#1  
Old 13-05-2015, 06:58 PM
Garbz (Chris)
Registered User

Garbz is offline
 
Join Date: May 2012
Location: Brisbane
Posts: 639
Weird PHD action

Well for the past year I've been battling some strange PHD results, and it's limiting my scope choice to a ED80 at the moment.

I've had these results since changing a few things: computer, remote access method, software, and windows upgrade. I'm ruling out hardware at this point because the problem has survived an entire hypertune on the mount too. While debugging I thought I'd ask you guys if you've seen anything of the sort. Now for a moment ignore the incredible amount of backlash in the dec axis, this I think is a hardware issue and I'll pull apart that axis on the weekend and see if I can't tighten it up.

Specifically has anyone experienced a perfectly happy dec tracking just suddenly take off for no reason. Look at the graph, it's perfectly flat and then shoots off into the positive for seemingly no reason.

I'm currently ruling out remote desktop being the source of the problem. Someone suggested they had similar issues with RDP and switching to Teamviewer resolved it, so that is my current line of thought. Failing that I may grab a different PC tomorrow and give it a run.

Thoughts? It almost looks like PHD is making corrections but it's taking a long time for those corrections to start feeding out to the mount.
Attached Thumbnails
Click for full-size image (Capture.JPG)
160.1 KB70 views
Reply With Quote
  #2  
Old 13-05-2015, 09:58 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,787
my EQ8 does this if the dec is too well balanced - I put it down to the gear/worm losing physical contact if wind etc should drive the axis into the backlash region. If the gear and worm are not in physical contact, phd2 cannot do anything to control the axis - it keeps applying corrections but nothing happens and then it eventually takes up the slack and over-corrects. Deliberate imbalance seems to do the job for my mount and since the gearing needs some backlash to work properly, I have not tried to tighten it up. spring loaded worms would be nice.

I use remote desktop as well - my understanding is that everything still runs on the remote machine, so I can't see how that software would cause such a problem.
Reply With Quote
  #3  
Old 13-05-2015, 10:03 PM
alpal's Avatar
alpal
Registered User

alpal is offline
 
Join Date: Jan 2012
Location: Melbourne
Posts: 2,808
You haven't said what mount you're using.
I assume it's an EQ6?
If so you have these options:
Guiding speed selectable from 0.25X, 0.50X, 0.75X, or 1X

The default is 0.5X - change it to 1X.
Too much play in the declination can cause your problem.
I had the same problem with mine until the Dec backlash was adjusted.

cheers
Allan
Reply With Quote
  #4  
Old 14-05-2015, 07:53 AM
Garbz (Chris)
Registered User

Garbz is offline
 
Join Date: May 2012
Location: Brisbane
Posts: 639
Sorry yes I have an EQ6. Guiding rate should make no difference on the pulse guider as the rate is accounted for during the calibration. The only time it makes a difference is if the rate is set to slowly AND the max pulse duration is too slow to make a difference. I've seen people suggest the guiding rate should be .1x and other say .9x. Mine is set at .9x

The DEC definitely has backlash. I'll try tightening it up a bit and see what happens. Ultimately I solved the DEC problem by guiding one direction only. That seemed to work well until later in the night when I was past the meridian.


Anyway that aside, my DEC issue wasn't actually what I wanted to talk about. I just posted a bad example of a picture. Instead look at the first picture below which gives a much better example. Okayish guiding for a long time and then suddenly the scope shoots off into the ether. I've had this problem for a good year now and it sucks a lot when it happens at the end of a 15min sub. Of interest is the second picture which shows it didn't happen at all yesterday. Normally it happens once or twice in a 30min period. Yesterday, nothing at all for over an hour.

I'm not yet ready to declare RDP the problem, but given how many hooks into the OS RDP actually has (sharing and remote access of COM ports is one feature) I have seen whispers of RDP causing issues with PHD before.
Attached Thumbnails
Click for full-size image (Capture.JPG)
125.5 KB38 views
Click for full-size image (Capture2.JPG)
126.5 KB35 views
Reply With Quote
  #5  
Old 14-05-2015, 04:33 PM
alpal's Avatar
alpal
Registered User

alpal is offline
 
Join Date: Jan 2012
Location: Melbourne
Posts: 2,808
Hi Chris,
I find that the guiding rate does make a difference.
Put it at X1 & get that Dec axis adjusted.

cheers
Allan
Reply With Quote
  #6  
Old 14-05-2015, 04:50 PM
LewisM
Novichok test rabbit

LewisM is offline
 
Join Date: Aug 2012
Location: Somewhere in the cosmos...
Posts: 9,326
What Allan saod. I set my guide rate at 0.75 or 1 and rarely get issues.

I once did have a dec spuke like yours (exactly at 3:30 each time, so exactly half a cycle). Mine was just too tifht backlash adjustment.
Reply With Quote
  #7  
Old 14-05-2015, 05:11 PM
peter_4059's Avatar
peter_4059 (Peter)
Big Scopes are Cool

peter_4059 is offline
 
Join Date: Jun 2007
Location: Brisbane
Posts: 4,046
Chris,

Have a look at this thread:
http://stargazerslounge.com/topic/11...uide-settings/

Chris Shillito is the creator of EQMOD. He recommends setting the guide rate at 0.5 - too high and the pulse duration is lower than what the PC can do (20ms) and you overcorrect. I've got mine at 0.2 and my guiding is working quite well.

Peter
Reply With Quote
  #8  
Old 14-05-2015, 07:40 PM
alpal's Avatar
alpal
Registered User

alpal is offline
 
Join Date: Jan 2012
Location: Melbourne
Posts: 2,808
Quote:
Originally Posted by peter_4059 View Post
Chris,

Have a look at this thread:
http://stargazerslounge.com/topic/11...uide-settings/

Chris Shillito is the creator of EQMOD. He recommends setting the guide rate at 0.5 - too high and the pulse duration is lower than what the PC can do (20ms) and you overcorrect. I've got mine at 0.2 and my guiding is working quite well.

Peter
And it also says:

Quote:
Set them to 0.5 to start with and see how things go. You don't want them too high as small errors will require very short guide pulses and these are less likely to be accurately implemented. On the other hand you don't want them so low that they cannot keep up with the max rate of change present in your mounts periodic error.

Chris.
It depends on your particular system.
Reply With Quote
  #9  
Old 14-05-2015, 08:03 PM
Slawomir's Avatar
Slawomir (Suavi)
Amateur Photon Collector

Slawomir is offline
 
Join Date: Sep 2014
Location: Proserpine
Posts: 3,002
Could also be cord catching or perhaps movement under gravity somewhere in the system?
Reply With Quote
  #10  
Old 14-05-2015, 10:34 PM
Garbz (Chris)
Registered User

Garbz is offline
 
Join Date: May 2012
Location: Brisbane
Posts: 639
Quote:
Originally Posted by alpal View Post
It depends on your particular system.
Indeed, but PE can be corrected for in other ways not just pulse guiding.

Also I don't buy that the correction is an issue when using a pulse guider as it's a series of commands sent to software. It may make an issue with ST-4 where the pulse is actually sent out to the mount, but I ask you the following: What's the difference between 20x1, and 100x0.2 ? It's just a software calculation and the resulting 20ms pulse gets sent to the mount in either case.

In any case I've tried both high mid and low settings and it hasn't made a difference.

Quote:
Originally Posted by Slawomir View Post
Could also be cord catching or perhaps movement under gravity somewhere in the system?
Yeah I've checked both of these and it's a recurring issue despite me having moved my scope all over the place in the past year. That spike in the graph above was the result while shooting pretty much straight up so very little impact with the cable. Also the magnitude of the spike would imply it's not a cable issue either. It knocked the star off considerably, the resulting picture didn't just have eggy stars it had sperm

Quote:
Originally Posted by LewisM View Post
What Allan saod. I set my guide rate at 0.75 or 1 and rarely get issues.

I once did have a dec spuke like yours (exactly at 3:30 each time, so exactly half a cycle). Mine was just too tifht backlash adjustment.
I've checked the period, there is none. Originally I thought the worm was catching on something, but if that were the case it would show up at either a regular interval or multiple thereof.
Reply With Quote
  #11  
Old 14-05-2015, 11:55 PM
RobF's Avatar
RobF (Rob)
Mostly harmless...

RobF is offline
 
Join Date: Jul 2008
Location: Brisbane, Australia
Posts: 5,507
+1 for backlash in DEC. Looks like you're very well polar aligned, so everything is fine until a bit of drift starts. You can see PHD2 throwing increasingly large corrections that wind in the backlash until BAM, it overcorrects way past zero the other way.

Either put the PA off to one side a fraction and set PHD to Resist Switch, or better, find time to tweak out that backlash. (I thought you were doing or had done a belt mod on the EQ6? Must still be looseness in the gear mesh?)
Reply With Quote
  #12  
Old 15-05-2015, 05:59 AM
peter_4059's Avatar
peter_4059 (Peter)
Big Scopes are Cool

peter_4059 is offline
 
Join Date: Jun 2007
Location: Brisbane
Posts: 4,046
Quote:
Originally Posted by alpal View Post
And it also says:



It depends on your particular system.
Like I said if you set the guide rate too high you end up losing control of the guiding because the pc can't deal with guide commands less than about 20ms.

