Peltier cooling of 1100D sensor chip

[pete_l]

The gap behind the sensor chip would need to accommodate not only the copper cold finger but some insulation so that it doesn't short out the chip pins plus a few thou for some thermal grease. I think the gap looks like about 0.7mm.

Have you been able to identify the sources of heat inside the camera?

The reason I ask is that it might be easier to remove the heat from them (suspects: DIGIC processor and the A-D converter/signal conditioner) than to cool the CMOS sensor, given how tight it all is. I guess that during an exposure the sensor doesn't generate much heat as the amplifiers/readout circuitry would be off.

You mentioned that you don't see appreciable noise with a sensor temp. below 37° which is considerably higher than ambient. So if you can remove the excess heat that's generated by the camera's electronics, that might be easier than cooling the sensor itself. You wouldn't be able to to "proper" TEC below ambient, but it might be enough.

[Gina]

Have you been able to identify the sources of heat inside the camera?

No, but I haven't tried yet. I am not running the camera with any part of the case off at present.

The reason I ask is that it might be easier to remove the heat from them (suspects: DIGIC processor and the A-D converter/signal conditioner) than to cool the CMOS sensor, given how tight it all is. I guess that during an exposure the sensor doesn't generate much heat as the amplifiers/readout circuitry would be off.

Good thought that I'll have a good look at the dismantled camera and suss out likely suspects.

You mentioned that you don't see appreciable noise with a sensor temp. below 37° which is considerably higher than ambient. So if you can remove the excess heat that's generated by the camera's electronics, that might be easier than cooling the sensor itself. You wouldn't be able to to "proper" TEC below ambient, but it might be enough.

Yes, that's what I was thinking. If I could get the sensor running at 20C or less I reckon I could use half hour subs at ISO 3200 which should bring out some pretty faint fuzzies. Of course, I need to get some good guiding, but that's a different matter which I'm also planning to tackle.

If the heating culprits are chips on the sensor or main boards, cooling should be quite straightforward. Of course, at present the camera is well cocooned in it's case and the temperature is bound to rise inside the case. Must admit, whole camera cooling with all the case on does seem rather strange.

We very rarely get high temperatures in this country, particularly at night so if I could get the sensor running at around ambient temperature, that would be sufficient. So heat sources thermally connected to an aluminium box with fins on should show good results. Add a fan and even better

Thanks for you thoughts - very useful

[Gina]

Here's a photo with two largish chips which could be heat generators, pointed out. On the back of the main board there's a screened section with the screen soldered on. Difficult to see but I think there might be a biggish chip under there.

Tomorrow I might take the back off camera no.2 and run it. May be able to feel if these are generating a lot of heat. If so I'll think of cooling them. But I think a fan blowing cool air at the main and sensor boards would be quite beneficial.

[iwatkins]

Sounds like a good plan Gina.

Maybe run the camera up and monitor temps over the bigger chips. A remote IR thermometer would help, I.e. measure temps without touching the devices themselves.

Cheers

Ian

[Gina]

Sounds like a good plan Gina.

Maybe run the camera up and monitor temps over the bigger chips. A remote IR thermometer would help, I.e. measure temps without touching the devices themselves.

Cheers

Ian

Thanks Yes a remote IR thermometer would be ideal but I haven't got one. I'll have to see what's available.

[Gina]

Found one on Amazon for 18 quid - should do the job Could be a generally useful tool.

Pellor Non-Contact LCD IR Infrared Digital Thermometer Laser Point: Amazon.co.uk: DIY & Tools

Got it on Amazon Prime - due to arrive Thursday

[iwatkins]

What? You don't have a full on FLIR camera setup?

Tsk, Tsk.

Amazon item is exactly the one I have, accurate as well. Use it for all sorts.

