I have recently moved over to using a finder guider with my Lodestar for my guiding and my guiding is now in a bad way, previous I would happily guide at sub 3-4s throughout the night and now it is sub 8s. My googling suggested the new focal length is 200mm, PHD2 calculates the calibration step to be 1250ms but I'm not sure that's right.
I have copied the setting from phd2 below any help gratefully received; hopefully it won't be a month before I get to test any changes out
Screen output
(https://farm8.staticflickr.com/7450/12801965193_03f876b412_z.jpg)
Brain setting - Global
(https://farm8.staticflickr.com/7361/12809622845_9c3c64c481_z.jpg)
Brain setting - Guiding
(https://farm4.staticflickr.com/3823/12810031724_e52ccfac8a_z.jpg)
Brain setting - Camera
(https://farm6.staticflickr.com/5537/12809622755_557b03f642_z.jpg)
Brain setting - Mount
(https://farm8.staticflickr.com/7340/12810031614_0cfeaac912_z.jpg)
I've had good results with that combo. Basically the further the guider moves for calibration the more resolution your guide movements will have. Ideally the guide star should move half a image width in each direction though I have gotten away with half that or less - see what works for you.
Robert
Ivor,
Unless you have a low PE mount and are doing passive guiding then the period you are using for the guiding subs is way to long.
I'd suggest you use 1 second subs and make sure the calibration moves the stars quite a bit so PHD can calculate the angles and movements correctly.
If you can use Drupal then you will suss this quickly.
Chris
Thanks for the replies and the vote of confidence, the advantage of Drupal though is I don't have to wait weeks before I next test things out.
The length of the guiding sub is an interest topic I've had varying suggestions, Ian King told me from the start there was no point going lower than 1.5s with an EQ6 Pro as the components weren't good enough to support anything less. However I've had a number of recommendations to reduce this further.
The thing I'm kicking myself the most enthusiasm to play with my new toys I forgot to record my current settings. Actually whilst writing I just had a eureka moment I moved to phd2 and I still have phd installed to I have the setting there yes bring it on!!
From now on I'm going to be systematic and record the changes so I can revert back.
Oh to another clear night...
Quote from: Ivor on Feb 27, 2014, 08:27:30
I have recently moved over to using a finder guider with my Lodestar for my guiding and my guiding is now in a bad way, previous I would happily guide at sub 3-4s throughout the night and now it is sub 8s. My googling suggested the new focal length is 200mm, PHD2 calculates the calibration step to be 1250ms but I'm not sure that's right.
I have copied the setting from phd2 below any help gratefully received; hopefully it won't be a month before I get to test any changes out
Finder guiders are usually around 200mm. I would guess you have too long a focal length entered in PHD2, I would expect around 350ms depending on guide camera pixel size.
Your minimum move is way too small, seeing variation will make your guiding overwork. I find guiding improves if I use the default 0.15 or above, on a bad night I will use 0.40 and even then PHD is trying to guide out the seeing!
BTW My experience for getting the best guiding is to use zero hysteresis and a aggression around 80%.
I'm still trying to improve my guiding, but as U2 would say "I still haven't found what I've looking for". I'm grateful for all the help I've received from numerous sources, what has been interesting is the variety in the recommendations I've been given which in some cases contradicts each other and in others appear to go against the known wisdom. Part of the problem here I think is a lot of people have married the finder guider with cameras that lend themselves to this method better than the Lodestar. My search of the internet has at last found a person (AndyUK from the Stargazerlounge) who got the Lodestar working this way, I've yet to try his suggestions and I need to get my head around his settings as they appear to go against my understanding of the parameters.
Below is the best graph to date from Saturday night, as you can see sub 2 arc secs, but not remotely smooth.
(https://farm8.staticflickr.com/7401/13021656864_19517caba0_b.jpg)
I've started to be more systematic about my approach to this problem, as "Push Here Dummy" just isn't working; I have been changing individual parameters by small increments and waiting 2 minutes to see the difference. Before trying the next approach I want to confirm the fundamentals and have the best possible methodology for the next clear night.
I'm using an EQ6 Pro which I understand should be capable of 1 arc sec guiding; this leads me to my first point of clarity:
1. Seeing in Kent is limited to 1 -2 arc seconds so there is no point chasing below 1 arc second guiding as I'd only be chasing the seeing. Is this a correct statement?
My normal imaging is between F5 – F7 and the SW finder scope has an aperture of 50mm and FL of 180mm giving F3.6, FOV 1.96° x 1.51° and a resolution of 9.4 "/pixel which is significantly different to my old guide scope (a Bresser 70 now handed down to my son) which gave a resolution of 2.42 "/pixel at F10. My next point of clarity:
2. Where I have been assured this is a viable option, a resolution of 9.4 "/pixel is surely going to make this very challenging to achieve, a lot of imaging nights are going to be lost trying to get this working is it really worth the effort?
Guiding setting
I use PHD2, below I've list the parameters I've been adjusting.
MIN MOVEMENT = 0.05
To calculate this I need to know the pixels per arc second which is 1/9.4 = 0.111. PHD2 needs sub-pixel movements so I've divided this number by 2 giving 0.05. This seems very small to me and surely on the edge of the limits of software and hardware.
