I think this passive cooler might be suitable.  Ideally with a copper plate from the underside of the Peltier TEC to the outside and this mounted on top for maximum convection but I think it might be possible to put it underneath, directly onto the TEC and mount the ASC on an arm.

ARCTIC Alpine 11 Passive - Silent Intel CPU Cooler for all Intel CPU'S up to 47 Watts

If this cooler has the same size pad as the other (fan cooled) units I have, the Peltier TEC will fit nicely directly onto the heatsink.  This arrangement would provide the best thermal conductivity to the fins and hence maybe the best cooling even though the airflow isn't optimum.  But the only "proof of the pudding is in the eating".

This cooling method will mean a completely new case but that is no problem - have 3D printers, can make any number of new cases ?  The Raspberry Pi will need turning from vertical to horizontal to get it in the space.

The alternative to having the heatsink directly against the TEC would be to use a copper or aluminium strip to conduct the heat from the TEC to the heatsink.  What I don't know is whether the increased cooling provided by having the base of the heatsink/cooler vertical, giving improved air flow results in better cooling that would more that counteract the reduced thermal conductivity from the TEC.  I guess the only way to tell is by trying both arrangements.  Unless someone who has more experience of cooling than I have can tell me.

These diagrams show the two arrangements.

I've taken the cooling apart on the ASC and realised a small problem with the first arrangement above - how to attach the camera to the cooler.  The present arrangement is that the round aluminium block has holes drilled through to take Nylon screws that screw into the four mounting holes in the camera body and counter bored to take the heads of the screws.  Whether it would be possible to drill the new cooler remains to be seen.  Otherwise I shall have to make a clamp that goes over the camera body to the top of it.  That wouldn't really be a problem as I could easily print something.  The other arrangement can simply have holes in the copper or aluminium strip for the nylon screws.  At the other end I can drill and tap holes in the cooler.

I some respects, the direct connection of Peltier TEC to cooler seems simpler though it means other components of the ASC need rearranging.  Using a copper or aluminium strip bent at right-angles means the cooler can be arranged in it's most efficient orientation with free air flow in at bottom and out at the top.  It may also mean that the other ASC components can be arranged much as now.  My intuition tells me that a better air flow over the cooling fins should outweigh the increased thermal resistance from using an intervening strip between TEC and cooler.

Cooler arrived but it was the wrong one - had a fan - so I'm returning it and have ordered again.   Just hope they send the right one this time. 

It's very rare that Amazon send the wrong item but now I have to rethink what I'm doing today.  I was hoping to have the ASC working today. 

OK let's get over that - another cooler coming tomorrow.  More thoughts on the design.  

I think I shall change the mounting for the ASC from the pole to mounting on the south side of the observatory roll off roof at the apex.  This would tie in with mounting the Peltier TEC directly on the cooler which is easier overall I think.  A 3D printed casing would fit over the camera and clamp the Peltier TEC between camera and cooler.  This casing would also carry the focus motor and electronics boards.  It would be screwed to the cooler and also be extended to form the mounting onto the wooden barge boards on the observatory roof.    It may also extend upwards and take a push on dome assembly.

I've looked into using an aluminium strip to connect the Peltier TEC hot side to the cooler but I think this has problems.  I have an aluminium strip 40mm wide and 3mm thick which I think would be hard to bend and also a thin piece about 90mm wide but only 1mm thick.  I'm not sure the shin sheet would conduct the heat well enough.

So far so good but I don't think it's worth spending time designing printed parts until I've done some tests to see if any of this would work well enough.  So it seem this project has ground to a halt until I get the right cooler.

Hi Gina, I've got some 120mm X 50mm X 3mm aluminium angle kicking around, I could bung a bit in the post if it helps, save bending if it's just a right angle needed.

Dave

Thank you very much for the offer

Do you mean this shape?

Or this shape?

The right cooler has arrived and is certainly a really large passive heat exchanger.  It also has some very useful attachment screws with springs to hold the cooler on.  I'm going to try the cooler directly on the TEC to start with and run it indoors (warmer environment) with the lens covered while capturing a long exposure giving maximum camera heating internally.  I'll run the Peltier TEC off a second bench PSU and see how the cooling goes.

