CEM120 Guiding RA vs DEC

[mark matzner]

Hi Group,

As previously posted, my CEM120 (non encoder) mount had a periodic RA excursion right out of the box.  iOptron replaced the RA worm package and the problem was corrected.  So back to imaging.

This is a really nice mount.  Once it was returned from repairs, I took advantage of the "cable management" feature and now have zero cables hanging off the backside.  Very clean, and much better balanced.  I initially had concerns since the USB2 ports in the mount's hub are not powered.  But all components (image cam, filter wheel, auto-focus, and guide cam) have played well together.  This was a pleasant surprise, since most all I've read and heard about USB hubs (even powered)  has been bad.  FWIW, the Lodestar guide cam is the one item out of the above four that gets its power through the USB2 port.  The rest have separate power sources.

I'm a faithful calibrator, use the PPEC algorithm along with the Guide Assistant recommendations.  Here's what I see.

1.  The iOptron's PEC feature simply does not work.  My attempts at PEC'ing and guiding result (IMHO) in poorer guiding.  I've tried both their simplistic "single cycle" method, and also a PemPro smoothed graph.  In both cases, the guiding is worse with the mount's PEC turned on.  There's quite a bit of chatter about this on Cloudy Nights.  I suspect that iOptron will eventually figure it out with a driver update.  But all that aside .....

2.  Guiding is pretty darn good.  Certainly better than my old Meade 12"LX200 Classic fork mount.  And better than the brief time I was using Celestron's CGE Pro.  DEC is reasonably good.  And RA isn't bad, just not as good as the DEC.  And this leads to slightly elongated stars.  I've seen a few comments by other non-encoder users along the same lines.

I can live with what I have, but would appreciate any input on how to finesse the RA to better match the DEC guiding.

Thanks,

Mark

The attached DropBox folder includes Calibration, GA and imaging of NGC5907 from July 24th.  This was shot over the city lights of Burleson, TX.  

Best I can tell, the guiding is fairly representative of all targets since that time, regardless of sky location.  I can supply more recent logs if need be.

https://www.dropbox.com/sh/t25bw13a5alce6b/AABphcd6Tkg-Kf7o9MKQszqfa?dl=0

Imaging Train:

12"LX200 Classic OTA, Feathertouch Shorty Focuser, Starizona SCT Corrector, Mitsuboshi OAG5, Atik EFW2 Filter Wheel, Atik 383L+mono Cam.

Focal Length:  2182mm

Scale:  0.51"/px

Guiding Train from OAG5:

0.50x Reducer, Lodestar Guide Cam

Focal Length:  1091mm

Mount Guide Speed:  0.75x

[Rainer Ehlert]

Hi Mark,

Please post the latest Guide Log in order to analyze it. 

Do you have elongated stars in RA or DEC direction ?

Thanks and regards Rainer

[mark matzner]

Hi Rainer,

Here is the latest Log and Debug files from August 7th.  The log sections are as follows:

Section 1:  M16 sop-west

Section 2:  M16 sop-east

Section 3:  NGC6781 sop-west

Section 4:  NGC6781 sop-east

1.  In this example, the curves show clearly that (for the most part) the RA amplitude is greater than that of the DEC.  Roughly speaking the RA RMS is about twice the DEC's.  And the "scatter" supports this graphically.  

2.  I've included a FIT file from each target.  I find it curious the "egging" of the stars is on a 45 degree angle.  This may be pointing out a separate problem, or just indicating that I don't understand the geometry of this setup.

BTW, I have enjoyed following your progress on CN.

Mark

https://www.dropbox.com/sh/75r5tw5q3act7rf/AAC5GotnnyH8bjCGxkE4EfV8a?dl=0

[Rainer Ehlert]

Thank you Mark,

Have you tried to guide with a normal RA guiding algorithm ?

I see in the latest log you are using Predictive PEC. Sometimes those fancy algorithms make our like more difficult then easier.

Did you get a printed out spec sheet of your PE ?

Diagonal stars normally are a sign of a misaligned mount against the Pole.  I only see the diagonal stars in one image. The egging in the other image I still have to think about that. I also have that sometimes even with a near perfect guiding.

Are you permanent or portable ?

regards Rainer

[bw_msgboard]

Hi Mark.  I’m not surprised you’re getting some elongated stars.  In my experience, as the difference between RA and Dec guiding RMS heads toward a factor of 2x, elongation is likely to appear.  Unfortunately, I don’t see a magic bullet to improve the RA guiding.  If we reconstruct a raw RA graph for the 43 minute guiding session, it looks something like this:

 

 

This shows the guide star moving by as much as +- 15 arc-sec from the lock point, and I don’t think it’s periodic in the sense of being a harmonic of the worm period.  Plus, the early stages at the beginning of the session have some very steep sections.  All of this means there is a lot of guiding work to do and PPEC probably isn’t going to be able to identify a consistent frequency for applying a computed correction.  One possibility is to reduce your exposure time to reduce the amount of lag in the corrections, particularly for the steep sections.  You have to be careful doing this because it increases the chances of chasing the seeing – you may need to adjust the min-moves upward when you see this happening.  If you have the time and interest, you might want to get involved in the beta testing Ken is doing with his Z-Filter algorithm.  That’s not going to be a quick fire-and-forget operation at this point, but it may allow you to better respond to the RA issues.

 

If you can’t get any improvement in the RA guiding, you may be left with only a couple of options.  You could choose to intentionally degrade the Dec guiding performance to bring it more in to line with RA.  I’ve had to do that at times in the past – my teeth were always clenched at the time, but it did reduce the amount of star elongation.  Intentionally setting a too-low min-move for Dec is one way to do this, as is using an algorithm like Hysteresis.  Neither are very appealing, obviously.  An easier option is just to fix the elongation during post-processing.  In PhotoShop, this is very easy to do and your viewers will be none the wiser.  Needless to say, I’ve done that too at times. J

 

Regarding the elongation in the main camera images, you may have a combination of guiding and optical issues.  These SCTs are well known for having lots of tube currents and big temperature gradients between the mirror and the air layer above it.  Since your scope is an LX200 classic, I doubt it has any fans.  One way to investigate is to take a few 10-sec exposures with the main camera and analyze those star shapes.  If you see elongated stars in those frames, you know it’s not a guiding problem.

 

Sorry I can’t be of more help,

Bruce

 

---

--
You received this message because you are subscribed to the Google Groups "Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guidi...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Steve Winston]

Mark,

Was this guiding done with PEC enabled or disabled?

And earlier you said:

>My attempts at PEC'ing and guiding result (IMHO) in poorer guiding.  I've tried both their simplistic "single cycle" method, and also a PemPro smoothed graph

What did PEMPro measure you base PE as?  And what was your measured PE after programming PEC and enabling it in the mount?

With over a 30 arc-s swing in RA it will be impossible to get good RA guiding results.  I wonder if there is some binding in RA that is causing the severs swings?

Steve

[mark matzner]

Rainer,

1.  I have not tried the Hysteresis algorithm.  PPEC has seemed to work "well" with this mount and "better" on the previous CGE Pro.  So there you go.  Once the sky clears, I will give Hysteresis a try.  But note Bruce's comments after yours.

2.  I did receive a new PE curve from iOptron after they replaced the RA worm set.  See attached.

3.  Thanks for looking at the images.  NGC6781 may be a bit better but, IMO, it's about the same.  FWIW, the M16 image was done lower to the south and just skimming over the trees.  I would have expected it to struggle more.

4.  Yes I'm in an observatory commissioned in 2009.  First the Meade, then the CGE Pro, and now the iOptron.

Thanks for your thoughts,

Mark

[mark matzner]

Bruce,

1.  Forgive my ignorance, my guide log doesn't show +/- 15 arc-sec variation for the 43 minute session (Log section #3).  How did you produce the graph you displayed?

2.  As you can see, I'm running at 3 second exposures.  I can try shorter (i.e. 2s, 1s) and see what happens.

3.  I will contact Ken and see if the Z-filter algorithm offers any advantage.

4.  I made a feeble attempt at decreasing the DEC MinMove, but thought I saw the beginnings of spiking in that mode.  Anyway, philosophically, it just doesn't seem right.

5.  I don't have Photoshop, and Juan over at PI casts a dim view on too much manipulating.:)

6.  I'll look into the optical behavior of the 12"LX200 OTA.  There may be something to that, at least in some situations.  Thanks for the suggestion.

As an afterthought, would bumping the mount's RA guide rate to 0.90x (currently at 0.75x) possibly help my situation?

Mark

[mark matzner]

Steve,

1.  All guiding on the sessions I've posted have been with mount PEC OFF.  On an earlier session I managed to convince myself that iOptron's PEC did not tolerate PHD2's PPEC very well.

