Auto Focus (Lakeside, JMI etc...)?

[kirkster501]

Hi all, a bit of a sanity check on auto focusing please?

Is this automated in the sense that the system itself can find the focus point automatically by measuring what it is seeing through the camera (via the PC of course) and then applying an adjustment to bring the image into focus?  Or does it work by you having to manually set the focus points beforehand for different scope/filter combinations and have them pre-loaded into your PC?  In other words, is this an intelligent system that works by real-time feedback or is it more of a "move here" to a predetermined position sort of arrangement?

Thanks, Steve

[Helen]

I use the Lakeside focuser with Focusmax and Maxim.  You train the software to know the variables in your system by running what are known as v-curves.  The software adjusts focus, takes a shot, measures result in FWHM and plots - the FWHM decreases until optimal focus and then starts increasing again (making a V ).  For best results the v-curves are taken multiple times to average out any other factors such as seeing.  That's really a do-once-forget job for each kit combination.  After that you just find an appropriate star, click focus and the software takes shots and determines the optimal focus.

Helen

[Davey-T]

Demo here

Dave 

https://www.youtube.com/watch?v=Ft2e203Rn1M#t=206.454744

 

[Gina]

The former - at least as far as KStars/Ekos/INDI is concerned.  I have never used an EQMOD/ASCOM auto-focuser.  It automatically finds a star near your target DSO and measures it, then adjusts focus to minimise FWHM plotting a V-curve and then setting the focus to the best.

[gnomus]

35 minutes ago, kirkster501 said:

Hi all, a bit of a sanity check on auto focusing please?

Is this automated in the sense that the system itself can find the focus point automatically by measuring what it is seeing through the camera (via the PC of course) and then applying an adjustment to bring the image into focus?  Or does it work by you having to manually set the focus points beforehand for different scope/filter combinations and have them pre-loaded into your PC?  In other words, is this an intelligent system that works by real-time feedback or is it more of a "move here" to a predetermined position sort of arrangement?

Thanks, Steve

Your second system of 'preloading' wouldn't work because it would be tedious and probably impossible to work out the settings for every temperature/filter combination.  I use a Lakeside with SGP.  How it works is as follows: SGP moves the focuser to a point that is well out of focus.  It takes a brief bin2 image.  It then meaures the 'half flux radius' (HFR) of a set of stars in the image.  It then moves the focuser inwards and repeats.  I set mine to use 9 steps.  As it goes through this routine, hopefully the stars get smaller and then get bigger again.  It plots this on a graph and draws two lines .  Where these lines intersect is the optimum focus point and it moves to this point, takes a final brief exposure and reports the optimum HFR.   

That is a simplified explanation of course.  My experience is that this achieves better focus than I can with my eyes and a Bahtinov.  I have set the focus to run every time there is a 1 degree drop in temperature.

One wrinkle is NB imaging.  For my LRGB I use 5 second exposures - 9 (actually 10) of them doesn't take too long.  But I need 25 second exposures for my NB filters.  This can eat up valuable time.  So, one evening I worked out offsets for all of my filters.  Now I focus using Lum (nice and quick) then SGP applies the offset.  Another advantage of this is that I can cycle through filters L-L-R-G-B-L-L without having to refocus on each filter change (unless the temperature drops, of course).

In re-reading the above, I noticed that I kept saying, 'I do ....'.  In fact, I do nothing.  SGP does it all for me and without me having to think about it.

I hope this is clear enough.  Once you get autofocus up and running, it is difficult to go back.

[PhotoGav]

Hi Steve,

I use a Lakeside focuser system with Sequence Generator Pro and it is a joy to watch it work its magic! Most importantly it can focus noticably sharper than I ever managed with a Bahtinov mask. The way it works is pretty much as described above, but you need to start roughly in focus and by roughly I mean very roughly, so that stars are points rather than bloated airey disks. It then does the whole V-Curve thing, starting by moving to an out of focus point that is the current focus point plus a pre-determined step amount (this is pretty easy to set up) and takes a picture. It measures the average HFR across the stars in the image. It then brings focus in by a step, takes another picture and measures. It plots the points on a pretty graph on screen that you can watch as it works. It continues like this for the nine points (user adjustable, but nine seems to give great results) so going from out of focus, through best focus and on to out of focus the other side. It then does the V bit of it all to establish the perfect focus point. It moves the focuser to that value and takes an image to confirm that this is as expected and good. You are then in sharpest focus and can get on with gathering subs.

The great thing with SGP is that you don't move away from your target, it uses the stars in your field of view to run auto-focus. It is very efficient.

Good luck and I hope that helps.

[PhotoGav]

I promise I didn't copy @gnomus's work...!

[gnomus]

Just to add to what Gav said, yes you need to start roughly in focus.  The stepper retains its position from the previous night, so most of the time you do start from rough focus.  Occasionally, there will be such a temperature shift from one imaging night to the next that the focus routine fails the first time.  It needs to have at least 3 out of focus positions before 'turning the corner' on the HFR curve.  If it doesn't do this then SGP will re-run its routine having moved even more out of focus.  This takes a bit longer than normal and, whilst it works, sometimes I intervene, by stopping autofocus, doing a GOTO to the  focus position where the graph did turn the corner, and then running autofocus from there.

