Equipment outside advertised specs

[codemonkey]

So, hypothetically speaking, were you to buy a rather expensive mount advertised as having +/- 5-7" periodic error "on average", but found yours to have at least 43" peak to peak, how would you react? Does anyone think that variation from specification is reasonable? Refund time? I'm talking about 7k to land the mount as well, so not cheap.

[Atmos]

Spending that sort of money I'd be asking serious questions. With that sort of "out of spec" results it could definitely come under there being a manufacturing issue at play. It would be under warranty I imagine.

[codemonkey]

The manufacturer is claiming that the primary reduction step in the mount is 122 seconds, and if you look at the graph I provided, then no period of 122 seconds is much outside the advertised 5-7" "on average" error.

Said manufacturer says not to worry about it, it can easily be guided out and that PPEC would be a waste of imaging time.

[glend]

So to use that mount you have to guide. Was that made clear in the details of the mount prior to purchase? If it was purchased through an Australian agent you should have remedy under Australian commercial and consumer protection law, if it was not made clear or you consider it unfit for purpose. I doubt a direct purchase from the manufacturer or an overseas agency is going to provide protection/warranty unlessvitcwas bought through a very reputable agency like "Teleskop-Express", etc.
Will they refund at this point? Who pays return shipping? Do you get relief from duty paid here? Many questions spring to mind. Do you have a written warranty and what does it say?

Just out of curiousity how long does one rotation of the worm gear take? Could it be slightly bent or flawed in such a way that rotation causes that graph pattern? The NEQ6 worm gear has an 8 min rotation period. Those peaks look worse than many of the NEQ6 graph images on the Net.

[issdaol]

If its a "Tier One" mount like a Takahashi, Astro Physics, SB etc etc then I would demand a replacement and shipping at their cost.

Too many purchasers pay anywhere from $10k to $20k for these mounts and when a problem crops up don't insist on the Manufacturer repairing/replacing and shipping at their cost.

This actually does a disservice to all future customers because it sets an precedent that allows manufacturers to get away with shoddy workmanship or QC and the purchaser ends up paying extra coin to ship or spending their own time to do field replacements/repairs.

So imagine going in to buy a new Car in Australia in that price range.........would you find it acceptable to discover that the paint work was poorly done? a faulty gearbox? the engine just won't start? upholstery faded/torn?

[rally]

Lee,

What have you used to generate the PE Curve ? . . . and under what conditions ?
Is it complete ?

I would have expected that the start point of the graph would also match the end point of the graph, (ie its a full cycle) so you havent captured a full cycle or alternatively there are other issues additional to and unrelated to PE in your capture.

If you had a complete PE cuve you should be able to shift Y axis zero and roughly calculate the area under the curve and get your own average tracking accuracy !
Then compare that to their claimed performance.

Assuming its correct, I wouldnt be accepting it on the basis of their claims and its sub standard for a mount claiming high performance tracking and accuracy.
I have friends with EQ6's that are better than this !

But their specs are all a little bit ambiguous and its difficult to determine if its gobbledygook marketing speal or just a language translation problem. But I'd have to say after reading their website there are a lot of odd or oddly worded statements and its hardly an industry standard way of presenting the mount's specified Periodic Error - and that might be a problem for you - what do they really mean by their statements ?

[Shiraz]

that's a difficult one Lee.

drift errors like this must be annoying and very disappointing for an expensive mount. However, Dec Backlash and RA gear jitter are way bigger menaces than a slowly varying RA rate drift like this if you are guiding - if you have little backlash and jitter and the RA drift guides out, then you will get results as good as any mount.

Looking at your graph, maybe the makers have a point - if you guide, it should be fine, since the error slew rate is not very high (almost always less than an arcsec every 3-4 seconds) and the jitter looks small. Although the extent of the drift is quite large, it varies slowly and smoothly - way better than an EQ6 for example. It isn't perfect in one respect, but if it does OK otherwise and that one problem can be managed, then you have a good mount. The makers suggest that it a very capable guided mount - they don't make any claim for unguided performance.

So the question is, does it have negligible dec backlash and can it routinely produce tight round stars with guiding? If yes, then you have a good mount - forget that it isn't quite perfect - it is good enough for the job and will possibly do better than a gear driven one due to the smoothness of the belt drive. If no, then you may need to go through the hassle of trying to get the supplier to give you a refund.

