wimvb

Not really. The very best have high end mounts that need perfect balance. And as @david_taurus83 already wrote, most mounts aren’t that sensitive to balance. In many a mount, the problem is solved with grease. But you may experience a difference in guiding. It’s a case of ”don’t sweat the small stuff”. If you do an automated meridian flip, you can’t even rebalance the mount. I know I don’t. If I find that for some reason the guiding is unacceptable on one side of the meridian, I image on the other side only and figure out what went wrong the day after. Or I collect RGB on the side with poorer guiding, and luminance on the good side.

Use the PHD guiding assistant first to see if you have a problem. "If it ain't broke, don't try to fix it."

That's actually a very valid question. East heavy implies that you need to adjust the balance of the scope/counter weights after doing a flip. When imaging towards the East, the scope is on the West side of the pier, and the counter weights on the East side. The counter weights should be slightly further out than perfect balance. After the meridian flip, you are imaging towards the West. The scope is on the East side of the pier and the counter weights on the West side. Now the scope side needs to be heavier, so the counter weights need to be shifted up the counter weight bar. In practice, you try to get a good balance with as little backlash as possible. You then accept a small difference in guiding/tracking accuracy between East and West. Or you can just do your imaging on one side. If you decide to image just East of the meridian, you switch to another target when the first one passes the meridian.

You can just see one RA allen screw near the bottom here (not indicated, but next to the cable).

RA is always engaged and moving in the same direction. That's why backlash is less important in RA. But it helps to have a slightly east heavy imbalance to help keep the gear cogs engaged. There's no need to take the scope off. But I wouldn't put a heavy load on it either. To use my method with Guiding Assistant, you need at least a guide scope with camera on the mount.

They are the top screws in your own image, and indicated in @david_taurus83 first image

And here was I thinking that using x-ray diffraction to measure the sag of silicon crystals under their own weight was advanced. Did that in the early nineties. Now, to get back on the IKEA track. I thought of measuring the sensor tilt by using 0 degree incidence and a laser collimator. That didn't work. But I did find a way to collimate a laser collimator with a 1.25" drain pipe. How's that for low tech? 😋

That's definitely past the IKEA level shown here.

my initial guess would be that the front plate, and the screen need to be abolutely parallel. With the reflection metod that is not a necessary requirement, since one is only interested in getting the reflection stationary during rotation along the camera axis.

In this case, it's the reflection of the main beam that is used for alignment. But the diffraction spots could be used as well, just a little more cumbersome.

Yes. You can apply calibration at any declination, but PHD doesn't like it to be too high. There is a difference between DEC and Altitude. DEC is a number of degrees north of the celestial equator. Altitude is the number or degrees above the horizon. For me at 60 degrees latitude, the celestial equator is at an altitude of 30 degrees above the horizon (due south). This is also DEC 0 degrees. You can safely do your calibration some 10 - 15 degrees north of the celestial equator (for me this is an altitude of 40 - 45 degrees, when looking South). This may even be better, because if you calibrate too low (above your local Southern horizon), you well get a larger effect of the atmosphere and more seeing related problems. Below your horizon sounds strange. in the Northern hemisphere, the altitude (degrees above the horizon) is always (90 - your Latitude) + DEC As I wrote before: for me an object is at altitude 30 + DEC, because I live at a Latitude of 60 degrees. If you live at a Latitude of 50 degrees, the altitude of an object is 40 + DEC. Normally you calibrate at or close to DEC = 0 (the celestial equator), which is at a local altitude of 90 - your Latitude. Objects with a Declination (DEC) that is higher than your Altitude, are always North. Zenith (pointing straight up) has a declination which is equal to your local Latitude. According to your Avatar information, you are located in or near London, which is at a latitude of 51 - 52 degrees. So, you would do your PHD calibration at an altitude of 38 - 39 degrees above the horizon. If you have bad local atmospheric conditions, you can calibrate at 45 - 50 degrees altitude, which is at a declination of 6 - 12 degrees.

Guiding Assistant is under the Tools menu in PHD2. You should run it at least once when you start a new profile,and every time you make a mechanical change in your setup. It will show you polar alignment error, settings for minimum movement (minmo) and measure backlash. MinMo is a setting that will depend on mount mechanics as well as seeing conditions.

Correct. These screws push the RA and DEC blocks where the worm gear is housed. These lock the RA and DEC block in place. They are much larger and you see them clearly (4 per axis).

That's a bit like insurance; you pay dear money for something you may never need. But better to not need it than to find out you have tilt.

