Strain wave gear PEC limitations


Eric Weiner
 

I seem to recall a series of posts a few months ago where we discussed the possibility of AP designing an ultra portable mount. During that discussion, if I recall correctly, Roland provided us with some information regarding strain wave gears, their limitations, and why producing a PEC for them is problematic. I tried searching for that post but can’t find it. Can someone with better search skills point it out? 

ZWO just announced a new mount, the AM5, and they’re making outrageous claims. 


Cheers,
Eric


dvjbaja
 

40 arc sec peak to peak is outrageous. And given that, strain wave gears have an erratic PE, if one would call it a PE at all. Best results with the RST-135 mount is accomplished with very fast tracking with a guide scope and multi-star guiding to compensate for the erratic peak to peak performance. Cheap products from China are disruptive to the market, but not always the best value for performance.

Buyer beware


Greg McCall
 

Out of curiosity, where did you get the PEC stats from? Is it peculiar to a brand or the motor?
I understand some mounts like RST135E have encoders on RA axis and planewave have encoders on both axis. 

Don’t get me wrong as I don’t doubt what you say and I’m VERY happy with my Mach2, I ask as I want to read up on the technology and learn. 


David
 

40 second peak to peak is not outrageous for a harmonic strain wave mount. It’s typical. I have a Rainbow RST-135, and no it’s not in the same league as my Mach2. BUT, that being said, it guides quite well with 0.5 second guide exposures for focal lengths under 600mm. And it does this for something under 10 lbs and no counterweights. This was with a WO Redcat and the RST-135. And no, it wasn’t guiding on a hot pixel. It’s nice to be able to lift the whole pier, mount, and scope setup which weighs less than 30 lbs.



David

On Dec 26, 2021, at 7:06 PM, dvjbaja <jpgleasonid@...> wrote:

40 arc sec peak to peak is outrageous. And given that, strain wave gears have an erratic PE, if one would call it a PE at all. Best results with the RST-135 mount is accomplished with very fast tracking with a guide scope and multi-star guiding to compensate for the erratic peak to peak performance. Cheap products from China are disruptive to the market, but not always the best value for performance.

Buyer beware






Christopher Erickson
 

For those that want to discuss the ZWO AM5 mount, there is a new groups.io  group for doing that.

https://groups.io/g/ZWO-Astro

-Christopher Erickson
Observatory engineer
Waikoloa, HI 96738
www.summitkinetics.com
   


On Sun, Dec 26, 2021, 5:46 PM David via groups.io <phrosty5=yahoo.com@groups.io> wrote:
40 second peak to peak is not outrageous for a harmonic strain wave mount.  It’s typical.  I have a Rainbow RST-135, and no it’s not in the same league as my Mach2.  BUT, that being said, it guides quite well with 0.5 second guide exposures for focal lengths under 600mm.  And it does this for something under 10 lbs and no counterweights.  This was with a WO Redcat and the RST-135.  And no, it wasn’t guiding on a hot pixel. It’s nice to be able to lift the whole pier, mount, and scope setup which weighs less than 30 lbs.



David

> On Dec 26, 2021, at 7:06 PM, dvjbaja <jpgleasonid@...> wrote:
>
> 40 arc sec peak to peak is outrageous.  And given that, strain wave gears have an erratic PE, if one would call it a PE at all.  Best results with the RST-135 mount is accomplished with very fast tracking with a guide scope and multi-star guiding to compensate for the erratic peak to peak performance.  Cheap products from China are disruptive to the market, but not always the best value for performance. 
>
> Buyer beware
>
>
>
>
>
>






Eric Weiner
 

As the originator, what I was interested in was the info Roland gave us a few months ago regarding the limitations of strain wave gear mounts, and the reason its difficult to correct the for inherent error using a PEC.


Christopher Erickson
 

I think there may be a way to do effective PEC with a strain wave gear. Basically, characterize at least 180° and maybe even all 360° of the gear with a tracking compensation data set. Not a big deal with modern microcontrollers and processors, which have oodles of storage capacity and horsepower. Create the original dataset in a special jig at the factory. Have a decent algorithm and utility for updating portions of it in the field.

