×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

New Internal Guider Features

  • Posts: 1119
  • Thank you received: 182

My first thought is that you may have set the proportional gain in the control parameters lower in RA than DEC. What are the values? I would suggest keeping both at 133.3%.

Jo
3 years 7 months ago #57600

Please Log in or Create an account to join the conversation.

  • Posts: 209
  • Thank you received: 33
Hi Jo,
Good idea but both values are set to 133.3%
(I would not have dared changing these values !)
As suggested by Hy I set the iteration number to the max 10 and this improved a bit the calibration plot (see attached)
3 years 7 months ago #57607
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
Jean_Claude,
That calibration (and guiding) looks great to me.
Hy

PS That other thing on your screenshot (how you lost track of the guidestar right after starting to guide) happens to me too :( I assume you're running the scheduler, and that what happens is it starts guiding, then interrupts guiding right away to focus. I've looked at the images and during my initial focusing, sometimes particularly at the start, it seems to "move the image" significantly. The result is that you may guide well, but your image is moved a bit from where the alignment put it. Is that your situation too? If so, curious what mount and focuser you have.
3 years 7 months ago #57610

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182
Jean-Claude:

I agree with Hy, the guiding looks excellent. I can only dream of that with my little mount.
If you are using 10 iterations, then most of them are cut off in your calibration plot. Hover over the image and use the mouse wheel to zoom out. Then all your calibration points should become visible.
As is, you are moving the guide star by an enormous number of pixels. 10 pixels per iteration. Only 2 iteration points in each axis per direction are visible in your image. That's way more than you need. I usually run 7 iterations, at 2000 ms, and that moves my guide star by about 12 pixels in total.



Jo
3 years 7 months ago #57611
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559

Steve,

It is the same concept as the "Reverse Dec ..." parameter in PHD2.
One could look at their doc, e.g. github.com/OpenPHDGuiding/phd2/wiki/Reve...-after-meridian-flip
but I can't verify this.
3 years 7 months ago #57612

Please Log in or Create an account to join the conversation.

  • Posts: 209
  • Thank you received: 33
I'm not running the scheduler but I guide before I finish carefully my focusing.
My mount is a Celestron CGE Pro (OTA 11" EdgeHD + Reducer x0.7) and my focuser is a Microtouch.
3 years 7 months ago #57613

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
Jo,

You are mistaken here. When you set it to 10 iterations, it uses UP TO 10, but once it crosses 15 pixels of movement it stops.
I could make that "15 pixels" a parameter some day, but we have so many already...
So, you can see that even though he had 10 set, the movement was sufficient in both axes to terminate the calibration after 3 or 4 steps.
[This was a change I made in the last few months, so you're probably relying on older experience with the calibration.]

Again, I recommend setting the number of iterations on the maximum (10) and letting is stop itself after 15 pixels of movement.

Hy
3 years 7 months ago #57614

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182

I see! I never realized that!

But wouldn't it be better in that case for Jean-Claude to reduce the pulse duration so he ends up with more data points on both axes? I would think that fitting a line through 5, 7 or 10 data points is going to be more accurate for calculating the angle than fitting it to only 2 points.

Jo

PS: Actually, Hy, that can't be right! When you are looking at JC's first calibration plot (the one with the short RA axis), the mount moved 25 pixels in DEC. Clearly more than 15 pixels.
How do explain that then?
Scratch that. You wrote "once it crosses 15 px". So it executes the next iteration and then stops. OK.
Last edit: 3 years 7 months ago by Jose Corazon.
3 years 7 months ago #57615

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
Jo,

1- I don't have a recommendation on the pulse size. You're probably right that 2000ms is excessive. Probably 500 - 1000 would make more sense, but just guessing.
FWIW, It doesn't actually fit a line, as it currently stands. It just draws a line between the first and last point. That's probably fine, but let me know if you disagree.

2- With respect to Jean-Claude crossing 25 on DEC (the vertical blue dots), I looked at that too when I first saw his plot, but you have to remember that it doesn't start at zero.
First there is some "backlash-clearing" movement. So, it looks like the backlash section (vertical while +'s) ended between 5-10 so it does make sense to me.

Hy
Last edit: 3 years 7 months ago by Hy Murveit.
3 years 7 months ago #57620

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182


Ah, in that case, 2 iterations would be sufficient then, as all the data points in between would be ignored anyway! Just increasing the pulse length so it equates to a deflection of ~15 pixels would be enough then.
But yes, I think it would be more accurate to calculate a linear regression on more data points, but when I am looking at my own calibration plots, it wouldn't really matter that much.
So I will adjust my own calibration now as well, increase pulse length to 10000 ms and run only 2 iterations. That should speed things up quite a bit then as long as I deselect the DEC backlash clearing, I presume.

Jo
3 years 7 months ago #57621

Please Log in or Create an account to join the conversation.

  • Posts: 1208
  • Thank you received: 559
Jo,

I don't know what would happen if you set it to 10000ms, but clearly that won't speed things up,
rather it will likely slow things down. It will probably guarantee that each iteration is a minimum of 10s long.
If I were you, I'd use a "normal value" like 500.

In general, please be conservative with changing all these various parameters.
Trying to save a few seconds here and there can cost you (and me) hours of debugging...

Hy
3 years 7 months ago #57622

Please Log in or Create an account to join the conversation.

  • Posts: 1119
  • Thank you received: 182

Understood!

Although, didn't one of your colleagues up there in Silicon Valley once say: "Move fast and break things"?

I think that was another company, though...

:-)
3 years 7 months ago #57623

Please Log in or Create an account to join the conversation.

Time to create page: 0.395 seconds