This is what I wrote on the Facebook group for EQ6 owners: I managed to get the backlash down from 4000+ ms to about 1300 ms. As it may be of interest to others, here's a description of what I did: 1. Point the telescope just east of the meridian and just north of the celestial equator. 2. Calibrate guiding, guide for about a minute just to stabelise the tracking, apply guiding assistant and measured backlash. 3. Park the scope (pointing north) and loosen the South grub screw. Tighten the North grubscrew about 1/32 of a turn. Repeat steps 1 - 3. When I tried to get backlash below 1000 ms, the gears started to bind. Just backing off a tiny bit, didn't really improve things. I also found that it's important to tighten everything in the right order. Tightening the South grub screw before tightening the allen screws that keep the DEC in place, changes the gear meshing just a tiny bit. Enough to change backlash. So I always tightened the allen screws first, before I tightened the South grub screw. I think there's also some hysteresis in the mechanics, and when the gears start to bind, it may be easier to loosen everything and start over, rather than trying to go back in small adjustments. And some time later: A final update: Tonight is a clear night after several days of rain, but with poor seeing. I opened my observatory to continue tuning backlash. Last time I tuned backlash too far and the mount started to bind. So I backed off a bit before closing up, expecting to have to repeat the entire exercise. But, lo and behold, without any adjustments, guiding assistant measured 248 ms (!). I let it guide for a while and ran guiding assistant again: 314 ms. I'm very happy with that. I will check backlash again when the air is more stable, but for now I'm very happy with this result. Note: the 248 ms and 314 ms backlash was measured at a guiding rate of 0.8 x sidereal. when I measured backlash a few days after I wrote this post, I did so at my normal guiding rate of 0.5 x sidereal, and it was 700 ms. This image is from during the first night of adjustments This image is after the last adjustments Btw, if you can feel a wobble in DEC or RA, your backlash is way off. At 4000 ms backlash, I couldn't feel it when I tried to move the axis by hand. Edit: One more thing you can do: Guide for at least half an hour. Don't change any settings. After that, download the guide log from PHD and look at it in PHD log viewer. Do a frequency analysis of the RA errors. This is the kind of graph I mean: The peak at 480 s is the main period of the worm gear. The 120 s peak is from the RA pulley on the worm side. The very small peak at 10 s is from the RA motor pulley. If you have a high peak here, it will be very difficult to guide out, and you will need to adjust the pulley. All the other peaks will be guided out without any problem. I had this problem two years ago and had to tighten the RA belt. That solved it.

Ok, silly suggestion maybe, since you seem to have tested everything: have you plugged the usb cable into another usb port on the RPi (preferrably with a clean setup)? Just to rule out hardware problems.

Yes it is, but like I wrote, with optimised download time. That should work, I think.

What is the image download time of an atik ccd? The lodestar is a dedicated guide camera, so I would imagine that the developers designed it to have fast download. But I'm not convinced that the Atik will. Ccd cameras generally have slower image downloads than cmos.

At this point I would resort to just reformat the sd card and set up linux from scratch. Sometimes this is faster than trying to fix the problem. After all, you said that it used to work.

I did a quick & dirty test of my idea, but I had too much wiggle room for it to work. The idea may work if the rig is stable enough, but it won't be easier or cheaper than the setup in the original post.

The image suggests tilt in the optical train. Start by imaging a totally different location, in the sky. If the pattern changes, gravity may be to blame. Especially, if you image with the camera hanging down or pointing straight up, and the stars look good, it's gravity pulling on the camera. If the pattern doesn't change, it's definitely tilt. Fortunately, Skywatcher focusers have tilt adjustment screws. They are at the base of the focuser (the pairs of screws at the base in the image).

Most of the time, atmospheric conditions determine star shapes and best sampling rate. But if undersampled, there's always the option to drizzle. https://www.stsci.edu/~fruchter/dither/drizzle.html

No, it shouldn't. The green pixels data goes into the green channel, blue pixel data into the blue channel, and red data into the red channel. Green and blue end up empty with the use of an Ha filter, apart from the camera's dark signal. During super pixel deBayering, there is no interpolation, so no green or blue is mixed with the red.

Not at all. Normal debayering ”fills in” the missing data by interpolation. But if you use superpixel debayering, the colour from 4 (2x2 ) pixels is simply combined into one colour pixel. Extracting the red channel from this would give the clean Ha signal without any deBayering artefacts.

I honestly have no clue about Bortle, but my darkest nights are at 20.7 mag, rural Sweden with Stockholms light dome more than 20 - 25 miles to the south. Osc is certainly doable here, even if a magnitude darker wouldn’t hurt. Osc means just one less device to worry about, and I seldom do narrow band imaging.