Dec Worm Adjustment


W. Christopher Moses
 

For no particular reason, I decided to adjust my dec and see if it had any effect.  It did.


All I did was use some feeler gauges and make the right side a tad bit tighter.  I then spun it around all the way by hand to make sure it wasn't too tight.


When I did a guiding assistant tonight, the dec was 0.51" for about 10 minutes.  That's unguided!  For me, that is a giant improvement.


Keep in mind that my guided performance was just fine before, so this was really just tinkering for the sake of tinkering.


Jeremy Parker
 

Hi Chris,

Of that 0.51", was there any drift component? When you think about it, if you have good polar alignment, the DEC should not have to 'guide' much at all, because there isn't PE to deal with. If there wasn't a breeze or other small physical movement of the mount, I would also get better unguided DEC performance than guided. So I wonder if your guided DEC performance has also improved? My guided DEC performance seemed to degrade because of backlash, and overshoot, until I backed off on DEC aggressiveness in PHD. So if backlash is reduced, I would imagine that your guided DEC performance should also improve.

Jeremy


W. Christopher Moses
 

Hi,
My PA was very good, but I've done equally good PA before and have never seen it so low unguided.  My guided dec performance was even better.  I'll have to check my notes, but my total RMS was 0.45"-0.55".  That is really good for me.


W. Christopher Moses
 

I uploaded the guide log to the Mount Related section.  Unfortunately, I was only able to get about 40m data before clouds rolled in.


W. Christopher Moses
 

Another great night of guiding...best I've ever had, actually.  Except for some gusts, the Total RMS was almost always below 0.5". 


Jeremy Parker
 

Hi Chris,

I looked at your guidelog data, and wow, you're PE is very low! It seems to be less than 6" peak-to-peak when looking at the unguided section, drift-corrected. I'm still stuck around 15" peak-to-peak. But there is a noticeable RA drift, so you could probably tweak your RA offset number. According to a presentation I watched on TIAC (The Astro Imaging Channel), RA drift does hurt RA guided performance.

But you're PE is so low, that it doesn't look like PHD is working too hard. If you exclude some of the frames where there might have been something going on (breeze, cables moving?), the RA error is 0.36" RMS, which seems excellent to me.

The DEC guided data shows that the corrections pules are almost always alternating north and south, with corresponding oscillations in the DEC trace. It also looks like you're using Hysteresis for the DEC algorithm? I think you could probably improve DEC even further by using the Resist Switch algorithm and back off the DEC aggressiveness parameter until the oscillations in the DEC trace are minimized. (I use a value of 50 for DEC.)

But your numbers are so good, at least compared to what I typically get, that you're probably getting pretty sharp/tight stars in your subs I'd imagine.

Jeremy


W. Christopher Moses
 

Hi,
Yeah. I  was thrilled with those numbers and they were even better last night.  I wanted to tinker, but it was so good I went ahead and imaged.

I thought I was using the PPEC algorithm.  I had assumed that was why I was getting so many corrections. I.e - it was frequently moving in anticipation of what was coming.  I'll double check that.

I'm tempted to try the same minor tweak to the RA axis, but if I can get sub 0.5 then I'm not going to touch it until I get some imaging done.  Last night I was regularly gettting sub 0.4.  Clearly the seeing has been very good, but still...

I did make one other change.  I spent a fair amount of time analyzing how much PA routines differ in results, both in an absolute sense and relative to each other..  The results are interesting.  Using that data, I think I truly had better PA than usual.  I'm working on putting the data in a spreadsheet.  I'll post it as soon as I do.


Jeremy Parker
 

Hi Chris,

Yes it looks like you were using PPEC for RA (called X guiding in PHD Log Viewer), but it was DEC that it looked you were using Hysteresis. Usually, Resist Switch is recommended for DEC, and it seems you have oscillations, so that might be an easy thing to try. And like I mentioned, you can use the Aggressiveness control to tweak it until you don't see oscillations/overshoots.

I also tried using PHD's polar alignment tools, but I found it wasn't agreeing so well with PoleMaster. So I'm sticking with PoleMaster for now, because I feel like my PA is always good. I'd be interested to see what your conclusions on PA were. Your DEC drift seems really low, but you might find it changes if you were to adjust your RA Offset Rate to minimize your RA drift.

In any case, yes, your numbers are impressive, and I think a big part is the low periodic error you have. You seemed to have really lucked out with a nicely machined worm, I guess!

Jeremy 


W. Christopher Moses
 

I just posted a summary of my initial PA testing:


Please let me know what other ideas you might have for testing PA techniques.


hubbell_jerry@...
 

Hi Chris,

Here something else to test. I am interested in how repeatable the alignment is for each method to understand the overall precision of the method. I would suggest that you do a polar alignment using each method at least 5 times and compare the results. Make sure you do your normal process whether you pick up your mount and  place it down on the same marks or move your mount to a different location when you start your alignment. Do you use a compass to do an initial gross algnment? I assume you level your mount when you place it in it's location and do not typically adjust the altitude until the PoleMaster or other program prompts you to.

If you can go through this process and get a series of alignment errors on the same night then we can get an idea what the real precision is on each method. If it takes you 15-20 minutes to do a polar alignment from scratch from when you first place your mount in its location then you will spend a couple of hours do this for each method.  I think it is important to understand the reputability of each method,

Jerry Hubbell
Director Electrical Engineering
Explore Scientific, LLC.


hubbell_jerry@...
 