CALIBRATION STEPS = 5200
This has been set based on the PHD2 calculation tool, I've watched the calibration process and there appears to be no issues, I'm not sure how to really decide if this needs adjusting.
EXPOSURE LENGTH = 2s
With my old F10 setup I had this set at 2.5s, by dropping down to F3.6 I could reduce this further. However I have been told previously the EQ6 can't respond to changes too quickly so an exposure of less than 2s wasn't going to make a difference.
3. Is there a way to judge whether the exposure is too long or too short?
MAX RA/DEC = 300
These parameter I think needs increasing and I'm also concerned this needs to be aligned to EQMOD. I've checked the PHD2 log file and there are a few times where the correction hits 300 so I'm inclined to double this. The EQMOD setting are as follows:
(https://farm8.staticflickr.com/7325/13092630564_bc29500a5c_b.jpg)
4. Should I change any of the EQMOD settings if I increase the MAX RA/DEC to 600?
RA AGGRESION = 70 HYTERESIS = 17
These are interactive settings which have to be the most debated variables and appear to be particular to an individual mount. The current settings provide the best results but my graph is still rubbish and so subject to further change. The problem here is everyone's setup is different with different scopes and mounts so you can't really just take someone else's numbers. Saying that AndyUK had success with 100/22 so I'll start my next test run at 90/17 and see what the adjustments result in.
5. Is there a way to calculate these variables other than systematic trial and error?
All thoughts and insights gratefully received :)
Your exposure length of 2s is far too long for the EQ6. By the time the exposure is taken and the mount correction is sent to the EQ6 it means a total delay of 2.5s or more in the feedback loop. It is not a question of how quickly the EQ6 responds to changes (I don't understand what on earth they mean by that in any case) but it is a question of how much periodic error and mechanical "jitter" can take place within 2.5s and I can assure you it is quite a lot on the EQ6 - I've generated the PE graphs and you could do the same - it might also tell you if there are any obvious mechanical issues with your EQ6 i.e. is it a good or bad example. Anyway, use an exposure of 1s or, on nights of good seeing, 0.5s.
Guiding of 1arcsec RMS in R.A. should be quite achievable under our skies on the EQ6 on an average night.
There is a commonly voiced argument about the risk of "chasing the seeing" when using short exposure times but this has to be balanced against the risk of falling behind the PE+Jitter curve.
I would leave hysteresis and aggression at their defaults until you have sorted out the major issues - I consider these settings as "fine-tuning".
The other thing is to guide on a bright but unsaturated star (but I guess you already do this). The reason being that if you guide on a dim star then there is too much noise in PHD's calculation of the star centroid and if you guide on a saturated star it again mucks up PHD's centroid algorithm.
I happy to help this Saturday if you feel like bringing your equipment over (assuming clear night sky) - PM me if so.
Mark
Thanks for the feedback Mark I will change the exposures down to 1s to see how I get on. I'd looked at PEC on my old setup and was planning to include it, however I made the move to this new setup beforehand and I didn't want to add another variable into the mix until I'd address the large issues.
Equally,thanks for the offer of help on Saturday, I might well take you up on that let's see how the weather pans out.
Don't play around with PEC i.e. correction of PE until the large issues are sorted. What I meant was that it could be worthwhile graphing your PE to check for any underlying mechanical issue - it could be done as part of a diagnosis effort if you are still having issues.
In fact don't play around with PEC on the EQ6 at all. PEC on the EQ6 can only correct the worm cycle but there are other period cycles generated by the rest of the gear train. If you don't correct for those other cycles when creating a worm cycle PEC (which is very difficult indeed) then PEC will make your guiding worse in almost every case.
I always guide at 0.5 secs with NEQ6 pro & HEQ5Pro
Looking closely at your chart, it would appear that you are already guiding at 1.5" RMS error in R.A. So you haven't got that far to go to reach 1" RMS error. To progress to sub 1" guiding would take a lot of stripping and tuning of the mount and may not make much difference anyway under most sky conditions and it also depends on where most of the errors are arising from i.e. which parts of the gear train.
Unfortunately that graph paints a slightly better picture than reality, the erratic nature of the graph leads to big jumps, here is a sub from this session.
(http://farm3.staticflickr.com/2038/13100983295_f2522b2bb2_b.jpg)
There's still a clear horizontal elongation of the stars.
I'm keeping a close eye on the weekend weather, as I looking likely to be calling your kind offer.
Is horizontal the R.A. direction?
Yes, I should have been clearer.
Are your cables getting stuck?
Could also be stiction or too much backlash (or a combination of both).
Quote from: RobertM
Could also be stiction or too much backlash (or a combination of both).
Stiction yes, backlash hopefully not (at least in R.A.) as long as the setup is weighted slightly heavier on the East to maintain load on the gears.
I used to get this sort of sub on my HEQ5 before I upgraded all the bearings. £75 well spent.
The reason I made the comment is that for most of the sub the guiding is ok, then there movement, like something catching and releasing, or as mark says a possible over balance.
I won't go into the other issue because that will cause a heated debate.
Last thing Ivor you might want to remove the location of your kit of this thread.
Thanks for the ideas I'm going to Mark's tomorrow night so hopefully we'll be able to sort it out then.
