Here is a short write up describing the process for the testing and observatory install of my CDK 14, in the hope somebody find this useful:
Step one was to assemble the system at home, inside, in a room with view on a distant power tower through a windows. Task included:
- established balance in RA and Dec, making sure I had adequate weights.
- prepare electrical and data wiring, with adequate slack for rotator. Design cable routing.
- Build a super light, simple, aluminum mount for the Pegasus Power Box V3. I considered 3D printing, but used an aluminum trim angle, cut and drilled precisely, then painted black.
- On the computer: driver install, various setups and profiles that can be done without stars.
- Preliminary dark libraries.
- Test cameras, focuser, rotator, fans, heaters, light panels.
- Design and 3D print in soft TPU a series of dust covers to protect the scope & camera during transport and install.
- Test the spreaders for shroud.
- Verify filter wheel configuration.
- Reach focus with both main and guide camera using the power tower as a target, with window open.
Two problems became apparent: the filter wheel carrousel had housing out of tolerance and extremely tight for the Chroma filters, the OAG did not secure the guide camera very well. The carrousel and the OAG will have to be swapped at some point.
Step two was to disassemble the scope, reassemble on a big tripod in the backyard, align the mount with the RAPAS, achieve focus on a starfield with the main camera, plate solve, sync, focus the guide camera, check rough collimation, and get a very simple NINA sequence to run. Then disassemble and pack the scope.
Step three. After that I was confident there was no major issues, and felt ready to install at the remote observatory. If the observatory had been located more than 3 hours away, I would have tested more carefully, but given the time necessary to set up to test versus the ease of the drive to the observatory, I choose to do the final tests and validation there.
So I packed everything and drove to location. It took about one afternoon to do a careful install. All the mechanical install was quick and smooth, as I had the gear organized in boxes, the necessary screws and tools prepared, and the procedure fresh in my mind. The biggest hurdle was the remote switch not playing well with the observatory gateway -blame the observatory, I guess. That delayed me a bit and I did not have time to do the meridian limits during daylight – so I used the Moana settings, which I guessed would work fine (a risky guess as it may cost a pier collision).
After install, as soon as dusk came, I aligned with the RAPAS, then drift aligned with phd2. It took me 45mn, as my initial alignment was, for some reason, very much off – much more than it should have been. In the end, after 3 iteration of drift align, I arrived where I needed to be: an excellent alignment, with a flat drift curve. Then I tried to to compute the alignment error, as there used to be an option for that in phd2, but either the option has been removed on the new version, or I could not find it. It was time however to move on. I synced the mount, a total pain with APCC, as if a sync is too far off, some misguided safety prevents the telescope to sync. APCC has gone a log way since its beginning, but ergonomics remains terrible. Then I fiddled with the OAG to get the guide camera exactly parfocal with the main camera. The limitations of the OAG became very apparent and frustrating. That part will be replaced soon, as already mentioned. Last I found a bright star, defocused widely, and collimated the secondary. That part was subtle and delicate, but not painful. I used the SCT method, putting one’s hand in front of the scope to find the screw to act on. For all its travel, the scope was very close to true, and I adjusted only one screw, maybe 1/16 or 1/20 of a turn. Felt more like applying slight torque to the screw than moving it. All that while perched on a high ladder at 1am in total darkness.
After that I went to bed, to avoid ruining the following day by being to tired to do anything productive.
During the week I used the scope remotely, computed a horizon file in Nina, fiddled with the Nina profile, tested focusing steps carefully, computed filters offsets, adjusted guiding parameters, etc…
It turned out the guide camera was not as parfocal as I thought, and I could not turn on the light panel.
On the following weekend I came back for the afternoon, fixed the light panel (one of the Anderson Pole was not clicked, resulting in a power breakage), took the time to do the meridian limits properly (a lengthy process), cleaned the wiring, and at night took the time to get the guide camera perfectly parfocal.
From there, I started remote imaging intensively. Star auto-select in phd2 was performing very poorly, and it took me a while to understand the 16 bit camera was set up as 8 bit for star amplitude, although the images saved by phd2 were 16 bits, which threw me off. The symptoms were: phd2 was consistently selecting extremely weak stars and guiding was poor. This was an easy fix once I understood the problem…
Then I had to go out of state for an extended period, leaving the following non-essential things the todo list:
- build the sky model (a pain as only the subscription based APCC, which I do not have, works with Nina, so it means going back to SGP).
- Get the rotator angle to adjust after platesolve. There is an intermittent bug in Nina 3, and “slew, solve and rotate” does not work consistently with templates. This is supposed to be resolved in the next built. Further, for esthetic purpose, on a rectangular sensor, I like my diffraction spikes aligned at sky angle 0, 45 or 90 degrees with the sensor, and assuming flats for each rotation angles, keep the flat library under control. Now working.
- Check Phd2 can guide at any angle without recalibration. Of course the OAG pick up prism mirror inverts the image, hence reverse the rotator angle seen by the guide versus the one seen by the main camera. I took advantage of the following full moon to verify things work as expected -they were, but it is always better to test intensively before letting the scope run on its own.
- Add an extra diffuser on the light panel. Now done.
- Perfect collimation & get an estimate of the system’s optical quality, using a wavefront tool (either Pixinsight or Starwave).
- Fiddle with the 2 tilt correctors, in front and behind rotator, although at f/6.5 there is very little to gain. If somebody has a well written procedure for that, I am interested!
- Reverse engineer (since Planewave keeps the info secrete) the scope temperature sensors, to read primary and secondary temperature with an Arduino.
While out of state I worked on automation, perfecting a set of python scripts to remove all the drudgery coming with dealing with big data.