Jump to content


Photo

Arc seconds per step for high end mounts.

This topic has been archived. This means that you cannot reply to this topic.
7 replies to this topic

#1 kingjamez

kingjamez

    Apollo

  • *****
  • Posts: 1000
  • Joined: 03 Oct 2006

Posted 10 July 2008 - 12:33 PM

Any idea on what size step most high end mounts use? I'm designing an equatorial mount and have run into a wall in balancing the dynamic range of my (servo) drive system. It seems I can either have a low arc seconds per step ratio, or fast slews but not both. So my goal is to find the maximum acceptable number for arc seconds per step that will give the appearance of 100% smooth motion.

Related question: If I have a camera that has 0.5 arc seconds per pixel resolution in my optical system (assume I'm on a mountain in Hawaii), will 0.4 arc seconds per step appear smooth or do I need at least 0.25 arc seconds per step?

How on earth does the AP1200 hit 1200x sidereal while still maintaining smooth motion at/below sidereal?

Thanks for any suggestions (hint hint:CCS_HELLO).

-Jim

#2 ccs_hello

ccs_hello

    Fly Me to the Moon

  • *****
  • Posts: 6618
  • Joined: 03 Jul 2004

Posted 10 July 2008 - 10:39 PM

Jim,

Can you provide a little bit of information on what you plan to build? Is it going to be DC perm-mag motor based or micro-stepping stepper based? Also, what type of controller (intelligent handbox) and protocol you plan to use?

The parameter "ticks per arc-second" is the resolver--gear train precision. For Autostar, a good range is somewhere between 2 and 6. (I would suggest 3.5, this is about 0.3 arc-sec/tick). It affects the GOTO precision and fine-tune in autoguiding.
Larger number means better precision with the trade-off that either the motor has to spin very fast during fast-slewing GOTO (coffee grinder) or a high-tick count encoder + higher-performance CPU has to be used (which is not the case for Autostar).

Use LXD75 as an exapmle:
- a DC perm-mag motor is used
- 50:1 mechanical gear box
- 37:35 further mechanical gear down
(above two mechanical net gear down is 52.857:1)
- 144:1 worm in mount's RA axis
- 108 slot encoder wheel, using quadrature to yield 432 ticks/rev

So its tick/arc-sec parameter is

432 * 50 * (37/35) * 144 / (360*60*60) = 2.537
(which is about 0.4 arc-sec/tick)

However, it may not affect the motor tracking smoothness, if it is DC perm-mag based. The servo PID loop determines that. Speed change is continuous. The ticks/arc-second value is just for math.

For microstepping stepper based, if it is the open-loop type (like EQ-G), the pseudo-ticks/arc-second value indeed indirectly reflects the finest microstepping size, e.g., 64 microsteps/step. The smoothness during tracking is partially governed by that. I.e., a 64-micropsteps/step has the potential better than a 16 microcteps/step stepper system, providing that the driving waveform has been tuned with the stepper motor's own characteristics to form a better final sinusoidal (instead of jittery stepping) shaft movement.

{Note: Speed adjustment in this case is not dependent on that pseudo-tick/arc-sec parameter either. It is done by internal timing loop.}

Use EQ6-Synscan as an example,
- a 200 steps/rev stepper motor is used
- later version has 64 microsteps/step (electronic gearing)
- 47:12 mechanical gear down
(the above mechanical net gear down is just 3.9167:1)
- 180:1 worm in mount's RA axis

So its pseudo-tick/arc-sec parameter is

200 * 64 * (47/12) * 180 / (360*60*60) = 6.963
(which is about 0.1436 arc-sec/pseudo-tick)


Clear Skies!

