Have newer AM5s had a period change; 288s vs the original 432s ?

Curious

Hello everyone, first post.

Over on CloudyNights someone is claiming that the latest AM5 shipments have had a spec change from a period of 432 seconds to 288 seconds.

The product sheet still says 432 seconds and I'm not inclined to believe everything I read so was hoping someone on here could confirm or deny.

Thank you :-)

Corsair

There's a post about it already in these forums: https://bbs.astronomy-imaging-camera.com/d/16233-pe-cycle-of-432-secs

For further information here are links to two posts from two different threads about it from Cloudy Nights:

https://www.cloudynights.com/topic/842943-zwo-am5-owners-club/page-43#entry12862839

https://www.cloudynights.com/topic/880277-improving-am5-mount-guide-performance/page-3#entry12763481

Curious

Corsair sorry I missed that; thank you.

w7ay

ZWO needs to correct this documentation flaw, otherwise it is false advertising.

There are many of us who pay great attention to things like the periods of the gears when it comes to autoguiding.

One reason the period could have been changed (speculation on my part) is that the new strain wave gears exhibited larger peak-to-peak amplitude than the original gears. (The change in gear could also have been caused by various reasons, e.g., inconsistent tracking (glitching), pricing, availability, etc.)

The primary reduction ratio of the strain wave gear has probably not changed (i.e., remained at 100:1)). The 430.82 second number comes when you add a further 2:1 reduction using belts and pullets to make the final 200:1 reduction ratio (the 430.82 number comes from a sideral day, about 23 hours 56 minutes, divided by this 200 number).

The belt and pulley also reduces the peak-to-peak periodic error amplitude of the gear/belt combination to 1/2 of the original p-p error of the strain wave gear itself. So, why not use a 10:1 belt to even further reduce the error? The imprecision of belt/pully systems of course. You are also adding backlash to a strain wave gear that has practically no backlash itself. I.e., strain wave mounts have backlash (read the ads really carefully to not get fooled), it is the strain wave gears that do not have backlash. But a mount can also include belts and pulleys, and other compromises such as stepper motor backlashes, if the mount uses steppers instead of servo motors (the reason why better mounts use servos).

Notice that when you increase the belt and pulley to 3:1, the peak-to-peak of the error curve now goes down by 50%. So, if the strain wave gear itself has a larger p-p amplitude error, you can make the specs on paper look good by using larger belt/pulley ratios. What they don't tell you is that the mount tracking has also become sloppier by increasing the belt/pulley ratio. (Remember the saying, "there are lies, damned lies, and statistics?")

When you change the belt/pulley to 3:1, the total ratio is now 300:1, or a period of 287 solar seconds (corresponds to 288 sidereal seconds). An FFT will show a peak at 287 seconds, not 288 seconds;, just as the FFT of a RainbowAsto RST-135 shows a 430.82 second period, and not 432 seconds.

Notice too that as you increase the belt/pulley ratio, although you have decreased the p-p amplitude of the PE curve (of zero importance when you are autoguiding), you have not changed ∆x/∆t one single bit, and ∆x/∆t is the most important factor to good autoguiding. I.e., you have reduced x, but you have also reduced t, by precisely the same ratio!

So, as far as autoguiding is concerned, you have not improved the performance a single iota, while adding problems from a pulley/belt system, just to make the p-p error in the spec sheet look good.

(If belt/pulley is a great way to increase gear ratios, you would just use a compound 200:1 belt/pulley system to start with!)

So, be very careful when you read spec sheets. They will for example say strain wave gears have no backlash, without telling you that the mount itslef has progressively poorer backlash as the belt/pulley ratio increases. Caveat Emptor.

Chen

w7ay

(I wrote...) if the mount uses steppers instead of servo motors

By the way, another Chinese strain wave geared mount manufacturer, Warp Astron, made a point of the fact that by using servos, they can guide away large periodic errors. You can check out the mount in this Cloudy NIghts discussion:

https://www.cloudynights.com/topic/877919-wd-17s-harmonic-equatorial-mounts-servo-motor-direct-drive/#entry12721638

Interestingly, they also tried to copy the mechanical design of the RST-135, up to the point where the altitude lock is virtually identical to the RainbowAstro's, and also where the connectors are located :-). Their azimuth adjustment looks pretty rough, though.

It almost sounds like Warp Astron is using the 100:1 gear directly (from a Japanese review) without going through any reduction belt (thus 862 second period, gulp), and not worrying if they have to publish poor peak-to-peak PE amplitude. If so, the peak-to-peak PE is probably of the order of 100 arc seconds :-), which we know has no bearing with autoguiding. This makes me think that their developers actually understand how to guide a strain wave mount, instead of putting lipstick on a pig.

Like the RST-135 (non-E), it uses a relative encoder with servos on both RA and declination axes (the RST-135E has absolute encoders).

Copy, copy, copy...

But this is the first Chinese strain wave mount that I have an even slight interest/curiousity in. And cheap enough to buy just to measure it for grins and giggles... I am slightly tempted. I am sure the (eBay/AliExpress) "warranty" from China is also useless when it develops a problem (one thing that you cannot copy is build quality).

Chen

Kevin_A

w7ay hi Chen, I had a message from a cloudy night member to pass this message on to you…

I am not a member of ZWO but was just grazing the forum re comments on the period change.
I see you are a member so maybe you can pass this on to Chen, who might not understand
the basic mechanicals involved properly, even tho he understands the math :-)
He posted
Quote
The primary reduction ratio of the strain wave gear has probably not changed (i.e., remained at 100:1)).
The 430.82 second number comes when you add a further 2:1 reduction using belts and pullets
to make the final 200:1 reduction ratio
and

Quote
It almost sounds like Warp Astron is using the 100:1 gear directly (from a Japanese review) without going through any reduction belt (thus 862 second period, gulp),
According to the specs, the original AM5 uses a 100:1 strainwave and a 3:1 belt drive
to give an overal ratio of 300:1
In a strainwave, to get 100:1, the flex spline has 200 teeth and the outer ring gear has 202 teeth.
One rev of the input shaft moves 2 teeth, and the period in the plots is based on the teeth
ie a 100:1 strainwave has the equivalent "period" of a 200 tooth wormwheel.
Belt harmonics do come into play with the total PE,
but the primary PE peaks appear to be based on the teeth.
Soooooo, for the new 288s period ( which equates to a 300 tooth wormwheel :-) )
it means they have probably got hold of a 150:1 strainwave head.
( ie much smaller teeth, hence lower PE but maybe rougher machining, hence the harmonics )

Only problem, is this isnt a std size, 160:1 appears to be the industry std.
All good fun, but it might change his idea on the math to determine
whats coming from where..

Andrew Johansen Melbourne Australia

w7ay

Kevin_A that means they have probably got hold of a 150:1 strainwave head.

ZWO's specs states otherwise, though: that the new strain wave gear is still spec'ed at 100:1. So, the change in the ratio has to come from somewhere else?

Further, this review of the WD-17 lists the ratios for different strain wave geared mounts (the most comprehensive table I have seen yet):

https://walachia.blog.fc2.com/blog-date-202302.html

Notice that it too lists a 100:1 strain wave gear ratio for the ZWO mount (Google translate works rather well even for the table:-). But "reduction ratio" is listed as 300:1. Not 150:1, which is needed (for a strain wave period) to produce a 287 second period (correcting my mistake that a 100:1 strain wave gear produces a period of sidereal day/100 instead of sidereal day/200).

(Thus, I think that 300:1 is a mistake and it is the reduction in periodicity of the strain wave gear, not reduction in gear ratio -- and the ZWO mount really has a 150:1 total reduction ratio.)

Notice that for the WD-17, the gears are listed as direct drive, with 100:1 gears and reduction ratio of 100:1 -- thus, because of the stain wave repeats, a period of sidereal day / 200. I.e., 430 seconds, and not 860 seconds (same as RST-135 period and the period of your version of the AM5). Makes it a little less interesting to buy one to measure, but I'm still tempted to pull the trigger.

(By the way, the author had a question right after the table about BLDC vs DC servo motors -- I believe BLDC simply implies a brushless DC motor, where there is no commutator, and the currents to the motor coils are simply modulated at different phases. So, still just a DC servo motor -- the one in the RST-135 is probably brushless too.)

We do know that there is a belt in the ZWO mount. There is also supposedly one in the RST-135, but I have not opened my mounts far enough to see it. This is what the RST-135E looks like after removing two layers of printed circuit board. Already past the warranty boundary :-). That thin barrel marked with [2.5] is not present in the non-E version of the mount:

Yes, my mistake on the 100:1 producing an 860 second period. Definitely my bad. Even the NASA paper that I linked to had mentioned that it takes two revolutions for the strain wave to repeat (I think the Wikipedia page also mentions it), so a 100:1 strain wave gear would repeat like a 200:1 "standard" gear ratio. I should have remembered that. Bloody rotting brains.

Kevin, I went completely the other direction last night; no tracking, no guiding; only used the mount to GOTO :-).

I was doing a practice run for the Perseids tonight, and simply parked the mount near the Zenith just a few minutes west of the Meridian.

I had shut tracking completely down -- but ASIAIR proceeded to turn it back on behind my back at some point (good thing I was watching) -- so I removed the mount from ASIAIR completely to get what I wanted, instead of what ASIAIR wanted.

I was running a 50mm lens (f/1.2 - yes, wide open with bloated stars and so on :-) on a full frame ASI6200C that is so wide that it would not allow me to track for even two hours because of the trees -- so I simply parked the mount and used 6 second exposures. I figured that except for a fireball, a meteor would not last for more than 3 seconds, and the 6 second exposure will minimize my sky background noise. The 6-second untracked stars also minimized the streaking.

(ASIAIR "auto-focus" is completely useless at f/1.2 of course.)

Anyway, I will be parking the lens a little east of the meridian tonight (as far as the trees will let me), and a declination a little north of Zenith (my Zenith declination is +45.45º), again as far as the trees will allow me, to get nearer to the radiant when it comes past.

So, no autoguiding, no tracking, no mount later tonight :-).

Chen

Kevin_A

w7ay i never use asiair to autofocus my fast glass either.

w7ay

Re: WD-17s, I'd said: Makes it a little less interesting to buy one to measure, but I'm still tempted to pull the trigger.

Eeek, I changed my mind after seeing the guide graphs in the Japanese reviewer's follow up post on the WD-17s:

https://walachia.blog.fc2.com/blog-date-202306.html

Clicking on the PHD2 guide graphs shows some pretty ugly reponses to the guide pulses. The errors would continue to drift for one or two arc seconds in spite of the correction pulses being aplied for a couple of dozen cycles, without the pulses even clipping. (It is possible that the calibration is wrong. It is not a MinMo problem either, since we can see the pulses being issued. But the mount is not reacting properly to the pulses.)

One more strain wave mount to not look into at least for now :-). I think everyone is trying to rush to get on the bandwagon without testing them completely. Everybody trying to grab your money instead of looking out for you. Weird thing is that I had bought by first RST-135 back in 2019, and for 3 years everything was quiet, then suddenly starting about 1 year ago, the market for strain wave mounts explodes.

A shame, because the build quality and indiustrial design appears to be the best after the RainbowAstro and Hobym mounts.

BTW, another interesting mount (looks like vaporware for now) is the SharpStar which also uses a 100:1 strain wave gear according to their specs. What is weird is they claim a 1000:1 reduction ratio, and yet a period of "432 seconds."

https://www.sharpstar-optics.com/Products_1/37.html

Chen