Cheers

Ian

[Gina]

Great! Thank you

P.S. I see you live in Gloucestershire. I was born in Wotton-Under-Edge, near Stroud

[Horwig]

Good to see someone else brave the innards of a cannon dslr. As has been asked already, what's the difference in noise between cooled and un-cooled for an 1100d. I did a test last year on my 30d at 20c and in the freezer at 0c, the difference was stunning, but are more modern chips quieter, I don't know.

I started modding a 450d last spring, and stripped it of the body, mirror and pentaprism, using it solely by usb. It was great, and it was used it like this whilst working on a cold finger, but I managed to rip the usb off the un-protected board. End of camera I'm afraid, but a valuable lesson learnt.

I wish you good luck to your mod Gina, I'll be following this thread.

Huw

[iwatkins]

P.S. I see you live in Gloucestershire. I was born in Wotton-Under-Edge, near Stroud

Ahh, so you are almost a Stroudle

I'm just on the southern outskirts of Stroud itself, but originally from the Lakes.

Cheers

Ian

[Gina]

Ahh, so you are almost a Stroudle

I'm just on the southern outskirts of Stroud itself, but originally from the Lakes.

Cheers

Ian

Both from hill districts then and still in hilly districts - I love hilly country

[Gina]

Back to topic...

I'm thinking it might be possible to get sufficient cooling by replacing the LCD display screen and it's plastic window with a fan and cutting a slot in the side of the back cover at the other end. If practical, this would save the weight of an aluminium box and cooling heatsink and fans. Plus the high current requirement of a Peltier TEC.

[iwatkins]

I love hills too

Indeed, back on topic.

What you propose should work. Only two concerns:

1. Enough airflow to the right parts, e.g. the items that get hot. If there is room consider thermal pasting some small heatsinks to them, radiative cooling alone through these would help. But maybe best to wait until you can do a thermal survey on the PCBs.

2. Airflow around the sensor. If you did go for a powerful fan, where there is positive pressure within the camera, I would ensure that none of the airflow can get around to the front of the sensor as I would suggest dust bunnies would become a constant issue.

Cheers

Ian

[Gina]

What you propose should work. Only two concerns:

1. Enough airflow to the right parts, e.g. the items that get hot. If there is room consider thermal pasting some small heatsinks to them, radiative cooling alone through these would help. But maybe best to wait until you can do a thermal survey on the PCBs.

Yes, I'm not opening up my working camera until I get the IR thermometer tomorrow. I thought of tiny heatsinks on the hot chips.

2. Airflow around the sensor. If you did go for a powerful fan, where there is positive pressure within the camera, I would ensure that none of the airflow can get around to the front of the sensor as I would suggest dust bunnies would become a constant issue.

Good point. I think the filter frame blocks the gap between shutter assembly and sensor but I'm not sure if it's airtight. I think there could be other gaps too. This is going to take some careful looking at. I'm hoping a fairly low airflow will be enough. I wonder it it would be possible to use a filter on the fan.

Thanks for those cautionary notes

[iwatkins]

HEPA Filter over the inlet side of the fan should help. But would also have a filter of some sort on the exhaust side of the camera as well.

Cheers

Ian

[Gina]

Why a filter on the exhaust? That would only increase the internal pressure and increase the likelihood of airborne dust getting to the sensor.

[Gina]

Here are some photos of the mods I've made to the back cover. LCD and window removed and several holes in the side away from the fan. Thought holes would be stronger and easier than a slot.

[Gina]

Photo showing a 50mm fan sitting on the LCD opening. I think this is a bit big and I've ordered a 40mm fan. The idea is to replace the window with a piece of board with the fan attached. Or I could just make holes in the plastic window and put it back.

[RAC]

I did some testing with my 1000d with a peltier on the back of the sensor. The 1000d has a metal shield over the circuit board that holds the sensor, all i did was fill the gap between the shield/circuit board and circuit board/sensor with heatsink past then mount the peltier on the shield. This takes the circuit board down in temp too, i had a thermocouple on the back of the sensor also.