[billdan]

I'm a little confused, where did the 122 secs come from? It seems you have one peak at the 681sec spot and the next peak at the 1641 sec spot, that's a difference of 960 secs.

Plus I think you should include the DEC graph as well, because sometimes movement in DEC can exaggerate the movement in RA. Those big swings could be caused by something moving in DEC (flexure??).

Cheers
Bill

[codemonkey]

Thanks for your input everyone, I appreciate it :-)



Yep, absolutely the manufacturer make a point of saying these mounts need to be guided, and I'm ok with that. I had some impressive performance when I first got the mount, but then it seemed to perform less well, so I started looking for causes. The mount also has big RA drift of about 5"/min, which, when corrected, seems to improve guiding significantly (small sample size though, so could have just been lucky). That makes me wonder what this PE is doing to my guiding and how much better would it be if I had a mount that was to spec.

There's no Australian agents, it was bought from Teleskop Express. Mounts are exempt from duty and I'm not sure about returns or shipping etc.



The mount's an Avalon Linear, so not "Skywatcher" class, but not "Astrophysics" class either. Definitely up there in cost, and I expected it to be within spec in every regard when paying this kind of money.



Cheers mate. The chart's done in Excel from a PHD tracking log - i.e. I kept PHD going for about 40mins but had guide commands disabled; standard way of recording PE for correction in PEC Prep.

Based on the pattern I believe there's two full cycles there, plus a partial at the beginning and another partial at the end. The first full cycle has a peak of about 43", the partial at the end closer to 50" and based on the patterns before, I don't think that had reached its peak when I stopped logging.

I've attached some screen shots of PEC Prep for both RA and DEC. Note that PEC Prep can't handle PE of longer than 1000s, whilst this mount seems to have a cycle length of about 1167s.



Cheers Ray. Note that in addition to the above (corrected for drift) chart, the mount also drifts in RA by close to 5"/min. You (and the manufacturer) could well be correct in that maybe it is guiding out fine, and I'm making an issue out of nothing, but how can I be sure? Maybe the poor performance I was seeing sometimes was just bad seeing, but I don't know how I can systematically determine what the issue is. What I do know is that this is very far from what is advertised and had I known it before I bought it, I may not have bought it.

The mount absolutely has negligible, maybe zero backlash, which really is amazing. Sometimes it seems to work well, other times it's junk and I can't get it to perform, and sometimes those "junk" sessions go for weeks and then I'll have a night when it's amazing. Maybe it's just bad seeing. Maybe it's bad seeing in combination with the above drift / PE?



Yeah, I'm also confused by that Bill. The manufacturer has stated that the primary reduction step is 122 seconds in length, which would be analogous to the "worm period" in a traditional mount, but there's clearly a much longer source of error.

I've attached a DEC graph as well, this time from PEC Prep so it's a bit more familiar.

[OzEclipse]

I had a similar thought when looking at the graph. This is quite strange. The manufacturer claimed 122s period doesn't correspond to a whole number of teeth on the worm wheel. The 960s period suggests a 90 tooth worm wheel. A high end mount wouldn't have a 90 tooth worm. If it did, I'd suggest it couldn't be made accurately enough to produce a ±5" PE. What are the mount specs for the number of teeth on the wheel?

Cheers

Joe

[codemonkey]

The mount has no worm, it's entirely belt driven, and has a 4 stage reduction system (again, all belts).

[glend]

I'd email Wolfi at Teleskop-Express to discuss your options, shoukd you wish to return it under warranty. I have found TS very good to deal with, and i know they have a no fault 18 day return policy, obviously way over that but there should be a manufacturers warranty as well. Will the manufacturer 'accept' a return on the basis uou have outlined?

[AndrewJ]

Gday Lee

I helped a user in the US when he was testing out a "Linear" a few years back.
The mount MUST be guided ( its in the spec sheet ) as the belts can produce quite large PE, and due to the gear ratios used, it cannot be fixed with a PEC model.
As noted, there is no "worm", as its fully belt driven.
The final gear has 66 teeth, soooo, about a 1,305 second period ( per tooth ), and was what caused the largest woofs in the PE.
Basically, ( based on my old notes ) there is
1) a 16 tooth pulley on the motor
2) then an 80-16 compound
3) then an 80-16 compound
4) then an 80-11 compound
5) then a 66 toother on the axle

1) has 114.88s period and 7.18s per output tooth
2) has 574.43s period and 35.90s per output tooth
3) has 2872.14s period and 179.51s per output tooth
4) has 14360.68s period and 1305.52s per output tooth
5) has 86164.1s period and 1305.52s per tooth

Based on what i learnt with my AZEQ5, belt tension can have a huge effect on how well the belt transitions between teeth.
Looking at yr data, i would surmise the final bet tension has changed and the transitioning of the teeth is causuing the bulk of yr problem.
Can you post the raw log files??? so we can have a play at extracting more data.