Quote from: Rocket Pooch on Mar 14, 2014, 20:39:52
Last thing Ivor you might want to remove the location of your kit of this thread.
Thanks missed that, I have updated it to a more secure location :)
Quote from: Ivor
Thanks for the ideas I'm going to Mark's tomorrow night so hopefully we'll be able to sort it out then.
If anyone else is at a loose end and fancies paying a visit this evening, you're more than welcome. Remember it's a full moon though.
Mark
OK, for the benefit of all, here is what we've found so far - mainly by looking at the PHD log file you generated with the Lodestar on the main imaging scope - 1.65arcsec/pixel.
1) With an guide camera exposure time of 0.5sec, there was anything from 0.75sec to 4sec between frames (average time between frames was 1.5 sec). 4 seconds is a serious lag in the feedback loop! I suspect this is waiting time caused by other applications.
Is your laptop a dual core? Is it possible that raising the priority of PHD and EQMOD might help prevent these waits?
With my old single core and underpowered laptop I often found that opening a downloaded sub-exposure from the main imaging camera would screw up the guiding.
2) Errors of more than 6 pixels (10 arcsec) in R.A. were quite common
3) We still need a PHD log done with guide pulses switched off. This will allow a PE graph to be generated which will check if the mount itself is behaving OK.
Mark
Additional comments & answers
1) my laptop is an i5, so upping their priority shouldn't be an issue. I normally have AA5, PHD2, CduC, EQMOD, Chrome, AVG and some times teamviewer. During recent tests AA5 has been running but not taking subs so the overheads should be low.
3) We discover that PHD2 log files are formatted differently to PHD ones, PECPrep can read them but the output is wrong. Hopefully the weekend will be kind and give me a sufficient slot to test this.
Quote from: Ivor
1) my laptop is an i5, so upping their priority shouldn't be an issue.
I wonder what is causing those "wait" periods then?
Any tabs open in Chrome ?
Google seems to spend most of it's time searching for ads. for "appropriate" goods for you
Download the following process monitor.
http://technet.microsoft.com/en-us/sysinternals/bb896645 (http://technet.microsoft.com/en-us/sysinternals/bb896645)
and run it, it will tell you which process is hogging all the cpu time.
I've also found out the following.
If something needs to be backwards compatable it will normally open a hidden command prompt.
And doesn't shut it down, this command prompt SLOWS windows 7 down to a crawl.
You might have one of these open.
Mac.
where is PHD 2 for downloading? Cant see it. Thanks
You'll find it here:
http://openphdguiding.org/
It's amazing what a cycle home does to clear the mind on a blustery night.
QuoteDownload the following process monitor.
http://technet.microsoft.com/en-us/sysinternals/bb896645
Thanks, I'll give it a go.
QuoteI wonder what is causing those "wait" periods then?
If you recall we found PHD2 was losing the star more frequently than we expected, I wonder if the longer gaps relate to that?
I've copied the configuration across to PHD ready but the gusts a bit too strong tonight so hopefully tomorrow. Just to check I've got it right to create the file for PECPrep I take the following steps:
- Run PHD
- Open brain and select "Disable guide output"
- Under tools check enable logging
- Connect to camera and scope
- select star
- Leave for a minimum of 40 mins
Anything else?
Start PHD
Connect to camera and scope
Enable logging
Calibrate on a star
Disable guide output
Select star and start "guiding"
40 min of data is fine
It is certainly possible that the "gaps" are linked to losing the star - I don't really know what PHD does under those circumstances.
Quote from: MarkS on Mar 21, 2014, 13:50:44
Quote from: Ivor
1) my laptop is an i5, so upping their priority shouldn't be an issue.
I wonder what is causing those "wait" periods then?
You sure they are wait periods caused by the operating system? PHD will not guide if the 'star mass' changes beyond the limit set in the brain so will miss a cycle or two. The idea is so that your mount will continue to track and a lost guide star (clod/crap/airplane) won't cause PHD to start chasing around issuing guide commands. PHD should then be able to require the guide star of the search box is set big enough to cope with your drift over the period.
You can avoid/negate this by unchecking the box or setting 'star mass' to 100. I find the air quality and seeing in Beckenham so variable that this will kick in even at a setting of 95.
Quote from: The Thing on Mar 22, 2014, 11:25:21
You sure they are wait periods caused by the operating system? PHD will not guide if the 'star mass' changes beyond the limit set in the brain so will miss a cycle or two. The idea is so that your mount will continue to track and a lost guide star (clod/crap/airplane) won't cause PHD to start chasing around issuing guide commands. PHD should then be able to require the guide star of the search box is set big enough to cope with your drift over the period.
You can avoid/negate this by unchecking the box or setting 'star mass' to 100. I find the air quality and seeing in Beckenham so variable that this will kick in even at a setting of 95.
It does look like a possible culprit I shall test it out tonight hopefully.
Interest difference PHD never loses the star where PHD2 does.
Here is the PHD log file https://drive.google.com/file/d/0B4aWTDNICIFvUjU2ZDNsWjdDek0/edit?usp=sharing (https://drive.google.com/file/d/0B4aWTDNICIFvUjU2ZDNsWjdDek0/edit?usp=sharing)
Hi Ivor,
Some confusing news here. Towards the end of your PHD log file there were quite a few Error Code 2 which are low signal-to-noise warnings. I wonder what caused that. A cloud perhaps? Or is it a manifestation of your previous problam of losing the guidestar?