There was no change in noise from 0degC down to -12degC so i just didn't bother and left the camera in my cooler box as i have no condensation problems like this. The sensor will need some good heating on the front to not frost over with a peltier right on the back. My one just froze over when testing.

One good thing was the speed of how fast it cooled down, less than 5min and it was at -10degC.

I might pick up another camera at some stage and do the same thing but get some heating to the front and just use air cooling of the peltier and not water cooling as i did in the above test as there is no need to go really cold.

[Gina]

I did some testing with my 1000d with a peltier on the back of the sensor. The 1000d has a metal shield over the circuit board that holds the sensor, all i did was fill the gap between the shield/circuit board and circuit board/sensor with heatsink past then mount the peltier on the shield. This takes the circuit board down in temp too, i had a thermocouple on the back of the sensor also.

Interesting, I hadn't thought of doing that.

There was no change in noise from 0degC down to -12degC so i just didn't bother and left the camera in my cooler box as i have no condensation problems like this. The sensor will need some good heating on the front to not frost over with a peltier right on the back. My one just froze over when testing.

One good thing was the speed of how fast it cooled down, less than 5min and it was at -10degC.

I might pick up another camera at some stage and do the same thing but get some heating to the front and just use air cooling of the peltier and not water cooling as i did in the above test as there is no need to go really cold.

After testing for noise with my 1100D I came to the conclusion that 20C or less will give a good enough noise level. At that temperature I can capture half hour subs at ISO 3200 with only minimal noise. In general this will be above ambient. Only a heatwave in the summer will get hotter at night and in those conditions I would probably want to be indoors with the air conditioning on.

So again I have come to the conclusion that cold finger cooling is not needed to get low enough noise levels. Must say, I'm quite surprised how much better the 1100D is than the 1000D or 450D. I could well imaging Canon's new generation not needing any cooling at all.

[Gina]

Hmm... Looks like I could do with starting yet another new thread about cooling my 1100D without using a Peltier TEC

I have bought pieces of copper sheet, pieces of aluminium sheet and Peltier TECs but now thinking of no Peltier, no cold finger, and no metal box! Gained a lot of knowledge and experience though.

[Gina]

Been looking into where air+dust might go with a fan blowing air into the back of the camera. There are gaps everywhere including into the shutter mechanism. If dust gets in there it can get on the sensor. Looks like yet another rethink required

[Gina]

Maybe that idea of filling the space between sensor chip and its PCB and between PCB and screen and then cooling the screen is a good one after all. Just wants less cooling so that the sensor temperature doesn't fall too low, causing frost or even condensation. A copper or aluminium plate could replace the LCD window and be thermally connected to the hot side of a Peltier TEC. Then have a heatsink and fan on the outside.

Another thought is to seal the back of the camera with say polythene sheet so that positive air pressure doesn't force dust into the rest of the camera.

Anyway, tomorrow I hope to measure the temperature of the various chips and try to determine the main sources of heat. That should give a much better idea of where to go next.

[Gina]

Just waiting for IR thermometer to come now.

Been looking at the possibility of preventing air+dust getting into the works if there's a positive pressure in the camera case. Far too difficult. The biggest problem is all the ribbon cables, I'd forgotten them when I was looking at the camera with the PCB removed.

Whatever cooling method I use I think the camera will want sealing from dust ingress. So the heat will want conducting out with copper finger(s) rather than using air flow inside the camera. I could make up a copper (or aluminium) plate to replace the LCD window and then cool the outside of that. Heat could be conducted away from the hot chip(s) using thinnish copper sheet finger(s) and thermal paste then connecting to the plate in the back cover.

[kookoo_gr]

a very interesting project you got there Gina as all of your projects. A 2p of my opinion. By watching the fan you placed at the back of the camera you might consider about dampening any possible vibrations from the fan that will cause trails at your image. I mention it because i had to install an on/off switch at my first ccd camera (an atik 16 ic) since i had vibrations, i had no cooling problems with that camera since the whole casing of the camera acted as a huge heat sink