Jump to content

Stargazers Lounge Uses Cookies

Like most websites, SGL uses cookies in order to deliver a secure, personalised service, to provide social media functions and to analyse our traffic. Continued use of SGL indicates your acceptance of our cookie policy.

sgl_2019_sp_banner.thumb.jpg.a0ff260c05b90dead5c594e9b4ee9fd0.jpg

Astro Projects

  • entries
    11
  • comments
    1,110
  • views
    3,285

Gina's Ultimate All Sky Camera

Gina

10,450 views

I'm hoping this is my final and hence "Ultimate" generation of all sky cameras.  Based on the ASI185MC CMOS astro camera and Fujinon fish-eye lens of 1.4mm focal length and f1.8.  Image capture is provided by a Raspberry Pi 3 in conjunction with INDI drivers.  This is used with KStars/Ekos client software running on a Linux Mint desktop indoors.  Communication is via Wi-Fi.  The astro camera is an uncooled version but I have added a Peltier TEC cooler.  This cools the camera down to something like -15°C for night sky imaging with longer exposures of around a minute.  Daytime imaging is also covered using the camera's minimum exposure and gain.  The colour camera differentiates between dark clouds and blue sky and also shows the colours of stars at night.

This Blog will describe the construction of the hardware and the special driver coding used to control dew heater, camera cooling and focussing.

 

  • Like 4


156 Comments


Recommended Comments



These photos shows the ASC mounted on my observatory roof giving a good all-round view.

1731593085_MountedonROR02.thumb.png.df80fd334150b8dbc0b289d9213b2e3a.png2010949010_MountedonROR05.thumb.png.5e068604e23c65335441b0d0c459a82e.png1639977645_MountedonROR04.thumb.png.e1d4bae06a347906538adc7a4fff384f.png

Share this comment


Link to comment

Now for some more detail.  These photos show the ASC before it was mounted on the observatory roof.  The casing was 3D printed in ASA filament which is unaffected by UV so fine for outdoor use in all conditions.  The dome is optical quality acrylic which is also supposed to alright outdoors though previous domes have shown yellowing after a year or so - we shall see.  The fish-eye lens can be seen inside the dome.  Below the main casing is a large fan-less cooler designed to cool by convection only.  This takes the heat away from the Peltier TEC hot side.  Behind the cooler and screwed to it is an aluminium plate which is used to mount the ASC on the wooden roof.  (The holes in the corners and a few other smaller holes, result from its use in a previous project.)  The different case colours result from using white ASA for the bottom part and natural ASA for the top.  It was a matter of what I had available.

1202077906_CasingTop05.thumb.png.43d9d5a6aab0a9e587e3eed0e19fe7dc.png582978019_CasingTop04.thumb.png.44a8849c86f9a5a0d8a764bef722a90b.png769048081_CasingTop03.thumb.png.8ac2ac838c2cc4e6036dcdfd45bfa4cc.png

Share this comment


Link to comment

Now for the inside.  Well, some of it.  This is mainly the focussing mechanism - gears and stepper motor, and the camera casing can also be seen.  This casing not only holds the camera but also provides thermal insulation to stop it getting heat from its surroundings.  It was printed in ABS with a criss-cross infill to provide air for better thermal insulation.

1091348893_CasingBottom02.thumb.png.bd3bd3ee5d072d0a0a9ee8277bac7352.png2016037482_CasingBottom04.thumb.png.e69f18f15326d6e02a93e284bee768a1.png900375463_CasingBottom3.thumb.png.704d612139034a0c6629758206925a3c.png

Share this comment


Link to comment
Gina

Posted (edited)

More details of the internal parts.  This shows a basic cross-sectional diagram of the camera, Peltier TEC and cooler.  The TEC cold side is on top to cool the camera body, which is internally thermally connected to the back of the image sensor providing the most efficient cooling.  The hot side is in contact with the "working" surface of the passive cooler.  Both sides were coated with thermally conducting grease before assembly.

