Stars drift with high resolution encoders


<Joerg>
 

Thanks!

When the weather is better I will make a video and a tracking log. Sadly the forecast says rain for many days. :-(

Joerg


Dan Gray
 

Joerg,
A tracking log will be a good tool.  You'll be able to see if it's servo position error for sure.
Do you have a video camera, or a fast cmos camera?  It would be really simple to make a video over time.  I would suggest the southern meridian for checking the azimuth, and east or west when checking the altitude.
Now you can see if it's interpolation error or not.

Dan


On Tue, Aug 24, 2021 at 2:14 PM Don W <westergren@...> wrote:
Hi Jesse,

Yes, that is an excellent point.  Use the Tracking log to see the numbers showing the motion of scope while tracking a target.
Don W


Don W
 

Hi Jesse,

Yes, that is an excellent point.  Use the Tracking log to see the numbers showing the motion of scope while tracking a target.
Don W


jmgoldba
 

On Tue, Aug 10, 2021 at 12:31 PM, wrote:
an issue with small star drift around the right star position.
Assuming the issue ultimately isn't encoder resolution (or interpolation) have you collected log files that give you more visibility of what the exact "drift around" issue is? Don will recall an issue we had with occasional jumps coming from our DEC axis encoders and the log files were helpful. -Jesse


<Joerg>
 

Jesse,

I tried a new PXP model but no difference to the old one.

Joerg


<Joerg>
 

Dan,

 

The diameter of azimuth disk is 1174.72mm, altitude is 1268.21mm.

At round about D disk =1200mm, 1arcsec conforms to 2.9µm on tape. 1 sec tracking at sidereal rate conforms to 0.0436mm on the tape.

The magnetic tape has a pole length of 1mm and the read head is MK1-2k-R.

That should give a maximal resolution of 0.5µm.

 

Joerg


jmgoldba
 

On Tue, Aug 10, 2021 at 12:31 PM, wrote:
The issue in "precise mode" or "cascade mode" can be switched off if "Track Only Sidereal (No PXP)" in Mount Parms has been activated.
How old is your PXP model? - Jesse


Dan Gray
 

Can you tell me the diameter of the disk for the magnetic tape?  In the spec sheet, it gives about +-0.004 mm per meter for magnetization accuracy + interpolation error .
If we know the diameter, we can figure out the arc seconds error.
We can also find the period of the interpolation if we know which readhead you have. The pole spacing is 1, 2, or 5 mm.
Dan


On Mon, Aug 23, 2021 at 3:52 AM <Joerg> <joeps@...> wrote:

Hi Dan,

 

The encoders have been discontinued but I found a link to the user manual.

http://www.enkodery.pl/download/instrukcja_obslugi_gc-mk.pdf

 

I have a pole length of 1mm and the highest resolution read head.

 

Now I made some additional tests. In all modes with activated scope encoders the drift is visible, also with tick management.

There is no large difference whether only one Az or Alt encoder have been activated or both at the same time.

When one or both scope encoder enabled the star follows a circular trail of about 3 times the star diameter. The time for a complete circle is several seconds.

What I don’t understand is the circular trail when only one encoder is enabled.

In my opinion in that case should be the trail a line if magnetization accuracy and interpolation error are blame for that.

 

Joerg


<Joerg>
 

Hi Dan,

 

The encoders have been discontinued but I found a link to the user manual.

http://www.enkodery.pl/download/instrukcja_obslugi_gc-mk.pdf

 

I have a pole length of 1mm and the highest resolution read head.

 

Now I made some additional tests. In all modes with activated scope encoders the drift is visible, also with tick management.

There is no large difference whether only one Az or Alt encoder have been activated or both at the same time.

When one or both scope encoder enabled the star follows a circular trail of about 3 times the star diameter. The time for a complete circle is several seconds.

What I don’t understand is the circular trail when only one encoder is enabled.

In my opinion in that case should be the trail a line if magnetization accuracy and interpolation error are blame for that.

 

Joerg


Dan Gray
 

Joerg, can you let us know the encoder part number?  The magnetic encoders have a lot of interpolation error.  We should look this up.
My guess, that's the problem.
Dan


On Wed, Aug 11, 2021 at 7:44 AM <Joerg> <joeps@...> wrote:

Hi Russell,

I'm unsure whether I tested Tick Management. Maybe not but as soon as possible I will do that.
But who knows what happens if "Track Only Sidereal (No PXP)" has been clicked?

The text box say: “This option will disable Offset Tracking Rates, and Pulse Guide ASCOM methods.”

Are scope encoders offline at that mode? Do sidereal rate comes only from motor encoder in that case?

 

Joerg


<Joerg>
 

Hi Russell,

I'm unsure whether I tested Tick Management. Maybe not but as soon as possible I will do that.
But who knows what happens if "Track Only Sidereal (No PXP)" has been clicked?

The text box say: “This option will disable Offset Tracking Rates, and Pulse Guide ASCOM methods.”

Are scope encoders offline at that mode? Do sidereal rate comes only from motor encoder in that case?

 

Joerg


Russell R
 

Hi Joerg,

Here is what happens with Cascade mode:

a. Calculate the position error based on the encoder.

b. Run the motor at the sidereal rate + a rate that will bring the encoder error down to zero in a given time.

c. When the given time expires, it will be running the motor at the sidereal rate again.

d. Repeat.

Cascade mode is recommended to have at least 13M ticks, so it could be the resolution of your encoders is not enough.

Have you tried Tick Management mode?

Russell R.


<Joerg>
 

Hi Don,

 

The resolutions of the axes are 7.4 and 7.9 million ticks/rev. At 7.4 I have a resolution of 0.175 arcsec / tick. In my opinion that should be more than enough for precise tracking. The seeing is seldom under 2 arcsec in my area so star diameter shouldn’t exceed not more than 2.35 arcsec at seeing of 2 arcsec.

What I don’t understand is if I use “cascade mode” AND "Track Only Sidereal (No PXP)" in Mount Parms is activated the star stands still in the right position without tumble.

In this mode tracking uses scope encoder without problems so the accuracy of the encoders should be o.k..

 

Joerg


Don W
 

What can you do? - Not much.  7 million tics is not very accurate - you can expect to track within a couple arc-sec - the diameter of a star.

Alt-Az mounts must track in both axis.  Cascade mode needs a lot more tics/rev with a Servo II to work.  Precise mode tries to keep the pointing accurate, but your resolution won't do it.  Polite mode seems to work because it only tries to correct occasionally.

realize that the Alt-Az mount tracks at varying rates across the sky, not a constant rate ever.

Don W


<Joerg>
 

Hi,


I have an issue with small star drift around the right star position when using high resolution encoders (on scope axes). It is an Alt / Az mount with magnetic encoder bands (ca. 7 Mill. ticks/rev.) that I have been using for many years.
The using has been only visual and now I tried to take photos. Even in short time intervals of 2 – 5 seconds exposures the effect is visible. Also centering a star with high power at star alignment is difficult and the drift is visible.    
It is only a small amount of drifting, right, left, up and down about 2/3 to one star diameter away.
Using encoder config to "ignore" or "polite mode" there is no issue. Using "precise mode" or "cascade mode" the issue comes up.
PID tuning isn't the blame in my opinion regardless I tried this without success.
The issue in "precise mode" or "cascade mode" can be switched off if "Track Only Sidereal (No PXP)" in Mount Parms has been activated.
Has anyone an idea what can I do?

cs
Joerg