Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Absolute Encoders vs. "Pre" Absolute Encoders - How much better?

  • Please log in to reply
46 replies to this topic

#1 BarrySimon615

BarrySimon615

    Pa Bear

  • *****
  • topic starter
  • Posts: 4309
  • Joined: 01 Mar 2004
  • Loc: New Orleans, LA

Posted 13 September 2019 - 08:27 AM

Back in the day we just called encoders "Encoders".  My G11 and G8 have encoders - I know because I installed them.  My Light Speed Wagon alt-az mount has encoders too that nicely give me sky coordinates and will do what the G11 and G8 encoders and display unit will do - guide me to objects.  Newer entry consumer level mounts also have what I believe we can call encoders as the hand unit displays can go to objects given their addresses in the sky.

 

So now enter Absolute Encoders and we cannot simply just use the word "Encoder" any more, we have to distinguish between the two.  My question is - How many people really need "Absolute Encoders"?  Are they really an accessory that will be valuable to all, or are they a high end item that can only be really fully utilized by a very few dedicated astrophotographers?

 

Answers to these questions will be informative for me, and perhaps several others.  I will admit that when it comes to a full understanding of the benefits of absolute encoders I am pretty naive.

 

Barry Simon



#2 psandelle

psandelle

    Mercury-Atlas

  • *****
  • Posts: 2790
  • Joined: 18 Jun 2008
  • Loc: Los Angeles

Posted 13 September 2019 - 08:51 AM

Absolute encoders tell the mount where it is...absolutely. And the mount corrects with that in mind. If a gust of wind uses the telescope as a sail...the mount corrects back (real quick; and we're talking where a guide star would be lost). If you take off the camera, add a spacer, put the camera back on, then do it a second time (as I did once), the star in the absolute center of the field will STILL be there absolutely centered (as I said, I did this). So, if you add an all-sky model of 25 to 40 points, you can track (within FL, say, up to 4000mm, which I've done, and rigidity, can't have flex or mirror flop while modelling)...unguided. Super-accurately. Clouds can go by, and it's still tracking accurately, too (for those in more cloudy climes). With Protrack on Paramounts and APCC Pro on APs, you can model and go unguided without them...but not as accurately (I've compared them).

 

If you're guiding, you don't need 'em as much, or at all.

 

So, for most imaging for most systems, most people really don't need them. But if you want to do unguided (or barely have to guide at longer focal lengths where things get tricky), they're great. And there are other reasons (it never gets lost, even if the power goes down in your remote obsy in Chile) as well, but that's the big one for me.

 

Paul


  • Stelios likes this

#3 TimK

TimK

    Vostok 1

  • -----
  • Posts: 161
  • Joined: 21 Jan 2006
  • Loc: Avon, OH

Posted 13 September 2019 - 09:14 AM

"Absolute" encoders send position information to the controller versus sending "tics" as the encoder rotates.

This does not in itself make the mount more accurate or more repeatable.

 

What is does allow is for the mount to be aware of its "absolute" position immediately after the mount is turned on without the need to reference or re-align the scope.  Assuming he clock is right.

Of course this all depends on how the controller is setup.



#4 bobzeq25

bobzeq25

    Hubble

  • *****
  • Posts: 16828
  • Joined: 27 Oct 2014

Posted 13 September 2019 - 09:22 AM

This also can be confusing because sometimes people use them term "absolute encoders" when their point is about "precision encoders", that can be used (with appropriate computer modeling) as an alternative to autoguiding.  I believe most (all?) precision encoders are also absolute encoders, but they're really two different properties of an encoder.  An absolute encoder remembers position, a precision encoder measures position very precisely.  It's not a requirement that the two properties are linked.

 

CN needs a glossary.  <smile>


Edited by bobzeq25, 13 September 2019 - 09:27 AM.


#5 Arie

Arie

    Ranger 4

  • -----
  • Posts: 361
  • Joined: 01 Sep 2015
  • Loc: Netherlands

Posted 13 September 2019 - 09:36 AM

The older ASA mounts had relative encoders. Very accurate though, but at the beginning of a session you had to tell it where zero is every time.

The modern ones, alike the 10Micron mounts, have absolute encoders on the actual RA and DEC axis.

With these indeed the mount knows at all times where it is pointing at. Even if you give it a manual swing (slipping through the clutches) or power it off ad back on again.

Obviously if you hit the tripod, it's all lost.

These encoders are a different class from the Ioptron or AP encoders....


  • psandelle likes this

#6 rgsalinger

rgsalinger

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 5020
  • Joined: 19 Feb 2007
  • Loc: Carlsbad Ca

Posted 13 September 2019 - 09:46 AM

Actually an absolute encoder system reads the position of the system at start up. Nothing is remembered. In order to do that the actual positional encoding method is different than just scribing marks on a wheel. 

 

This is pretty easy reading about the matter. 

 

Rgrds-Ross


  • psandelle and Der_Pit like this

#7 WadeH237

WadeH237

    Aurora

  • *****
  • Posts: 4686
  • Joined: 24 Feb 2007
  • Loc: Snohomish, WA

Posted 13 September 2019 - 09:48 AM

So now enter Absolute Encoders and we cannot simply just use the word "Encoder" any more, we have to distinguish between the two. 

There are more than two types of encoders in astronomy :)

 

To address your specific question, I think that context is important.

 

If you are adding push-to or goto to an existing mount, you will need encoders.  These are relatively low resolution and low cost, and are good for finding objects in the sky.  If you are trying to improve the tracking of a mount, then you might use high resolution encoders.  These range from high cost, to really high cost and have precision that can range into tiny fractions of an arc second. They are mainly of benefit to people looking for very high precision, which is typically deep sky astrophotographers.

 

I'm not going to go deep into the difference between absolute and relative encoders, except to say that not all high resolution encoders are absolute encoders.  As such, using the terms "encoder" vs "absolute encoder" does not differentiate between the two main use cases above.

 

And finally, encoders are sometimes associated with unguided imaging, but they are really two different things.  Robust unguided imaging requires both tracking precision, and variable tracking rates to compensate for various errors that don't involve the mount.  The compensation is done through software that samples a bunch of points in the sky to build a tracking model.  The precision required can be achieved through (absolute or relative) high resolution encoders, but it could also be achieved through mechanical precision without additional encoders.

 

So to sum up, there really isn't a single term that disambiguates which use of the word "encoder" is meant.  You need some context to know for sure.


  • Tonk, EFT, bobzeq25 and 1 other like this

#8 OldManSky

OldManSky

    Surveyor 1

  • *****
  • Posts: 1709
  • Joined: 03 Jan 2019
  • Loc: Valley Center, CA USA

Posted 13 September 2019 - 10:43 AM

Just to clarify something in the OP...

 

Most entry-level GOTO mounts don't have encoders.  Of any kind.

They GOTO because they use stepper motors to drive the axes.  Once they're star-aligned, they GOTO objects by moving the motors a defined number of steps.  Which works great, as long as neither of the axes slips or stalls.  If they had encoders, with slipping/stalling the encoder feedback would let the computer know that even though a step-the-motor command was issued, the axis didn't move as it was supposed to, and it could correct for that.  But they don't have 'em, so stepper motor steps is what they've got to work with.


  • Der_Pit likes this

#9 Lola Bruce

Lola Bruce

    Apollo

  • *****
  • Posts: 1168
  • Joined: 22 Sep 2014
  • Loc: Southern California

Posted 13 September 2019 - 11:39 AM

So much miss information and good mixed with assumptions. In machine positioning the terms incremental and absolute are used . Incremental is from position to position. Absolute is from assigned zero to each or any position.

Our mounts are machine positioning devices, but being that we are special we do not follow convention. All of my powered mounts have had encoders. Celestron cgem, celestron SE, Celestron CPC, Losmandy G11 and GM8, and AstroPysics Mach 1. All of these have tick counters on the motors (Yes true encoders). Only the Mach 1 claims to know where it is at any time it is on. On all of these mounts if you loosen the clutches and run the servos the mounts report axis movement that is not true. On what we call absolute encoders the devices are mounted to the components of the axis it's self. As a result if the clutches are disengaged and the servos set in motion the mount does not report movement as the axis has not moved. These so called absolute encoders can be augmented with limit switches or reference ticks thus eliminating the need for setting a zero position. Contrary to some opinions these axis hard mounted encoders also count ticks (the days of analog pots are long dead). These if high quality are immune to manual movement, gear lash, over speed slews, and a host of other position error reporting. They can not give reliable position information to the target on a sloppy mount, a sloppy tripod, or other flexure problems. Even though they are of great precision, on a quality mount, and on a wind less night they are not truly absolute as they can not compensate for variations between the  mount and the target. That is what guiding is for.

 

Spend as much as you wish for accuracy but the image in the sky is prone to capricious movement.

 

Bruce


  • psandelle, Wildetelescope and OldManSky like this

#10 555aaa

555aaa

    Vendor (Xerxes Scientific)

  • *****
  • Vendors
  • Posts: 1400
  • Joined: 09 Aug 2016
  • Loc: Lynnwood, WA, USA

Posted 13 September 2019 - 12:05 PM

You can have incremental encoders on either the clutched or unclutched side of the mount. Absolute encoder mounts should not require limiting or homing switches or sensors.