367157543_ASCCooling01.png.ab1482629e014ae76a2cc10efb857ee6.png

The next two screenshots are from the design in SketchUp of the camera casing and bottom plate - both printed in ABS.  The camera casing was designed to closely fit the body of the camera, with a slot for the USB connection.  The bottom plate fits round the Peltier TEC with slots for the connecting wires.  This has two functions, provide the best possible thermal insulation between the warm cooler and cold camera body, and to attach the camera to the cooler.  The bottom plate was glued to the camera casing after the camera was placed in its casing.

1877811047_CameraInsulation01.png.96f20f505a029b8a4148a15cb2a5b164.png1116285708_ClampPlate01.png.94f1a27b7dd865af22cb6079a476dad4.png

The photo below shows the camera casing with attached bottom plate and spring loaded screws to hold the assembly of camera, TEC and cooler in good contact without crushing the TEC.  The passive cooler was lifted above the table to allow air circulation over the fins.  I was testing the cooling for long exposures at night - hence the camera cap.

1561889069_CameraOnCooler01.thumb.png.5ce493c8e82d51b9a50deb437c845513.png

Edited by Gina

Share this comment


Link to comment
Gina

Posted (edited)

The dew heater consists of resistors arranged around the lens and attached to the bottom part of the cover.  I don't have a photo of this.  The wires for the dew heater and Peltier TEC are taken out through a hole in the bottom together with the USB cable which connects to the camera.  These were fed in through a ventilation gap up under the barge boards and into the inside of the roof.  Here they connect to the control box with a 4 pin connector.  The control box is screwed to the inside of the roof.  Inside the control box, at the top is a buck converter to drop 13.8v down to 5.1v for the Raspberry Pi.  Plugged onto the GPIO pins of the RPi is a HAT (Hardware Attached on Top) which carries components to control focussing, camera cooler and dew heater - more of which later.  The left hand twin cable is the main power input from the observatory main supply.

1565576776_ControlBoxMounted03.thumb.png.e0fb2ae33f747f43d4292b2c6d8e0956.png2079751247_ControlBoxMounted01.thumb.png.0381109c78b194b6dc70a6248065ddc5.png1925601896_ControlBoxMounted021.thumb.png.8e94a38197bcdcf5ad6b06652a3fc8b7.png

Edited by Gina
  • Like 1

Share this comment


Link to comment

Excellent prints and design!

Puts my basic (and so far unused) dome to shame.

What camera are you using.

Which version of the ultimate camera is this? <ducks>

  • Thanks 1

Share this comment


Link to comment

ZWO ASI185MC as I think you'll find in the Blog description.

This is meant to be the first and last ultimate camera!  ?  I can't see any modification other than an expensive upgrade to the astro camera such as the ASI385MC-Cool which costs the best part of 700 quid!!  Maybe a future possibility but I have better things to spend my money on.  I'd have to arrange an air intake Hahaduct from inside the observatory - I wouldn't want to use a 700 squid camera out in the weather and wet.

Share this comment


Link to comment

No, don't go and hide - your contributions to SGL are much appreciated :)

  • Like 1

Share this comment


Link to comment
Gina

Posted (edited)

I'll have to see if I can find the model number.  It's 30mm square and about 5mm thick.  Rated at 15v and about 20W so relatively low power.  I bought it from Farnell element14 and I'm pretty sure it's this model :- MCPE-127-10-25 - Peltier Element, Thermoelectric Cooler, 19.6 W, 6.9 ohm, 2 A, 15.7 V, 75 °C

I tried various Chinese Peltier TECs but found they were very inefficient - a lot of power input for little actual cooling.  Cheap but not very cheerful!!

Edited by Gina

Share this comment


Link to comment

Before continuing with the construction details I'll take a pause to give some results.

So far I have found the camera cooling to be quite adequate and I doubt the ambient conditions would get much worse.  Last night the breeze had died and there was little if any air movement across the cooler meaning that all the cooling was by convection.  Cooling was significantly less than when there was an easterly breeze but still adequate to kill noise and hot pixels.  I would imagine the visual results will be a lot better when the weather is cooler and we get real astronomical darkness on a clear night.  As for daytime use - no problem even in full sunlight, though the sun does spread somewhat.