Andrew

[Shiraz]

What does it do when in "junk" mode Lee? - does it only misbehave in RA?

[rally]

I took a copy of the blurb off a “Technical brochure”
http://www.baader-planetarium.de/avalon/download/Avalon-Linear-Mount.pdf

This blurb is full of self contradictions, superlative laced claims and a few technical errors.
So on the one hand they make the claims :
“ . . . to completely fix any play or backlash issues, eliminating the excessive costs of the traditional mechanical devices”
and
“. . . eliminating any play and noise, at the same time assuring a very high precision.”
and
“. . . because no play or unpredictable behaviours are seen”
and
“so the motion is transmitted by the tooth belt in a very gradual and smooth way, without “peaks” and accelerations/decelerations”

All leading the unwary to believe that this is what they are getting when in fact what they are getting is a whole set of new tracking errors which are greater in overall magnitude, incapable of being tweaked out mechanically and unable to be described mathematically and removed with traditional PEC software.

Yet on the other hand they also state –
“ . . . the tracking error generated by the transmission devices inside the LineAR is large in absolute terms (typical, mean values are of +/- 5, +/- 7 arc-seconds)”
and
“The LineAR transmission, based on four stages of reduction, introduces tracking errors which can be larger than those shown by traditional mounts in the same price range,”
and
“ . . . it should be underlined how the PEC is useless with the LineAR, as well as the software commonly used to handle the periodic errors.”
and
“ . . . the LineAR mount requires an autoguider”

There is a lot more to comment on but this post is already extremely long !

I understand that the nature of the “periodic” error problems are different between their mount and say other mounts due to their multistage belt drive system, but in the similar price range mounts I think the comparisons and claims made simply don’t fly in the face of real world performance.
The fact that you can apply PEC to almost all of the other higher end mounts and virtually eliminate it whereas you cannot do so on the Avalon Linear is a huge issue.
But the sheer quantum of error that you will get in a Tak, Paramount, AP etc is not as high or as bad as they want you to accept either.

They appear to be conveying the notion that Periodic error only comes from gear teeth contact errors, which is of course erroneous – bearings can have concentricity and alignment problems, non concentric gears and pulleys caused by normal machining tolerance in the manufacturing processes or imperfect sizing or alignment of the shafts or bearings, eccentric loads caused by belt tensions, housing stresses, belt dimensional variations, dimensional errors in the machining process – eg incorrectly positioned bearing recesses and shaft support bores from one side of the case to the other etc etc
These exist in their system, and the more complex the gearing system, the more components are needed, the more errors and tolerance buildup you will get – ultimately resulting in more “periodic” error (-like) problems.
The fact that they can smooth these errors out a bit is not necessarily a solution – its just a difference ! It may mean that sharp changes – that are faster than a guider can guide out might disappear, but really we aren’t comparing their mount with cheap Chinese mounts are we !?
How many PE errors are so fast in other equivalent high end mounts that they cant be guided out ?
Can you guide it out using their mount – Yes !, and it would seem that is the only solution.

I wonder what a 20-40 minute unguided image would look like if it happened to coincide with one of the big 43 arc second peak to peak Lee found – I would think completely unuseable, yet there are plenty of mounts when accurately polar aligned that could yield superb unguided results.

I’d say it’s a case of marketing hyperbole.

A revolutionary technology: Tooth belt transmission
Designers at Avalon Instruments wanted to completely fix any play or backlash issues, eliminating the excessive costs of the traditional mechanical devices used on equatorial mounts (worms and gears) and making a new transmission system, based on pulleys and tooth belts. A four-stages reducer is used.
Those devices are used with great success for industrial mechanical applications, as they provide a perfect transmission of movements in a very linear way, eliminating any play and noise, at the same time assuring a very high precision.