I loaded the file into PECPrep but it really didn't come up with a sensible analysis. After correcting for drift, I plotted it in Excel (up to the point where the errors started appearing) and here is the result:
(http://www.markshelley.co.uk/webdisk/IvorPEGraph.jpg)
The total error range is 20-25 pixels peak to peak. I think you were guiding at 1.6arcsec/pixel, which gives a range of 32-40 arcsec which is a bit on the high side. But worse is the fact that it is very difficult to discern any regular cycles in this data and the changes tend to be fairly rapid - these rapid movements would be difficult to guide out.
Compare it to the graph of my EQ6. I was told this was quite a bad graph but at least you can see some obvious cycles and the changes are not so rapid:
(http://www.markshelley.co.uk/webdisk/EQ6BeltMod1.jpg)
So it's beginning to look to me that the mount is a big part (probably the major part) of your problem. The other part of the problem is to determine why PHD/Lodestar loses the star.
Mark
I agree the graph doesn't make sense, what confuses me further is the subsequent guiding I did after this was very good. Other than one odd blip which saw a maximum dy movement of 4.67 the vast majority of movements were under 0.25 during 4 hours of guiding. The time delay between pulses appear pretty good with 30% are under 2s and 80% under 2.5s. Looking through the guide log there are no error codes and the movement is smooth for
The guiding graph wasn't erratic and looked back to normal. The guide log can be accessed here
https://drive.google.com/file/d/0B4aWTDNICIFvQ2tzMmVHVVRWYjQ/edit?usp=sharing (https://drive.google.com/file/d/0B4aWTDNICIFvQ2tzMmVHVVRWYjQ/edit?usp=sharing)
Below is a 360s luminance sub on the FLT110, as you see the stars are looking pretty tight.
(https://farm8.staticflickr.com/7061/13376186654_cd09c7d3da_b.jpg)
NOTE: THIS IS GUIDING WITH THE BRESSER NOT THE FINDER GUIDER.
This suggests there is something odd going on with PHD2 so I'll park that and focus on PHD.
Shall I try and reproduce the PEC file again, it seems odd that the guiding file looks so normal in comparison?
Are you sure it's PHD2 that's the problem and not the finderguider set up as you say you were using your bresser?
Carole
Quote from: Carole on Mar 24, 2014, 18:28:59
Are you sure it's PHD2 that's the problem and not the finderguider set up as you say you were using your bresser?
Carole
When I was at Mark's place we set the scope up with the Bresser and when using PHD2 we were still getting a noisy graph plus it was losing the guide star this didn't happen with PHD.
It was so nice to be able to produce a clean sub I stayed up until 03:00 and now have 6 x LRGB @ 360, if I can get another 10 I'll have enough to produce an image; I've almost forgotten what to do when I do have the data :)
Sounds to me like you could benefit from a strip down and re-grease of your mount. Maybe also replace all the bearings. This is a relatively easy and cheap procedure.
What is the focal length of the Bresser? The RMS error in that log is around 0.7-0.8 pixels so with the Bresser focal length we can convert that into arcsecs.
I've tried reverse engineering a PE chart out of the data (it relies on a few assumptions) but again it looks very jumpy with no obvious periodicities.
The star mass is reasonably constant (unlike previous runs) and there are no error codes such as low signal-to-noise.
Your sub certainly looks better than the last one you posted.
Mark
Quote from: Ivor on Mar 24, 2014, 18:42:32
When I was at Mark's place we set the scope up with the Bresser and when using PHD2 we were still getting a noisy graph plus it was losing the guide star this didn't happen with PHD.
Did you do darks in PHD2? It improves the star images a lot. FYI If you do a dark at one exposure time PHD2 will scale the dark for other exposures (though you can take more darks) and it stores them in a single file and can use them automatically next time. (You don't need to do File, Save Dark).
Thanks for the replies.
Quote from: Mike on Mar 24, 2014, 20:02:40
Sounds to me like you could benefit from a strip down and re-grease of your mount. Maybe also replace all the bearings. This is a relatively easy and cheap procedure.
Yes I think it going to be on the cards, but if you're going that far is it worth just doing the belt conversion? How much would the strip down cost?
Quote from: MarkS on Mar 24, 2014, 20:39:15
What is the focal length of the Bresser? The RMS error in that log is around 0.7-0.8 pixels so with the Bresser focal length we can convert that into arcsecs.
The Bresser is F10 FL700mm Aperture 70mm.
Quote from: The Thing on Mar 24, 2014, 21:59:03
Did you do darks in PHD2? It improves the star images a lot. FYI If you do a dark at one exposure time PHD2 will scale the dark for other exposures (though you can take more darks) and it stores them in a single file and can use them automatically next time. (You don't need to do File, Save Dark).
Yes I took darks, the only configuration difference between the PHD2 run at Mark's and my run at home with PHD is I've taken your advise and increase the star mass tolerance to 0.95.
Quote from: Ivor
Thanks for the replies.
Quote from: Mike on Mar 24, 2014, 20:02:40
Sounds to me like you could benefit from a strip down and re-grease of your mount. Maybe also replace all the bearings. This is a relatively easy and cheap procedure.
Yes I think it going to be on the cards, but if you're going that far is it worth just doing the belt conversion? How much would the strip down cost?
I did my belt conversion mainly to gid rid of an obvious periodic error caused by the transfer gear. In your mount the usual known frequency periodic errors caused by the various gears are being obscured by something more random - maybe sticktion in the bearings or maybe the mount binding somewhere. So a strip down of the R.A. axis looks like the first place to start.
With the Bresser f/l of 700mm and a Lodestar pixel of 8microns your guide camera gives 2.34arcsec/pixel. So your latest guiding is around 1.6-1.9 arcsecs RMS. That's not good but at least you didn't have those wide excursions this time.
Mark
I've looked at Astro-Baby's tutorial on this, I'll try and get everything together so I can do it at the weekend. Assuming no major issues, how long does this take to do?
QuoteSounds to me like you could benefit from a strip down and re-grease of your mount. Maybe also replace all the bearings. This is a relatively easy and cheap procedure.
I'll give you a tip for free....
Getting the bearings out and then back in, can be a royal pain in the arse.
When i replaced all of my bearings, i found when you come to stick the main bearing back its such a tight fit its problematic. you will know the one's i mean. If you pop them
in to the freezer overnight, and leave your mount in your nice warm house, when you come to slide the bearing in, it does just that,
slides in very easy, mind you have to be pretty quick in doing so.
As for time, mine took about six months, another story. but realistically, allow a good day,
also you will need emery cloth & wire wool to remove all of the overpaint,
and if you are going to do the supertune either get the shim kit, or as i did a nice thick sheet of plastic,
something like the overhead projector transparencies
Mac.
Is there anything I particularly need to look for in my Lithium grease? There appears to be quite a broad price range and I want to buy the correct quality.
http://www.ebay.co.uk/bhp/lithium-grease (http://www.ebay.co.uk/bhp/lithium-grease)
I got mine from Evans Cycle Shop:
http://forum.orpington-astronomy.org.uk/index.php?topic=9291.0
Thanks, duly bought. Alas I have to wait until the weekend for the emery paper and pliers it's not something Canary Wharf really caters for.
Do you think this would work for the guide scope removal?
http://www.toolstation.com/shop/Oil+Filter+Wrench/p67000 (http://www.toolstation.com/shop/Oil+Filter+Wrench/p67000)
What grit value emery paper is best 120?
I used a Boa Constrictor. They're readily available on eBay (not the live snake kind!).
Not sure if I used 120 or 180 grit for the removal of oversprayed paint.
Has Robert Dyas gone from Canary Wharf?
It was there last week !
Yes but they don't sell either.
Stripped R.A. axis on the EQ6PRO generally it was fairly clean there appears to be a Vaseline like grease inside. I can't get the slotted nut out from the worm nut so I'm going see it I can find a better tool than the snipe pliers.
The big problem is one of the bearings it is significantly stiffer than the other two, it clearly needs replacing. All the bearing appear to have the same markings so I guessing there are the same. Now to find where to get it from google, google, where for out thou?
OK slightly confused the bearing appears to have 6008 RZ on it but people appear to refer to to rs2? Confused from Tunbridge Wells
Quote from: Ivor
Stripped R.A. axis on the EQ6PRO generally it was fairly clean there appears to be a Vaseline like grease inside.
Mine was similar - a good quality grease not full of all the swarf and black gunk that one often reads about. My stripdown and grease replacement made no difference whatsoever.
Your stiff bearing could well be the problem. I haven't replaced mine.
Mark
Ivor,
Which bearing? I assume you have taken it out before deciding its stiff?
Chris
It's the dec one in the mount base, see in the final image of step 1 http://www.astro-baby.com/EQ6%20rebuild%20guide/EQ6%20RA%20strip%20down.htm (http://www.astro-baby.com/EQ6%20rebuild%20guide/EQ6%20RA%20strip%20down.htm)
By stiff I mean it is running noteable less smooth than the others, I have to apply 100% more force to move it compared to the others. This does highlight another question I have which is how free running should the bearings be, the better ones move freely when I apply force but they don't run freely in a skate wheel kind of way is that the level movement I should expect?
I decided to just repack the main bearings with castrol lithium grease and found that made things much better. The freedom is such that if I release the clutches and lift the c/w shaft to horizontal I get about three swings past vertical before it stops. The other main issues and those that made the best benefit were a) removing the excess paint from the moving surfaces (snagged as the axes were rotated) and b) replacing the worm bearings (with SKF).
Robert
Wow mine's never been like that, so the replacement is a no brainer. Now do I bite the bullet and replace all of them?
Is this the correct part? http://simplybearings.co.uk/shop/p6362/SKF+60082RS1+Rubber+Sealed+Deep+Groove+Ball+Bearing+40x68x15mm/product_info.html (http://simplybearings.co.uk/shop/p6362/SKF+60082RS1+Rubber+Sealed+Deep+Groove+Ball+Bearing+40x68x15mm/product_info.html)
I decided to do what I saw as the essentials and the main bearings weren't on my list - but it's your choice...
My SKF's were from Bolton Bearings on eBay as I found them the best cost/rating - £12.84 inc delivery for all four which I didn't think was too bad. The larger bearings are likely considerably more so that may temper your enthusiasm somewhat ;)
Robert
I received the SKF bearings last night, clearly a quality product compared to the ones supplied with the mount and I've fitted them into the chunky brass bits (I'm sure there's a more technical name for them).
So far so good, one of the other problems areas is the RA Worm it makes a lot of noise when rotating http://vimeo.com/90792441 (http://vimeo.com/90792441), I've really struggled to remove it and the bearing to the left appears to stuck; a more forceful approach will be taken tonight to extract it and work out what is causing the noise.
I'm putting the mount back together and I'm unclear when to tell I've adjusted the RA and Dec correctly. The term "play" is used so in the case of the DEC axis does that mean when I hold the saddle I can move it slightly back and forth?
Quote from: Ivor
I'm putting the mount back together and I'm unclear when to tell I've adjusted the RA and Dec correctly. The term "play" is used so in the case of the DEC axis does that mean when I hold the saddle I can move it slightly back and forth?
Yes, you need the gear to carefully mesh with the worm. Not too loose so there's backlash (rocking the saddle back and forth) and not too tight so that it binds. It is often a compromise between the two.
Similarly the worm bearing needs to be tightened so the worm cannot move end to end but is still able to move freely.
Mark
I'm struggling to get the RA axis back on, before I take the DEC section apart again can someone confirm whether the two copper bits for the RA and DEC are the same size?
Thanks
Which copper bits?
Here's the rebuild guide:
http://www.astro-baby.com/EQ6%20rebuild%20guide/EQ6%20RA%20rebuild.htm
The brass worm gear, section 5 picture 1
I've no idea. My guess is that they're identical, in order to keep down manufacturing costs.
I found it very tight to reassemble the RA axis - it is essential that the shaft goes in absolutely straight otherwise there is a risk the brass gear can jam against the walls of its housing.
So I put the taper bearing in first and then ensured the spindle went through the centre of the taper bearing as the mount base was installed onto the RA axis. This ensures the spindle plus the brass gear go in perfectly straight.
Mark
Thanks, I thought the same. I took the same approach as you I just struggle to get it to go the last 1 cm and be flush with the worm housing.
I shall put the brass section in the freezer ( in a air tight container) for a bit and try again.
I'm still struggling to get the base back on the EQ6 I must be doing something silly but I just can't see what it is. I've checked the RA & DEC worm gear are identical so I'm at a lost as to why there is a depth discrepancy between the hole in the mount and the height in the worm gear. I can't see how you can put this together wrong so I'm perplexed as to why this has occurred.
(https://farm8.staticflickr.com/7167/13745864323_9f0f902140_c.jpg)
RA worm gear depth
(https://farm8.staticflickr.com/7276/13745818223_553c8b6de1_c.jpg)
mount base depth
(https://farm8.staticflickr.com/7195/13745804045_2a7c432fea_c.jpg)
mount depth flush
Any thoughts?
That looks a complete mess, how did the brass get that scored?
The scoring is probably from the clutch mechanism so it shouldn't be a worry.
Have you got an RA Scale ring lying around somewhere? My guess is that you might be omitting it! I would think that takes up most of the depth difference you are seeing.
Mark
Quote from: Rocket Pooch on Apr 10, 2014, 07:28:16
That looks a complete mess, how did the brass get that scored?
That just grease it cleans up nicely.
QuoteHave you got an RA Scale ring lying around somewhere? My guess is that you might be omitting it! I would think that takes up most of the depth difference you are seeing.
Thanks, knew I forgetting something simple, I should stop trying to do this late at night but I want to look at Mars while there is a gap in the cloud.
I've got the mount back together but not tighten everything up yet, it all moves smoothly and feels controlled not at all sloppy. If I put a 5kg counterweight on and from horizontal let go it will swing past the vertical once then return to the vertical and rest. Is this what I should expect?
OK I've got there in the end, Graham at AstroTec was a great help lending me a homemade part to sort out a problem with the worm nut.
There is minimal backlash in Dec which I can't seem to tune out and in RA I've got the sloop down to <1mm but I get a strange clicking sound you can hear it in the video below at 10-11 secs and 24-30 secs.
http://vimeo.com/92186320 (http://vimeo.com/92186320)
Any thoughts what it could be?
Is this the same problem as Fay's? Also I think the link I posted up for Fay was this same guy for mending mounts.
Sorry I can't help Ivor, no good at this sort of thing.
Carole
Well at long last I got to test my rebuilt mount, it is lovely and quiet in comparison to before except for one of the worm bearing which I commented on above. The mount appears to move smoothly however the RA axis is stiffer than before when the clutch is off which would make balancing the scope difficult if I didn't already know the right spot.
I guided whilst imaging M13, the log file can be accessed here https://drive.google.com/file/d/0B4aWTDNICIFvV1VLUHJMUTZadEU/edit?usp=sharing (https://drive.google.com/file/d/0B4aWTDNICIFvV1VLUHJMUTZadEU/edit?usp=sharing)
I'd set the max duration 1000ms and during the two hours of imaging it hit that 109 times the larger subsequent adjustments push the average guiding up and I'm getting error 2 and 4 which appears to collates to a lower star mass but I can't find any reference to what they actually mean. There were a few clouds so I'm not sure it that is the problem.