ccs_hello

#3 gnabgib

gnabgib

    Viking 1

  • *****
  • Posts: 508
  • Joined: 05 Jun 2005

Posted 10 July 2008 - 11:38 PM

Jim;
I tend to agree with CCS although I have found in my own case the .4 sec/tick is somewhat coarse and would suggest trying to achieve .2 sec/tick. Stepper motors by their nature have limited dynamic range as compared to ac or dc servo motors. Systems I have built in the past have gone fron tracking rate (at 8 encoder counts per arc second) to 180 degrees per second slew rate. All using a single dc drive motor. Granted 180 degrees per second is excessive it can be done but NOT with stepper motors! So my advice would to strive for something in the range of 6 to 10 encoder counts per arc second to achieve a smooth rate. By the way the 180/sec scope was part of a project to try and image the visual wavelenght component of a gamma ray burster so getting on target in the shortest time was paramount. Personally I find a slew rate of 5 to 10 degrees/second very adequate for 99.9% of astronomical projects!
Kevin

#4 Kaizu

Kaizu

    Messenger

  • -----
  • Posts: 472
  • Joined: 17 Aug 2007

Posted 11 July 2008 - 05:33 AM

I would say that if the camera resolution is 1", then you should move the dot one step left when defining the tracking resolution, ca. 0.1" /step. If you use steppers it is that step and if you use servo, it is the reslution of the encoder. I had earlier a home made mount with steppers and I noticed that 4 steps/ arc sec was to coarse and caused vibrations. In that case the the max slewing speed OK, it was 144x trackig speed but I had to drop that to get better tracking resolution.

Kaizu

#5 ccs_hello

ccs_hello

    Fly Me to the Moon

  • *****
  • Posts: 6618
  • Joined: 03 Jul 2004

Posted 11 July 2008 - 06:28 AM

Kaizu,

Can you share your stepper system parameter such as how many microsteps per step?

As I've stated in my previous post, smoothness is highly dependent on the tuning of microstepping driving with the actual stepper characteristics to emulate sinusoidal motor output shaft movement. (I.e., emulate an AC synchro motor.)

Clear Skies!

ccs_hello

#6 kingjamez

kingjamez

    Apollo

  • *****
  • Posts: 1000
  • Joined: 03 Oct 2006

Posted 11 July 2008 - 11:40 AM

Thanks for the help! I knew I could count on you guys. I'm planning on using a brushless servo motor for the first time because I need high dynamic range. I've done stepper systems in the past, and was concerned about the smoothness issue. But it makes sense that a servo based system would be smoother at is by design a continuous motion rather than an emulation of continuous motion like a stepper.

So I'll shoot for 4 or more ticks per arc second, and a well tuned servo.

Thanks again!

-Jim

#7 Kaizu

Kaizu

    Messenger

  • -----
  • Posts: 472
  • Joined: 17 Aug 2007

Posted 17 July 2008 - 01:38 PM

Sorry about the delay, I had a one week stint to Japan.
I do not use microstepping, only the half steps. For the speed correction I have three stage switch - stop/ 1X/ 2X.
the 1X is tracking by halfsteps and the 2X is forward correction by full steps. I just have the gear ratio so big that I dont need the microstepping. Old stepper controller shows how I have built the controller. This is the very first version where the tracking was too coarse for imaging with long focal length.
Afterwards I multiplied the gear ratio by ten and removed the fast move unit. I search the targets manually and had the fine adjusting screws for final aligning.
The text is in Finnish but the schema shows the principal.


Kaizu

#8 ccs_hello

ccs_hello

    Fly Me to the Moon

  • *****
  • Posts: 6618
  • Joined: 03 Jul 2004

Posted 19 July 2008 - 04:12 PM

Kaizu,

Thanks for the info. Hope you enjoy the Japan trip.

Basically your old design illustrates the required dynamic range need to be built-in as an astro-GOTO/fast-slewing controller. Actually I would feel jumping from 12x to 144x in one single step actually may not work with mount under load. This is why microstepping with ramp-up and ramp-down is used in modern stepped-based designs.

Clear Skies!

ccs_hello






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics