Peltier cooling of whole 1100D DSLR

[Gina]

Having virtually given up on cooling the image sensor direct due to very difficult access, I'm now considering cooling the whole camera. Firstly though, I'm doing test runs to see how good the results might be.

I'm running long dark subs using APT. I've worked up to 20 minutes @ ISO 3200 with T=37C and the noise is now starting to show. The temperature started at 28C and has risen to 37C in a couple of hours. The earlier subs were captured with shorter times and lower ISO values and didn't show any appreciable noise but now that I know at what exposure the noise starts to show I can do a temperature run with that setting.

I thought of cooling the camera in the fridge or freezer but thought I might get condensation forming inside on critical components. With a purpose built cooler box I can seal the box and include some bags of silica gel.

I already have all the components to make a cooler for the whole camera.

1. A large enough diecast aluminium box (170x120x110mm).
   
2. Approx 20W Peltier TEC - 30mm square.
   
3. 80mm square finned heatsink and fan for the hot side.
   
4. 50mm square finned heatsink and fan for the cold side (to circulate cold air inside the box).
   

The box is heavier than I would like but just big enough (allowing for air circulation around the camera) and would make an easy to use box to check out the efficacy of this idea. I could replace it with a box made up from lighter gauge aluminium sheet and I might get away with smaller and lighter fan/heatsink combos.

[username]

Hi Gina,

I haven't got any experience of cooling AP equipment but at work often have to remove moisture from air.

One effective method I have used is to use anhydrous calcium chloride on the air inlet to the vessel I want to ensure has a dry environment; the vessel is protected from the calcium chloride by a cotton wool plug.

Things to note is that the calcium chloride will dissolve in collected water and it's difficult to re-use but it is more effective than silica gel (when the same surface area of dessicant is compared) and it's cheap.

If you want to stick with silica gel (as it can be recharged) maybe it would be better to make sure it is on the air inlet of a sealed system anyway ensuring that minimum moisture enters your sealed unit.

[Gina]

Thank you

[ncjunk]

Gina,

I saw a spanish article where they did this with an older Canon and they put silica Gel packets in to stop the condensation (obviously a requirement, I assume they would need to be regenerated every time or use the ones that change colour 5 pounds for 100g). I think they used a copper sheet and cut out a box shape rather than alluminium and then encased it in polystyrene.

I can look for the article if you want.

[ncjunk]

You`ve probably seen this but the design on this site is similar to the spanish one I saw.

Peltier Cooling of Modified Canon Digital Rebel XSi (450D) - Version III -by Gary Honis

http://www.astrofotografiadesdesoria.com/caja-fria-refrigeracion-camara-dlsr-con-celula-peltier/

[Gina]

You`ve probably seen this but the design on this site is similar to the spanish one I saw.

Peltier Cooling of Modified Canon Digital Rebel XSi (450D) - Version III -by Gary Honis

Caja fría: refrigeración cámara dlsr con célula peltier | Astrofotografía desde Soria

Thank you for the links Neil I have been reading Gary Honis's article and mine is similar. I'll post some photos shortly.

I've also found some others, such as this one :- Peltier DSLR Coolbox

[username]

Hey

Sorry Gina, I thought you would be drawing air over the cooling element hence my idea. Wrong end of the stick

Jim

[Gina]

Have had the darks schedule (20m subs at ISO 3200) running for nearly two hours now with an ambient temperature of around 7C (as opposed to the lounge room temperature of 22C). 5 dark subs completed with sensor temperature readings of 15, 24, 25, 26, 26C. It seems the sensor has probably stabilised at 19C above ambient. The camera is in the open air but sheltered from any breeze, so it's mainly cooling by convection. I imagine the cooling would be better with a forced draught and probably improved further by having a metal bolt in the tripod bush thermally connected to the cold box.

So I now have three or four darks at 26 and 35C and one at around 15C. The one finishing at 15C probably started at about 9 or 10C, the ambient temperature the camera started at, so this is not usable as a working dark.

[Gina]

Hey

Sorry Gina, I thought you would be drawing air over the cooling element hence my idea. Wrong end of the stick

Jim

Both fans blow air over their respective heatsinks. On the hot side the air is heated and takes the heat away and on the cold side the fins cool the air blowing over them and round the sides of the camera. I don't know if it would make any difference which way the air flows.