A couple of images.  Firstly taken last night just before midnight UTC (1am BST) - 60s exposure and camera gain of 150.  With 4m delay between captures the camera temperature dropped to -9°C and rose to -5.5°C by the end of the exposure.  HIGH level cooling with 13.8v on the Peltier TEC.  This is a screenshot of the KStars FITS Viewer with cropping and contrast enhancement in GIMP.  

1119897574_Screenshotfrom2018-07-0600-07-32.thumb.png.0504a6e636019b1ff808f62a6baf031a.pngµ

The second image is a daytime shot taken in the morning a few days ago with virtually no cloud - contrails instead.  32µs exposure and gain of 0.  LOW level cooling with 5v on the Peltier TEC.  This is a full scale image as written to disk, converted from FITS format to PNG in PixInsight with no other processing.  This also shows the coverage of the fish-eye lens image on the camera sensor.

Light_001a.thumb.png.2e52c7f6f31d960236e34775ad578486.png

  • Like 1

Share this comment


Link to comment
Gina

Posted (edited)

Some more details of the control circuitry.  As mentioned previously, power from the main observatory supply comes in via the twin core cable.  The full voltage is used for the dew heater, for the HIGH level of camera cooling and for the stepper motor, which is controlled by the daughter board where basic controls are converted into drive currents for the two coils of the stepper motor.  From the main power supply 5.1v is derived with a buck converter (seen in the left hand of the box in this photo) to supply the Raspberry Pi board  and for the LOW level camera cooling.  On the power side, the 13.8v has yellow wires and the 5.1v wires are red.  Power Gnd is wired in green.  The orange wire connects to the dew heater and the blue wire to the Peltier TEC for camera cooling.

On the data side, the control form the RPi (green PCB behind the HAT) is connected to the HAT (big red PCB) via a double row connector at the top.  On the HAT, the stepper motor is controlled using DIR (direction), STEP turns the motor shaft by one step, and SLEEP (which powers down the motor when not needed).  These connections are the white, yellow and blue covered wires to the stepper motor driver module (small daughter board at left of HAT).  Power for the stepper driver logic of +3.3v is derived from the RPi and coloured red.

The rest of the circuitry on the HAT controls the camera cooler (Peltier TEC) and dew heater.  Because the RPi is easily destroyed by input currents into it's data lines (GPIO), opto-couplers are used to separate the data lines of the RPi from the power circuitry that drives the dew hearer and TEC.  These opto-couplers can be seen as the small black rectangular components under the orange wire.  From these, resistor networks and power MOSFETs control the relatively high currents for the camera cooler TEC and dew heater.

682796583_WiringCircuitry01.thumb.png.f88475c06c218bdd70d1ac7fc037fed5.png

Edited by Gina
  • Like 1

Share this comment


Link to comment
Gina

Posted (edited)

The motor driver module is a standard A4988 type and the connections are pretty much standard.  Motor power is provided from the 13.8v main power and logic power from the RPi +3.3v line.  The motor side is isolated from the logic/data side and they have separate ground connections.  Very basic use of the A4988 module is shown below but in this application the speed control lines (MS1,2,3) are pulled to logic 1 to give 16x microstepping and the ENABLE and RESET lines are also held at logic 1.  The SLEEP input is used to control motor power.

1873858009_A498801.thumb.png.adbddd3830fb452e346013deb365abfd.png

Edited by Gina

Share this comment


Link to comment

Next to the camera cooling and dew heater control circuits.  In both, power Gnd is green and logic/data Gnd is black.  The only route from the data Gnd back to the incoming power Gnd is back through the RPi.  Isolated grounds avoid a ground loop and subsequent problems. 

