PHD2 Drift Align - trend line doesn't change

[frugal]

I have been having a bit of a problem with drift alignment in PHD2 this evening. I leveled the tripod, balanced the scope etc. I then polar aligned using the polar scope on the mount. Polaris was slap bang in the middle of the "polaris goes here" circle. Apparently this little target circle is 1 arc hour across, so I should be well within 60' accuracy.

So I start up PHD2 and start to drift align. The Azimuth star is just off of due south, at the intersection of the meridian and the galactic equator according to Carte Du Ciel. I start drifting, and I leave it for 50 iterations (at 2 seconds an interation), so 100 seconds. The line is trending down and it claims I am about 30' out of alignment. I then adjust the azimuth bolts on the mount and move the star to the edge of the magenta circle that PHD2 is showing me and have another go. After another 50 iterations I am apparently still 25' out and the adjustment circle is just as big. So I adjust again, once against moving the star to the edge of the circle. Start drifting and again I am still 25' out, and the line is still trending down. After one more iteration of this I am confused, the Polaris is going to be way off by now given the amount I have adjusted the zaimuth bolts. So I go back the other way (just once), and the line gets almost perfectly flat. hmmm... suspicious.

So I move to Altitude, pointing the scope due West, and close to the galactic equator as I could (allowing for the house in the way). Exactly the same problem, moving to the edge of the calibration circle that PHD2 gives me does not seem to change the trendline or the polar alignment error. I would assume that if I am 25' off alignment, that if I adjust the correct direction, then I should end up at 0', and if I go the wrong direction I should end up at 50' out, I can not for the life of me figure out why it is not changing.

After much faffing I managed to get it to about 2' error. However when I double checked this with the polar scope, the polar scope was showing Polaris significantly inside the circle on the polarscope that shows Polaris's track through the sky. Currently PHD is guiding with an RMS error of 1.05", as opposed to the 0.7" I normally get.

Does anyone have any ideas why the trendline does not get altered when I think it should? Is 100 seconds enough time to accurately calculate the drift? I do get a sinusoidal pattern that I assume is periodic error in the mount, but that is obvious after 50 iterations. 

[michael8554]

Hi frugal,

I can never work out which side of the circle to move towards the first time, it's hit and miss, sounds like you missed with your az adjustments, but eventually reversed direction and started getting a good result.

Also when the error is large I found PHD2 too time consuming, I drift align first and then fine tune with PHD2.

As to the difference in the Polaris position after adjusting with PHD2 - if the polar scope was that accurate you wouldn't need further alignment - I'd be tempted to get good PA and then adjust the polar scope to match!

Michael

[frugal]

If it had got worse I could understand that I had gone the wrong way. What is freaking me out slightly is that adjusting the Azimuth bolts significantly made no effect.

[higgins]

I have not used the phd2 drift align routine yet but don't take it for granted that the polar scope came from the factory perfectly aligned. If you haven't checked, this video is how i did mine, which was out.

     

[frugal]

This is pretty much what I saw through the polar scope. Can the polar scope really be that much out of alignment?

 

[higgins]

12 minutes ago, frugal said:

This is pretty much what I saw through the polar scope. Can the polar scope really be that much out of alignment?

 

I would say yes. I'm still a beginner but know for sure this hobby is about dotting the i and crossing the t.

 

[Anne S]

The manual says to point the mount axis at a fixed object and rotate the RA axis from one side to the other. If the polaescope is aligned the object should stay in the centre of e polarscope. I did try this with a CG5 once but the only end result is that I lost one of the grub screws from the polarscope!

Anne

[frugal]

I am sure that I did the polar alignment routine a while ago (I remember that the grub screws were a right pain to adjust). I will have another go just to make sure that it is still good.

[frugal]

Ok, I have just checked the polarscope alignment against a spire on a local roof and it looks good. A full 360 degree rotation did not lead to any detectable wobble in the target.

[kens]

Did you rotate the mount so that Polaris was in the right position on the circle?

[frugal]

Just now, kens said:

Did you rotate the mount so that Polaris was in the right position on the circle?

I used the EQMod polar align, which had Polaris at exactly the 3 o'clock position, which coincided with Polar Finder on the phone. So I am fairly confident that I had it in roughly the right position to begin with.

[Thalestris24]

Hi

One thing with PHD2 drift alignment - be 100% sure you're doing the azimuth adjustment when pointing south and the DEC when pointing E-W! It's easy to make a mistake...  I would drift for 3-5mins, depending on your focal length. You need to be fairly close-ish with PA to start with, and you need to be balanced correctly and have achieved a good calibration. Not sure if any of that's of any help...

Louise

[frugal]

59 minutes ago, Thalestris24 said:

Hi

One thing with PHD2 drift alignment - be 100% sure you're doing the azimuth adjustment when pointing south and the DEC when pointing E-W! It's easy to make a mistake...  I would drift for 3-5mins, depending on your focal length. You need to be fairly close-ish with PA to start with, and you need to be balanced correctly and have achieved a good calibration. Not sure if any of that's of any help...

Louise

For the Azimuth I use Carte du Ciel to make sure that I am pointing as due south as I can, and where the due south meridian crosses the galactic equator (not the southern horizon as I first thought). I also make sure that I am on the correct side of the meridien so that the mount does not hit its' limits during the drift (did that once, took a while to figure out why the stars kept moving).