2.  See attached the PemPro curves that came off the mount, and what went back into it.  I'm an absolute novice with PemPro so can't really comment.

3.  As I mentioned to Bruce, I do not understand the +/-15 arc-sec swing (as to where it came from).  FWIW, this mount was repaired by iOptron by 

replacing the RA worm package and that did eliminate (to my mind) the periodic excursions.

Thanks for your input.

Mark 

[bw_msgboard]

Hi Mark, see below.

 

---

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of mark matzner
Sent: Sunday, August 12, 2018 9:49 AM
To: Open PHD Guiding
Subject: Re: [open-phd-guiding] Re: CEM120 Guiding RA vs DEC

 

Bruce,

1.  Forgive my ignorance, my guide log doesn't show +/- 15 arc-sec variation for the 43 minute session (Log section #3).  How did you produce the graph you displayed?

 

This was done using a feature of the LogViewer.  It essentially “backs out” the RA guide corrections to estimate what the tracking would have looked like with no guiding at all.  It’s only an approximation but your log didn’t include an extended run with the Guiding Assistant.  Have seen an extended (e.g. 10+ minute) GA run since you had the mount repaired?

 

2.  As you can see, I'm running at 3 second exposures.  I can try shorter (i.e. 2s, 1s) and see what happens.

 

3.  I will contact Ken and see if the Z-filter algorithm offers any advantage.

 

4.  I made a feeble attempt at decreasing the DEC MinMove, but thought I saw the beginnings of spiking in that mode.  Anyway, philosophically, it just doesn't seem right.

 

5.  I don't have Photoshop, and Juan over at PI casts a dim view on too much manipulating.:)

 

6.  I'll look into the optical behavior of the 12"LX200 OTA.  There may be something to that, at least in some situations.  Thanks for the suggestion.

 

As an afterthought, would bumping the mount's RA guide rate to 0.90x (currently at 0.75x) possibly help my situation?

 

I think this is generally a good idea because it reduces latency and can help to overcome some kinds of resistance in the gear train.  But in your particular case, I wouldn’t expect to see a lot of benefit.  

 

Bruce

[Rainer Ehlert]

Hi Mark,

That is a nice PE curve. Very sinusoidal and not nervous.

regards

[mark matzner]

Hi Bruce,

Thanks for following up.

If I understand correctly what you are asking for ......

Here is the link to the files including the GA run for July 24th. This was done after the RA worm drive repair.

Let me know if you need something else,

Mark

https://www.dropbox.com/sh/t25bw13a5alce6b/AABphcd6Tkg-Kf7o9MKQszqfa?dl=0

[bw_msgboard]

Hi Mark.  Here’s a look at the unguided RA tracking after the repair:

 

 

So you had  a range of about 11 arc-sec in unguided performance which isn’t too bad.  I think the challenge is the steep sections in the curve, places where the RA tracking is changing at a high rate.  Those will be times when guiding will need to be pretty intense and it will be easy for PHD2 to fall behind.  Buried in the GA output is an item called “drift-limiting exposure time”, something we rarely talk about.  That number comes from looking at the steepest section in the graph and computing how long it will take to trigger a guide correction equal to the min-move.  In your case, it was 0.8 seconds.  So if you were using 0.8 second exposures during this time – not something we’d recommend – you would still get a guide correction for every exposure.  With an exposure time of 3 seconds, the guide star could move as much as 4 x min-move between corrections.  With your set-up, that would be about 0.75 arc-sec and I think you’re looking for something better than that.  So I think this supports the idea of trying shorter exposure times, maybe 1.5 sec, to see if things improve in RA.  You may also want to try reverting to Hysteresis for RA guiding because PPEC isn’t addressing any of this behavior at this point.

 

Hope this helps,

Bruce

 

---

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of mark matzner
Sent: Monday, August 13, 2018 6:27 AM
To: Open PHD Guiding
Subject: Re: [open-phd-guiding] Re: CEM120 Guiding RA vs DEC

 

Hi Bruce,

--

[mark matzner]

Thanks Bruce,

All I need now are some clear skies.  We've had a few days of rain lately (thank God!), and it will be a few more before it clears out.

Appreciate the recommendations.

Mark

[Al Moncayo]

Mark,

Have a look at the stuff I just posted over in the Z-filter thread.

Like you, I am dubious that the CEM120 PEC record/playback is fully functional. I've already raised this to iOptron as a bug report. They've not confirmed nor denied it though. In spite of my doubts, I've found a small but noticeable improvement when I record while PHD's GA is enabled. Then I enable playback while GA is still running and wait for at least two worm cycles. I've seen a slight reduction in peak-to-peak variation. I'd expect more, however. With the playback enabled, then I go to guiding.

Al

[Al Moncayo]

Bruce,

My CEM120's RA looks as follows unguided.

The first two cycles were without mount PEC. The second two were with. If I squint a bit, I can maybe say that there was a slight reduction of low frequency peak-to-peak variation by playing back PEC curve on the mount. I'm not convinced that it is being done properly, however. I really should not have been antsy and should have waited a few cycles more pre and post.

Also note that my CEM120 exhibits similar high rate-of-change regions in unguided RA, although if I look over the same timer interval it seems like Mark's slopes may be a bit greater.

Al

[mark matzner]

Original poster here with a follow-up.

I've tried two algorithms besides the previously employed PPEC since my last post.  

a)  Ken's ZFilter (Bessel 4, Corners 4 & 8, exposure 1.5s).    I freely admit I did not know what I was doing with the ZFilter algorithm and it's probably worth more exploring.

b)  Hysterisis (exposure 1.5s).

I found the Hysterisis to produce ever so slighly better results than the PPEC and ZFilter.

General comments:

1.  The previously tried PPEC was with a 3s exposure.  Probably should have gone back  once more with the 1.5s for a fair comparison.

2.  All three algorithms suffered from nights of sporadic clouds and hazy skies.

3.  The last two algorithms, ZFilter and Hysterisis, saw an increasingly bright moon.

Hysterisis comments (GuideLog_2018-08-25_205533):

1.  This may be due to the conditions of one night only, but it appears that the Hysterisis produced slightly lower magnitudes of RA and DEC, and a bit rounder scatter.  

2.  The attached Hysterisis log file also shows better performance (IMHO) on Log Section 2 vs. Log Section 1 (see closing comment).  

3.  The RA periodic excursion, for which the mount was serviced by iOptron, is still present.  At least in Log Section 2 (sop-east).  It is substantially reduced from prior to the service, but still there.

FWIW, there was a sizeable difference in guide star  magnitude between the two sections.  Log Section 1 (sop-west) had a very dim star field to choose from (SNR 31).  Stars were very difficult to see in the PHD2 window.  

OTOH, Log Section 2  used a very bright star for guiding (SNR 131).  

I welcome any thoughts and/or questions.

Mark

https://www.dropbox.com/sh/hrxfi1ovsi2qqrt/AAAVHdYVY1KlBQPRc9qFcJTXa?dl=0

[Rainer Ehlert]

Hi Mark,

I saw this on your last Guide Log.

Mount = iOptron ASCOM Driver for Mount,  connected, guiding enabled

while in my Guide Logs I have this

 Mount = iOptron ASCOM Driver for CEM120,  connected, guiding enabled

I recommend to upgrade to the latest ASCOM driver from iOptron 

I see this values for your guiding and I think you are guiding too agressively

X guide algorithm = Hysteresis, Hysteresis = 0.100, Aggression = 0.700, Minimum move = 0.180

Y guide algorithm = Resist Switch, Minimum move = 0.150 Aggression = 100% FastSwitch = enabled

Interesting is that you use Backlash compensation. Make a test and disable that. From my point of view a mount like the CEM 120 with the floating worm system does not need Backlash compensation even if the Guiding Assistant from PHD2 gives you a backlash value.

Backlash comp = enabled, pulse = 785 ms

I ran the Guiding assistant and got backlash values out of the world varying between 340ms and 3000ms, on the same mount over idfferent nights, and I just politely ignore that and my guiding is very tight.. My Aggression values are never more then 40%.

Rainer

[mark matzner]

Thanks Rainer,

My current Ascom driver for iOptron is v6.02.1.  So, in goes the v6.1 and we'll see what happens tonight.

I believe you're recommending my reducing the RA Hysterisis aggression from the 70% to something less, i.e. maybe 40%.  We'll give it a try.  

Keep in mind that my mount is a non-encoder version if that makes a difference.

I'll wait on the Dec suggestions until there's improvement on the RA.  Currently Dec by comparison is pretty good.

Appreciate your help,

Mark

[Rainer Ehlert]

Hi Mark,

Could be that my mount reacts a bit better then yours due to having encoders in RA but when it comes to the DEC, also with encoders, I do not see what the encoders would do right now in regard to correcting anything as the DEC axis does not move at all actively and the ecnoders do not dinamically adjust any position of the DEC axis except perhaps compensate refraction issues but these are long term corrections.

So when I saw that DEC RMS was always much bigger then RA I also went more and more aggressive and it did not improve and so I said, I do not loose anything if I go the opposite way and did set values below the GA recommendation and very low aggressivity and guess what. Now sometimes my DEC is under RA RMS a dn I ahve to change the values and go up in order to balance it.

When we apply,at last my theory too high values in MinMo then we are waiting until the error comes and we have to pull it back aggressively a long way, but if we use low MinMo values we do not need to pull it back a long way and can use low aggression values and we pull back more often but the CEM 120 do react very good to this as far as I have seen.

It does not cost us anything except invest a bit of time and sometimes we think that doing this costs us a lot of imaging time. I do not think that way, the time I invest now in learning to know how my mount reacts is experience and the time is paid back many times with less problems during imaging.

Rainer  

[mark matzner]

Hi Group,

Here's the latest PHD2 guidelog from last night:  Guidelog_2018-08-26_204033.

This is for the same target (M27) from previous runs, with the same settings save one:

RA Hysterisis, Aggression 70%, exposure 1.5s

Following Rainer's suggestion the latest iOptron ASCOM driver (v 6.1) was installed and the first sub of Log section 1 was set with aggression at 40%.  All subs for the rest of the session were then set back at aggression 70%.  

I felt the lower aggression setting (i.e. 40%) produced a higher RA magnitude.  Although a single sub (especially the first) may not have been sufficient to make a value judgement.

That aside, the rest of the evening was very consistent with the previous night's run.  Log Section 1 was at sop-west with a very weak guide-star field (SNR 35).  Log Section 2 was at sop-east with a much brighter guide star (SNR 130) and showed both the RA and Dec RMS improved.

Out of curiosity I checked the setup's balance in both the sop-west and east positions for this target.  In both postures there was an ever so slight bias to the west.

For tonight, I intend to raise the mount's RA guide speed from 75% to 90%. 

 

Thanks Rainer for your input and I welcome comments from the group.

Mark

https://www.dropbox.com/sh/29jjfmoo54q5mxq/AABb_3OIL3XyZCfqt5yzKkVBa?dl=0

[Rainer Ehlert]

In both postures there was an ever so slight bias to the west.

Hi Mark,

I was thinking about above ...

¿ how is that possible ? If one leg is heavier then the other it should always go down to the side in which this leg is ...

... and how much is ... " an ever so slight " :-) 

I can check your Guide Log in the afternoon. I can not do that in the Office :-)

[Rainer Ehlert]

Hi Mark,

I just looked at your Logs.

The first test with aggression 40 was made at an altitude of 

RA = 20.01 hr, Dec = 22.8 deg, Hour angle = -2.43 hr, Pier side = West, Rotator pos = N/A, Alt = 56.5 deg, Az = 97.5 deg

The second test with your 70  was made at an altitude of

RA = 20.01 hr, Dec = 22.8 deg, Hour angle = 0.13 hr, Pier side = East, Rotator pos = N/A, Alt = 80.1 deg, Az = 190.6 deg

and so I doubt about the validity of the tests. I have guiding logs were I started high and as the telescope was nearing the horizon the RMS values got bigger and bigger without having changed the parameters. 

There is a difference of 24° and that is quite a lot ...

Anyhow the values look better now and are not 3 times higher just the double

All in all I do not like what I see in your RA axis. Have you commented this to iOptron ?

Looking into the analysis you have a PE of +- 4.8". Far too much for a CEM 120. They specify in their website +- 3.5" 

[mark matzner]

Hi Rainer,

1.  Regarding the balance:  The behavior is as I described it.  The reason for checking this was to determine if there was a substantial imbalance in the setup.  I do not see that.

2.  Regarding the two Aggression settings:  The session started out at 20:43:01 with the aggression set at 40%.  The first sub commenced imaging around 20:46.  At the completion of that first sub there was an AF, and I took that opportunity to change the aggression from 40% to 70% (default).  The next sub commenced at 20:59 and should have featured the 70% setting.  I believe the affect is immediate, but stand corrected if not.

3.  Regarding the RA behavior and iOptron Support:  I will explore a few more options for improving the RA. Then we'll see.

Thanks for your comments Rainer.  Mine is not the sharpest of minds (at least at times), and all help is appreciated.

Mark