Max pixel peak to peak error is 15.93, which is better than before but I'm not sure how to judge the overall guiding due to the errors and 1000ms issues. The image still have some elongation in the stars but it is better than before.
(https://farm6.staticflickr.com/5498/13953839747_e73808c898_c.jpg)
Another couple of nights is needed to tweak the parameters however I like to get to the bottom of the errors codes as well, I'm worried there is a problem with the camera.
Unable to access the log files (from work). Have you done one with the guiding switched off so the PE of the mount can be graphed? I remember the previous one was very erratic (see graph earlier in this thread).
Error code 2 is where the S/N ratio of the guide star has dropped too far to be used.
Mark
No not yet, that's part of the next session, it was just nice to image again I'd almost forgotten what I brought the telescope for.
Any thoughts about the error 4?
Quote from: Ivor
Any thoughts about the error 4?
I've got a note of it at home, from looking at the source code.
Mark
Here we are - this snippet defines the error codes:
enum FindResult
{
STAR_OK=0,
STAR_SATURATED,
STAR_LOWSNR,
STAR_LOWMASS,
STAR_TOO_NEAR_EDGE,
STAR_MASSCHANGE,
STAR_ERROR,
};
So 4 equates to Star Too Near Edge
Mark
Thanks surprising not well documented I would have thought the error code would have included which edge of the box it left to help with adjustments.
Tonight looks hopeful so I'll try and produce the PEC file, mind you it's my 18th wedding anniversary today and I don't think I can get away with the line "have a romantic night looking at the stars" as my wife is wise to this now :)
Why or why are the only clear nights in a row at the full moon so unfair!!!!
I produced the PHD file for PEC analysis last night it can be accessed from the below link
https://drive.google.com/file/d/0B4aWTDNICIFvaE1UZjFoWDZJWjQ/edit?usp=sharing (https://drive.google.com/file/d/0B4aWTDNICIFvaE1UZjFoWDZJWjQ/edit?usp=sharing)
I will have to watch the PEC tutorials again so I can understand what's going on and maybe a little help from my friends :)
Once this as done I went back to M13 to take a few more subs and test out the guiding adjusting the max pulse up to 2000ms as it max'ed out last time. This time the guiding only max'ed out 3 times so I will increase it to 3000 for the next run. No guiding errors this time so I'm the last time was just clouds. The overall stats came out better than before which is pleasing, I feel I'm getting there.
Peak2Peak 13.56
RMS 0.66
arcsec/pixel 2.42
Average guiding (RMS*arcsec/pixel) 1.59 * I'm guessing this is the right calculation
Max swing 6.95
However the stars are still slightly elongated, whilst pondering this on the way into work I wondering whether this could be blooming, there is a setting AA5 for it but I don't it enabled. I shall try the next batch of shots with it set to medium.
(https://farm3.staticflickr.com/2935/14003340309_24d9da2903_z.jpg)
Check the subs for M13 - I think you'll find most have that elongation - it's guiding not blooming.
Here's your PE graph from the unguided run:
(http://www.markshelley.co.uk/webdisk/IvorPEGraph2.jpg)
This is slightly better than before but it's a terrible PE chart for an EQ6 - it's difficult to see the usual periods in that chart.
This is confirmed by the PECPrep Fourier analysis:
(http://www.markshelley.co.uk/webdisk/IvorPECPrep2.jpg)
The peaks on that chart correspond to amplitude of the various periods it finds in the data.
The usual culprits for an EQ6 are visible:
478.7sec Worm Drive
366.6sec Transfer Gear
182.2sec Transfer Gear 2nd harmonic
239.4 Worm 2nd harmonic
However the usual culprits are swamped by some other very high amplitude periods:
The biggest has a period of around 132sec
Another peak at 66sec
Another at 33sec
Notice the peiod halving - 2nd and 4th harmonics of 132sec
These periods do not correspond to anything I know of in an EQ6 but they are giving your PE chart its extraordinary profile which is more or less impossible to successfully guide away. They weren't present in your previous PECPrep but your previous one was complete garbage anyway - very difficult to get anything out of it.
Just possibly this is some random effect but if you perform another unguided run and those frequencies are still present then it is definitely telling us something.
Out of interest, the replacement bearings you put in - did they have the same number of balls as the original EQ6 bearings? I wonder if that is a possible explanation.
Mark
Thanks for this Mark, I went through the PECPrep video tutorials last night and I was perplexed by the results as well so I'm glad I'm not going mad but concerned by the confusing results.
I used the SKF 6008 Series 2RS bearings as they were sealed I didn't check the number of bearings however these were recommended to me by Graham at AstroTEC so you'd think they would be right.
In an effort to isolate the cause, below is the list of issues/concerns I feel are outstanding with the mount.
- I've not replaced the worm bearings and at least one is a bit ropey, I was previously advised this wouldn't be a problem as it was away under load.
With the clutches off the mount doesn't move as freely as before, with a little bit of force I can move the telescope but if misbalanced it wouldn't move on it's own.
I think I'll order the worm bearings today it's not going to do any harm replacing them and it will rule out of of the possibilities.
