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

CEM60 or CEM60-EC

  • Please log in to reply
743 replies to this topic

#726 starflyer

starflyer

    Lift Off

  • -----
  • Posts: 8
  • Joined: 07 Dec 2011

Posted 17 October 2019 - 03:11 AM

Gday Ian

Can you post a PHD debuglog for the calibration sections

as it contains a lot more info than the guide logs

Might provide an insight into whats going on

 

Andrew Johansen Melbourne Australia

Hi Andrew,

 

You should find everything for the last two sessions in that Drive folder.

8th Oct is with the Tak

16th Oct is with the CT10



#727 OzAndrewJ

OzAndrewJ

    Mercury-Atlas

  • *****
  • Posts: 2879
  • Joined: 30 Nov 2010

Posted 17 October 2019 - 03:49 AM

Gday Ian

 

Now I have it I either need to get it working reliably or come up with some proof that SDE is smearing my images

Something is very odd in the calibration with DEC vs RA being so far out,

but i have no idea what that is

Based on your earlier 500ms unguided log, i doubt SDE is the culprit in your case

Will look at yr data tomorrow morning when the brain is fresh

 

 

Andrew Johansen Melbourne Australia



#728 starflyer

starflyer

    Lift Off

  • -----
  • Posts: 8
  • Joined: 07 Dec 2011

Posted 17 October 2019 - 04:02 AM

Based on your earlier 500ms unguided log, i doubt SDE is the culprit in your case

Will look at yr data tomorrow morning when the brain is fresh

Thanks again Andrew,

 

The first run was at a load of 6.4KGs with the Tak, the second is at 15.7KGs with the Newt.  I thought I'd read of SDE showing up more at higher loads / being more visible at higher resolution / oversampling?

Tak = 332mm fl / 2.82"/px

CT10 = 1200mm fl / 0.79"/px

 

Cheers,

Ian


Edited by starflyer, 17 October 2019 - 04:12 AM.


#729 Der_Pit

Der_Pit

    Viking 1

  • -----
  • Posts: 855
  • Joined: 07 Jul 2018
  • Loc: La Palma

Posted 17 October 2019 - 08:20 AM

I did see some strange calibration, too, at times.  I thought it was specific to my setup (using INDI/EKOS), and it seemed that on initialization the guide rates got set improperly, so I had to enter them again afte the initial connect.

With proper speed setting (using default 50:50) I'm getting a nice calibration with (almost) identical speeds for RA and DEC.

 

If you have the commander logs you should be able to look for the ":RGnnnn#" command, e.g., ":RG5050#" to set both to 50% sidereal.  It should confirm with an "1".



#730 starflyer

starflyer

    Lift Off

  • -----
  • Posts: 8
  • Joined: 07 Dec 2011

Posted 17 October 2019 - 03:38 PM

If you have the commander logs you should be able to look for the ":RGnnnn#" command, e.g., ":RG5050#" to set both to 50% sidereal.  It should confirm with an "1".

Thanks for the pointer, something else to check off.  I have the logs and I do see the guide rate successfully set and and confirmed, with equal rates for RA and DEC.



#731 OzAndrewJ

OzAndrewJ

    Mercury-Atlas

  • *****
  • Posts: 2879
  • Joined: 30 Nov 2010

Posted 17 October 2019 - 04:03 PM

Gday Starflyer

Just having my weeties but after coffee, i plan to compare the commands in your commander log

against the commands in the PHD "debug" log

( you can match them by timestamps )

There is a lot more re settings in the debug logs but they are a pain to wade through :-)

 

Andrew Johansen Melbourne Australia



#732 OzAndrewJ

OzAndrewJ

    Mercury-Atlas

  • *****
  • Posts: 2879
  • Joined: 30 Nov 2010

Posted 17 October 2019 - 07:18 PM

Gday Ian

I think i know whats going on now lol.gif

I had to modify my app a bit as your commander log uses older commands than the 120s

but once plotted, it looks identical to what we saw in one log from a 120EC

where it was recovering from a big uncommanded glitch.

If i am correct, it means that the IOptrons selectively change how they apply RA guides

ie what rate they use to blend guides in.

This explains lots of things ( ie runaway oscillations at short guide cadence )

Up until your debug and commander logs, there was no evidence from the 60EC,

but the new commander logs give the clue.

More in an hour or so

( Hint for Der Pit )

Get a commander log from when you had oddball glitches when calibrating

Plot the RA during the calibrate stage and see what "rate" it applies at

 

Andrew Johansen Melbourne Australia



#733 OzAndrewJ

OzAndrewJ

    Mercury-Atlas

  • *****
  • Posts: 2879
  • Joined: 30 Nov 2010

Posted 17 October 2019 - 09:46 PM

Gday Ian

I have the logs and I do see the guide rate successfully set and and confirmed

I think this lies confused1.gif ( ie guide rate )

If you or someone else who uses PHD to do a calibrate, and can collect a commander log at the same time is willing, i have a test lol.gif

Can you set the guide rate to 50%

Set the calibration step pulse to 400ms in PHD and do a calibrate

Now set the step pulse to 800ms and repeat

Now set the step pulse to 1200ms and repeat

ie just do 3 calibrates in a row with the only change being the step pulse time, nothing else.

Now set the guide rate to 80% and repeat

Bundle up the logs and post em.

I suspect you will get several very different results

 

Andrew Johansen Melbourne Australia



#734 starflyer

starflyer

    Lift Off

  • -----
  • Posts: 8
  • Joined: 07 Dec 2011

Posted Yesterday, 10:07 AM

Bundle up the logs and post em.

Hi Andrew,

 

Thanks again, next time I get out I'll get this done and provide the logs. The weather's not looking great for the next week though.

 

Could you look at the PHD log if you have time please? I'm interested in seeing if the performance is worse with the heavier load of the newt.  I'm getting fed up with spending clear nights imaging and getting eggy stars.

 

 

Thanks

Ian



#735 OzAndrewJ

OzAndrewJ

    Mercury-Atlas

  • *****
  • Posts: 2879
  • Joined: 30 Nov 2010

Posted Yesterday, 04:26 PM

Gday Ian

Here is the first blush plot of the "reported" data via commander for the last calibrates in both your logs.

In all cases, only pulse guides are used but you can see the "reported" movement doesnt appear to go at the defined guide speed, it "appears" to get applied at an averaged variable rate.

Also note the total distance "expected" based on no of pulses sent is never achieved

( eg in top plot, theoretically, 51 arcsec of pulse is sent for RA but it only moved 26 arcsec )

Based on this, i suspect the system may be storing the pulse requests and issuing them sequentially

vs deleting any existing error when a new pulse comes in.

This would explain some of the underrunning and overrunning.

Also, in the second plot ( not annotated ) you can see that when the blue trace ( RA ) was being brought back to start, some pulses "appeared" to be ignored, thus giving a staircase effect, again proof that pulses arent being treated as a simple "move at guide speed for DDDms".

I need to dig a bit more into that, but this data clearly shows how pulseguides dont appear to go at a constant rate ( according to commander data )

This is something controlled bench testing would also highlight

Andrew Johansen Melbourne Australia

Cal2.jpg Cal1.jpg


Edited by OzAndrewJ, Yesterday, 04:29 PM.


#736 OzAndrewJ

OzAndrewJ

    Mercury-Atlas

  • *****
  • Posts: 2879
  • Joined: 30 Nov 2010

Posted Yesterday, 06:23 PM

And the thot plickensconfused1.gif

Just digging through the associated PHD debug log

During calibration, there are lots of comments re pulse finishing early

ie they are testing the "IsGuiding" status and it is returning zero, even tho the mount is moving.

I cant actually figure out the real timings as the log is fabricated

but i assume that is the root cause

 

Andrew Johansen Melbourne Australia



#737 Shinebsr

Shinebsr

    Sputnik

  • -----
  • Posts: 30
  • Joined: 19 Apr 2016

Posted Today, 07:02 AM

Well, the 60EC went back and I got the plain Jane 60 delivered and under relatively clear skies last night. RMS was about 0.5 with RA and DEC both around 0.36-0.40. Peak RMS were both under 1.40. So, this mount can seem to guide pretty well, even under the fair seeing last night. What bothers me however, is the RA tracking unguided is horrendous. I got an image shift (30s subs) taken over an hour of 67”. Image scale is 0.67”x0.67”.

This is pretty much identical to the EC version. My CGX had RA drift over time, but this almost seems worse. At this point I guess I need to know what is “normal” RA drift in a mount in this price range. I usually am guided to keep things tightened up, but there is no way I can shoot unguided with this thing. Tracking was awful in sidereal and king. I may have way too high of expectations at this image scale. The 80mm may have just “hid” the RA drift in both mounts and it is only now evident at 1420mm FL.

#738 Der_Pit

Der_Pit

    Viking 1

  • -----
  • Posts: 855
  • Joined: 07 Jul 2018
  • Loc: La Palma

Posted Today, 07:32 AM

Hi Andrew,

 

interesting analysis waytogo.gif

 

TBH, this is what I do expect from such a system.  If the guide pulse would just do that, switch the speed for a given time, you could as well just stay with the ST-4 way of guiding.  With pulse guide you don't do that, you don't even send a pulse, you send how long the pulse would be if sent via ST-4.  So for a 2000ms "pulse" you don't have to wait two seconds, you immediately know that you'd need a 2s × guide rate arcsec correction.  And only this will allow you to do proper calculations like mixing guiding with PEC (or encoder) signals.

 

I wonder how such a plot would look for other mounts (i.e., non-iOptron).



#739 Shinebsr

Shinebsr

    Sputnik

  • -----
  • Posts: 30
  • Joined: 19 Apr 2016

Posted Today, 07:41 AM

One other thing, if RA drift has a PE component, why would the RA drift CEM60 owners “see” tend to go off indefinitely in one direction? What I mean is that I show image translation that has no periodicity in it. The first and last image are translated in a perfect, linear vector where stars are 67” apart over an hour. My PE graph that comes with the mount looks great. PE of +/- 3 arcsec and the positive and negative deflections follow the zero axis nicely (the graph has an overall slope of around zero). Shouldn’t this just mean the stars will all just be enlarged by the periodic error and not trailed off indefinitely due to this drift? I am thinking this is for sure RA tracking rate error.

#740 Shinebsr

Shinebsr

    Sputnik

  • -----
  • Posts: 30
  • Joined: 19 Apr 2016

Posted Today, 07:45 AM

Hi Andrew,

interesting analysis waytogo.gif

TBH, this is what I do expect from such a system. If the guide pulse would just do that, switch the speed for a given time, you could as well just stay with the ST-4 way of guiding. With pulse guide you don't do that, you don't even send a pulse, you send how long the pulse would be if sent via ST-4. So for a 2000ms "pulse" you don't have to wait two seconds, you immediately know that you'd need a 2s × guide rate arcsec correction. And only this will allow you to do proper calculations like mixing guiding with PEC (or encoder) signals.

I wonder how such a plot would look for other mounts (i.e., non-iOptron).


I found this interesting post on eqmod that basically illustrates what I think you’re talking about as far as variable speed corrections. I wish the CEM was compatible with eqmod.

http://eq-mod.source...net/eqspeed.htm

#741 ChrisWhite

ChrisWhite

    Aurora

  • *****
  • Posts: 4721
  • Joined: 28 Feb 2015
  • Loc: Colchester, VT

Posted Today, 09:02 AM

Well, the 60EC went back and I got the plain Jane 60 delivered and under relatively clear skies last night. RMS was about 0.5 with RA and DEC both around 0.36-0.40. Peak RMS were both under 1.40. So, this mount can seem to guide pretty well, even under the fair seeing last night. What bothers me however, is the RA tracking unguided is horrendous. I got an image shift (30s subs) taken over an hour of 67”. Image scale is 0.67”x0.67”.

This is pretty much identical to the EC version. My CGX had RA drift over time, but this almost seems worse. At this point I guess I need to know what is “normal” RA drift in a mount in this price range. I usually am guided to keep things tightened up, but there is no way I can shoot unguided with this thing. Tracking was awful in sidereal and king. I may have way too high of expectations at this image scale. The 80mm may have just “hid” the RA drift in both mounts and it is only now evident at 1420mm FL.

 

The CEM60 is not really a mount to be used un-guided.  And if you do decide to run unguided you should create a PE model using pepro or at the very least with the hand controller.  Just be aware that you cannot use PEC at the same time as guiding, so if you decide to guide you must disable PE. 

 

1420mm is not a very forgiving focal length.  You really need to get good PA, and without question Guide.  I don't think PEC correction will give you anything near usable unguided at this focal length. 

 

When guiding, try PHD2 PPEC algorithm for RA.  It works very well with this mount after the first couple of worm cycles are used to calculate PE.  FWIW I am imaging at 2350mm FL at .54"/px image scale and the mount guides perfectly and I get tight round stars at 600seconds. 


  • psandelle and Dynan like this

#742 Shinebsr

Shinebsr

    Sputnik

  • -----
  • Posts: 30
  • Joined: 19 Apr 2016

Posted Today, 09:17 AM

The CEM60 is not really a mount to be used un-guided. And if you do decide to run unguided you should create a PE model using pepro or at the very least with the hand controller. Just be aware that you cannot use PEC at the same time as guiding, so if you decide to guide you must disable PE.

1420mm is not a very forgiving focal length. You really need to get good PA, and without question Guide. I don't think PEC correction will give you anything near usable unguided at this focal length.

When guiding, try PHD2 PPEC algorithm for RA. It works very well with this mount after the first couple of worm cycles are used to calculate PE. FWIW I am imaging at 2350mm FL at .54"/px image scale and the mount guides perfectly and I get tight round stars at 600seconds.


I figured as much about unguided. The dilemma I have been dealing with (since the EC version I used) was unguided and huge drift or guided and drift is about 1/3 as bad. I clearly have some diff flexure somewhere and since there is NO chance of movement with my guide scope the way it is mounted, I clearly have mirror movements in the edge 8 even with mirror locks engaged. I will have to resort back to the COAG to deal with the flexure (focus and good guide star shapes in the OAG were driving me nuts). Oh well. Back to the drawing board.

#743 OzAndrewJ

OzAndrewJ

    Mercury-Atlas

  • *****
  • Posts: 2879
  • Joined: 30 Nov 2010

Posted Today, 03:51 PM

Gday DerPit

 

TBH, this is what I do expect from such a system.

If it "blends in" the move over time, then it can really upset other processes that rely on the time.

esp as the IOPtron "appears" to be telling the pulse sender that the guide is done when it isnt.

This kills PHD, esp the way it does its calibrations.

In the second unannotated plot above, you can see the RA moved 90arcsec according to commander

PHD had sent enough pulses to move 430 arcsec in the same time.

It then sent 430 arcsec worth of reverse pulses to get back

PHD doesnt actually care what the "selected" guide rate is, it uses its own internally calculated rate to come up with pulse times, so a bad cal really affects things.

 

I wonder how such a plot would look for other mounts

For my Meade mounts, pulses apply at the rate specified if the std motor commands are used

or at a higher rate for less time, if the specialised motor commands are used.

 

In the Syntas, we have been experimenting with using "gotos" to do DEC guiding

ie we convert the guide rate * pulse time to equivalent microsteps and issue a command

It allows easy application of backlash and is deadly fast/accurate

2000ms of lash can be applied in about 100ms :-)

Recently, Roland from AstroPhysics posted the pulse guide characteristics for the new Mach 2

He posted data for tracking with encoder on then off then on, and also response to guides.

It appeared to be deadly accurate.

Andrew Johansen Melbourne Australia



#744 OzAndrewJ

OzAndrewJ

    Mercury-Atlas

  • *****
  • Posts: 2879
  • Joined: 30 Nov 2010

Posted Today, 03:58 PM

Gday Shinebsr

 

I found this interesting post on eqmod that basically illustrates what I think you’re talking about as far as variable speed corrections.

With the syntas, you basically send a command to the motor to go at a set speed.

EQMod ( and a few other apps ) simply follow that process, ie they send a speed command

When doing PC based PEC, EQMod merely tweaks the speed for the given bin distance

If a guide comes in, it just tweaks the speed to the guide rate for the given time.

This doesnt work properly in EQMod if you are using the mount based PPEC, as the mount simply tweaks the speed for you on a bin rollover ( and this overrides any external guide inputs other than ST4 ). The new Greenswamp driver disables PPEC when guiding pulses are in process, so it works quite happily with the PPEC

Lots of different mounts / softwares do it different ways

 

Andrew Johansen Melbourne Australia




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