I'd rather understand the root cause and address that than simply hit it with a bigger hammer.
Reply With Quote
  #13  
Old 15-05-2015, 06:23 AM
alpal's Avatar
alpal
Registered User

alpal is offline
 
Join Date: Jan 2012
Location: Melbourne
Posts: 2,808
Quote:
Originally Posted by Garbz View Post
Indeed, but PE can be corrected for in other ways not just pulse guiding.

Also I don't buy that the correction is an issue when using a pulse guider as it's a series of commands sent to software. It may make an issue with ST-4 where the pulse is actually sent out to the mount, but I ask you the following: What's the difference between 20x1, and 100x0.2 ? It's just a software calculation and the resulting 20ms pulse gets sent to the mount in either case.

In any case I've tried both high mid and low settings and it hasn't made a difference.

Hi Chris,
Fair enough.
I was only giving you ideas - I didn't know that you had already tried that.
While PHD is calibrating it is obviously doing so when there is not
a rapid excursion problem -
so it is not seeing that problem & sets in the gain of the system
for a normal situation.
It then cannot handle the sudden rapid excursion which
is probably caused by a Dec backlash issue. ( or some other mechanical issue )
It could be anything: a bit of grit on the worm wheel or
a maladjusted Dec axis.

When my NEQ6 was adjusted the technician connected a 2 meter long steel bar
on the top of the mount & he could check the play that way.
Apparently he used a section of cigarette paper
to place between the worm drive gear & cog to measure the gap.
It needed a tiny gap so that it wouldn't bind & he managed
to obtain far less play without it binding.
It might be something you could do yourself?


cheers
Allan
Reply With Quote
  #14  
Old 15-05-2015, 06:50 AM
LewisM
Novichok test rabbit

LewisM is offline
 
Join Date: Aug 2012
Location: Somewhere in the cosmos...
Posts: 9,326
When adjusting the backlash, remember not to only adjust spur gear (motor) to spur gear (worm), but worm to worm gear too. Inserting cigarette paper with the spur to spur is easy - doing so between worm and worm gear is impossible (unless you want remnants of the paper ruining your guiding ). So, worm to worm gear mesh is CRITICAL, and also difficult without binding as most of it is done by feel (not sure with the EQ6 if you CAN readily see the mesh between worm and worm gear, but on my Vixens it is DIFFICULT).

Vixens have 5 adjustment bolts (4 space/tension bolts, 1 tension adjustment-balance/relief screw) with worm meshing. Not sure on Skywatcher.

Doing spur meshing is only half the procedure. I eliminated ALL backlash in my spur gear engagement, only to find a SIGNIFICANT backlash in the worm to worm gear - 2.5 hours of fiddling later, I was happy (and guiding, as shown below, was pretty darned good). That's with ST4 guide (guiderate set to 0.75 in the handset), not pulse guiding.

http://www.iceinspace.com.au/forum/a...e.php?a=181261
Reply With Quote
  #15  
Old 15-05-2015, 07:15 AM
peter_4059's Avatar
peter_4059 (Peter)
Big Scopes are Cool

peter_4059 is offline
 
Join Date: Jun 2007
Location: Brisbane
Posts: 4,046
Chris has a belt drive modded EQ6. There is ņo spur gear.
Reply With Quote
  #16  
Old 15-05-2015, 07:33 AM
Garbz (Chris)
Registered User

Garbz is offline
 
Join Date: May 2012
Location: Brisbane
Posts: 639
Quote:
Originally Posted by RobF View Post
I thought you were doing or had done a belt mod on the EQ6? Must still be looseness in the gear mesh?)
I have. Present theory is there may be slop in the belt. The final installation of the stepper against the belt is completely blind. There's every chance that when I put it in the belt was riding along the top ring on the motor gear and slipped after a bit of using. It's easy enough to check so that's on my todo list tomorrow.

Quote:
Originally Posted by alpal View Post
Hi Chris,
Fair enough.
I was only giving you ideas - I didn't know that you had already tried that.
While PHD is calibrating it is obviously doing so when there is not
a rapid excursion problem -
so it is not seeing that problem & sets in the gain of the system
for a normal situation.
It then cannot handle the sudden rapid excursion which
is probably caused by a Dec backlash issue. ( or some other mechanical issue )
It could be anything: a bit of grit on the worm wheel or
a maladjusted Dec axis.
No, I appreciate you throwing ideas at me. Thanks.

To be clear I'm not trying to correct for the excursion problem in PHD. Looking at the magnitude it's not a correctable issue, I'm trying to identify the source of the problem, and I'm not convinced it's not a software issue. By that I mean computer doing something stupid, not a specific setting like gain (because close inspection of the graph shows no guide commands preceding the spike).

