PDA

View Full Version here: : How good is your pointing / tracking?


g__day
21-11-2007, 03:58 PM
Ladies and Gentlemen,

I thought this might be an interesting thread to develop over time. Please answer two questions and lets see what is possible by various sort of gear in real world conditions.

1. After good polar alignment - how well can you mount point (on say 3 well spaced stars)?

2. How well will your mount track - what have you been able to keep right on target? (Start on the middle of the moon (30 arc minutes width - how long does it take to move the centre of the eyepiece to the edge of the moon ~ 15 arc minutes or 900 arc seconds)?

For example, given alignment with 30 arc seconds of the pole, 1 star align, thru Maxpoint 30 star modeled sky, Vixen Atlux carrying 3 scopes:

1. Pointing +/- 1.7 arc minutes
2. Tracking < 10 arc seconds per 10 minutes - so under an arc second per minute fast

Be real interested understanding what folks gear can achieve!

Thanks,

Matthew

montewilson
21-11-2007, 07:36 PM
With the Takahashi EM-200 I get, according to TPoint, an all sky RMS pointing of about 20". That is after a 100 or more point Automapper II run.

Once it is aligned to within about 15" of the pole with a sigma of about 10", I will get a periodic error of about 5" which fits with what the manufacturer advertises. There is no drift.

I am sorry I can't give you a drift answer in the format you requested.

In all with TPoint I am pretty happy with the results.

Lee
21-11-2007, 07:37 PM
Don't know numbers..... My G-11 now it is pretty well polar aligned (approx 2 arc minutes off apparently) puts objects in the centre of the DSI CCD everytime.
Tracking - unguided 1 minute good, 2 minutes hit/miss (600mm F/L); guided - round stars at 10 minutes so far......

jase
21-11-2007, 08:16 PM
Mount: Losmandy Titan (not the 50:1 gear ratio version that gives more torque), but running 18v 5amp to it, so it has plenty of juice.

After an automated 10 point model run I achieve the following (transcript output from ACP):
13:29:08 Start slew to Map-010...
13:29:10 Az = 358 Alt = 41
13:29:10 (wait for slew to complete)
13:29:36 (slew complete)
13:29:36 Acquiring mapping point...
13:29:48 (taking 10 sec. exposure, Clear filter, binning = 1)
13:30:09 (HFD = 3.09)
13:30:09 (avg HFD = 2.88)
13:30:09 Plate-solve pointing image.
13:30:10 233 image stars found
13:30:10 1742 catalog stars found
13:30:11 Solved! 156 stars matched.
13:30:11 Average residual is 0.87 arcsec.
13:30:11 Pointing error is 1.023 arcmin @ angle 345.82
13:30:11 True focal length is 53.1 cm.
13:30:11 True image center (J2000): 00h 19m 52.8s 10 42' 00.98"
13:30:11 Imager sky position angle is 74.8 deg.
13:30:24 [sync] pointing model updated
13:30:25 Training run completed at 13:30:25 UTC

This is after 10 mapping points. 1' (arcmin) is pretty high (in my opinion). My model now consists of around ~20 points, but I haven't had time to review the logs. I know its considerably less than 1' now, but can't tell you the exact value until I head down to the remote observatory again - sat comms are presently down so can't access the observatory computer.:rolleyes:

Tracking - I've been able to do ~10 minute unguided (tracking only) exposures with the FSQ operating at a focal length of 530mm (native). This is quite a forgiving wide field focal length (unlike 2000mm+). The STL11k delivers a 3.5 arcsec/pixel with the FSQ, so the periodic error (PE) ideally should be below this value. In my case, the measured PE (from PEMPro) on the titan is around -/+ 1.5 arcsecs well within the range. If we are talking guided (which isn't what I consider "tracking"), then the longest exposure I've taken is 20min with no signs of field rotation. This was sometime ago and I've since revisited polar alignment. I feel sure I could go for 30min now, but haven't had the opportunity to test this.

Things to note:
If you're using a SCT or other "sliding" mirror focusing system, there will be some image shift (even if you can't visually detect it). This will produce erroneous pointing results.
Don't eyeball the pointing model alignment stars, this will result in model inaccuracies. As you can see from the ACP output above, I use a CCD camera to automatically take an image, plate solve it against a known catalogue, then automatically slew to center the star - subsequently the pointing model is updated. In fact, every time a slew to an object and take an image, the pointing model is updated. This ensure accurate pointing across the sky.
Good polar alignment is key to avoid field rotation at extremely long exposures. Don't think that because your autoguiding, you images will not be subject to field rotation. You'll find with poor polar alignment the image will rotate around the star your guiding on.
Reduce you mounts PE and backlash as much as possible...and don't forget good balancing of both RA and DEC axis. Software tools are great for PE improvements. Don't forget evil flexure - the hidden enemy.
You're CCD camera has better vision than you. Use it to your advantage. Use it to polar align as it will detect drift much faster and with greater accuracy. Use it to plate solve and build pointing models (automatically if you choose). Even using software generated cross-hairs on the image output from the camera improves accuracy.Hope this helps.:)

gary
21-11-2007, 10:28 PM
Hi Monte,

That's an excellent result. How many terms are in your model at present?

Best Regards

Gary

ballaratdragons
21-11-2007, 10:54 PM
I can't give fancy figures coz I don't understand what the fancy figures mean, but I can tell you how my mounts behaves.