Like this, if you need it let me know the width and I'll cut a bit off.

Dave

Thank you Dave, I'll let you know.  That would be ideal for the more remote mounting but I think if it works well with the TEC straight on the cooler I'll probably go with that.

I have all the bits connected up and trying to connect to the RPi.  Initial attempt failed but I think the reason may be that I've installed a WiFi Access Point since I ran it last.  I also need to look back through stuff to find out how I did things ?

I think the way to go is to connect it with Ethernet cable and see what I've got the WiFi set to.  WiFi in the router is turned off and I'm using the AP for WiFi as it has a much greater range.  Anyway, I don't want too many things to get working at the same time so I'll stick with the Ethernet to test the cooling.  I don't need WiFi until I mount the ASC on the observatory and even then I could use a wired connection if I mount it on the side of the roof.

Working fine with cable connection.  Running camera with short exposures and the sensor temperature is around 27°C without cooling other than convection from the exposed camera casing.

With the camera covered with a booklet to try to keep the light out and exposure increased to one second the temperature rose to nearly 30°C and lots of noise was seen.  There is a bit of light leakage.  Now to add cooling.

Cooling on and camera temperature dropping nicely but it takes a while.

Now running cooling at 9v and 1A and camera temperature is 9°C.  Ice is beginning to form on the camera casing (currently open to the atmosphere - no thermal insulation).  ATM I can't get the camera light tight and a 1s exposure is showing light.  The noise is much less.  I think I need to wait until dark to stop light getting into the camera unless I can find something to plug the camera with.

I have a feeling that adding extra aluminium into the thermal path is going to be counter productive because there is a noticeable difference in temperature between the aluminium of the cooler near the Peltier TEC and further out and yet more down the fins.  The aluminium is about 6mm thick on the top plate so I would expect it to conduct well - far better than several inches of 3mm aluminium.

Temperature difference between cold and hot sides of the TEC have meant I have been able to increase the voltage to 10v while still maintaining 1A of current.  Think I'll swap my PSUs over so that I can run the TEC above 1A.  The rest of the unit is drawing less than 0.5A.

Been out for the evening but now back and have turned off the light in my living room (alias workshop!).  Now capturing with 10s exposures and getting plenty of noise.  But have cooling on at 12.4v and 1.2A and the sensor temperature is coming down.  I think I need decent thermal insulation around the camera as it's drawing water out of the air and freezing it and the latent heat is stopping the camera cooling down more.

I've put the zoom lens with aperture control on the camera and reduced the aperture to almost closed.  There is light on the ceiling from the computer monitor, my moon clock and several LED displays.  I'm taking 30s exposures with a gain of 144. Sensor temperature is about 14°C with cooling compared with 30°C without cooling and lots of noise.  As can be seen, the noise is pretty well gone, with just a few hot pixels.  The focus is out because the spacing was set for the bigger lens.  (C v CS mount lenses.)

I don't think there's much more I can do tonight - I need to design and print a housing for the camera to hold it against the TEC and to insulate it both from the environment and particularly from the hot cooler face which is just 4mm away from the bottom of the camera casing.  I estimate that the cooler surface is around 50-60°C.  ATM only the weight of camera and lens is holding the camera against the Peltier TEC, probably sufficient but if the parts are held together I could turn the assembly on its side and see how much difference proper orientation makes to the efficiency of the cooler.

Been working on the clamping system design.  I was originally thinking of fully insulating the cooler from the rest of the ASC but heat from the cooler could help with dew heating.  The most important thing is to insulate the camera, it helps to keep everything else warm.

I'm just concentrating on the camera insulation and clamping plate ATM.  I'll sort out the main casing and supports for the PCBs and focus motor later.

Here are designs for the camera insulation and clamp plate.  I'll print them in ABS so that they can be glued together.

Camera thermal insulation and clamping plate completed and fitted.  Now testing again.  It will be some time before I can post results because this thermal process is extremely slow.

The camera temperature seems to have settled down to about 11°C with continuous 10s exposures so I've turned the rig to give what should be better cooling.

Initial camera temperature 12°C (crept up a bit from previous reading) and now leaving the rig running with repetitive 10s exposures and everything the same except for the orientation of the cooler.