In the camera control, the +5v (actually 5.1v) is connected to the Peltier TEC through a diode to provide the LOW level of cooling when the J438 P channel power MOSFET is off.  This occurs when GPIO13 is at logic 0 and the opto-coupler is off, leaving the photo-transistor not conducting.  In this state the J438 gate is held at source potential and the MOSFET is OFF.  When the GPIO13 line goes high the EL817 turns on, it's photo-transistor turns on and pulls the gate potential to Gnd ie. -13.8v with respect to the source and the J438 turns on providing the full 13.8v to the Peltier TEC.  The diode form the +5v rail is reverse biased.

497465444_CameraCoolerControlCircuit01.thumb.png.aa3948a9e98b0f45bd93a4a2ff9fcc5a.png

The dew heater control starts off similarly but with the EL817 opto-coupler controlled by a different GPIO line.  In this case the GPIO line is normally high - logic 1 - +3.3v and the current flowing into the LED switches the EL817 on, the collector connection is pulled down to 0v and the IRLZ44N channel power MOSFET is OFF.  This provides the OFF state of the dew heater.  When the GPIO line goes low - logic 0 - 0v, the opto-coupler turns off and the collector is no longer pulled to Gnd and +5v is applied to the MOSFET gate, turning it ON.  This provides the ON state of the dew heater.

2137012702_DewHeaterCircuit03.thumb.png.0bd151e6f05e76f353a85b0067e0daf4.png

Share this comment


Link to comment
Gina

Posted (edited)

Now to the software.  This is Ubuntu MATE operating system and INDI drivers to use remotely with KStars/Ekos indoors.  I modified a 3rd party INDI driver for the controls for focussing, camera cooler and dew heater.

The Raspberry Pi uses a micro SD card as system drive instead of a Hard Drive and the first job was to install Ubuntu MATE onto it.  For this an SD card reader was needed on the main computer and software to write the disk image to the card.  In Linux I used Etcher but if using Windows there's Win32DiskImagerEtcher has the advantage that it uncompresses the downloaded package as it writes the image to the card.  With Windows the package has to be uncompressed first.  Ubuntu MATE was downloaded from the Ubuntu MATE download page choosing the Raspberry Pi 3 version (takes two clicks) and downloading.

834390694_UbuntuMate01.thumb.png.2d34b6af90b5af34bbc104e3086d664f.png 

The disk image was then written to the micro SD card by choosing the downloaded image, checking the the right drive was chosen for the card and starting the software going.  The micro SD card was left in the main computer to edit a file before putting it in the Raspberry Pi.  This will be covered next.

Edited by Gina
  • Like 1

Share this comment


Link to comment

Great use of a Pi!

You know, they like these sorts of projects for the magazine.

You can use Etcher on windows too btw, I prefer it to Win32DiskImager too.

  • Thanks 1

Share this comment


Link to comment

The great storm of a few days ago did a lot of damage to my ASC.  Ripped the top off which landed the other side of the observatory upside down on the grass, seems mainly undamaged apart from a tiny mark on the dome.  The rest didn't fare so well.  The dew heater resistors and some of the wiring was strewn across the observatory roof and everything was full of water.  That is the worse part - lens and camera were soaked.  Eventually, after a couple of days of warming, the lens dried out and is now alright.  The camera did not survive! :eek:  Wet had go in and shorted something and destroyed it.  There were signs of wet corrosion on the image sensor board.  Cleaning this off didn't make any difference.

276341639_ASI185MMFault01.png.ff20ee57565833861a18d5b24712b315.png

Share this comment


Link to comment

I bet it's fixable. Those resistors are just 0R links and the caps should have survived. That corrosion needs cleaning off though. I bet the transistor is all that needs replacing at most.

Share this comment


Link to comment

Unfortunately U4 is not a transistor but an IC with 6 connections.  I guess only ZWO know what it is/was.

Share this comment


Link to comment

They might be willing to tell you if you ask.  I  take it whatever writing is on the top of the IC is no longer legible?  Worth trying to get a macro shot to see if it's any clearer, perhaps?

James

  • Thanks 1

Share this comment


Link to comment

Guest
Add a comment...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.