Jump to content


Photo

30K Encoders - How Much Improvement?

  • Please log in to reply
12 replies to this topic

#1 Jeff Morgan

Jeff Morgan

    Fly Me to the Moon

  • *****
  • Posts: 5788
  • Joined: 28 Sep 2003
  • Loc: Prescott, AZ

Posted 08 October 2013 - 10:26 AM

StellarCAT (maker of ServoCAT) has floated the concept of 30K encoders as a replacement to the standard 10K units. Provided good 2-star alignment and a well-made mount, the improvement in pointing accuracy would be 9 times that realized by the standard 10K encoders. Of course they not be cheap.

It would seem that such a jump would be a major time saver for centering small targets, especially small targets in rich fields. Milky Way planetary nebula and doubles come immediately to mind, although faint galaxies hunting would probably be a good application too.

Are there any mount owners (any type) using 30K encoders, and does the increase in resolution make a difference for you?

#2 EFT

EFT

    Vendor - Deep Space Products

  • *****
  • Vendors
  • Posts: 2564
  • Joined: 07 May 2007
  • Loc: Phoenix, AZ

Posted 08 October 2013 - 01:03 PM

With standard encoders (i.e., not absolute encoders) it's really a question of at what point does the resolution of the encoders exceed the resolution of the mount mechanics and the level of magnification you are using. The higher the precision of the mount in the first place, the more use it can make of higher tick encoders. But if the mount is of very high precision, then it may work just fine with lower count encoders, particularly at lower magnifications.

#3 Relativist

Relativist

    Skylab

  • *****
  • Posts: 4297
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 08 October 2013 - 02:14 PM

From a controls perspective, the more accuracy the better (within reason of course) especially when trying to have a system that 'learns' the quirks of a mount.

#4 andysea

andysea

    Surveyor 1

  • *****
  • Posts: 1611
  • Joined: 03 Sep 2010
  • Loc: Seattle, WA

Posted 08 October 2013 - 02:36 PM

Is this for visual observing?

#5 orlyandico

orlyandico

    Fly Me to the Moon

  • *****
  • Posts: 5720
  • Joined: 10 Aug 2009
  • Loc: Singapore

Posted 08 October 2013 - 07:53 PM

the 10k encoders have 2' per tick. so this would give ~ 4' resolution per axis. with all the systematic errors.

Now 4' to 5' pointing is pretty good.. going to 30k encoders would reduce that by maybe a factor of 2 at best. not sure if it would be worth it considering that most dobs aren't that rigid or orthogonal anyway...

#6 Jeff Morgan

Jeff Morgan

    Fly Me to the Moon

  • *****
  • Posts: 5788
  • Joined: 28 Sep 2003
  • Loc: Prescott, AZ

Posted 09 October 2013 - 12:22 AM

Is this for visual observing?


Yes. Currently with Argo Navis (on a ServoCAT dob) I can land a target within the inner half of a field having a diameter of 21 arc minutes 100% of the time in early trials.

While that is good enough for large/bright Messier targets, for the small and/or faint targets in rich fields that still involves referencing the field to a star atlas to get a positive ident. (And perhaps that is my answer right there.)

My thinking on this was that increases in mount accuracy would be a real time-saver in terms of getting on-target, resulting is more objects seen per night. But I was not sure how common the 30K encoders are, so I was wondering if anyone is using them on other mounts and what the results have been.

But perhaps they are too new and there is no experience base with them yet.

#7 Relativist

Relativist

    Skylab

  • *****
  • Posts: 4297
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 09 October 2013 - 12:51 AM

I would take a first step of trying to stiffen things and see if that improves your goto significantly, or at least measure your errors.