Edited by 555aaa, 13 September 2019 - 12:07 PM.

  • Tonk and EFT like this

#11 dr.who

dr.who

    Voyager 1

  • *****
  • Administrators
  • Posts: 13812
  • Joined: 05 Jan 2012

Posted 13 September 2019 - 12:46 PM

Bruce is spot on. Furthermore absolute encoders have been around for quite a while just not in our hobby. Lastly to give you a (somewhat) short answer: absolute encoders are something that benefits an astrophotographer significantly in terms of mount positioning and accuracy. They do not replace the need for good polar alignment. So if you are heavy into AP and can afford the $7,000+ entry price (the upcoming AP Mach2 starts at this price I believe) then it will benefit you. If you are not that into AP or if that price is not realistic for you, a lower tier lower cost mount with good polar alignment and good guiding will do fine. Lower tier mounts, in spite of the irrational exuberance and rabid evangelism by some reporting parties, with certain encoders have a potential oscillation issue that leads to suboptimal shaped stars  because of rapid overcorrection. Thus it is actually better to save the money and not get them this dos not appear with shorter focal length scopes.



#12 Lola Bruce

Lola Bruce

    Apollo

  • *****
  • Posts: 1168
  • Joined: 22 Sep 2014
  • Loc: Southern California

Posted 13 September 2019 - 03:44 PM

Note yes encoders can be on either side of the clutches and are. But the term absolute encoders as used in astronomy dictates they must be fixed to the axis components and report true position regardless of anything except lack of power. Well designed and implemented ones you can put the mount in any position (not a crash) then turn on power and by seeking a marker it can know where it is exactly.

 

Bruce

 

Note: Part of my profession was designing and or building one off machines. I always preferred axis mounted encoders but time and money more often than not dictated the use of encoders on the device providing the power.


  • psandelle likes this

#13 555aaa

555aaa

    Vendor (Xerxes Scientific)

  • *****
  • Vendors
  • Posts: 1400
  • Joined: 09 Aug 2016
  • Loc: Lynnwood, WA, USA

Posted 13 September 2019 - 04:04 PM

If the mount has to move to determine its position then it doesn't have absolute encoders.
  • Tonk and EFT like this

#14 Lola Bruce

Lola Bruce

    Apollo

  • *****
  • Posts: 1168
  • Joined: 22 Sep 2014
  • Loc: Southern California

Posted 13 September 2019 - 06:15 PM

Machine world yesish but we in astronomy don't conform to a standard. The CNC equipment that I dealt with when woke up had to do a position finding sub routine. With markers integrated into the encoders they moved very little. With systems like machine centers they liked to use a probe and a high precision reference point and they are absolute based machines. Position known from a fixed reference (0,0,0) point. All of these can be parked at a known position and revived taking up where they left off. But if you power them down and manually move them you have to perform a re-zero.

 

Bruce



#15 WadeH237

WadeH237

    Aurora

  • *****
  • Posts: 4686
  • Joined: 24 Feb 2007
  • Loc: Snohomish, WA

Posted 13 September 2019 - 07:47 PM

Whether an encoder is absolute or relative is a physical characteristic of the encoder.  It does not matter how the encoder is installed or used.  For example, the add-on encoders that Astro-Physics uses are Renishaw absolute encoders.  They have what are effectively microscopic codes on the encoder ring that provide 62 million unique addresses around the ring.  The read head can determine its position over the encoder ring without have to count ticks.  They do not get lost no matter what you do, including power cycling.  You do not need to re-zero them.  Ever.*

 

The encoders are installed on AP1100 and AP1600 mounts by attaching to the worm wheel.  As long as the clutches remain locked, the mount cannot get lost.  If you release the clutches and move the axes, the worm wheel does not turn, so the encoders don't "see" that movement.  The new Mach2 installs the encoders on the axes themselves and will "see" movement, regardless of whether the clutches are locked or released.  The Mach2 cannot get lost no matter what you do, including releasing the clutches.

 

The Renishaw encoders are well documented by Renishaw, so there is no need for speculation (I have tried to summarize what's already been thoroughly discussed regarding their use with an Astro-Physics mount).  The encoders used in iOptron EC and EC2 mounts are relative encoders and have different characteristics.  I am not aware of the manufacturer and don't remember the specs (although some of the specs have been discussed on Cloudy Nights).  The encoders used on the 10Micron mounts are proprietary to 10Micron, and I'm not sure how much technical information that they provide.

 

I am not familiar with the machining world or language at all.  I only know what I know about the encoders used on the Astro-Physics mounts because this stuff has been hashed out a number of times, both here on Cloudy Nights, and also on the Astro-Physics user group, complete with references to the Renishaw data sheets.  I also know the behavior of the Renishaw absolute encoders on an Astro-Physics mount because I own an absolute encoder-equipped AP1600.

 

I'm sorry that this has gone way, way beyond the point that OP raised at the start of the thread, but there are lots of statements above that are just plain wrong - specifically in the context of high resolution encoders used on astronomical mounts.

 

* Note that Astro-Physics does have a utility that is used to "home" the encoders.  This is not because the encoders require it.  It is because of how the encoders don't "see" movement of the axes with the clutches released.  The purpose is to establish the position of the axes relative to the worm wheel (the worm wheel is tracked by the encoder).  This utility does not apply to the Mach2 because the encoders are directly attached to the axes, as above.


  • EFT likes this

#16 OzAndrewJ

OzAndrewJ

    Mercury-Atlas

  • *****
  • Posts: 2819
  • Joined: 30 Nov 2010

Posted 14 September 2019 - 03:37 AM

Gday Ed

but absolute encoders are mostly all high-precision (although it is entirely possible make absolute encoders of any resolution needed).

Whilst i agree with the fundamentals here, it really needs to be reduced to a simpler description, that is independent of resolution.

The use of incremental encoders doesnt change based on how they are fitted.

The use of absolute encoders can change, esp when a clutch is involved.

ie

a) All relative encoders need to be "zeroed" in some way when initialised.

b) If an absolute encoder is in use, and the encoder can move relative to the Saddle plate ( due to clutches ), then it needs a means to be zeroed.

c) If the absolute encoder is fixed to the output axle/saddle plate, and isnt affected by clutches, then once set in the factory, it should never need zeroing, it always knows where its output axle is pointing.

 

That said, the home sensor in the Synta mounts is technically an absolute encoder that complies with definition c)

but its accuracy is limited to only 2 positions :-)

All good fun

 

Andrew Johansen Melbourne Australia


  • psandelle likes this

#17 Der_Pit

Der_Pit

    Viking 1

  • -----
  • Posts: 806
  • Joined: 07 Jul 2018
  • Loc: La Palma

Posted 14 September 2019 - 09:25 AM

If the mount has to move to determine its position then it doesn't have absolute encoders.

This is wrong.

If you switch on they have to calibrate, usually in finding an index marker.  Only after that their reported position is guaranteed to be correct.



#18 rgsalinger

rgsalinger

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 5020
  • Joined: 19 Feb 2007
  • Loc: Carlsbad Ca

Posted 14 September 2019 - 09:37 AM

Actually, if you read the Renishaw literature there is no mention that movement is necessary for their encoders to figure out their position. That's consistent with my experience with this brand of encoders. If I lose power on the mount that I use that has them (not an AP mount) then all I have to do is get the power back on. Period. Even though the mount is not tracking, the planetarium shows and accurate RA/DEC. Now I cannot prove that there is absolutely no movement but with the tracking off, it's hard to see how any can occur. 

 

The AP mounts with absolute encoders seem strange to me because they lose the advantage that absolute encoders have over relative encoders. Still, I can tell you that using two AP1600's one with and one without, the tracking was clearly better with the encoder mount. 

 

Rgrds-Ross


  • psandelle likes this

#19 Der_Pit

Der_Pit

    Viking 1

  • -----
  • Posts: 806
  • Joined: 07 Jul 2018
  • Loc: La Palma

Posted 14 September 2019 - 09:52 AM

Actually, if you read the Renishaw literature there is no mention that movement is necessary for their encoders to figure out their position. That's consistent with my experience with this brand of encoders. If I lose power on the mount that I use that has them (not an AP mount) then all I have to do is get the power back on. Period. Even though the mount is not tracking, the planetarium shows and accurate RA/DEC. Now I cannot prove that there is absolutely no movement but with the tracking off, it's hard to see how any can occur.


Hmm, I now our Heidenhahn absolute encoders definitely do need this.  Don't know the Renishaws, maybe they have a completely different working principle.  But it of course depends on the density of index marks.  Might be the mount does this automatically when powering on, and the needed movement is so little you don't notice it?


  • psandelle likes this

#20 psandelle

psandelle

    Mercury-Atlas

  • *****
  • Posts: 2790
  • Joined: 18 Jun 2008
  • Loc: Los Angeles

Posted 14 September 2019 - 10:05 AM

Hmm, I now our Heidenhahn absolute encoders definitely do need this.  Don't know the Renishaws, maybe they have a completely different working principle.  But it of course depends on the density of index marks.  Might be the mount does this automatically when powering on, and the needed movement is so little you don't notice it?