[username]

Yeah I understand now, earlier when I suggested about putting the desiccant over the air inlet valve to remove moisture from air before the cooling element I thought you were doing it that way (pumping air over the cool side of the peltier, like an air conditioning unit) but seeing the design that was linked to earlier I see that the fans are just for circulation and you're not drawing air from outside through a port.

I just shot off an idea before considering your actual design.

Jim

[Gina]

A couple of photos, not up to the usual standard I'm afraid.

[iwatkins]

Hi Gina,

Are you going to remove as much of the camera case as possible? That would speed up cooling the actual sensor.

I'm guessing there isn't a way to just remove the shutter box, mirror, chassis etc. leaving you with just a sensor and PCBs and power for the camera to still image? I've got a 10D with a busted shutter I was going to try this with.

Other concern would be weight, either causing sag on the focuser, or slipping focuser when aimed zenith direction.

Considered water cooling to remove heat from the system which would still allow a sealed system? Hanging a big Zalman or similar combined reservoir/pump/radiator from your pier/tripod with only a couple of hoses/transfer blocks on the camera/scope itself might help weight too.

Cheers

Ian

[ncjunk]

Removing the plastic body comes down to how you plan to operate. I've seen a lot of websites here in spain with this mod and the idea is to retain the camera as a fully functional dslr whilst providing the ability to cool it. I think sticking it on for a few hours beforr hand should result in it cooling enough.

If you dont care about being able to use it as a dslr then chop away!

I keep meaning to get an old 350/450 and do this...just for fun really. I'll try any tinkering once

Enviado desde mi GT-I9003 usando Tapatalk

[Gina]

Hi Gina,

Are you going to remove as much of the camera case as possible? That would speed up cooling the actual sensor.

Thanks for your reply Ian This is what I originally had in mind when I was going for the direct sensor cooling. The other 1100D worked alright without the back, which would give almost direct access to the image sensor assembly. It was when I went further that something went wrong and the camera stopped working.

I'm guessing there isn't a way to just remove the shutter box, mirror, chassis etc. leaving you with just a sensor and PCBs and power for the camera to still image? I've got a 10D with a busted shutter I was going to try this with.

One of the "HyperMods" showed this approach (almost) removing as much of the camera that wasn't wanted as possible. This is what I'd really like to do - cutting down the size of box required and hence the weight and the amount of cooling and heat removal.

Other concern would be weight, either causing sag on the focuser, or slipping focuser when aimed zenith direction.

Precisely - see above.

Considered water cooling to remove heat from the system which would still allow a sealed system? Hanging a big Zalman or similar combined reservoir/pump/radiator from your pier/tripod with only a couple of hoses/transfer blocks on the camera/scope itself might help weight too.

Yes, I considered water cooling but it's rather complicated and the hoses would be even more troublesome than cables.

I'd worked out that if I could remove the heat directly from the right place I could use a tiny Peltier TEC and probably just connect the hot side to the filter wheel casing to get rid of the heat. See my thread on Peltier cooling the image sensor.

[Gina]

Removing the plastic body comes down to how you plan to operate. I've seen a lot of websites here in spain with this mod and the idea is to retain the camera as a fully functional dslr whilst providing the ability to cool it. I think sticking it on for a few hours beforr hand should result in it cooling enough.

If you dont care about being able to use it as a dslr then chop away!

I keep meaning to get an old 350/450 and do this...just for fun really. I'll try any tinkering once

I love tinkering which is sometimes my undoing

I only want the Canon DSLR for astro imaging - I have another DSLR for other uses. As long as the camera works from the USB connection, the shutter works and live view works, nothing else matters. I will be making an 8v power supply for it and can switch that to provide the on/off function. I was hoping to remove a lot of unwanted heat generating and space taking stuff.

ATM I have one working, complete, unmodified 1100D body and another modded, stripped down many times and put back together again and NOT WORKING. Question is, do I have the courage to strip down my one working 1100D body? Do I have two and a half ton of spare funds?... NO. Can't even afford a new focuser ATM.