#8 EFT

EFT

    Vendor - Deep Space Products

  • *****
  • Vendors
  • Posts: 2564
  • Joined: 07 May 2007
  • Loc: Phoenix, AZ

Posted 09 October 2013 - 12:52 AM

Outside of absolute encoders in some GEMs, I can't recall hearing anyone say that they were using anything above 10K tic encoders. On a good machined alt/az style mount, a 4096 encoder can be very accurate. But a dob mount would have to be quite stiff in many respects to benefit from higher resolution encoders. It's not really the same as a machined alt/az or GEM that is manufactured to a much tighter tolerance than the average dob. If you are already using 10K encoders, I'm not sure that you will get a lot more from 30K encoders which will typically be a fair amount more expensive. If you are using less than 10K encoders then, you might just upgrade to 10K and see how much better your system does (or doesn't) do.

#9 ErikB

ErikB

    Mariner 2

  • -----
  • Posts: 201
  • Joined: 05 Feb 2006
  • Loc: Central Arizona

Posted 11 October 2013 - 12:38 AM

The higher the encoder resolution, the higher the frequency of digital state changes in the optical sensor and in the logic that interprets those changes as scope movement. For various reasons the sensor and logic will have a limit to how high that frequency can be before state changes are missed. If that limit should be too low, you would be forced to slew the scope very slowly in order not to miss pulses and lose control of current position. Reasons for a frequency limit may include filtering to avoid electrical noise from being interpreted as scope movement, and if the pulses are sensed by a microprocessor, that processor may be polling an input only so often and relying on the fact that not more than one state change take place between polls. I have not experienced this problem myself with my 4k encoders, but since 10k has been the upper limit for years, I would be wary of assuming that I could just change to 30k and be able to move the scope as quickly as before.

#10 Jeff Morgan

Jeff Morgan

    Fly Me to the Moon

  • *****
  • Posts: 5788
  • Joined: 28 Sep 2003
  • Loc: Prescott, AZ

Posted 12 October 2013 - 04:48 PM

The higher the encoder resolution, the higher the frequency of digital state changes in the optical sensor and in the logic that interprets those changes as scope movement. For various reasons the sensor and logic will have a limit to how high that frequency can be before state changes are missed. If that limit should be too low, you would be forced to slew the scope very slowly in order not to miss pulses and lose control of current position. Reasons for a frequency limit may include filtering to avoid electrical noise from being interpreted as scope movement, and if the pulses are sensed by a microprocessor, that processor may be polling an input only so often and relying on the fact that not more than one state change take place between polls. I have not experienced this problem myself with my 4k encoders, but since 10k has been the upper limit for years, I would be wary of assuming that I could just change to 30k and be able to move the scope as quickly as before.


Interesting. Gary Myers at StellarCAT is trying to put together a group purchase, I'll have to ask him about the frequency sensitivity.

#11 Startraffic

Startraffic

    Viking 1

  • *****
  • Posts: 802
  • Joined: 12 Feb 2006
  • Loc: Lat. 39.143345, Long. -77.174802

Posted 12 October 2013 - 05:24 PM

Jeff,
Erik is right, there is a limit on how fast you can move before the pulses get missed. I put a set of 50k's on my G11 with a Sky Commander 4. I found that I had to push very slowly or the SC couldn't keep up. I've since moved over to Gemini-1 & had had to slow the slew rate down to 400 from 600 to get a consistent slew. I also have a Gemini-2 on my HGM-200, but currently encoders aren't supported yet. They are supposed to be in the next revision release.

Clear Dark Skies
Startraffic
39.138274 -77.168898

#12 John Carruthers

John Carruthers

    Skiprat

  • *****
  • Posts: 3552
  • Joined: 02 Feb 2007
  • Loc: Kent, UK

Posted 13 October 2013 - 03:32 AM

As Orylandico points out, precision of construction may be a constraint. Using 2 encoders per axis and averaging their readings would give better precision, especially if the axes/encoders are at all eccentric.
Incorporating a 3:1 gear might give a similar effect?

#13 CounterWeight

CounterWeight

    Cosmos

  • *****
  • Posts: 8237
  • Joined: 05 Oct 2008
  • Loc: Palo alto, CA.

Posted 13 October 2013 - 05:27 AM

How about one or two of these DFS series? The quadrature numbers are insane.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics