Onstepping Eq5 and skywatcher 200p


reddish7575@...
 

Hi, I'm Chris from the UK.

I'm looking at onstepping my Eq5 mount and skywatcher 200p and was wondering if there was a recommended option for this? 


George Cushing
 

0.9° 1.6A NEMA 17 steppers, 48T:16T pulleys, 138-142mm belts.

Any of the ESP32 based controllers. Go to the Wiki's Showcase and scroll down to the EQ% section.


 

I use 0.9° 1.68A NEMA 17 steppers on my CG-5 (which is same as EQ5) but 60T:15T GT2 pulleys because there is just enough space for 60T pulley so I decided to go with that. With RAMPS build and dynammic microstepping mod (1/32 tracking and 1/8 GoTo) I have excellent performances (up to 2°/sec goto speed and decent tracking resolution). With MaxESP3 i have even faster goto - up to 4°/sec but same tracking resolution (I have no dynamic microstepping there, 1/32 microstepping for both goto and tracking).


George Cushing
 

The generic EQ-5 is a good mount and IMO the minimum size for serious work, but I don't feel that 5:1 GR1 is going to get much more realized tracking resolution out of the mount than 3:1. At least to justify the loss of slew speed.





Gear Reduction
Est. tracking

STEPS/arc-deg Stepper-Steps MICROSTEPS GR1-gears GR2-Mount TR
RA 25600 400 32 5 144 0.29
RA 15360 400 32 3 144 0.48
I guess I should have said 1.5A+/-. Folks have said they have had no problems with TMC2130s with motors rated up to 2A.


 

60:15 is 4:1
60:16 is 3.75:1
48:16 is 3:1


George Cushing
 

Ouch, doing math in my head again.


ETXcetera
 

Is it important to have an integer ratio when it comes to PEC? My first conversion I did with 60:16 (3.75) and .9A steppers as I thought thought the larger ratio would help tracking accuracy (60 tooth barely fits). I have another mount that I have 48:16 (3:1) with 2A steppers. My plan was to use 64microsteps with the larger motors and integer ratio but keep similar tracking accuracy.  Definitely needed the 20V power supply for the 64 microsteps.


Nick Bramhall
 

Hi Chris. I'm also in the UK and have upgraded my old EQ5 mount with OnStep. I mostly use it with a small refractor for imaging but have also had my 200P on it for visual and OnStep had no problem driving the bigger reflector. I have a thread here for my build and also a page on my website. If you have any questions feel free to reply here or on the build thread.


 

You only must have integer number of steps per worm rotation, or rounded up to nearest whole number (AXIS1_STEPS_PER_WORMROT is integer), but that has nothing to do with gear ratio of 3.75 because these two decimal places will disappear by multiplying by 200 or 400 steps per motor revolution anyway:

400 (motor steps/revolution) * 32 (microsteps per each full step) * 3.75 (GR1) = 48000 steps per worm rotation

I would get rid of PEC for sure. It's very delicate, requires special attention and it's potential source of many problems if used incorrectly. And can't cope with autoguiding. I rather make shorter subs but many of them if i don't use autoguiding. PEC will not help you a lot anyway and when it does, it only works with smaller focal lengths like camera lens up to 300mm.

With example above (400steps, 32 microsteps, 3.75 GR1) you will have 53 microsteps per each arc-second on EQ5, that's plenty for AP. Going with 1/64 over 1/32 microsteping may not be always better idea. Depends on tons of factors (motors build quality, current and voltage settings, stepsticks advanced features, MCU oscillator accuracy) and use scenario you may even have worse performance. In edge cases, there will be more skipped steps with 1/64. And most of them happen when motor is accelerating or decelerating (changing speed, starting, stopping - autoguiding). So, before you go 1/64 you first get rid of edge cases - increase GR1, use interpolated drivers, motors with more steps per revolution, etc... But then You already increased angular resolution by other means and not need 1/64 anyway. My unguided tracking worked just as same with 1/64 as with 1/32, but with autoguiding I had better total RMS when using 1/32, until I brought all the other things to perfection, now it's all the same for me.


ETXcetera
 

Thank you Vladimir for detailed reply. I have also since moved to Autoguiding for most cases in which case I do not use the PEC feature. When traveling though, I do enjoy avoiding the setup hassle and extra cables and computers and just using a DSLR to capture images. On my C8 at 1280mm, I can keep more subs when using the PEC feature. Below is an overlay of star trails for a 600 second (1 worm rotation) exposure. The white line is the mounts natural PE (white) and the black line is with PEC enabled. The goal is to get that line straight as possible. I think maybe if I don't have an integer ratio for my gears I might not get repeatable PEC.


 

It's hard to tell how much of PE there js just by looking at that line (specially without knowing whole methodology) because there are other error present for sure. Some maybe made by imperfect mechanical parts (other than worm), some by PA offset. I even had bad tracking clock on one of my builds which was off by only 0.14Hz, the worst thing was that made onstep PEC fully useless for me until I had that tracking rate adjusted.

And very PEC is where I liked EQMOD better, back in the days while I didn't had an autoguider. Using PecPrepare I was able to analyze and apply series of filters, mainly low pass to the PEC curve in order to indentify and isolate only PE out of everything and after applying it was much better. OnStep only applies wightened average rate 3:1 on recorded guiding correction data set all the time. So my impression is that PEC onOnStep, can help, but not as much as people expect and then I question if it's worth the effort.


Howard Dutton
 
Edited

On Wed, Nov 30, 2022 at 01:56 PM, Vladimir wrote:
And very PEC is where I liked EQMOD better, back in the days while I didn't had an autoguider. Using PecPrepare I was able to analyze and apply series of filters
... Sky Planetarium can export to and import from PECPrep, I've used it many many times.

Best to record under optimal seeing/high in the sky and lots of data using PHD2 (not guiding.)  Then generate the curve and upload to OnStep using Sky Planetarium.

If serious about PEC implement the sensor so you don't have to worry about exact integral worm periods or loosing the index.

If not as serious make sure you record several times using the traditional method of auto-guiding as the data are averaged and noise is reduced.


hitosi sato
 

PHDlogviewer allows frequency analysis of PHD2 guiding logs; guiding corrected data and raw data can be compared.
The attached image shows two different ways of doing this: raw and corrected by guiding.
You can see where the peaks are in the periods to give you an impression of the mechanical review.
The image shows a large peak related to the 144-tooth worm wheel.



Hitoshi


ETXcetera
 

On Wed, Nov 30, 2022 at 01:56 PM, Vladimir wrote:
Some maybe made by imperfect mechanical parts (
How dare you disgrace the mechanicals of my 30yo Vixen and precision aliexpress pulleys!

Actually for this test I was just trying to see if the PEC actually did anything because I had already invested the time to figure out the sensor and all that. This was also before realizing that off axis guiding was really the only way for me to take 60s or longer exposures and consistently get to keep them. I do think that PEC increased the number of subs I was able to keep at the time. Now with autoguiding I get to keep most all of them except for starlink photobombing.


ETXcetera
 

On Wed, Nov 30, 2022 at 02:07 PM, Howard Dutton wrote:
If serious about PEC implement the sensor so you don't have to worry about exact integral worm periods or loosing the index.
I have the sensor, but when it goes around once, the little pulley goes around 3.75 times and any "oblongitudeness" (out of round) my pulleys are or slight bends in my motor shafts might be different the next time the big pulley goes around. I'm sure it makes no real world difference though and probably not worth the experiment (but that probably wont stop me). 


Howard Dutton
 

On Wed, Nov 30, 2022 at 07:06 PM, ETXcetera wrote:
I'm sure it makes no real world difference though
Sure sounds like it would.


 

On Thu, Dec 1, 2022 at 04:06 AM, ETXcetera wrote:
the little pulley goes around 3.75 times and any "oblongitudeness" (out of round) my pulleys are or slight bends in my motor shafts might be different the next time the big pulley goes around
With integer GR1 ratio the error you mention become entangled with worm PE. That's why at the time I switched 60:16 with 60:15. But then I gave up because of the belt; If you think that deeply, then consider the belt is not homogenous, somewhere stretchy somewhere little wider, etc. That could also affect PEC. Then you could find a belt with total count of teeth divisible by big pulley teeth count. With 60T big pulley this is not possible, since 120T belt is 240mm long and that's way too long to fit on EQ5. Even with 48T:16T (integer GR1: 3) you could only use 192mm belt with 96 teeth but that is also long. You could introduce tensioner or idler pulleys to fix this problem, depending how persistent you are in your journey. I was not. When obtained autoguider I just got rid of anything PEC.


 

I just realized that even though all errors are periodical, small pulley eccentricity error period is shorter than worm cycle so it will be contained into PEC. Think of it like PE withing the PE, while belt error will have longer period and so it will be stretched across multiple worm cycles. Like PE outside of PE. Not sure how good PEC can counteract this.


ETXcetera
 

Next built 48:16 gears and 192 tooth belt with idler. Perfect PEC, Zero PE, Retire mount once perfected.