Help Needed with PHD2 Drift Alignment

[Gib007]

Tonight I thought I really had the drift alignment procedure down. I set up my mount, with three bolts that attach the mount head to the tripod pier adapter a bit loose (so I could actually tweak the Azimuth bolts). I polar aligned with EQMod as usual and then focused the optical system. Now the drift alignment. I started PHD2 and slewed my telescope to a star to the South, very near the Meridian and about 30° above the horizon. I calibrated on a guide star and within about 14 steps per direction, it was guiding. 

Now, I open the Drift Align tool and click Drift. I watched as the red Dec line stabilised and pointed up at a steep angle. The polar alignment error was being quoted as around -50' (I assume arcminutes is the unit as degrees and arcseconds would be preposterous). I click Adjust, tweak the Azimuth bolts and click Drift again to check. Eventually after a bit of tweaking, I got the red Dec line to be quite, quite flat and the polar alignment error was being quoted as around 0.15'. I then slewed the telescope East to a star about 30° above the horizon. I click Drift again and the red Dec line, surprisingly, is already pretty flat as it stands. The polar alignment error is quoted as 0.17' and the circle around the guide star appears quite small and stable. So, I don't bother adjusting the Altitude bolts. This is when I tighten the three bolts that attach the mount head to the tripod pier adapter and park the telescope. 

Time for an imaging test run. I slew to the Veil Nebula and capture a preview exposure. Unfortunately it seems I'm severely off-target, which is odd when your polar alignment is apparently so good. AstroTortilla centres the target quickly but this is already a warning of things to come. PHD Guiding (the original one, not PHD2) is started, a guide star is selected, it is calibrated on and guiding starts. I set off a test 15 minute exposure. This is the PHD Guiding graph:

And the following is the disaster of field rotation that pops up after 15 minutes:

Wow, right? That is unholy field rotation. Keep in mind the focal length here is only 367.7mm, so that level of field rotation is pretty impressive. What happens next baffles me to death. I slew back to the South, again to a star very near the Meridian and about 30° above the horizon and start PHD2. I calibrate on a guide star and start the Drift Align tool. Clicking Drift, guess what happens? The red Dec line is again really steep and the polar alignment error is quoted as about -60'. Unfortunately a blast of fog from the sea comes over Gibraltar at this point and sends me home packing. 

However, why or how would the polar alignment be SO far when PHD2 says it was so good? I really didn't understand this, and still don't. Unfortunately, until I get this done right, I don't want to seriously image as I would like to be able to capture 15 minute exposures unhindered. I definitely want to get drift alignment sorted and actually working. Did I do something incorrectly? Should I really be monitoring the red Dec line in PHD2 when drifting? Is there something else I need to look at? Could it have been tightening those three bolts on the mount that threw off my drift alignment so badly? But really as badly as -60'? All ideas and help welcome! Thank you!

P.S. There was no wind whatsoever tonight and the tripod was leveled well and everything was nice and stable. The mount is my Avalon M-Uno, loaded with the Borg 77EDII at f/4.6. Imaging was done with an ATIK 383L+ through a Starlight Xpress USB Filter Wheel with a Baader 7nm Hydrogen-Alpha filter. The autoguider is a Starlight Xpress Lodestar popped into the Starlight Xpress Slim OAG for the filter wheel. 

[rfdesigner]

Field rotation is not really a function of focal length ( I forget the maths, but the size of image is what really counts.. i.e. megapixels), you can get it with an 18mm lens photographing the milky way.  However I too would be disappointed with your result after your efforts.

Also levelling a German Equitorial doesn't buy you anything (other than to minimise the chances of it falling over), if the polar axis is aligned you're good to go, it matters on corrected Alt-Az mounts though (meade/Celestron Schmidt Cass without wedge).

Guess1: the DEC tracking turned itself back on?

Guess2: you have slop somewhere in your mount.

Your basic procedure was more or less what I've done to get 1 hour shots on film many years ago, you're dealing with a bug, not a basic misunderstanding.

Derek

[Gib007]

Thanks for your reply! Last night I figured something out. You're meant to point to stars very near zero Declination coordinates when you drift align apparently. This means pointing towards the equator. I was pointing near the horizon and at least for South, I was pointing at a star with Declination of about 30 to 40 degrees rather than near 0 degrees. Surely this would play a role in how accurate my drift alignment is... Oops!

[Gib007]

Just an update - I nailed it. It was indeed what I said in the above post. Last night I even captured a 30 minute exposure of the Veil Nebula in Hydrogen-Alpha, to prove to myself I really did nail it. Here it is, gratuitously!

[ollypenrice]

Good.

So, using this method, do you guide in Dec only when drift testing to the south to adjust azimuth and RA only when testing to the east to test Altitude?

And, if so, does PHD sort out which axis should be guided during these two operations? I have downloaded PHD2 but not played with it yet.

Olly

[Gib007]

PHD2 has a really intuitive Drift Align tool for this. You point at a star at the intersection of equator and meridian (South), calibrate on it and then open the tool. When you click Drift, it drifts in Declination and then you see two trendlines after a couple of exposures. All you need to pay attention to is the red Declination trendline. You just click Adjust, tweak Azimuth and then click Drift again to see the changes. This is done until your red Declination trendline looks nice and flat after a number of exposures. You'll see a calculated "polar alignment error" in the graph as well and a circle around your guide star that gets smaller as you get closer to alignment.

Next, you click Altitude and now slew to a star on the East or West along the equator line. You repeat the process but this time tweak the Altitude on your mount. Mind you, the mount will still be guiding on RA and drifting in Dec. You still only pay attention to the red Declination trendline. Oh and you don't have to recalibrate on the guide star as PHD2 makes the corrections to the calibration.

You can repeat the process again but I found my polar alignment to be fantastic after just tweaking Azimuth and Altitude once. My overall reported polar alignment error was about 0.15 (arcminutes I believe, but could be degrees - can't see the unit well on the graph). More than enough accuracy by the looks of the 30 minute guided exposure.

[gingergeek]

Looks fantastic, good job on the drift align issue. It actually inspires me to look at learning this now

[rfdesigner]

Good.

So, using this method, do you guide in Dec only when drift testing to the south to adjust azimuth and RA only when testing to the east to test Altitude?

And, if so, does PHD sort out which axis should be guided during these two operations? I have downloaded PHD2 but not played with it yet.

Olly

No

Guide in RA in both cases.  DEC always drifts...   you're watching the difference between perfection (RA only guiding) and how your mount's setup.

N/S polar alignment error in mount produces no drift when pointing south, but drifts significantly pointing east/west, and vice-versa

Thanks for your reply! Last night I figured something out. You're meant to point to stars very near zero Declination coordinates when you drift align apparently. This means pointing towards the equator. I was pointing near the horizon and at least for South, I was pointing at a star with Declination of about 30 to 40 degrees rather than near 0 degrees. Surely this would play a role in how accurate my drift alignment is... Oops!

You are right and not-so right.  Yes you want to point to zero declination when pointing south.. however if you're not, so long as you go round the whole loop several times, you're fine (I had to do this from a very restricted location)

Derek