Hi Chris,

I have looked at your most recent log files and loaded them into PECPrep to examine the period distribution and in both log files the peak periodic error was not the fundamental 240 second period, but a value of 75.8 and 75.5 seconds. This is kind of a weird period. It is not an integer division of the fundamental worm period. The most interesting part, and is most likely just a coincidence, is that when you divide the fundamental 240 second period by the 75.5 second period you get a value close to pi, which is basically an angular motion equal to 1 radian. Really curious. 

I am trying to wrap my head around what would cause such a PE.period in the G11. I am wondering if it is related to the coupler storing energy if there is a small amount of residual misalignment in the worm/motor coupling. Perhaps you can take a video of your worm and coupler with the cover removed and see if you can detect any small amount of movement 3 times during each worm rotation that could be causing this. To be sure, this is a very small amount of error and may not be worth the trouble to figure out, but I find it interesting.

Jerry Hubbell
Director Electrical Engineering
Explore Scientific, LLC.


W. Christopher Moses
 

Jerry,
I like all your ideas.  That was my initial goal, but when I saw how good my guiding was (.4-.5)  I cut things short to collect some image data.  Now that I have a few nights of images, I'll spend a couple full nights testing PA routines.

I will also be adding the new PdD2 static polar align to the mix, as I expect many people to start using it.

Please let me know any other ideas you have.


W. Christopher Moses
 

Hi,
My gross PA is just memory as I usually set up in the same location.  I then level it fairly well, in order to make the adjustments easier. I do not do ant adjustments until the programs prompt me too.


W. Christopher Moses
 

Thanks for taking a look!

I did notice the 70s spike but did not notice the ratio.

Taking a video sounds like a good idea.  I had it about as smooth as I could get it by feel before tightening things down.

I tried to put the Ruland couples on, just for fun, but they would not fit.  The hole going onto the worm gear was just a little bit too small. Oh well.

Overall, the adjustment seems to have helped, with the exception of that spike, which is visible on the guiding chart.  In fact, I was regularly getting .4"-.45". Obviously, it was very good seeing.

The only other change I made was to how tightly I screw down the clutches.  I used to really crank them down, but after turning the worm by hand as I adjusted the clutch tension, I decided to ease off significantly.

Whatever the cause of the improvement, I'm sure my images will look very good.

Thanks,
Chris


hubbell_jerry@...
 

Yes, it sounds like you have your mount tuned up the best it can be probably. Now start imaging and forget your mount!! It should disappear.

Jerry Hubbell
Director Electrical Engineering
Explore Scientific, LLC.


Jeremy Parker
 

Hi Guys,

The 76" error is a known issue with G11s, if you look on the forums/studies people have done. It is close to a 3rd harmonic, but not quite. Usually, it is associated with worm bearing misalignment and bearing race issues.


I have noticed it on my setup, especially if there is too much compression force applied between the two bearing blocks (to remove the side-to-side play in the worm) when tightening them, either on the stock setup or OPWB.

I just can't believe how low your 240s PE is though, and I'm left wondering what it would take for me to get even close that level. My stock worm in the OPWB is giving me 16" peak-to-peak, at best. The 2nd OPWB with new worm is giving much less peak-to-peak error, but the waveform is very strange, with many unrelated frequency components. I'll break this out to a seperate thread, or post it onto the OPWB thread when I have more data to share.

Jeremy


Mark Lamb
 

According to PHD2 GA, my DEC has backlash problems.  A few nights ago it was ~4100ms, and after loosening and tightening, it was 1568ms.  This seems to vary from night to night even without any adjustments (measurement error?).

Today, I just got some feeler gauges and checked the gap between the bearing blocks and the flats on the mount.  I checked the DEC, and the motor side bearing block had a  noticeable wider gap that the far bearing.  With my feeler gauge, I used the 2 widest consecutive blades, totalling 0.047", and adjusted both of the blocksto this gap.  This was definitely tighter on the motor side, and I did not record the pre-adj distances.

I then checked the RA worm distances, and they are tighter.  But the motor side has the smallest gap at 0.033" and the far block at 0.037".  I did NOT make any changes on the RA worm.

Questions:
Are there any recommended block gaps?  If so, what?

Is the small difference between my RA block gaps OK?

Should the gaps in the RA and DEC worms be the same?

What amount of backlash is OK?

Thanks!

Mark


W. Christopher Moses
 

I'm just pulling out my mount.  I'll double check my gaps, but my motor side was smaller than the other side.

I dont think the gaps in the ra and dec should have any relationship upon each other.

Backlash - that is a good question.  I've wondered that also. I'm pretty sure 1500 is close to reasonable and 4500 is large.  But, my GA failed to measure mine the last two nights and I had good guiding.  So, I wish I had a better answer but ultimately if the guiding is ok....



W. Christopher Moses
 

Jeremy,
Very, very interesting about the error.  I will look into that.

I wish I could tell you why my guiding improved so much.  All I did was
1. Remove, clean, and regrease the worm. (it had been open in a really buggy, pollen-filled area)
2. Try to get everything situated as tightly as possible without binding.
3. Switch to not tightening my clutches as much.  If you turn the worm by hand while tightening the clutch, you can feel how much of a difference it makes.

Have you posted any of your recent logs?  Please do


hubbell_jerry@...
 

Thanks Jeremy for finding more about the 76 second periodic error. I wonder if Scott Losmandy has done any work to improve this with the OPWB as stated in the linked article about the error. It sounds like that was the plan.

Jerry Hubbell
Director Electrical Engineering
Explore Scientific, LLC.