With the 10Microns I never saw/heard anything move upon power-up (like Ross), but, again, "if an encoder moves in the forest so small amount you don't notice, did it actually move?" Unless a mount/encoder company fills us in, the world may never know.

 

As to Ed: yeah, I sometimes forget it's not just "absolute" encoders, but the ones I've used tend to be "high-resolution" as well. The mounts I've had that have them are kinda one-and-the-same in that regard, so to me "absolute" means "high-resolution" as well. It's a really good point though (and I won't go back and correct my earlier post, just to look like I know what I'm talking about grin.gif, but I acknowledge that I should have mentioned it).

 

Still, I think a key to the OP's question is: if you want to do unguided, high-rez/absolute encoders coupled with a well-designed sky-modelling software make it really easy and fun, but if you're guiding and don't want to spend the extra bucks...you really don't need them, BUT, as Ross pointed out, he can see just a bit of difference with the absolute/high-rez encodered mount being that bit better at guiding. That's a point for me: with the abso/high-rez encoders, you can do guided OR unguided, and without, unguided is not so easy/accurate. Best of both worlds, if you have the dough.

 

Paul



#21 chadrian84

chadrian84

    Ranger 4

  • *****
  • Posts: 325
  • Joined: 03 Jun 2018
  • Loc: Wisconsin

Posted 14 September 2019 - 10:15 AM

I’m still waiting for someone to test guided performance in phd2 by going back and forth with encoders on vs. off. I’ve seen zero evidence encoders improve guiding. I don’t see any improvement in phd2 with a near perfect PEC curve enabled vs. disabled. It seems to me that seeing is far and away the limiting factor in how well your mount can guide.

#22 psandelle

psandelle

    Mercury-Atlas

  • *****
  • Posts: 2790
  • Joined: 18 Jun 2008
  • Loc: Los Angeles

Posted 14 September 2019 - 10:32 AM

I’m still waiting for someone to test guided performance in phd2 by going back and forth with encoders on vs. off. I’ve seen zero evidence encoders improve guiding. I don’t see any improvement in phd2 with a near perfect PEC curve enabled vs. disabled. It seems to me that seeing is far and away the limiting factor in how well your mount can guide.

I'm not sure you can turn the encoders off on most of the mounts I've used. And, again, if you're guiding you don't "absolutely" need encoders by any stretch (and having a near-perfect PEC curve and excellent PA and a finely machined mount, you probably can't tell the difference - I only point out Ross' observation, but encoders will cover a multitude of sins if you DON'T have a perfect PEC curve or excellent PA, etc.). To me, they're for other things, but they can help if it gets windy, cloudy, things that a guider won't necessarily recover from. And, of course, unguided. Most people don't need them for the type of imaging they're doing, but they do have their uses.

 

Paul


Edited by psandelle, 14 September 2019 - 10:34 AM.

  • Der_Pit likes this

#23 Lola Bruce

Lola Bruce

    Apollo

  • *****
  • Posts: 1168
  • Joined: 22 Sep 2014
  • Loc: Southern California

Posted 14 September 2019 - 10:39 AM

If you stop, turn it off, don't move it, and turn it back on it picks up from the last position. As precise as the mount may be it does not error check position with out being commanded to do so.


Edited by Lola Bruce, 14 September 2019 - 10:44 AM.


#24 chadrian84

chadrian84

    Ranger 4

  • *****
  • Posts: 325
  • Joined: 03 Jun 2018
  • Loc: Wisconsin

Posted 14 September 2019 - 10:44 AM

AP mounts can be turned on/off. It’s puzzling why no one has done this simple test on what’s an upsell that nearly doubles the cost of mounts. IMO the null hypothesis should be that encoders do not improve guiding in any meaningful way.

@Bruce. Have them on, select a star, and guide. Turn them off, select the same star, and guide. Do it many times and compare the results.

Edited by chadrian84, 14 September 2019 - 10:48 AM.


#25 Lola Bruce

Lola Bruce

    Apollo

  • *****
  • Posts: 1168
  • Joined: 22 Sep 2014
  • Loc: Southern California

Posted 14 September 2019 - 10:58 AM

A side note. Encoders on the motors may be low resolution per rotation of the motor but due to gear reduction they can be high resolution to the axis rotation. Servo drives that mid and above mounts have provide exceptional torque and positional accuracy for their price. The fact that the objects in the night sky generally march in a fixed direction at an even pace further enhances their value.




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics