All Sky Camera Revisited

[Gina]

32 minutes ago, Chriske said:

Direction unpredictable, how can this be..?

I don't know.  All I know is that when I operate the focussing control from the control panel on my main PC the motor turns the right way mostly but sometimes the wrong way.  The direction is controlled by a data input to the A4988 driver module.  Logic "0" makes the motor go one way and logic "1" the other as I expect you know having used these or similar drivers in 3D printers.  The only way I can see that this could happen would be if the signal from the RPi was between the logic "0" level of <0.8v and the logic "1" of >2v.  My DMM shows voltages of 0.5v and 3.3v on the Direction pin depending on the set direction.  If the problem was a dry joint I would not expect the 3.3v state to give the problem as the o/c state of the A4988 is logic 1 I believe.

Edited September 29, 2019 by Gina

[Chriske]

Most probably loose connection somewhere..?

[Gina]

Yes, that's what I thought but can't find one.  The direction going o/c would not affect both directions.  Weird.  Might try resoldering the driver connections though.

[Chriske]

I'd replace the wiring completely, just to be on the save side

[Gina]

There's only one wire for the direction.  All of 10mm long.

[Gina]

As a result of reading another thread about binning colour images I've been thinking about my ASC.  Currently I'm using a colour camera (ASI178MC) but if focus and lens resolution are very good, each star is likely to fall on just one pixel so star colour will depend on which colour pixel the star lands on and my intention of using a colour camera to show star colour is thwarted!  Moreover, sensitivity is going to be much reduced compared with the same image sensor in mono form (ASI178MM which I also have).  I was considering binning to increase sensitivity but it all depends on how good focus and lens resolution is.

My thought is - should I change to the mono camera for nighttime imaging?  I should see more stars but the result will not be as "pretty".  I could use a separate ASC for daytime with reduced requirements.

[Chriske]

mono, I would...

[Gina]

As a change from the high power image processing in PI I've come back here and just looked through this thread.  I'm thinking I'm setting myself too much of a task trying to get high sensitivity out of a colour camera.  Also, whether it's the camera or the software/firmware, I sometimes get strange colours.  This, plus the reply above leads me to use the ASI178MM camera.  I've concluded that sensitivity is more important than colour for night sky.

The fact that I have put this project aside for so long indicates my dislike of the current setup.  I don't like the rectangular box - OK I know the look of the thing may be less important than the functionality but those are my feelings.  The water cooling is very efficient but I don't like those flexible tubes coming out the side - very untidy.  I have waterblocks with the pipes coming out the bottom which may be better.  The tubes are larger than necessary too but that's easily fixed.  OTOH water cooling uses a lot of parts.  If I use a mono camera it may not be necessary to get the camera as cold as with the colour one. 

ATM I'm using a single stage cooler - manufacturers of cooled astro cameras use double stage cooling and air-cool the hot side.  I could do the same if needed.  I'm using an f2 lens so there's quite a good amount of light available.  Cooled astro cameras as designed for use with telescopes where the aperture is more like f5 or more.  Countering that is that DSO imaging uses stacking which I'm not using in the ASC so need single subs to be good.  Yet again, I'm dealing with stars rather than glowing clouds of gas and also wideband imaging so capturing more light.

[Gina]

Overall I think I would prefer the RPi etc. where I can get at it easily - ie. inside the observatory.  This fits in with making the ASC as compact as possible and reducing the volume of air to keep dry to avoid condensation inside the camera and lens.  OTOH it also means I shall need waterproof connectors (or cables though grommets for sealing) for USB, power to Peltier TEC and focus motor plus possibly dew heater (to heat dome and stop dew on the outside).  If I were to use a dual stage cooler, the hot side could be well above dew point, maybe even above ambient temperature if I used air cooling and would eliminate the need for a dew heater.

Of course, with the RPi etc. in the ASC, only the 12v power wires need to pierce the casing.  Comms to/from the control computer would be WiFi.  Apart from the access problem this means the RPi may need some sort of cooling whereas in the observatory there can be plenty of air cooling.

Edited September 30, 2019 by Gina

[Gina]

Thinking about the enclosure/casing/container/etc...  3D printed seems to be porous unfortunately which is a big disappointment!  An empty paint tin of a litre or so might be a possibility - that would definitely be waterproof.  If I don't put the RPi inside I can use metal.

[AngryDonkey]

Hi Gina,

I printed my housing with 4mm wall thickness and 100% infill in PETG. Did extensive testing on waterproofness (i.e. drown in bucket for numerous days) and all was well 🙂

[Gina]

Thank you Mike - that's very interesting.  Maybe I'll have another go.

[Starlight 1]

As I do not have a 3d printer,  I use a stainless steel thermal flask for the out side and cut down to size.

[Gina]

Regarding the camera cooling, I've been using a 30mm square 19W Peltier TEC but the camera base will take a 40mm one quite easily and I think this size may work better.  I've ordered a TEC1-12703 module to try.  Probably not as efficient as the £30 ones from Farnell Element14 etc.  I may go for dual stage cooling with air cooling of the hot side.  I've decided water cooling is too complicated.  With air cooling, the hot side will be running much warmer than with water cooling which held the hot side just a degree or two above ambient.  This means I can use some of the heat from the warm side for dew heating.

[Gina]

Been thinking about getting the heat from the cooling hot side up to the dome for dew heating.  One possibility is an aluminium tube from the 100mm square passive cooler/heatsink up to the dome.  Another is an internal finned heatsink and small fan.  The aluminium tube idea is simpler so I've ordered a 4" OD x 100mm x 1.5mm thick tube from ebay with 1st class post delivery.  This will also increase the heat radiating area on the outside to improve cooling.  This tube is too long for use with the smaller lens but I think would be about right with the Fujinon.  There is no focussing facility on the Fujinon lens but I might try fixed focus using shims with "trial and error" manual focus.  A further simplification.  Alternatively, I could shorten the tube in my lathe.

