Dec inaccuracy with Vixen GP #EQ5


Sam
 

Hi all, I converted my Vixen GP with a MiniPCB v2. I control the rig via Astroberry. I've been using the Ekos-ASTAP platesolver with an accuracy of 30 arc seconds. Notice that the platesolver needs to iterate up 10 times or fail because DEC doesn't get within the required accuracy. RA will give me 20 or even zero but Dec is always problematic. It averages at 50 to 40 arc seconds. I might get lucky and get around 20+ or 17 arc seconds.

The platesolver captures an image, solves, and slews. While RA gets within the required accuracy, Dec always overshoots or what.

I tightened the belt to remove slop. I even tightened the worm mesh a little but to no avail. My payload isn't even that heavy.

Do I have to turn on refraction? Configure backlash? Add microsteps?


Khalid Baheyeldin
 

I use KStars/Ekos with the plate solver, and have the accuracy set to 5 or 10 arc seconds (forgot which one).
It takes 2 or 3 iterations to be within that error, consistently.

Yes, I had to configure backlash, using the easy method (terrestrial object) described on the Wiki.
On DEC my backlash is 33", and on RA, it is 161".


Sam
 

On Wed, Oct 20, 2021 at 04:42 PM, Khalid Baheyeldin wrote:
I use KStars/Ekos with the plate solver, and have the accuracy set to 5 or 10 arc seconds (forgot which one).
It takes 2 or 3 iterations to be within that error, consistently.

Yes, I had to configure backlash, using the easy method (terrestrial object) described on the Wiki.
On DEC my backlash is 33", and on RA, it is 161".

Khalid, I did as you told although I'm not sure what "jump" means or "lags" in this context (pardon me, English isn't my first language.)

I do notice that when North and then reversing South, the Dec snaps back to North (Right) a very little amount before going South (Left). I'm using a Nikkor 180mm ED AIS.


Khalid Baheyeldin
 

On Thu, Oct 21, 2021 at 09:08 PM, Sam wrote:
I did as you told although I'm not sure what "jump" means or "lags" in this context (pardon me, English isn't my first language.)
Lag means when you reverse the direction (e.g. from W to E), there will be some time with no motion, then motion starts.
That happens if your backlash value is too low.

Jump is the opposite: when you reverse the direction, OnStep will move very fast before moving at the rate you are using. That means that the backlash value is too high, and OnStep over compensated for it.

When adjusting backlash, you should use the live view function, and zoom in at the highest magnification allowed in the Nikon, so you get the best accuracy.


Sam
 

On Thu, Oct 21, 2021 at 06:13 PM, Khalid Baheyeldin wrote:
Jump is the opposite: when you reverse the direction, OnStep will move very fast before moving at the rate you are using. That means that the backlash value is too high, and OnStep over compensated for it.

Huh, i though OnStep compensated backlash by accelerating a bit before going into normal slew. So, motion must be uniform without any noticeable acceleration or lag.

 

I am on live view with max zoom. As I have described, when reversing from North to South, the mount snaps a little to the North before going South. But, when I do the opposite and reverse from South to North, the mount moves North immediately without snapping south first. Does this indicate an issue with the drive setup? motors? gears?


Khalid Baheyeldin
 

On Thu, Oct 21, 2021 at 09:21 PM, Sam wrote:
Huh, i though OnStep compensated backlash by accelerating a bit before going into normal slew. So, motion must be uniform without any noticeable acceleration or lag.
Exactly ...

But only if you compensated the correct amounts, which is practically impossible on a first try.
If you have approximate values, there will be non-uniform motion when reversing direction, either no-motion for a while, or fast motion.
Depending on which one it is, you decrease or increase the value until you get uniform motion.

I don't know what is causing the North movement that you describe.
Maybe Howard will comment ...


Sam
 

Would it be possible to accurately dial down backlash compensation through Astroberry?

I'm thinking of using the platesolving feature to determine how much the backlash really is.

For example, I'd platesolve to determine where the scope is pointing at. I'd issue a move command with exact arcseconds in either N/S or E/W direction. Then apply another move command in the opposite direction. Platesolve, then determine the difference (backlash) between the two points?


Khalid Baheyeldin
 

You can try that method.
Note that the sky is always moving, but at least that is at a fixed rate.
Let us know if it works.

I have also used PHD2 guiding assistant, but from what I remember it only measures
backlash in DEC, and my results were mixed or inconclusive (only tried it one night).

You can try that method too, and see if you have better results.


Sam
 

I don't have a guider yet, the ASI120 I ordered and Svbony 30mm is yet to arrive this week. So, can't try the PHD2 stuff yet.

I got the accuracy down to tolerable levels. What I did is that I slewed back and forth between Aldebaran, M45, and Orion's Belt. After slewing I plate solved. I checked the iterations and how far accuracy was with each iteration before it reached the 30 arcsecond required accuracy. At first, accuracy in Dec was all over the place with RA always in the ballpark. I was still 500 or more arcseconds far from target at the 5th or 7th iteration. I added about 10" to the Dec and RA backlash with each platesolve.completion.

 

Now, I get 2k at first iteration, 200-100 at second iteration, 50 or within 30 seconds at third. Before I was able to get within 30arcseconds after 8-10 iterations. Now, my maximum is 4-5 iterations. My RA backlash right now is 240" with accuracy of around 2-7 arcseconds, Dec backlash is around 330" but am now able to get within 20-10 arcseconds consistently.

 

I plan to do this for some more nights to increase accuracy. I had to stop as the sun was already rising where i am at.