Benefits of the tooth belt
A tooth belt, which has no play for itself, is better than other mechanical devices because of its working principle. In the gear-worm systems, there is just one, tangent contact point between the working parts.
Under these conditions, the lever arms and the efforts produced by the telescope-mount system introduce some small non-linearities in the transmission of the motion; they are visible as small shifts, evident in the images obtained with those traditional devices. Those small effects are proportional to the focal length, so they are particularly visible with telescopes having a long focal length. The tooth belt transmission does not introduce these troubles, as there is no contact among pulleys and motion is transmitted by the belt with no play at all. The belt has an extended contact with the pulley, so the motion is transmitted by the tooth belt in a very gradual and smooth way, without “peaks” and accelerations/decelerations, typical of traditional gear-worm systems and giving troubles while taking images, even with modern autoguiding devices. Autoguiders, indeed, often are not fast enough to compensate for those sudden anomalies coming from the gear-worm coupling.
As every experienced astroimager knows, well calibrated autoguiders are unable to provide perfectly tracked pictures with traditional mounts, even when they employ gear-worm systems with small tracking errors. These anomalies are not recorded by the guiding telescope, as they originate between the guiding device and the mechanics of the instrument.
Thermal variations, flexures and torsions, especially close to the meridian, transfer plays and mass to the gear (which is in equilibrium) and the resulting effects introduce several oscillations, creating unwanted shifts and sudden movements.

Images always perfectly tracked
Since the very first tests, the manufacturer noticed how easy it was to capture very well tracked, long exposure images, matching the primary goal of the project. While visual observing with high magnifications, the absence of any noise and a very fast damping time are clearly evident, as well as the lack of any play or backlash and the immediate, smooth response to the electronic commands:
We call it Fast Reverse because no play or unpredictable behaviours are seen, even with high payloads. The LineAR mount must be perfectly balanced –which is very easy to do, as the LineAR axes have no friction – because it has not the play troubles originating during the mass transfer while crossing the meridian.
The LineAR transmission, based on four stages of reduction, introduces tracking errors which can be larger than those shown by traditional mounts in the same price range, but the absence of peaks and sudden accelerations makes possible perfectly tracked – autoguided - long exposures, even with more than 20kg of payload, competing with more expensive, higher class mounts.
It must be stressed that the LineAR mount requires an autoguider (but it can be easily guided by hand, thanks to its smooth, gradual errors and the quick response of the commands, event at the lower - 0.125X –rates). Furthermore, it should be underlined how the PEC is useless with the LineAR, as well as the software commonly used to handle the periodic errors.

The LineAR mount redefines the Periodic Error concept
While the tracking error generated by the transmission devices inside the LineAR is large in absolute terms (typical, mean values are of +/- 5, +/- 7 arc-seconds), this is no reason of concern: the mount, which uses a four-stages reducer with tooth belts, has a very slow and smooth error, well spread over the time, so that it is very easy to handle with any autoguider, even while doing images with long focal length (2-3 meters) telescopes.
The LineAR must be used with an autoguider and a slow correction rate, as 0.125X, but it still makes possible to take unguided images with 500-600mm focal length telescopes. Furthermore, because of the technology used, having four different, very slow tracking curves, the PEC function is not useful (as well as the software generally used to optimize the PEC performances, which do not contemplate such a complex case).

[AndrewJ]

Gday Lee
Just for fun, i dug out the plots of my AZEQ5, which showed bad ripple in RA when tracking. The diag shows the before and after effects of adjusting the tension of the belt.
The belts in the Linear are far more coarse than my AZEQ5, so wrong tension could create quite large ( but slow and smooth ) PE woofs.
Best bet is to try and leave it logging data for several hours if possible, as that will really highlight whats going on.

Andrew

[billdan]

Lee,

I had a search around SGL last night to see if anyone else has been complaining about the Avalon Linear. All I saw was lots of praise and very flat guide graphs usually around 0.25 arcsec RMS guide error. The only person that had a complaint (Olly Penrice), who believed they suffer a bit when tracking in the wind (belts stretching).

As the mount is designed to be guided only and if it does a good job doing that, unguided PE graphs are just a distraction to make you worry.

Cheers
Bill

[codemonkey]

Thanks again everyone, appreciate all the help! I've attached the original guide log for those that are interested. The one used to generate the above graphs is the last guide log in the file.