[Gina]

Been looking through my earlier ASC threads.  The new Fujinon lens is definitely poor at the edge of field so I think I'll stick with the Arecont type.  Still feel sick at smashing the first Fujinon lens!   The front lens element has never shown up!  Might use the Fujinon for a daytime ASC if I bother.  The ASI178MC is now not doing anything but OTOH it might be useful for planetary imaging.

[Gina]

Another thing I'm looking at is the mounting.  Two possibilities here.  The aluminium pipe attached to the north side of the observatory or attached to the apex of the roll-off-roof.  The former favours a cylindrical ASC housing and the latter is suitable for a rectangular box shape.  The large passive cooler favours the box shape - it is not easily mounted atop a round mast. 

Actually, the passive cooler doesn't work too well with the cooling face horizontal - I think it was designed to have the face vertical so that air could easily rise between the vertical fins by convection.  I've just tried it with a Peltier TEC and the camera.  The camera temperature started at 25°C dropped to 6°C then rose to 12.5°C as the cooler heated up until it was almost too hot to touch!  Think it needs a fan.

Edited October 1, 2019 by Gina

[Gina]

I'm not happy using a fan outdoors where it can get drenched with rain.  Another idea would be to take pretty much dry air from the observatory and blow it out to the ASC but this is getting complicated again - maybe not quite as much as water cooling.

[Gina]

I have another idea.  Two enclosures (sort of).  An inner enclosure just containing the camera with a seal around the bottom of the lens and around the Peltier TEC(s) with the USB cable and TEC wires sealed.  This would have desiccant bags to dry the air in the camera.  Then below the TEC would be a finned heatsink.  Air would be blown through the fins and out at the sides up to the dome.  An outer casing would enclose the whole assembly up to the dome with vents just below the dome to let the warm air out. The bottom of the outer casing would be extended to go through the side of the observatory roof.  I'll draw a CAD model.

Edited October 1, 2019 by Gina

[Gina]

Here's a rough cross-section of the ASC.  The left hand tubular section goes into the observatory to draw in dry air.  The fan blows the air through the fins of the heatsink and out the sides, then up towards the dome and out of a ring of holes just below the dome.  A 3D printed housing surrounds the camera to provide thermal insulation and a sealed enclosure.  Some silica gel desiccant bags will go inside the camera enclosure to dry the air and prevent condensation in the camera.

Edited October 1, 2019 by Gina

[CedricTheBrave]

slightly on / off topic

anyone tried using one of these as an all sky camera? https://www.amazon.co.uk/LEFTEK-3072x1728-3-5-10-5mm-Distance-Weatherproof-white/dp/B0785J1H6V?ref_=fsclp_pl_dp_2

i guess it depends on what control of the camera you have?

[Gina]

Wouldn't cover the whole sky though.

[DaveL59]

9 minutes ago, CedricTheBrave said:

slightly on / off topic

anyone tried using one of these as an all sky camera? https://www.amazon.co.uk/LEFTEK-3072x1728-3-5-10-5mm-Distance-Weatherproof-white/dp/B0785J1H6V?ref_=fsclp_pl_dp_2

i guess it depends on what control of the camera you have?

Would depend what you'd like to see, daytime sky or night and stars. Likely it'll not be able to slow the shutter below 1/25s so would need video capture and stacked to get much of an image. 3.5mm won't give a wide fisheye and the lens likely will be relatively slow (f2.x or slower). Add that the dome is tinted and that'll wipe any chance of capturing stars other than the really bright ones at best.

Also and perhaps more importantly, CCTV domes aren't intended to be used dome upward, even if IP66 rated, so you'd probably need to adapt it to suit to ensure it resists water ingress. I rebuilt a pendulum dome type and so far that's working pretty well, tho the smaller dome similar to this is now showing condensation inside after I moved it recently, one more job when weather permits.

[CedricTheBrave]

3 minutes ago, DaveL59 said:

Would depend what you'd like to see, daytime sky or night and stars. Likely it'll not be able to slow the shutter below 1/25s so would need video capture and stacked to get much of an image. 3.5mm won't give a wide fisheye and the lens likely will be relatively slow (f2.x or slower). Add that the dome is tinted and that'll wipe any chance of capturing stars other than the really bright ones at best.

Also and perhaps more importantly, CCTV domes aren't intended to be used dome upward, even if IP66 rated, so you'd probably need to adapt it to suit to ensure it resists water ingress. I rebuilt a pendulum dome type and so far that's working pretty well, tho the smaller dome similar to this is now showing condensation inside after I moved it recently, one more job when weather permits.

interesting though!

[DaveL59]

yeah hence am sorta following that route too, I have a thread "El cheapo sky camera" if you can find it, shows what I put together.

I'd say at £96 that one's pretty expensive for what it is, given a cctv IP board camera is in the £24 region with a fisheye lens, would just need the housing. OK you can't pan/zoom but would you need to? Mine stay in a fixed aim unless I decide to adjust that manually and give 150-170 degree view from a 2.5mm or 2.1mm  lens. That lens likely would only go to 120 degree and the extra elements to enable the zoom range would probably not be good for astro, too much loss if nothing else.

If that one is something you think worth a go, factor in swapping out the dome for a clear one if you can find a suitable one, else you'll see almost nothing at night