[Gina]

The way the power is controlled is complicated and combined with other controls in a multi-way fine conductor ribbon cable or two, or three... Fault-finding this beasty ain't easy. In the old days of thermionic valves with wires etc. it was relatively easy.

[Gina]

Back to the 1100D noise performance...

These are the results for dark subs of 20 minutes at ISO 3200 and temperatures of 25 and 35C. There is a significant difference IMO. These are the JPG images scaled in GIMP with no other processing and their respective histograms. A 10C lower temperature seems to produce about a fifth of the noise going by the histograms so well worthwhile. I shall be taking more measurements but I think I can say the cooling will be worth doing.

20m @ ISO 3200 at 25C

20m @ ISO 3200 at 35C

20m @ ISO 3200 at 25C

20m @ ISO 3200 at 35C

[Gina]

Planning to do more noise/temperature tests today. I'll increase the exposure time to see the noise better at lower temperatures. Upping the ISO to its limit of 6400 seems to bring up noise generally even at shorter exposure times so I'll stick with 3200.

As for the construction of the cooler, I think I'll probably remove the back of the camera and blow cold air directly at the back of the sensor assembly. That should help the cooling - I'm sure it must be possible to reduce the 19C differential between sensor and ambient.

I think I need to improve my method of detecting and measuring the noise to get a more scientific and accurate assessment of noise levels. I guess I can start by applying fixed amounts of stretching in GIMP but I think there must be software that will do this better without costing money.

[ncjunk]

Do you know if the 1100 has less noise than the 450d? Has there been an improvement in noise with each model or is it around the same?

Just wondering as i haven't really looked at dslrs since i bought my d50 and its been a great camera i've not thought of upgrading.

I take it for astro work the canon software is still the best?

Sorry for the questions i'm just thinking out loud.

Would the noise measurement be down to looking at the average pixel value over a certain area rather than a visual method?

Enviado desde mi GT-I9003 usando Tapatalk

[JamesF]

As far as I recall, when I looked into the noise issues I found a web page with samples from the 1100D, 350D and 450D showing the sensor noise for each under the same conditions. I think they came out in that order in terms of decreasing sensor noise, though I may have the 1100D and 350D around the wrong way.

James

[RogerTheDodger]

Gina, I've just had a look at some images I took on the 18th Feb - outside temp was about 2 deg C - over a series of about 30 x 80s exposures the temp reported by the camera went from 11C (this always seemed to be where the temp dropped back to at rest, but switched on) to 16C. Standard 1100d.

Might give you an idea on how much cooling in still air you may get.

[Gina]

Do you know if the 1100 has less noise than the 450d? Has there been an improvement in noise with each model or is it around the same?

Just wondering as i haven't really looked at dslrs since i bought my d50 and its been a great camera i've not thought of upgrading.

It seems the 1100D has a considerably lower noise than the 450D. Comparisons have been done by Gary Honis. The 1100D has the newer Digic 4 processor which gives 14 bit data depth as opposed to 12 of the Digic III.

I take it for astro work the canon software is still the best?

It's good but doesn't do anything like as much as APT (or BackyardEOS, I believe). APT has scheduling and you can set it up to take multiple subs at various camera settings.

Sorry for the questions i'm just thinking out loud.

No problem

Would the noise measurement be down to looking at the average pixel value over a certain area rather than a visual method?

That would probably be a better way of taking proper measurements. I haven't investigated that (as yet).

[Gina]

As far as I recall, when I looked into the noise issues I found a web page with samples from the 1100D, 350D and 450D showing the sensor noise for each under the same conditions. I think they came out in that order in terms of decreasing sensor noise, though I may have the 1100D and 350D around the wrong way.

James

I think the order is - form best to worst :- 1100D, 450D, 350D. The 1100D is certainly better than the other two.

[Gina]

Gina, I've just had a look at some images I took on the 18th Feb - outside temp was about 2 deg C - over a series of about 30 x 80s exposures the temp reported by the camera went from 11C (this always seemed to be where the temp dropped back to at rest, but switched on) to 16C. Standard 1100d.

Might give you an idea on how much cooling in still air you may get.

Sounds about right

[Gina]

Haven't managed much on this project today - other things have cropped up (the pleasures of being a housewife ).