I have even dug out the manual to ensure that Azimuth really is the left/right turning bolts (as opposed to the up down lever).

I am certain that it is User Error, but I can not figure out what I have done wrong.

[frugal]

Louise, That is a good point about calibration. What do people calibrate on when doing drift alignment? Do you calibrate on the Due South star before you start to do Azimuth drifting. Do you then re-calibrate when you move to the Westerly star for Dec drifting?

Re-calibration seems to be a contentious issue. Some people say that unless you change your set up, you should never re-calibrate; others say calibrate every session; others say re-calibrate every time you change what you are pointing at.

[kens]

You mean celestial equator don't you? That is Declination 0 degrees.

If you change your polar alignment by a  substantial amount you should recalibrate and check with another drift alignment. You don't really need to recalibrate in the middle of a drift alignment.

[frugal]

2 minutes ago, kens said:

You mean celestial equator don't you? That is Declination 0 degrees.

If you change your polar alignment by a  substantial amount you should recalibrate and check with another drift alignment. You don't really need to recalibrate in the middle of a drift alignment.

You are quite right, celestial equator. The galactic equator is something completely different. 

For the avoidance of doubt, here is where I point the scope for the Azimuth (left / right) part. I pick a star somewhere the area that I have selected, PHD is pretty good at finding a good guide star wherever I point it.:

For Altitude (up/down) I have a hill and a house in the way of where the celestial equator meets dues West, so I pick a star as close as I can, somewhere in the yellow marked area, so somewhere about 15-20 degrees above the equator:

From the PHD2 documentation: "Changing targets will not require another calibration because PHD2 can know where the scope is pointing and automatically make adjustments to the guider calibration.  Most users get a good calibration and then re-use it until they make hardware changes of some kind.". SO I guess in this case, moving the Az or Alt during polar alignment counts as a "hardware change".

[Thalestris24]

Hi

What I meant was that in phd2 to make sure you're in the right direction mode when in the drift align tool. It always starts with azimuth (scope south) but if you click on it it goes to DEC. You don't want to try and adjust the azimuth when the scope is actually pointing e/w! I know it's not likely that one would make such an error but anything's possible

Yes, you need a good calibration to start with. If you've had to adjust the mount a lot during drift aligning it's probably worth recalibrating and rechecking PA after. 

Louise

[frugal]

Ok, so my plan of action for next time there is a break in the clouds is:

 - Setup

 - Point the scope at the Azimuth target star (Due south, close to the celestial equator).

- Calibrate on that star

- Drift align for Azimuth

- Slew to West

- Drift align for Altitude

- Slew back to the Azimuth star and re-calibrate

Then keep looping to see if the polar alignment / calibration is stable.

[Owmuchonomy]

It shouldn't really be necessary to do a re-calibration during the routine.  Just make sure the initial calibration is good.  I wouldn't switch from drift align to 'checking' the position of polaris in the polarscope either; that's just an eternal loop you will never resolve in my opinion.  To be honest I would ignore the polarscope completely.  Just align the mount roughly then do a polar alignment from the handset.  I find this gets very close to the PHD2 drift align sub-routine limits.  Just fine tune your PA then in PHD2.  I find seeing plays just as big a part in error as anything else.

[frugal]

I know I probably should not need to re-calibrate, but this was causing me such a head ache last night that I am going to go for a belt and braces approach next time to try to narrow the problem down. Then I can start to remove steps gradually

[Thalestris24]

3 minutes ago, frugal said:

I know I probably should not need to re-calibrate, but this was causing me such a head ache last night that I am going to go for a belt and braces approach next time to try to narrow the problem down. Then I can start to remove steps gradually

Hi

How are your calibrations - much backlash? Can you post a screen grab?

You probably don't need to recalibrate after pa - I just suggested it as a double-check  Calibration only checks the relationship between the guide scope and the mount.

Louise

[frugal]

15 minutes ago, Thalestris24 said:

Hi

How are your calibrations - much backlash? Can you post a screen grab?

You probably don't need to recalibrate after pa - I just suggested it as a double-check  Calibration only checks the relationship between the guide scope and the mount.

Louise

I do get quite a lot of backlash on the Dec axis. Grabbing a very guide file off of the hard drive (from the end of March):

[Thalestris24]

Looks ok - 7-8 steps are supposed to be sufficient but I might be tempted to adjust the step size so as to get 12-15 steps, though not essential.

Louise

[frugal]

3 hours ago, Thalestris24 said:

Looks ok - 7-8 steps are supposed to be sufficient but I might be tempted to adjust the step size so as to get 12-15 steps, though not essential.

Louise

Whenever I do a calibration I always use the Brain icon to tell me the step size based on my setup and declination of the calibration star. Maybe next time I will take that number and then use 1/2 of it, that should get me a lot more steps.

[Thalestris24]

9 minutes ago, frugal said:

Whenever I do a calibration I always use the Brain icon to tell me the step size based on my setup and declination of the calibration star. Maybe next time I will take that number and then use 1/2 of it, that should get me a lot more steps.

I don't think it's critical. I'm afraid I don't know why your trendline isn't moving when it should! It's probably something obvious! Maybe try drift aligning the traditional way - without using phd2 and give it 5 mins.

Louise