Jump to content

Narrowband

PHD2 graph - Comments please


Recommended Posts

OK, scope tracked for just over an hour, and all seemed as it should.

I've attached screen captures of EQMod settings.  The driver correctly had the right information when opened, which was confirmed by re-selecting the HEQ5/Sirus 4:1 option.  The two images on the right were the before and after the run which lasted just over the hour.  I've also included a screen capture of CdC showing it was locked on the position throughout the test

Comments on a postcard please..... :)

Hopefully if we get a clear night I'll try the other suggestions, having attempted finer focusing on the star as suggested

 

eqmod settings.png

cdc.png

Link to comment
Share on other sites

Looked at latest guidelog, which will be with the old settings:

Guidespeed only 6arcsec/sec, up it to 12arcsec or more, with a new Cal

Calibrating at Dec 21.4, should Cal at Dec 0 is where the stars "move" fastest, so PHD2 has most movement to Calibrate on.

Guidecam HFD is 6 pixels, use the HFD reading in the Star Profile to get it down to minimum, 3 to 4 is good.

RA rate is half the Dec rate.

Here's the last Cal in the log:

 

Cal12th.JPG.da4d72ea476dae92eb6d756e0a4ac09a.JPG

I'm beginning to notice the blue East overshoots E0 to E5 on the return journey

I'm not sure if this is because the RA rate is half what PHD2 is expecting, or because of RA backlash - there are 4 to 5 RA backlash steps in the Cal

As before, Dec is 12 steps at 7.5 arcsecs/sec which is fine, but RA is 22 steps at 3.7 arcsecs/sec which is about half the setting of 6 arcsec/sec.

Now RA:

 

RA12th.JPG.fd4338933718309874e1d447587af192.JPG

 

The star mass and SNR at the top of frame is bouncing around, poor seeing ?

The horizontal blue dotted line is where your Min Move is set, at 1 arcsec.

Note how the large up spike above the RA label is the result of over correction of the downward spike before it (which itself might be due to chasing the seeing) - so too much Aggression ?

So this Min Mo setting means RA isn't being corrected until the error is bigger than 1 arcsec !

Now Dec:

Dec12th.JPG.cabf179e56877214f8f27f85eee681a4.JPG

 

PA is good, but Dec is bouncing around, perhaps due to the poor seeing suggested by the Mass and SNR graphs, so perhaps your balance and PA are too good?

PA doesn't need to be any better than 5 arcmins, this will give a gentle oneway drift that can be easily guided out, and slight off balance will stop Dec yoyo-ing

I notice in the image of your Advanced Setup/ Guiding, the Use Dec Comp box is unticked. This is preset to on, so that PHD2 can compensate the RA steps at higher Decs. Did you untick this?

So see earlier post for suggestions for your next session.

Michael

Link to comment
Share on other sites

Hi Michael,

Yes the seeing wasn't great last night, the sky was fairly "milky" and then clouded over altogether as you can see by the end of the log file.

The DEC Comp box was checked by default.  I unchecked it to see if it made any difference.  I'll reinstate it for the next session.

When you say that calibration should be at Declination zero degrees, I was always under the impression you calibrated on a star in the field of view of the target rather than on some random star at zero dec ?

 

So planned course of action for the next session.

  • Focus finder / guider using PHD2's star image to get a low HFD
  • Load up PHD2 and locate a gude star on or near the meridian at zero Declination
  • Check the stars HFD
  • Confirm DEC Compensation is checked
  • Reset all settings under the Algorithms tab of the settings.
  • Run the guide assistant and accept any settings it advises
  • Calibrate and run a guide session for 30 minutes (regardless of any warnings when calibrating ?)
  • Re-calibrate and run the same guide session, this time with PE enabled 

Should there be any other points to add to the list ?

Link to comment
Share on other sites

28 minutes ago, malc-c said:

When you say that calibration should be at Declination zero degrees, I was always under the impression you calibrated on a star in the field of view of the target rather than on some random star at zero dec ?

Run Guide Assistant at the target.

If you have a ST4 type setup with no mount position info being sent back to PHD2, you should Cal on every new target, every night.

Because the amount of guiding required by and large changes as Dec changes.

You Guidelogs are reporting the mount position, so with that Dec Comp checked, and after a Cal at Dec zero or thereabouts, with that single Cal PHD2 will compensate for any target every night, if the guidescope orientation isn't altered.

Re your plan, the PHD2 Developers recommend starting a new Profile rather than tweaking what you have a) it's too easy to miss a setting b) you have a old one to go back to or to compare.

Cal first at Dec 0, then run Guide Assistant at your target, or at the position you intend to test at.

I don't think you need to Cal again after PE is enabled, it doesn't change the guide rate.

If you have time I'd like to see the effect of Cal and guiding with those different Guide Rates for Dec and RA !

Michael

 

Link to comment
Share on other sites

35 minutes ago, michael8554 said:

 

If you have time I'd like to see the effect of Cal and guiding with those different Guide Rates for Dec and RA !

Michael

 

Do you mean setting the sliders in EQMod... if so what rates would you like me to set and would 30 minutes per log be OK

Link to comment
Share on other sites

I managed to grab an hour or so in late twilight to do some testing.

Started with a fresh profile, ran through the wizard let it select the defaults, none of the settings were manually tweaked, and let the applications take dark frames with the lens cap still on the camera.  Removed lens caps and covers and set the scope up to zero declination and then rotated the scope in RA until the scope was pointing south and luckily the camera found a bright star.  Due to it still being quite bright the exposure was as low as 100ms with the screen contrast as far right. With a 200mS exposure the star had FWHM values between 0.85 and 1.75 - seldom over 2.  The star graph showed a sharp thin spike rather than fat based rounded image.

I kicked off guiding (log entry 1) and it took around 10 steps in each direction and reported no errors.  I didn't save the the result.  The mount started guiding (log entry 2).  I let it run for around 10 minutes.

I then ran the guiding assistant (log entry 3), but wasn't sure if I had done things right, so stopped and restarted it (log entry 4) .  This time the form was populated with data and I left it running until the PA stabilized around 0.0 to 0.1 acr min.  Just as I stopped the assistant the star was lost as the clouds had completely covered the sky.  The results of the guid assist are attached, I accepted the settings, and then closed up for the night.

One thing I noticed, is that with 1 second exposure the HFD of the star was 6-7, but with faster exposures the value dropped to 1.5-1.8.  Although the guide assistant did report the focus could be improved.  I might try searching the net for a bahtinov mask for the SW finder so that I can see if I can get better focus.  The thing is the computer and scope are in two different rooms making it hard to adjust focus whilst being close enough to the screen to see the results.

PEC was not activated in this short test session

guide assist.png

PHD2_GuideLog_2020-04-13_200715.txt

Link to comment
Share on other sites

A quick report as I'm imaging.

The change in RA and Dec Guide rates has given a good Cal, equal steps on both, none of those RA overshoot on the east return,  but there's that Dec Backlash.

Must be a reason why the drives are so different ? A 50% error can't be wrong size cogs. Was the mount behaving before the belt mod ?

That's 6 secs to change Dec direction. If you can't improve that on the mount try PA to about 5 arcmins and guide in the direction that pulls the Dec back to a straight line, with Min Mo a bit more than usual.

The Periodic Error on RA is about 50 arcsecs peak to peak, if the mount has PEC would be worth applying.

But more anon

Michael

Link to comment
Share on other sites

Hi Michael,

I have no idea on how the mount performed when geared as the changes were done in 2011 and I no longer have any log files.  As mentioned the belt drive was effectively the first belt drive mod for the HEQ5 - it uses the same toothed pulleys for both RA and DEC. 60t on the axis and 15t on the motor.  EQMOD set to 4:1 ratio as shown above.  I could try and take up the backlash by adjusting the worm backlash adjustment on the mount.

The idea will be to use PEC for guiding.

Link to comment
Share on other sites

Sorry, didn't spot you had PEC off for the test.

Having seen the latest Cal I'm beginning to realise that RA rates must be correct - your Sidereal tracking test confirms that.

RA true rate 1.367 at 13.5 setting

Dec true rate 1.170 at 6.0 setting (could go up a bit, try 6.5 or 7 ?)

So it's Dec that is the wrong rate - but GoTo's are correct ? Do they go through EQASCOM or by the mount's firmware ?

I've found the lengthy 2011 discussion on doing the mod, I shall have to plough through it tomorrow.

I did notice that early on the Tracking Offset (whatever that is) for 4:1 was -211, you have -189 ?

Perhaps you could do me a short precis of how the ratios are worked out from the motor pulse rate onwards ?

Michael

 

Link to comment
Share on other sites

13 hours ago, malc-c said:

the HFD of the star was 6-7

Hi. If the camera doesn't have one, put a uv-ir filter on the guide telescope.

11 hours ago, malc-c said:

The idea will be to use PEC for guiding.

Just change the RA algorithm from hysteresis to PPEC. It should flatten the sine in a few minutes.

 

11 hours ago, malc-c said:

adjusting the worm backlash adjustment on the mount

+1. Absolutely. Get out the allen keys!

Rowan supply belts and pulleys to obtain the same ratio belt driven as with the original direct drive.

Cheers and HTH

Edited by alacant
Link to comment
Share on other sites

10 hours ago, michael8554 said:

Sorry, didn't spot you had PEC off for the test.