I think this simply because there so far is no discernible period. Also it did it before and after my hypertune / belt mod and I was very VERY careful to inspect the worm for damage.



Anyway it didn't happen yesterday. Weather permitting I'm going to rule whether the spiking issue is resolved by imaging half the night then undoing my change and imagining again and seeing if it causes a problem.

Also I'm going to check the steppers against the belt tomorrow, and if that doesn't fix it I think the DEC axis is coming apart again next week to re-check shim spacing. The DEC axis has almost no play that I can feel by hand so one thought is that the gear on the worm doesn't have its grub screws in all the way. Which will suck because that's a complete mount disassembly, so last resort.
Reply With Quote
  #17  
Old 15-05-2015, 11:03 AM
Terry B's Avatar
Terry B
Country living & viewing

Terry B is offline
 
Join Date: Mar 2006
Location: Armidale
Posts: 2,759
I have also had times where the guider will suddenly give a big correction. This is using both PHD and CCDSoft for guideing.
It is usually when I'm also doing other things on the computer.
I have noted that occasionally I will get interference lines on the guide image or a cosmic ray hit. The interference happens when the puter processor is busy. This tricks the guider into thinking that the guide star is not where it actually is and gives a big kick to the mount. it then over corrects back the other way as you see on your graph.
Reply With Quote
  #18  
Old 15-05-2015, 01:45 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,787
re second graph posted, whatever it was affected both axes. Clutching at straws, but possibilities may be:
- sharp wind gust?
- sudden scope movement, maybe slipping slightly in tube rings or tube rings not rigidly attached to bar?
- Guide camera able to move very slightly?
- mount not locked down tight on tripod?
- are the three dovetail puck screws completely tight on the EQ6?
- power supply noise spike - eg when a fridge or aircon switches on/off.

If this was taken with an OAG through your C8, it could be mirror flop - I see similar sharp motion when the mirror in my Newtonian slips sideways on the supports.
Reply With Quote
  #19  
Old 15-05-2015, 02:44 PM
Garbz (Chris)
Registered User

Garbz is offline
 
Join Date: May 2012
Location: Brisbane
Posts: 639
Quote:
Originally Posted by Terry B View Post
I have also had times where the guider will suddenly give a big correction. This is using both PHD and CCDSoft for guideing.
Was it actually making a large correction? My problems are characterised by large changes with NO correction. I.e. it bounces off into the distance but according to the history graph in EQMOD and in PHD2 there was no guiding command sent to do it. That's why at first I thought external: bird, wind etc.

Quote:
Originally Posted by Shiraz View Post
re second graph posted, whatever it was affected both axes. Clutching at straws, but possibilities may be:
- sharp wind gust?
- sudden scope movement, maybe slipping slightly in tube rings or tube rings not rigidly attached to bar?
- Guide camera able to move very slightly?
- mount not locked down tight on tripod?
- are the three dovetail puck screws completely tight on the EQ6?
- power supply noise spike - eg when a fridge or aircon switches on/off.

If this was taken with an OAG through your C8, it could be mirror flop - I see similar sharp motion when the mirror in my Newtonian slips sideways on the supports.
The telescope is in a box with 4 sides, so no wind.
I'll check the tube rings. Unlikely as this has been a problem a lot not just with my current setup.
Everything else is tight, I learnt my lesson from the dovetail puck screws on the EQ6 at astrofest 2 years ago At least that was an easy fix.

The powersupply theory is interesting. Very interesting. I'll try and replicate the spikes and then I'll switch to running from battery. The only massively inductive load we have in the house is a fridge but interestingly I have managed to get the PC to play up by turning equipment on and off. I know I need to change the rail capacitor design in my little electronics project so this one here is actually quite credible. Thanks.
Reply With Quote
  #20  
Old 23-05-2015, 04:32 PM
Garbz (Chris)
Registered User

Garbz is offline
 
Join Date: May 2012
Location: Brisbane
Posts: 639
Well re-tuned the mount today. No sign of sky tonight so I won't know how it went.

Did find one problem though, the Dec motor wasn't engaging the belt properly. There was backlash between the motor and the worm gear. This was serious as you can't resolve this by weighting the mount and thus would explain the problems with Dec backlash.

I still have a tiny bit of slop between the worm and the main gear, but that can be taken care of by weighting.

Looking forward to another clear night. I want to test my software theory. I'll guide one night using remote desktop and another night using teamviewer and see if I can see differences.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 07:07 PM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
Advertisement
Bintel
Advertisement
Celestron Australia
Advertisement
OzScopes Authorised Dealer
Advertisement
Meade Australia
Advertisement
NexDome Observatories
Advertisement
Lunatico Astronomical
Advertisement
SkyWatcher Australia
Advertisement
Astronomy and Electronics Centre
Advertisement