Of course if the behavior changes significantly under diff loads, everything could become more interesting.

I have an RST-135E, a Mach1, an 1100GTO and getting a 1600GTO in January. And I just ordered the ZWO mount and when I get it, I'll be tearing it apart and studying it carefully.


-Christopher Erickson
Observatory engineer
Waikoloa, HI 96738
www.summitkinetics.com
   


On Sun, Dec 26, 2021, 7:01 PM Eric Weiner <weinere@...> wrote:
As the originator, what I was interested in was the info Roland gave us a few months ago regarding the limitations of strain wave gear mounts, and the reason its difficult to correct the for inherent error using a PEC.








Roland Christen
 

One problem with very fast guiding is intermittent loss of guide star due to very short guide exposure. Some places you might get bright enough guide stars for a .5 sec exposure. Other places you might get nothing. When I'm guiding my 130 and 160mm refractors with off-axis guiders and Lodestar guide camera, sometimes I have to use 5 second guide exposures to get a bright enough star to guide on. Things aren't any better with CMOS guide cameras.

Rolando

-----Original Message-----
From: dvjbaja <jpgleasonid@...>
To: main@ap-gto.groups.io
Sent: Sun, Dec 26, 2021 7:06 pm
Subject: Re: [ap-gto] Strain wave gear PEC limitations

40 arc sec peak to peak is outrageous.  And given that, strain wave gears have an erratic PE, if one would call it a PE at all.  Best results with the RST-135 mount is accomplished with very fast tracking with a guide scope and multi-star guiding to compensate for the erratic peak to peak performance.  Cheap products from China are disruptive to the market, but not always the best value for performance. 

Buyer beware







--
Roland Christen
Astro-Physics


Roland Christen
 

The specs for the high end strain wave gears list the periodic error as being 90 arc seconds max. So it's possible to get that large an error, although most might be less. However, it's not the max error that's a problem. The real problem is the slope of the error. If the slope is 1 arc second per second, you would need to guide much faster than once per second to keep the stars round.

Rolando

-----Original Message-----
From: Greg McCall <emailgregnow@...>
To: main@ap-gto.groups.io
Sent: Sun, Dec 26, 2021 7:41 pm
Subject: Re: [ap-gto] Strain wave gear PEC limitations

Out of curiosity, where did you get the PEC stats from? Is it peculiar to a brand or the motor?
I understand some mounts like RST135E have encoders on RA axis and planewave have encoders on both axis. 
Don’t get me wrong as I don’t doubt what you say and I’m VERY happy with my Mach2, I ask as I want to read up on the technology and learn. 

--
Roland Christen
Astro-Physics


Roland Christen
 

How many arc seconds was each pixel?

Rolando

-----Original Message-----
From: David via groups.io <phrosty5@...>
To: main@ap-gto.groups.io
Sent: Sun, Dec 26, 2021 7:46 pm
Subject: Re: [ap-gto] Strain wave gear PEC limitations

40 second peak to peak is not outrageous for a harmonic strain wave mount.  It’s typical.  I have a Rainbow RST-135, and no it’s not in the same league as my Mach2.  BUT, that being said, it guides quite well with 0.5 second guide exposures for focal lengths under 600mm.  And it does this for something under 10 lbs and no counterweights.  This was with a WO Redcat and the RST-135.  And no, it wasn’t guiding on a hot pixel. It’s nice to be able to lift the whole pier, mount, and scope setup which weighs less than 30 lbs.



David

> On Dec 26, 2021, at 7:06 PM, dvjbaja <jpgleasonid@...> wrote:
>
> 40 arc sec peak to peak is outrageous.  And given that, strain wave gears have an erratic PE, if one would call it a PE at all.  Best results with the RST-135 mount is accomplished with very fast tracking with a guide scope and multi-star guiding to compensate for the erratic peak to peak performance.  Cheap products from China are disruptive to the market, but not always the best value for performance. 
>
> Buyer beware
>
>
>
>
>
>






--
Roland Christen
Astro-Physics


Roland Christen
 

With a strain wave gear you can correct the average periodic error to some extent. Problem is that one PE curve does not fit all the errors because the error is a variable from one tooth to the next. If you measure 5 teeth you will get a PE curve that won't fit the next 5 teeth. So, you might be able to bring the PE down by half the original value at best. It could also make the PE worse at some other point on the gear angle.

Placing a high resolution encoder on the output shaft also does not completely eliminate the PE because the output of a strain wave gear wobbles back and forth between each tooth contact. It's a tiny cycloidal motion that gets translated into a runout error in the encoder output. You will still have to guide, although it allows for longer guide exposures versus a non-encoder strain wave mount.

One thing to keep in mind: high end strain wave gears are very expensive. They generally use expensive dual pre-loaded bearings that provide smooth rolling action. They have to be extremely precise bearings because they have no shaft length. They are placed back to back and heavily pre-loaded. There are cheap strain wave gears also available from other Asian sources (approximately 1/3 to 1/2 the cost of the Japanese ones). I would immediately question how "bumpy is the ride". In other words, is the tracking smooth or are there moment to moment glitches and wiggles in the tracking, which would indicate cheaper bearings. Might work ok where the image scale is greater than the wiggles.

Another place where wiggles originate is in the stepper drives. Open loop steppers have repeating errors on the order of several arc seconds that occur fast enough to mess up high resolution images, and they can't be guided out. They occur so fast that you can only see them at video frame rates. A typical 2 second guide exposure doesn't show it at all, but the stars are nevertheless wiggling back and forth in the Ra direction.

Rolando

-----Original Message-----
From: Eric Weiner <weinere@...>
To: main@ap-gto.groups.io
Sent: Sun, Dec 26, 2021 9:01 pm
Subject: Re: [ap-gto] Strain wave gear PEC limitations

As the originator, what I was interested in was the info Roland gave us a few months ago regarding the limitations of strain wave gear mounts, and the reason its difficult to correct the for inherent error using a PEC.








--
Roland Christen
Astro-Physics


Roland Christen
 

Forgot one thing: these strain wave gears are not made with stainless steel and the flexible cup gears will eventually rust out.

Rolando

-----Original Message-----
From: Roland Christen via groups.io <chris1011@...>
To: main@ap-gto.groups.io <main@ap-gto.groups.io>
Sent: Mon, Dec 27, 2021 11:56 am
Subject: Re: [ap-gto] Strain wave gear PEC limitations

With a strain wave gear you can correct the average periodic error to some extent. Problem is that one PE curve does not fit all the errors because the error is a variable from one tooth to the next. If you measure 5 teeth you will get a PE curve that won't fit the next 5 teeth. So, you might be able to bring the PE down by half the original value at best. It could also make the PE worse at some other point on the gear angle.

Placing a high resolution encoder on the output shaft also does not completely eliminate the PE because the output of a strain wave gear wobbles back and forth between each tooth contact. It's a tiny cycloidal motion that gets translated into a runout error in the encoder output. You will still have to guide, although it allows for longer guide exposures versus a non-encoder strain wave mount.

One thing to keep in mind: high end strain wave gears are very expensive. They generally use expensive dual pre-loaded bearings that provide smooth rolling action. They have to be extremely precise bearings because they have no shaft length. They are placed back to back and heavily pre-loaded. There are cheap strain wave gears also available from other Asian sources (approximately 1/3 to 1/2 the cost of the Japanese ones). I would immediately question how "bumpy is the ride". In other words, is the tracking smooth or are there moment to moment glitches and wiggles in the tracking, which would indicate cheaper bearings. Might work ok where the image scale is greater than the wiggles.

Another place where wiggles originate is in the stepper drives. Open loop steppers have repeating errors on the order of several arc seconds that occur fast enough to mess up high resolution images, and they can't be guided out. They occur so fast that you can only see them at video frame rates. A typical 2 second guide exposure doesn't show it at all, but the stars are nevertheless wiggling back and forth in the Ra direction.