I honestly can't recall too much in the way of specifics in terms of its behaviour when performing poorly. The other long guide log in the attached file was one night where I wasn't happy with its performance. It's been a couple of months since I've even done any imaging. Hopefully in a few weeks I might be in a position to start doing some testing again.

When the mount is performing well I'm getting about 0.5" RMS (total), but I was having times where it was getting close to 1" and my subs were looking very soft, but not elongated. If I was getting 0.25" RMS I'd be stoked.

Andrew, thanks very much for your post. The adjustment on your AZ-EQ5 shows a massive difference in PE. The PE on the Linear I have is also clearly a lot better than that of the AZ-EQ5 before modification, as it's much smoother. I might talk to Avalon about re-tensioning the belts (without voiding my warranty hopefully).

[Shiraz]

that sound to me like it's only seeing Lee. My guiding varies from about 0.4arcsec in rare good seeing to well over 1 in crap seeing. bad seeing softens the stars out, but no elongation. Never seen 0.25 arcsec.

If you aren't getting elongation, your RA is working well enough. Having no backlash and no rapid gear jitter is a huge bonus - you would pay a lot more for any other mount with this characteristic.

[AndrewJ]

Gday Lee
Something looks odd in the last calibration, as you got nowhere near a 90deg result for RA vs DEC??????
( This also shows up as the plotted DEC is not a straight line )
I would have expected better for a mount of this class, but then again there is still a high drift rate in RA ( around 4.5 arcsec per min?? ) that may be affecting that. The originbal unit tested by the chap in the US several years ago also had rather large RA drift, so maybe thats normal ??????
I think i need to butcher the file ( a little ) to manually recalc the positions based on the best fit camera angle, to see what it looks like then.

Andrew

[AndrewJ]

Gday Lee

I just chopped off the last section of yr data ( as it was affecting the calcs ) and reran the FFT and got the following.
The biggest peak is at Fundamental 1.1 ie 1182 seconds ( i think this is really the 1305s associated with the teeth on the final belt ), but we need more data to confirm that.
The next biggest is Fundamental 2.27 ie 575s and this looks like eccentricity of the first gear after the motor ( as it has a true 574.4s period )
The third biggest is Fundamental 7.36 ie 177 seconds and this lines up nicely with the teeth on the third belt from the motor ( true 179s ).
The next is at Fundamental 10.68 ie 122 seconds and this looks like minor eccentricity of the gear on the motor itself ( true 114.9s )
There is also a small hump at Fundamental 36.5 ie 35.8s which again lines up nicely with the teeth on the second belt from the motor ( true 35.9s )

To confirm this better, you really need to grab a longer dataset if possible, as that will allow the FFT to do its magic.

Andrew

[codemonkey]

Thanks Ray, appreciate that. So it might just be I've had continual runs of bad seeing. Dunk was over yesterday and gave me some ideas on that which I'll use to get a better idea in future. The reference to 0.25" wasn't representative of my expectations by the way, someone mentioned that they saw people consistently getting 0.25" with Linears.





Thanks Andrew, appreciate all the effort you've put into this. That tracking log was obviously captured a while ago, so I could be wrong, but I don't recall PHD warning me about the angle as it usually does if it finds it suspect, so I didn't even look at that. That does explain the DEC graph, which I hadn't looked at before posting this thread.

My gear is all out of action at the moment pending the new obs build, so I'll not be able to get more data for a little while, but I'll report back when I do.

Thanks again!

[billdan]

That was me that said a user was getting 0.25 RMS arcsec from this CN link.
Which was actually 0.14 but that could be in pixels and not arcsecs.
This graph was with a 8in SCT -F10 and an OAG plus lodestar 8.2microns.

http://www.cloudynights.com/topic/490662-my-time-with-the-avalon-linear-fast-reverse/

[gregbradley]

I take it other factors are not causing the large errors. Is the polar alignment bang on? No cable drag, autoguider properly calibrated and using a non saturated single star (not a double star).

Scope is balanced etc etc

If so then I would confirm the PE after double checking all this and especially if you are using tried and true gear and software that works on other mounts then I agree you have a mount way way out of spec and it would seem unusable.

My AP1600 about 2 arc secs and at times .9 arc secs with PEC
I had a PMX with about 3 or less arc secs PE with PEC
My PME about 2-3.5 arc secs with PEC. Perhaps less.

Greg.