This is using a Toucam and ED80, so it is equal to testing with a 6mm EP in an ED80 @ f7.5

The mount is a Skywatcher EQ6 running v2.5a

I have polar aligned by drift aligning for over 10 mins on each axis x3 times and get Zero drift (using crosshairs in Guidemaster)

On 3 star alignment, all 3 stars within 1/3 of centre of the finderscope camera.
After 3 star alignment, 90% of GoTos are dead centre of 6mm FOV, 10% are slightly to the side but still in the 6mm FOV.

Tracking is slightly erratic due to PE, but objects stay within about 10 arcseconds of centre FOV. Only once has it gone as far as about 30 arcseconds.

Guiding with GuideDog I was getting constant accuracy of within +/- 5 arcseconds.

But now Guiding with Guidemaster I get constant accuracy of within +/- 1.5 arcseconds.

Hope that is the type of info you are after.

jase
22-11-2007, 07:39 AM
ok, comms are back up so can access the remote observatory again....

15:14:08 Start slew to Map-022...
15:14:10 Az = 254 Alt = 56
15:14:10 (wait for slew to complete)
15:14:26 (slew complete)
15:14:26 Acquiring mapping point...
15:14:29 (taking 10 sec. exposure, Clear filter, binning = 1)
15:14:40 (HFD = 2.00)
15:14:40 (avg HFD = 1.80)
15:14:40 Plate-solve pointing image.
15:14:40 590 image stars found
15:14:41 219 catalog stars found
15:14:41 Solved! 82 stars matched.
15:14:41 Average residual is 0.59 arcsec.
15:14:41 Pointing error is 0.621 arcmin @ angle 344.18
15:14:41 True focal length is 53.1 cm.
15:14:41 True image center (J2000): 02h 55m 23.7s 60° 24' 40.02"
15:14:41 Imager sky position angle is 92.5 deg.
15:14:51 [sync] pointing model updated
15:14:52 Training run completed at 15:14:52 UTC

This is after 22 mapping points. I feel certain if I performed a similar 100 point mapping run like Monte, I'd get reasonably similar results. In my situation, I don't see any benefit going beyond a 40 point model. Considering I'm plate solving to determine the precise image center, the telescope will simply slew to the image center coordinates I request if its above the defined tolerance threshold.

jase
22-11-2007, 01:54 PM
Matt,
Have you tried SmartTrack to improve your tracking/pointing? It uses your maxpoint model as its tracking engine. This operates in a similar fashion to Software Bisques ProTrack for their Paramount ME.

http://winfij.homeip.net/maximdl/smarttrack.html

You'll need MaximDL.

montewilson
22-11-2007, 04:28 PM
Hi Jase - I agree with you, that there is not a lot gained by going for more points. In your case this is certainly true. I go for a lot more, over 150 as it helps me get reliable polar alignment information. With Automapper I can go and do something else and it will knock over about 1.5 points per minute.

This is important to me because I don't have a permanent set-up so I need to be efficient in polar aligning and very accurate. If I don't, as you know, I will get field rotation my large FSQ frames.

In your case, you are already well aligned so the mapping run helps with more with pointing than anything else.

montewilson
22-11-2007, 07:28 PM
Hi Gary - When I wrote that, I was a long way from the black box so I had to try to remember the value. Today I checked it and I found a run of 159 points gave me an all sky RMS of about 29" not 20" sorry for the off the mark guess. Anyway I am happy with that.

That was with about 15 terms. I could have added more but as you told me too many terms and some will start eating each others lunch and their contributions will be indistiguishable from each other and the whole model would become noisy.

g__day
23-11-2007, 12:09 AM
It really interesting to see what is possible in the big leagues! Jase how do you install smart tracker once you download and unzip it - it just says Runtime error 429: ActiveX can't create object

jase
23-11-2007, 12:54 PM
Not sure Matt. I ran it a while back. All I did was place the executable in the Maxpoint directory. It executed fine. I don't run Maxpoint now, so I can't comment if anything has changed. May want to ping the author an email or get yourself a copy of ACP (which already has this tracking routine integrated).

g__day
23-11-2007, 06:00 PM
D'oh! Tried it on a PC - that didn't have MaxPoint loaded! Installs fine on one that does - now is there a manual?

jase
23-11-2007, 06:08 PM
From the website I provided, not exactly a manual so you'll need to join the dots...

Usage:
The SmartTrack application relies on the MaxPoint model being accurate enough to calculate small errors in the mount pointing. In order for this to be true, you need to ensure it has enough measurement points on each side of the meridian you will be tracking. I would recommend having at least 30 points on each side of the meridian, or more if your mount has a large amount of error reported by MaxPoint. This is easily achieved by using the auto-mapping feature of MaxPoint in combination with MaximDL.

Once a sufficiently accurate mount model has been built up, simply slew to your desired target, set the desired maximum allowable error in arcsecs and click the “Start SmartTracking” button.

This will change the status line from “Normal Sidereal tracking” to “SmartTrack Tracking” and the current deviation from ideal tracking in RA/DEC will be displayed.

Whenever these errors go above the configured allowable error, a pointing adjustment will be sent to the mount to reduce the error.
To stop the SmartTrack corrections, simply click the button again and the mount will revert to normal Sidereal tracking.

If SmartTrack detects the mount is slewing, it will suspend making corrections until the slew completes, after which it will update its internal target coordinates to maintain the mount pointing at.

This means you can start SmartTrack at the start of your imaging session and leave it running while you image and slew, with no manual intervention required.