Rolando

-----Original Message-----
From: Eric Weiner <weinere@...>
To: main@ap-gto.groups.io
Sent: Sun, Dec 26, 2021 9:01 pm
Subject: Re: [ap-gto] Strain wave gear PEC limitations

As the originator, what I was interested in was the info Roland gave us a few months ago regarding the limitations of strain wave gear mounts, and the reason its difficult to correct the for inherent error using a PEC.








--
Roland Christen
Astro-Physics

--
Roland Christen
Astro-Physics


Eric Weiner
 

Roland,

Thank you for those details.  I find all of the mount discussions both interesting and educational.

Regards,
Eric


dvjbaja
 

The tracking error graphs were published by ZWO on their website.  ZWO has not published anything about periodic error correction.   


On Sun, Dec 26, 2021 at 5:41 PM Greg McCall <emailgregnow@...> wrote:

Out of curiosity, where did you get the PEC stats from? Is it peculiar to a brand or the motor?
I understand some mounts like RST135E have encoders on RA axis and planewave have encoders on both axis. 

Don’t get me wrong as I don’t doubt what you say and I’m VERY happy with my Mach2, I ask as I want to read up on the technology and learn. 


Roland Christen
 

If you go to the strain wave manufacturer, they publish the specs and they are twice that high. The published specs are of course the max allowable, and typically they could be better.

Rolando

-----Original Message-----
From: dvjbaja <jpgleasonid@...>
To: main@ap-gto.groups.io
Sent: Mon, Dec 27, 2021 1:25 pm
Subject: Re: [ap-gto] Strain wave gear PEC limitations

The tracking error graphs were published by ZWO on their website.  ZWO has not published anything about periodic error correction.   

On Sun, Dec 26, 2021 at 5:41 PM Greg McCall <emailgregnow@...> wrote:
Out of curiosity, where did you get the PEC stats from? Is it peculiar to a brand or the motor?
I understand some mounts like RST135E have encoders on RA axis and planewave have encoders on both axis. 
Don’t get me wrong as I don’t doubt what you say and I’m VERY happy with my Mach2, I ask as I want to read up on the technology and learn. 

--
Roland Christen
Astro-Physics


Christopher Erickson
 

Besides my 1100, Mach1 and soon to be 1600 (which I all love,) I also have an RST-135E (which I also love) and I have ordered the ZWO AS5 directly from their web site. They charged me immediately and are saying shipments will start in April. I dissected and documented my first RST-135 in macro pics and I'll do the same for the AS5.

My past experience with ZWO tells me to say to hold off getting excited about the AS5 until I have studied it extensively internally and under the stars.

-Christopher Erickson
Observatory engineer
Waikoloa, HI 96738
www.summitkinetics.com
   


On Mon, Dec 27, 2021, 11:41 AM Roland Christen via groups.io <chris1011=aol.com@groups.io> wrote:
If you go to the strain wave manufacturer, they publish the specs and they are twice that high. The published specs are of course the max allowable, and typically they could be better.

Rolando

-----Original Message-----
From: dvjbaja <jpgleasonid@...>
To: main@ap-gto.groups.io
Sent: Mon, Dec 27, 2021 1:25 pm
Subject: Re: [ap-gto] Strain wave gear PEC limitations

The tracking error graphs were published by ZWO on their website.  ZWO has not published anything about periodic error correction.   

On Sun, Dec 26, 2021 at 5:41 PM Greg McCall <emailgregnow@...> wrote:
Out of curiosity, where did you get the PEC stats from? Is it peculiar to a brand or the motor?
I understand some mounts like RST135E have encoders on RA axis and planewave have encoders on both axis. 
Don’t get me wrong as I don’t doubt what you say and I’m VERY happy with my Mach2, I ask as I want to read up on the technology and learn. 

--
Roland Christen
Astro-Physics