In SGP you need to determine what your step size will be - stepper motors, focusers, telescopes are all different.  However, the manual tells you how to do this and it is very easy.  And you only need to do it once.

[Gina]

Seems to be pretty much the same as KStars   I haven't had a chance to test this with a scope but it worked with my all sky cam though it chose to use a distant light rather than a star as that was brighter.

[kirkster501]

Brilliant stuff thanks so much for the explanations folks - I understand it now!

Ideally then, you'd also need a motorised focuser for *each* of your imaging scopes as well as I'd imagine it'd be a PITA to remove and keep re -attaching to different scopes all the time.... ?

I may buy a Lakeside for my TEC as that is not quite parfocal with my filter set and needs a tiny touch on blue. 

I'd like to start getting more complete data in the can on the rare nights the UK allows for AP of late by doing two subs on each filter, move filter, refocus, another two subs next filter etc and cycle through my filters like that multiple iterations.  SGP facilitates this easily and allows to get something in the can in limited imaging time.  I've got about three hours red and rreen and six hours Luminance on M33 last autumn and then it clouded over for weeks and I need the blue!

Great tips about using NB Steve!

 

[gnomus]

Yes, the focusers are bolted on and it would be impractical to be moving it from one scope to the other.  So one motorfocus per scope it must be.  

I also have a Feathertouch unit on my Esprit 120.  If your TEC has a Feathertouch focuser you might want to look into that.  The FT unit has a clutch that can be disengaged if you wish to focus manually.  It is a lovely piece of kit, but, being FT, it ain't cheap.  There's also a unit made by Sharpsky that you might want to look at.

[sloz1664]

Or Steve if you fancy a challenge 

https://sourceforge.net/projects/arduinofocuscontrollerpro/

Steve

[kirkster501]

11 hours ago, sloz1664 said:

Or Steve if you fancy a challenge 

https://sourceforge.net/projects/arduinofocuscontrollerpro/

Steve

I think I'll give that a miss thanks Steve in favour of a turn-key solution!  I have enough trying to get dodgy code to work in my day job!

 

[swag72]

Another Lakeside user here.... and I use SGP for focus. It's already been described how it works with V curves so I won't dwell on that! I set mine to a refocus at every 1 degree temp change.

Also one thing worth noting is that the Lakeside and SGP play nicely working in dual as well...... With two instances of SGP open, then the Lakesides always opens in the same com port so you always know that the right Lakeside is connected to the right scope...... That's been very useful.

[sloz1664]

I have built three Myfocucers for my scopes all work perfectly well in conjunction with SGPro. If anything did go wrong with them you can fix them.

 

 

 

 

 

[Gina]

I make my own remote focusers as detailed in various threads on here.  The latest works with RPi with INDI server and KStars/Ekos on client computer indoors.  I'm just finishing off a tutorial in the Software forum.

[gnomus]

Ah - if only we mortals had the skills.  The last time I tried soldering something, my solder joint ended up looking like a bit of sweetcorn.......... (After someone had chewed on it.)  

[Gina]

I hardly think I'm immortal   But maybe I should keep my big mouth shut

[Davey-T]

46 minutes ago, gnomus said:

Ah - if only we mortals had the skills.  The last time I tried soldering something, my solder joint ended up looking like a bit of sweetcorn.......... (After someone had chewed on it.)  

Hope you're not a surgeon Steve 

Dave

[gnomus]

30 minutes ago, Davey-T said:

Hope you're not a surgeon Steve 

Dave

Who told you that surgeons solder people back together, Dave?  Was it the BBC?

[Davey-T]

2 hours ago, gnomus said:

Who told you that surgeons solder people back together, Dave?  Was it the BBC?

No but you need steady hands to solder and drill holes in folks heads 

Dave

[kirkster501]

I have bought a Lakeside for my TEC.  The central focuser spindle is circular and 6.25mm in diameter on the FTF3545 that my TEC (build date March 2016) has.  I have been talking to Ian King about it who has been very helpful.

I will take some pictures during fitting it.

Actually, at F7, the TEC140 is a doddle to focus even manually - it is the most perfect scope I have owned.  The FSQ85 with reducer is much very much difficult to focus. Guess what I am going to be buying next?

 

[sloz1664]

1 hour ago, kirkster501 said:

I have bought a Lakeside for my TEC.  The central focuser spindle is circular and 6.25mm in diameter on the FTF3545 that my TEC (build date March 2016) has.  I have been talking to Ian King about it who has been very helpful.

I will take some pictures during fitting it.

Actually, at F7, the TEC140 is a doddle to focus even manually - it is the most perfect scope I have owned.  The FSQ85 with reducer is much very much difficult to focus. Guess what I am going to be buying next?

 

What another TEC140 

[gnomus]

17 minutes ago, sloz1664 said:

What another TEC140 

Could be.  My guess was a TEC 180.

[Dave In Vermont]

Just a note on the JMI MotoFocus: While it isn't 'practical' to move the focuser-unit from scope-to-scope, the hand-boxes are generally identical. Easy to un-plug and plug into another.

I've had these on various scopes for many years without a problem.

Dave