All Sky Camera Revisited

[Gina]

Been thinking about whether I can combine the dehumidifier with the camera cooling to improve power efficiency. 

Looking at temperatures :-

• Camera may be -20°C
• Cold dehumidifier plate/heatsink - a degree or two above freezing
• Hot side of dehumidifier maybe 20-30°C (or more)

Both temperature differences are easily obtained with Peltier TECs.  TECs could be stacked with an aluminium plate between with the dehumidifier air cooler mounted on the aluminium plate.

[Gina]

Here's a CAD drawing of aluminium plate, camera, 2 Pelter TECs and 2 heatsinks.  Air would be blown from left to right, being cooled below dew-point in the LH heatsink, condensing out the moisture and then being warmed by the RH heatsink before being directed up into the main ASC body and dome.  Heat is removed from the camera body by a 30x30mm TEC and transferred to the ali plate.  Below this is a 40x40mm TEC which grabs heat from the plate (and hence top TEC) and left heatsink and transfers it to the right heatsink.  Hence, this multifunction unit removes heat from the camera, moisture from the air and heats the dome to prevent dew on the outside.  Next to add the air ducting and fan.

Edited April 19, 2019 by Gina

[Gina]

I'm trying to work out the best place for the fan and what sort to use. 

Three possibilities of where to put the fan :-

1. Where the air comes in.  The air is damp.
2. Between the heatsinks.  Air is dry but cold.
3. After the hot heatsink.  Air is warm and dry.

The sort of fan is axial or radial.  Radial fans are more efficient but I'm not sure how to arrange it.

[Gina]

The aluminium plate will be 100mm x 40mm x 3mm rather than as shown above (thought I had a piece that size but couldn't find it) so the heatsinks could be separated by 20mm and the fan could go in the gap.  Trouble is the format of the duct through the heatsinks is about 40mm x 20mm.  I guess a 40mm axial fan might be possibly.

Another possibility might be to use a radial fan in the observatory and take the air through a pipe up to the ASC.  Standard 12v radial fans have an output port of 18mm x 13mm ie. an area of 234 sq mm.  That's equivalent to a circular shape of about 17mm diameter.  Since I will no longer be using water cooling, the air pipe could replace the two water pipes.

Another thing I'm not sure about is whether the 40mm square heatsink with 11 fins will give sufficient heat transfer - it might get very hot.  I have a bigger one with much thinner and many more fins but it's 60mm square complete with axial fan so much bigger.  I think that's probably overkill.

[Gina]

I've been examining an axial fan which wasn't working and it doesn't look like a bit of damp would cause it any problems.  This leads me to simply putting an axial fan in the bottom of the ASC casing.  I can duct the air up to the cold heatsink so that it flows through the fins then further duct it into the end of the hot heatsink.  The TEC1-12706 Peltier TECs are due to arrive tomorrow so then I may be able to start experimenting.  Unfortunately, the ali plate is on order and not due until around Thursday.  I may have a hunt through my stuff in the meantime and see if I can find anything.

[Gina]

Just stripped down the ASC and there was moisture in the sealed camera casing.  No free water but damp and all the silica gel beads were dark green - saturated.  My ability to produce a watertight seal is evidently non-existent!  This confirms that the air in the ASC needs dehumidifying continually so that a dry environment is maintained.  It needs to be dry enough that no damp condenses in the camera when it's cooled down well below freezing.

Edited April 19, 2019 by Gina

[Gina]

CAD drawings of the heatsinks etc. and with the air duct added (shown in transparent blue).  In the second drawing the air will be blown into the bottom left and emerge on the right where it will be diverted up to the dome.  Condensate will be removed through the hole in the bottom of the air duct.

[Gina]

Hmm...  Sticking out a long way to the left...  Actually the smaller TEC doesn't need to be centred on the larger one and could be a bit nearer the centre of the plate.  Also, the heatsinks could be closer together.  Anyway, that's all mainly cosmetic, I've got to see if it works yet!

[Dr_Ju_ju]

Instead of having everything bolted together as a monolithic block, have you considered using heat pipes (https://en.wikipedia.org/wiki/Heat_pipe) to conduct the heat to different areas to meet you needs ?

[Gina]

I have CPU coolers that use heat pipes but these are too big for this purpose.  I don't know if it's possible to buy heat pipes in other forms than computer/laptop coolers.

[Gina]

The Peltier TECs have arrived and I'm doing some testing.  I have a TEC attached to one of the heatsinks with thermal paste with a 50mm axial fan blowing air into it.  With 10v and 2.7A supplying the TEC I estimate the heatsink temperature as about 70-80°C and the cold side just above freezing.  The efficiency of these £4 TECs is certainly far lower than the £30 odd 30x30mm one I bought from Farnell Element14.  Might be worth paying more for something decent.  I did wonder about stacking the 12706 TECs but I don't think so.

Later... The cold side has drops of condensate at the bottom so this is dehumidifying.

Edited April 20, 2019 by Gina

[Gina]

I've found a piece of aluminium 125mm x 65mm x 3mm which I can try.  Much bigger than required but worth a try as it's a "bird in hand".

[Gina]

Testing with the large ali plate.  Running TEC at 12v 3.33A.  Hot side around 80°C cold side - not very yet.  I have the cold heatsink also on the plate and it's not cooled below the dew point.  There is a considerable thermal gradient along the plate.  I'm thinking a rearrangement is indicated.  Maybe put the cold heatsink on the other side of of the plate to the big TEC.  The small TEC hot side doesn't need as low a temperature as the cold heatsink.

[Gina]

This might work with appropriate air ducting.

[Gina]

An another idea would be to have the 60mm square CPU cooler on the big TEC and the two 40mm heatsinks in a row under the plate, giving more air cooling area.

[Gina]

Set up a test rig with the 60mm CPU cooler on the big TEC on the aluminium plate with a 40mm heatsink underneath.  The cooling on the plate and small heatsink is sufficient for condensation which can be seen in the photos.

[Gina]

I've seen what might be a bit of a problem with this - nothing must go above the level of the lens otherwise it would be in the FOV and that includes the air duct to the fan.  Of course, the air doesn't have to go into the top of the heatsink, it can go in one side and out the other with a "lid" to keep the air in between the fins.

[Gina]

From tests of the ASC with the ASI 178MC camera, it doesn't seem necessary to reduce camera temperature as far as -20°C and something like -5°C would seem quite adequate.  This leads to another possible arrangement for cooling and dehumidifying.  The camera temperature is not much below that needed for condensing the moisture in the air so an aluminium plate between TEC and camera casing extended out to the side with a heatsink on it may work as a dehumidifier.

Edited April 21, 2019 by Gina

[Gina]

To get enough cooling from a TEC-12706 it needs an efficient heatsink - the 40mm ones aren't sufficient so I shall use the 60mm x 65mm one.  For the condenser I'm not sure but the same size would seem logical (with the same air capacity) so I've ordered another of these CPU coolers.  Due to arrive tomorrow.  At least that gives me another option to try.

[Gina]

I have the larger heatsink on top of the big TEC on the plate and the two small heatsinks underneath.  Then, just for testing, I have a 50mm axial fan blowing through both heatsinks.  There is not enough cooling of the hot side of the TEC to get the cold side down below the dew point so it needs more air blown through it.

[Gina]

This is proving problematic and I'm wondering whether to separate the dehumidifier from the camera cooling.

The only part that actually needs dry air is the camera, the lens and dome can have warm air blown over them to prevent condensation.  Maybe I should go back to the idea of sealing the camera and including desiccant like the cooled astro cameras do.

[Gina]

I think I shall go for the simplest approach to this and try a sealed chamber for the camera again.  I can use a 3D printed plastic adapter ring for the lens to thermally insulate it from the cold camera body and hopefully this may let the lens warm up above dew point.  I've concluded that a standard axial fan should not fail from damp air as the motor part will be warmed by the current flowing through the motor coils and the air RH will probably be above dew point anyway. 

A standard CPU cooler on the hot side of the Peltier TEC may produce enough cooling whilst the warmed air can be fed up into the dome.  Clearly this will be less efficient than water cooling but may be sufficient.  It's worth a try.  If the 60x65mm CPU cooler is not enough I have a heat pipe type with larger fan and cooling fins.

Edited April 23, 2019 by Gina

[Gina]

The CPU cooler I ordered arrived today but the fins are further apart and fewer than I thought but deeper.  Whether this is as good, better or not as good as the one I already had I don't really know.  I guess totalling up the surface areas of the fins should tell me.  Base is the same size.

Here's a photo of them both - new one on left. 

 

[Gina]

The length of the fins is the same for both so the number x height will give the comparison.   LH 16x32 = 512 and RH 27x20 = 540 so the RH one seems slightly better.   I could test both.

[Gina]

Now testing with the new cooler replacing the waterblock in the ASC.  The camera temperature is slowly coming down.  The fan is pretty noisy run from the full 12v.  I could do with rigging up some digital thermometer chips for reading temperatures with this sort of testing.  With the aluminium adapter ring for the lens and no warm air blowing over the lens it has steamed up.  Camera image sensor has not yet got down to freezing - still a degree or so above.

Edited April 23, 2019 by Gina