Re: Mach 1 and Absolute Encoders


Michael Fulbright <mike.fulbright@...>
 

It is interesting but with PEM on my present Mach1 giving me < 1 arcsec PE, and the unlikelihood I'd ever spend time building a model (presumably required to take advantage of the 2 axis encoders?) each time I setup and breakdown the next day it wouldn't be something I'd invest in.

If I had a permanent setup I would have bought an AP1100 + encoders instead of a Mach1.

Michael Fulbright


On 1/24/2018 7:06 PM, chris1011@... [ap-gto] wrote:
 


Well, when you put it this way -- um -- don't think so!
Anyone else have anything to add? What would you pay over and above the basic mount price, and what functionality (1 or 2 axis control?)

Rolando


-----Original Message-----
From: 'Woody Schlom' woody@... [ap-gto]
To: ap-gto
Sent: Wed, Jan 24, 2018 5:48 pm
Subject: RE: [ap-gto] Mach 1 and Absolute Encoders



Roland,
 
Well, when you put it this way -- um -- don't think so!
 
Thanks so much for going into so much detail.  It explains a lot of things.
 
Woody
 
-----Original Message-----
From: ap-gto@... [mailto:ap-gto@...]
Sent: Wednesday, January 24, 2018 3:37 PM
To: ap-gto@...
Subject: Re: [ap-gto] Mach 1 and Absolute Encoders

 
That all depends if there is any market for this.

Here's what we're up against. Yes, you can add a low cost optical encoder to a mount to tame the periodic error. Or you can add a more expensive absolute encoder which would make the mount foolproof so you would always know where it is pointing. The cost of a decent million+ bit encoder is going to run the manufacturer $500, absolute encoders run in excess of $800, and that does not include installation and testing labor or the parts needed to attach it to the axis. By the time you're done, you have increased the consumer price of the mount some $1200 - $1800 per axis.

Second problem is that just about all encoders that I have studied will have 1 million bits derived from a 10K to 40K discrete ticks on the optical disc. This means the high resolution is achieved via interpolation. It just so happens that typical interpolation errors (sub-divisional error or SDE) will run on the order of +-2 to +- 8 arc seconds. These errors are fast moving, occurring on a time scale of 5 to 15 seconds, and are difficult or impossible for a guider to guide out. Yes, you can have an overall RMS tracking error of 0.2 to 0.5 arc seconds, but the peaks will still be there and the stars will bobble back and forth by the amount of SDE that the encoder produces. For that reason I have ruled out any and all low cost or even medium cost encoders that are presently available. One could theoretically map out the SDE in a PE memory of sorts and then play it back in opposite direction, but the cancellation is never perfect and can produce more problems than it's worth.
For our 1100/1600 mounts we use the Renishaw RESOLUTE absolute encoder with a ring that produces 2^26 individual addresses around the periphery, which means we have 67 million counts to work with. The encoders are guaranteed to have SDE of less than +-0.1 arc seconds. Each address/count represents a shaft position of .0193 arc seconds. Therefore we can control the actual telescope position to better than 1/50 arc second with less than 1/5 arc sec of back-forth motion. The Renishaw encoder system is the gold standard and nobody else comes close. They are of course expensive, and we are always trying to work out a better deal with the manufacturer. The problem is they won't drop the price without a large order, and we can't order a large number unless there is higher demand. Typical chicken and egg.

One nice thing about the Renishaw rings is they are self contained and have only two parts. The rings allow thru the mount wiring and use of polar scopes for quick setup. They are totally immune to dirt and contamination, and are easy to set up and adjust, even for a user.

So, the question is, how much demand would there be for 1 axis or full 2 axis control of a Mach1?

1 axis would basically give you error-free tracking, 2 axis control would produce a mount that you could always know where it is pointed and allow full remote operation with sub-arc sec tracking accuracies of both axes.

Roland Christen


-----Original Message-----
From: geoffrey.cook@... [ap-gto] gto@...>
To: ap-gto gto@...>
Sent: Wed, Jan 24, 2018 3:59 pm
Subject: [ap-gto] Mach 1 and Absolute Encoders



Is Astro-Physics planning to offer absolute encoders to go with the Mach 1 in the near future?





Join main@ap-gto.groups.io to automatically receive all group messages.