If we can find what mechanical component gives the 132second period then the problem will be diagnosed.
Nothing in the gear train corresponds to this; the main bearings turn too slowly; so that leaves the worm bearings. But 132sec does not correspond to the usual worm bearing ball period. Is it possible that your worm bearing is not standard and has smaller balls? Who knows?
If you look at the PE graph knowing to expect a 132sec period then it is actually very obvious in the graph.
Mark
QuoteI think I'll order the worm bearings today it's not going to do any harm replacing them and it will rule out of of the possibilities.
Definitely do the worm bearings, in my opinion they make a massive difference.
I also can't understand why the axes appear so tight - are you sure something isn't rubbing and that there is a tiny gap between the two faces that separate the axis with body (if that makes sense).
Robert
The bearing are ordered and shipped so if I'm lucky they'll arrive tomorrow so I can fit and test the same day.
Quote from: RobertM on May 16, 2014, 16:08:47
QuoteI also can't understand why the axes appear so tight - are you sure something isn't rubbing and that there is a tiny gap between the two faces that separate the axis with body (if that makes sense).
If you mean between the RA Scale ring and the main body then yes. It feels like the worm gear isn't slipping against the outer casing, neither gear has any grease on it as I thought they weren't meant to is this a correct understanding?
Quote from: Ivor on May 16, 2014, 09:27:18
I used the SKF 6008 Series 2RS bearings as they were sealed I didn't check the number of bearings however these were recommended to me by Graham at AstroTEC so you'd think they would be right.
I did my worms first ('you can get a pill for that' said the missus) and it made a big difference, convinced me to do the rest. Did you change the thrust bearings as well? In my HEQ5 they were the worst bearings and were bumpy and sticky which caused the guiding to jump every so often.
And I put lithium grease on all the contacting surfaces and squeezed it into the thrust bearings.
I'm pretty sure the 132sec period and it's higher harmonics are caused by the balls in the worm bearing.
This thread (http://stargazerslounge.com/topic/100401-eq6-belt-mod/page-4) reckons the ball period for the 608-2RSH (worm) bearing to be 139sec and NOT the 174sec period indicated by PECPrep. I used the bearing frequency calculator on SKF's website to confirm the 139sec figure given a period of 478.8sec for the worm itself. I'm not sure why yours differs slightly (132sec vs 139sec) but it is close enough to be significant and all the higher harmonics will be due to the fact there are multiple balls.
I've also decided to replace the worm bearings on one of my mounts.
Mark
Thanks for investigating further, it makes me feel happier knowing there is now a quantifiable reason for my problems. I've only replaced the RA worming bearings at the moment so I can isolate the results; Thursday looks promising so hopefully I can put it through its paces then.
I have to admit I'm struggling visualise how the harmonics work on the worm bearings as it has multiple parts playing roles. Each ball bearing within the worm bearing has a rotation which combines to give the rotation of the inner ring of the worm bearing itself, how does this relate to 139secs?
I really need to find the time to read up on harmonics theory it's too many years since I did it at Uni.
Think of it like this: there is an inner and outer part to the bearing with balls filling the space between them. It could be designed to have tiny balls or big balls. For a given rotation speed of the bearing as a whole, tiny balls would have to rotate very fast but large balls would rotate a lot slower. Any irregularity in the balls or debris inside would lead to irregularities in the rotation of the bearing as a whole. The reason for the harmonics is a lot more difficult to explain.
I've just had a look at the log of my belt modded EQ6 using PECPrep and I can see distinct periods of 133sec and 68sec which I reckon match the periods you are seeing (given the accuracy of the frequency analysis). But mine have a fairly low amplitude - it is the worm and its harmonics that dominate that dominates. However, since they are definitely visible I think it may well be worth me replacing those worm bearings.
Patch cloud here so still undecided about whether to do a test run.
I got my worm bearings from here http://www.qualitybearingsonline.com/608-2rs-skf-8-x-22-x-7mm/ (http://www.qualitybearingsonline.com/608-2rs-skf-8-x-22-x-7mm/)
OK I managed to produce a PHD log file last night (https://drive.google.com/file/d/0B4aWTDNICIFvS1R1SW1zM19HU0U/edit?usp=sharing (https://drive.google.com/file/d/0B4aWTDNICIFvS1R1SW1zM19HU0U/edit?usp=sharing)) from my quick look this morning before I head off the graph appears smoother in PECPrep with clear peaks and the 133 harmonnic has disappeared. However I don't think my setup was right as both RA and DEC errors are larger. It looks like the markers I have in the ground have moved as polar alignment was off and the log file shows RA duration maxing out alot, I also don't think I had the balance right I still have this stiffness with the clutch so balancing is an educated guess (clearly wrong this time).
(https://farm4.staticflickr.com/3921/14170625807_8d7acda596_z.jpg)
I will replace the DEC bearings tomorrow as Saturday night looks hopeful; I feel like I'm a little closer to solving this.
To me, the plot still looks all over the place.
The chart you posted shows it dominated by the 182sec 2nd harmonic of the transfer gear with 133sec, 66sec. A 44sec period (3rd harmonic of 133) has also appeared.
The worm bearing is still showing us its ugly characteristic.
Mark