All Sky Camera Revisited

[Gina]

Just realised I have the answer all along - I have water cooling and the waterblock can cool the RPi as well as the hot side of the TEC.  At first I thought of an aluminium box for the RPi but then I realised that would stop the WiFi from working ?

I can put the RPi below the camera and cool air from the waterblock can descend onto the RPi by convection.  Now to redesign the enclosure...

Edited April 25, 2019 by Gina

[Gina]

The new design will use the 4" dome to give enough room for the dew heater resistors.  With the outer casing being sealed there is no need to seal the camera housing - just provide thermal insulation.

[Gina]

Camera housing for thermal insulation.  Shown in transparent blue.

Edited April 25, 2019 by Gina

[Gina]

Camera housing base - thermally insulates the camera from the waterblock.  Square hole takes the Peltier TEC.

Edited April 25, 2019 by Gina

[Gina]

Been watching a science programme on BBC4 which mentioned the Challenger Shuttle accident and the cause, which turned out to be freezing cold weather affecting rubber O rings.  Now I wonder if frost was the reason my ASC seals failed.  Bathroom sealant is used at room temperature and not designed for large temperature range, particularly cold.

[tonyowens_uk]

16 minutes ago, Gina said:

Been watching a science programme on BBC4 which mentioned the Challenger Shuttle accident and the cause, which turned out to be freezing cold weather affecting rubber O rings.  Now I wonder if frost was the reason my ASC seals failed.  Bathroom sealant is used at room temperature and not designed for large temperature range, particularly cold.

I don't think so Gina. Remember that low outgassing grades of silicone adhesive and potting (not RTV types admittedly) are extensively used in aerospace for mounting large refractive optics. Silicones are fine at low as well as high temperatures. If the adhesion of the sealant is not in question, I'd look carefully at permeability of the printed parts. Layer adhesion (a function of internal voiding) of PETG is purported very good - less so some other filaments, and its vapour barrier properties are well-known from food packaging uses in PET foil form.

1. Would it be practical for you to reverse-pressurise and waterbath/bubble test the enclosure you require to be hermetic (without the camera inside obviously)? I'm thinking about how to simply prove the seal integrity and identifying cracks/porosity if any

2. Crude and a cop-out perhaps - but would a 3mm plastic tube run into the enclosure and used either to positively pressurise it to a couple of inches of water head, OR a flow/return pair of such tubes solve the problem permanently? The air pump could be a small high head DC fan and the dessicant for the recirculating system could live in the vessel along with the fan somewhere indoors near the pylon...

3. If disassembly is not a requirement, there are alternatives to RTV silicone sealant and encapsulant which may work better for you. RS Components for a quick review, and Easycomposites UK for a more focused assessment of some of them.

Keep going!

Tony

[Gina]

Thanks for your reply.  No.1 is definitely possible and a good idea once I've built the new enclosure.  I thought of No.2 but it's more complicated and the desiccant would want replacing pretty often.  No.3 doesn't apply because disassembly is a requirement.

[Gina]

Having said that, I do wonder whether applying positive pressure to the inside to the enclosure might have some benefit - not circulating air but just a positive pressure through a small tube with some way of checking the pressure to confirm no leaks.  I'll give it some thought.

[Gina]

This could be the finished design of the camera base and support.  The focus stepper motor will be mounted on it too.  It will fit inside the main enclosure and sit on the base, providing room for the RPi and buck converter plus the waterblock and heatsink attached to the Peltier TEC under the camera.  There is a bracket to design and make for the latter.

[Gina]

The ASC assembly less main enclosure, camera, cooling parts and RPi etc.  Focus motor will be fitted under the camera support with the shaft projecting above and with a small 9t pinion on.  This will engage with the focus quadrant gear attached to the focus ring of the lens.  The gear ratio is 8:1.

[tonyowens_uk]

18 minutes ago, Gina said:

The ASC assembly less main enclosure, camera, cooling parts and RPi etc.  Focus motor will be fitted under the camera support with the shaft projecting above and with a small 9t pinion on.  This will engage with the focus quadrant gear attached to the focus ring of the lens.  The gear ratio is 8:1.

Nice and minimalist!
With a pinch clamp to fix the quadrant to the focus ring and a bit of axial length to that clamp to react the radial component of pinion contact pressure, through the quadrant, and into the focus ring. Alternatively, provide a lower bearing land for the lower skirt of the quadrant to ride on. Pinion/quadrant mesh clearance adjustment via slightly oversized camera fixing screws. Use of Nye Lubricants PTFE damping lubricant to kill any stick-slip in the gearing and pinion bearing.

I assume you are connecting the motor to the pinion shaft via a timing belt given the offset position of the shaft?

 

Tony Owens

[Gina]

The focus ring turns quite easily, particularly at a radius of about 40mm.  The focus quadrant should be a push fit on the focus ring as the hole is the same size as the earlier focussing setup which worked fine.  I agree that a longer contact with the focus ring might be better or I might bring the motor further up though it's about the same as before.  The stepper motor is the ubiquitous 28BYJ-48 with about 64:1 internal gearbox and the pinion directly drives the quadrant.

[tonyowens_uk]

56 minutes ago, Gina said:

The focus ring turns quite easily, particularly at a radius of about 40mm.  The focus quadrant should be a push fit on the focus ring as the hole is the same size as the earlier focussing setup which worked fine.  I agree that a longer contact with the focus ring might be better or I might bring the motor further up though it's about the same as before.  The stepper motor is the ubiquitous 28BYJ-48 with about 64:1 internal gearbox and the pinion directly drives the quadrant.

I hadn't seen that motor Gina - with the offset output shaft. It will fit perfectly as-is and no more transmission needed!
 

[Gina]

Heatsink clamp.  This holds a 40x40mm heatsink onto the waterblock (40mm x 40mm x 10mm copper).  Fits between the fins and screws into the M4 holes in the camera body.

[Gina]

Photo showing the camera support, waterblock and cooler for the RPi.

[Gina]

I have an idea for pressurising the enclosure which I'm going to try - a bellows for increasing pressure and a diaphragm for checking that pressure has been maintained.  The diaphragm could operate a micro switch connected to the RPi to show pressure GOOD/BAD and be in the bottom of the enclosure so no wires need to go through the sealed vessel wall.  I would have preferred a linear readout of pressure so that I could see if the pressure changed but that would involve much more complication.  I guess if I pressurise through a tube, I could have a pressure dial in the observatory which I could check every so often.

[Gina]

The obvious pressure gauge is a water manometer but that could introduce damp into the system.  Been doing some reading and it seems the commonest liquids used in U tube manometers are water, oil and mercury.  It only needs a very small increase in pressure above atmospheric, maybe an inch or two of water.  I guess the amount of water vapour given off by a surface of a couple of mm diameter would be negligible compared with the volume of air in the enclosure and of the silica gel packs used to desiccate the air.

 

Edited April 27, 2019 by Gina

[Gina]

A photo of the assembly so far with a partial print of the enclosure to confirm I can get the RPi and buck converter.

[Dr_Ju_ju]

I'd also add a layer of mylar ? or some other form of non-conductive film, between the water pipes and the electronics underneath, just in case there is any condensation that drips….

[Gina]

The water pipes and waterblock plus the heatsink will be at or slightly above ambient temperature and the enclosure air will be dried with silica gel sachets so this should not occur.  The enclosure will also have a positive pressure of a couple of inches of water which will ensure that nothing enters through the plastic casing or seals.  I usually give all electronics a spray of ACF50 anyway.  I take your point though.

[Gina]

Currently printing the main casing - about a 4 hour print with about an hour to go.  Clamping ring for dome designed and the base nearly finished.  Wants a pipe connection for the pressure and maybe under-pressure sensor plus holes for the power connections.  The two projections are for the water pipes, which should be a tight fit in the holes/tubes.

Edited April 27, 2019 by Gina

[tonyowens_uk]

Just a thought Gina if you are leaning towards overpressure monitoring.

Don't overlook washing machine pressure switches. Cheap, reliable, switching pressure generally in the range 3" - 7.5" H2O. Many are adjustable. Most are a bit large but certainly not all. All the diaphragm and microswt stuff taken care of. Best to go get one from a local appliance repair business.

 

Tony Owens

[Gina]

I can get one like your photo from Amazon - £6.37 with delivery Tuesday.  Might try that.

CDA Tecnik Washing Machine Pressure Switch. Genuine Part Number 49579500

 

[tonyowens_uk]

1 hour ago, Gina said:

I can get one like your photo from Amazon - £6.37 with delivery Tuesday.  Might try that.

CDA Tecnik Washing Machine Pressure Switch. Genuine Part Number 49579500

 

I envy you the fast cheap delivery. We dont see a lot of that in Ireland.

Two sets of contacts = DPDT? I think one screw controls contact closure pressure and the other controls opening pressure. It would be interesting to know what sort of leakrate you will be looking at.

Re use of water manometers. The vapour pressure of water is what it is, given a closed system pressure and temperature. A gradient can exist along the capillary, though it will equilibrate quickly enough. Use a dessicant in the camera, and the water column in the manometer will shorten. But as reading these is differential (LHS column height - RHS column height), that would not affect accuracy. But it will kill the dessicant needlessly. To maintain non-condensing conditions exist the camera volume during cooling periods probably best not to use a water manometer anywhere (except for initial switch calibration perhaps). One thought would be to pump air using a bicycle pump into a reservoir 2L PET water bottle with screwcap) a quarter-filled with dessicant, and to connected this to the camera via a 4mm soft PU pneumatic tube. The pressure switch would be teed off this line, near the bottle.  And when the switch resets (or the PET bottle gets a bit flaccid) just give it a bit more air. With some determined initial purging, that might actually work, depending on the leak rate!

Tony Owens

[Gina]

Yes, I was thinking I might use a bicycle pump to pressurise.  I hadn't thought of a separate air reservoir with desiccant, so thank you for that.  But why so big?  Even quarter filled with desiccant is quite a lot of desiccant.  I have some clear plastic 1L bottles.

I've ordered the pressure switch.  I wondered why there are so many connections - testing with DMM ohmmeter will be interesting!