor of a ferrite ring with wire threaded through about 5 or 6 times and a 100nF capacitor as RFI filter.  I've been using the standard analoguewrite for the PWM.  I got a delta T of -30C with the Farnell TEC with the most recent 1100D system.

[russellhq]

There are some drawbacks of using just a mosfet (without filter) with the PWM output from the arduino to drive the TEC.

If you run it directly from the mosfet to the TEC, you will at 50% duty cycle for a 2A TEC at 12V generate 12W of heat.

With a filtered output (using an inductor) you'll have 6W of heat, and the same efficency for heat transfer.

If you use analogwrite of the arduino, the PWM frequency is very low and you need a very high inductance to get a DC level out, and to keep EMI low. Using the built-in timer and PWM of the arduino, you'll get a much higher switching frequency.

That's why I used a DC/DC converter with a 300 khz PWM frequency and built-in filter instead. Bit more complicated to solder, however. Price-wise more or less the same.

Whether of not this will affect the performance a lot in the end is difficult to say. The closer you run to 100% duty cycle, the less of a difference it will make.

Thanks, but unfortunately I don't have a clue about electronics and would need a schematic with parts list before I could attempt anything like that The idea of halving the heat generated by the TEC though is a compelling one!

[Gina]

I would have thought that unless you go for high frequency switching you would need a large inductancebut I haven't done any of this sort of calculation for a long time.  I guess you could run off 12v say until the desired temperature and then switch to a lower voltage so that the on/off ratio was more efficient.  Whether the extra trouble is worth it I don't know

Generally speaking, if the simpler solution works I go for that.

[elenhinan]

I tried to simulate the results with a DC input and a PWM input to see the difference. Not sure if I've gotten the math right, I just did a quick change of my script based on intuition, so it might be off. And a theoretical approximation, so some real-world effects might not be taken into account. The difference will probably be larger for a higher current TEC, and will wary depending on the cooler.

Arduino analog outputs have a frequency of 490Hz or 980Hz depending on the board and pin number, so the RFI filter would need a very low cutoff frequency to be efficent enough to get a DC level out. High frequency components would be cut off as normal though, which are the ones which might cause problems in the first place where noise is concerned.

DC input:

PWM input:

Anyway, I tend to overcomplicate things to either perfection or total failure, so unless you're like me you should probably just ignore this. In the end, if it works well and cools down to the desired temperature you're ok.

[russellhq]

Making slow progress on this one. I've now insulated the box and made a new cold finger from 0.55mm thick copper sheet. This allowed my to isolate the cold finger from the back of the circuit board with a piece of 0.5mm plastic thus eliminating the chance of a short circuit.

Now I'm at the stage where I need to consider the cables that I want to go into the box. The first cable will be USB, this will connect to the camera. The second cable will carry power and data lines for the arduino. For this cable I think I need the following:

1: +12V (TEC power)

2: +8V (Camera power)

3: +5V (DHT22 temperature & humidity sensor)

4: Switched ground (TEC PWM control)

5: Common ground

6: Data1: Cold Finger Temp

7: Data2: Box temp & humidity

My plan would be to have this cable connect to a control box attached to the outside of the coldbox. The control box would only have a 12v in so I just need to run a 12V line and USB line up the scope. It would then output the 7 lines mentioned above.

Having little experience of this, I'm open to any improvement suggestions as that sounds like a big cable

[Gina]

You could drop the volts for camera and DHT22 in the box.  In fact the DHT22 meeds so little current that a resistor and zener diode are quite adequate.  I used a 4k7 resistor and 5.1v zener as I recall.  For the camera I used a small switch mode voltage regulator for Hong Kong via ebay.  These will provide up to 2A.  I've also used an LM317 analogue voltage regulator in the past - the TO220 version with the tab thermally coupled to the heat sink.

I also like to have the power MOSFET as close to the Peltier TEC as possible to reduce the risk of RFI problems. 

In my latest cooled DSLR incarnations I'm having an Arduino Nano beside the camera sealed box - within the outer casing.  I may also include a USB hub so that the camera unit only needs one USB connection.

[russellhq]

I bought some LM2596 boards to do my voltage conversion:

http://www.ebay.co.uk/itm/370975230952

I was hesitant to put the MOSFET and voltage converter inside my box as it would only add to the generated heat. But maybe it's an insignificant amount.

If I put the MOSFET and the LM2596 inside the box, I would then need to feed the following cables:

1: +12V (Common power)

2: Common ground

3: Data1: Cold Finger Temp

4: Data2: Box temp & humidity

5: Data3: MOSFET Gate PWM

 

I'm also thinking maybe I should have the TEC on it's own socket (phono) and have the data and voltage lines on another socket (e.g. 4 PIN DIN). I like your idea of the zener though. I'll need took look those up to see how to use them.

 

I'm running the box on test at the moment. Currently getting RH 14% and delta T of 35C I'm happy with that.

[russellhq]

To those that have stripped down the 1100D and removed parts like the viewfinder unit, flash capacitor and autofocus unit, is there anything to watch out for that will stop the camera working?

I've practised stripping down the non-working 1100D, but now it's getting about time to do the same to my working camera

[Gina]

Maybe it's time I produced an abbreviated thread of my 1100D strip down for astro, including removing all the unwanted parts and modifying the top section ribbon cable so that none of the top section is used other than part of the ribbon cable.