Having seen the latest Cal I'm beginning to realise that RA rates must be correct - your Sidereal tracking test confirms that.

RA true rate 1.367 at 13.5 setting

Dec true rate 1.170 at 6.0 setting (could go up a bit, try 6.5 or 7 ?)

So it's Dec that is the wrong rate - but GoTo's are correct ? Do they go through EQASCOM or by the mount's firmware ?

I've found the lengthy 2011 discussion on doing the mod, I shall have to plough through it tomorrow.

I did notice that early on the Tracking Offset (whatever that is) for 4:1 was -211, you have -189 ?

Perhaps you could do me a short precis of how the ratios are worked out from the motor pulse rate onwards ?

Michael

 

 

At the time of developing the belt drive my experiments with commercially available pulleys resulted in the 4:1 ratio as being the best compromise for speed and torque.  Smaller pulleys resulted in failure having enlarged the hole to fit the motor.  Chris who developed EQAscom modified EQMOD to handle the 4:1 ratio (in fact he added additional 5:1 and more ratios as well)  He did all the calculations and recompiled the application.  Now whilst there were a lot of people using EQMod to control their scopes, the belt drive was not an option for those who used the hand controller, so this is were Dave stepped in.  He owned a business that has CNC machines, so milled the motor pulley from one piece of material so as to maintain the default ratio without the motor pulley failing.  With the PA as good as it is, my goto's have been very good.  The HEQ5 was purchased second hand in 2011, but was stripped and serviced in 2012.  I'll get the tool box out and have a look at tightening up the worm to see if that improves backlash.

As mentioned, I've downloaded the latest versions of these applications.  I could re-install older versions, but to be honest I personally doubt if that would make any difference.  Let me revisit the scope balancing, finder focus and adjust the DEC worm and see if I get any better results.  

 

Link to comment
Share on other sites

1 hour ago, michael8554 said:

So you've no ideas as to why the guide rates are so different ?

Michael

 

Nope !

There is no goto error in DEC.  Originally I released the clutches on the mount, set it to park in the traditional home position with the weights down pointing at Polaris.  Once parking was complete eyeballed the position as close as I could and locked the clutches.  Picked Procyon as a target, slewed and it was close, just needing the RA clutch to be slackened off and the mount rotated slightly westwards.  DEC was spot on.  The star was synced in CDC, the scope parked, and the star reselected... it was bang on.  Selected Pollox - slewed and bang on.  I repeated this for around four other targets and they all appeared within the field of view of the d400, most fairly central.  So the ratios and EQMOD settings would appear to be OK.

One thing I have done today is (hopefully) sorted out the DEC freeplay/backlash.  With the scope ballance checked again I noticed a very, very slight movement in the DEC axis when the clutch was locked off.  When locked with the tube horizontally, pushing with the finger on the mirror cell, you could feel a slight bit of free play.  It was almost impossible to visually see the movement, but you could feel it, so I spent an hour with the allen keys out adjusting the worm guides until the mount had no movement at all, but the gears still engaged and didn't bind or slip.  The mount was then driven in all rotations on both axis to ensure  there was no slipping, or stalling.   I've now just got to find a means of being able to fine tune the guider focus and hopefully this nice clear blue sky will remain so I can test the results later. 

Link to comment
Share on other sites

OK, things are not going well....

Managed to get better focus of the finder.... aligned the finder up with the main scope and the red dot finder so everything is pointing at the same thing.

Set the DEC to 00.00 and RA to around 9/10 hrs and opened PHD2 - cleared previous calibration data and let it select the start.  The good news is that on 1sec exposures the HFD is low so the effort to focus the finder seems to have paid off.

But, no matter what I do I now get error messages stating there is a difference between the axises as a result of huge backlash errors - but the mount is now tight where as I found movement this afternoon.

 

I've run the guide assistant (see attached) and accepted its findings - I've set it running with PPEC enabled in the settings (disabled the EQMOD option).  I'm running a data run at the moment and will load the log files later, although with the constant stop starts whilst I tried getting the guide assistant to run (user error) its a bit of a mess !

 

Maybe the mount is just worn... ??

guide assist 2.png

Link to comment
Share on other sites

36 minutes ago, malc-c said:

Maybe the mount is just worn

JTOL...

Or probably sticking after over adjusting DEC? Try the north adjuster 1/8 turn anticlockwise, south the same clockwise.

Try 2s guide exposures.

Cheers

Link to comment
Share on other sites

Well RMS readings on your last two runs were 0.7 to 0.8 arcsecs  - RA and Dec are similar, so you should have round stars. 

The star Mass and SNR were all over the place indicating poor seeing, so not bad for 1 second exposures that were probably "chasing the seeing" ??

The RA overshoot on the east return of Cal is back, maybe balance has changed, RA no longer always meshed ?

2 secs Dec backlash isn't so bad, does anyone get it perfect ? I'd leave it alone.

I'd say you're nearly there.

Michael

Link to comment
Share on other sites

Thanks for the reassurance.  

The thing I don't get is that having found there was movement in the DEC axis (backlash or free play) I tigherned things up so the movement went away but didn't bind the axis.  The RA wasn't touched, but the scope was rebalanced.  Yet now we have E/W overshoot on the RA axis, and PHD2 is complaining about backlash after a calibration run, which it didn't do previously.  Yet the actual guiding seems more precises looking at the target reticule and graph.

I didn't take any images to see what the stars looked like so can't comment on that.  The seeing was deteriorating, with a noticeable foggy glow around Venus, and the sky had a thickness about it, making it hard to see fainter stars.  You would think in this current lockdown, with less cars on the roads, and most airlines grounding the planes that the atmosphere would be a lot more transparent and stable.

I'll check for end float on the gears, but won't try and improve any other alignment.  I may also look at the balance of the scope, maybe I missed something (lightbulb moment ! ) - One thing I did was to balance the scope with all the covers off.  Maybe previously when balancing the RA axis it had the covers on and it the weight of the cover causing the CG / balance to be just slightly out, but just enough to induce the RA overshoot ?? - Nothing on the RA was touched today.  The balance was DEC only with the CG checked (placing the mount with the weight bar horizontal, loosen the DEC clutch and place the OTA vertical.  If it then wants to swing rotate the OTA until it balances and remains in any position around the DEC axis).

Comparing the previous guide assist report with tonights, it's heading in the right direction, I've reduced the backlash error by over 4s :)

Link to comment
Share on other sites

The typical backlash in either axis can't be felt by rocking, it's just to small to detect.

In other words if you can feel it it's miles out !

3 minutes ago, malc-c said:

the weight of the cover causing the CG / balance to be just slightly out, but just enough to induce the RA overshoot ??

Wrong way round - in your case the RA balance NEEDS to be slightly out so that the gears are always in mesh = no backlash.

Yes a great improvement in Dec backlash.  PHD2 will tell you about the Dec Backlash but that doesn't always mean you can't go ahead and guide.

Once you have a good or acceptable Calibration you use it all the time.

Michael

Link to comment
Share on other sites

Well I've just spent the past couple of hours rebalancing the mount, re-routing cables so (hopefully) the don't impact the results, and adjusting the worm gear tension / mesh.  It will either improve things or make them worse.. All being well if it remains clear and the high level thin cloud goes away I'll give it another test tonight.

Link to comment
Share on other sites

Guys, I followed the same process at other nights...

  • Slewed to Procyon, released clutches and centred the star in the finder and locked them into place
  • Set DEC to zero degrees (or as close as I could to within 00:00:03
  • Set RA to around 9 hrs
  • Opened PHD2 and created a new profile
  • Set the exposure to 1.5s
  • Auto selected star
  • Calibrated - received a warning that RA was 104% of DEC rather than 100% - closed and let it begin guiding
  • Fell of my observatory seat ! (see attached)

The mount is running with the PEC file generated through EQMODs PECPREP 

I've attached a couple of screen captures whilst its gathering data, and naturally I'll post up the log files later.  I'll run the guide assistance later, but I really want to get an hour or so's data

 

121012105_guiding15-4-20.thumb.png.08127b70f629b5a4090d8f9a85f52e0a.png

 

 

Link to comment
Share on other sites

1479750385_guideassist1.png.2bc104b791dfe7df70beb0466ef56d7c.png

 

Here's the guid assist results.  The mount was running PEC, which may be the reason why the Polar align went from 0.8 through 0.0 and back up to 0.6?  But I've reduced the backlash from 6305 to 701 which is the main thing.  As things are going so well I didn't apply the recommendations.

I've now moved to a target (09:33 RA, +21:24 DEC) and the guiding is just as smooth.  Running off 20x 4 minute subs and the stars are nice and round :)

Link to comment
Share on other sites

Excellent work on the backlash, I'm impressed.

Looking at the Guidelog, every comment starts with "good".

Good HFD = 3.09

Good PE

Good backlash.

However, to rain on your parade:

You didn't enter the focal length of your guidescope, so the Cal and guiding figures are, well, not Calibrated.

Explains why Cal has jumped up to 26 steps.

So the guiding results of RA = 0.15 arcsecs, Dec = 0.10 arcsecs, are out of this world.

If you use the Profile Wizard you can't miss any settings.

Also in your screen grabs, the graphs are set in Settings to show pixels instead of arcsecs.

You need to see arcsecs  if you're aiming to tweek settings for the magic sub 1 arcsec guiding.

Almost there though !

Michael

Edited by michael8554
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.