Robofocus settings - Attention Jase

[montewilson]

Hi Jase and anyone else using an FSQ with RoboFocus.

I am keen to know if others are finding the motor too coarse to be easy to use with the rack and pinion system on the Tak.

I recently found the default in the software is four pulses of the motor for each pulse of the software. I have changed it to 1:1 and I am hoping this will help. The motor and gearbox combo is 7.5Deg per pulse with a 75:1 reduction before the output. Hurst make a 300:1 but maybe my problem was software configuration related and I don't need to get a lower geared motor.

Any thoughts?

[Zuts]

Hi,

I just installed a robo focus on my TV85. Of course that was last tuesday and it's been raining since then so I cant try it out.

An imager at the SPSS had a 106 FSQ and claimed it worked very very well with that so on the strength of that I got one. I would be very interested in knowing if the settings change has solved your problem. I am planning on using FocusMax integrated with CCDSOFT.

Cheers Paul

[jase]

I doubt you’ll need a lower geared motor Monte. You should have no problems getting the Robofocus to repeatedly hit within the FSQ Critical Focus Zone (CFZ) of 55 micron. Don’t tell me your still running @focus or @focus2. FocusMax all the way my friend, start being productive under the dark skies instead of troubleshooting your woes.:P

Just so we are on the same page. One pulse = one microstep. What you can configure is how many microsteps make a step. A step is what is counted by the Robofocus controller, so in another words what you are configuring is the number of microsteps per count.

The default step size is 4. So, the stepping motor goes through four microsteps before it completes one full electronic/magnetic cycle. One full cycle is small fraction of rotation of the output shaft. Typically a microstep is 0.0254mm (depends on the gearing used in the particular motor).

To explain: If the command to the Robofocus controller is for a movement (i.e move 42 steps IN), the controller begins sending pulses of current to the stepping motor. Each pulse will move the stepping motor one “microstep”. Meanwhile, the RFCP (software) has told the controller to consider that it will take some number (e.g., four) microsteps to make one step or count that the RFCP will consider. Thus, the RFCP command to move 42 steps will make the motor move 42x4=168 microsteps. The controller counts these steps, and after 168 microsteps will send the current position back to the RFCP for display. Note you can change the speed of the motor by changing the “pause/microstep” setting in the RFCP.

So the question is…what have you configured as your microsteps per step?
A 1:1 is rather extreme. Try it. If you can't hit the CFZ using this setting, something is seriously wrong! I’m currently running it as 3 (which is very fine adjustments) - provides ~4 steps (12 microsteps) in the CFZ (until I can see HFD and/or FWHM changes occuring). This may not seem like a lot of play, but with the repeatability of FocusMax HFD routine, it nails the CFZ everytime - even if through a Ha narrowband filter which shifts the focal plane (if not parfocal with your LRGB set).

:)

[montewilson]

Hi Jase - As usual your reply is so weighty and considered that it merrits a serious reply and it will get one but in the mean time, Paul/Zuts you made no mistake in getting a RoboFocus, they are great! Our fingers can't do even 10x above the resolution that the RF can.

Feel free to PM me or send to secretary ayt asnsw dort com

[robgreaves]

Monte, you need more focal length and make that CFZ wider! ;)

Somehow, I nail focus every time, with careful thumb/forefinger manipulation on the '150.

Regards,
Rob

...although I do have a robofocus system lurking somewhere in a box. Maybe when it's observatory mounted :)

[Zuts]

Hi Rob,

On my TV85 I have 68 microns CFZ with reducer. I am getting better at 'nailing' it with a TV fine focus. It's not that I'm lazy it's just that when I do get it, it has usually taken around 30 minutes and I am never sure if it is correct or that if i went a bit longer could I make it better?

Also, because it takes so long I am usually reluctant to refocus because of temp change or take out the reducer or even say add a UHC filter because all this changes the focus.

If the robofocus gives me consistant, good focus in say 5 minutes, allows me to swap in reducers/filters with ease and some people say its possible then I will be a very happy chappy indeed.

Hi Monte,

Thanks for the offer of help.

Paul

[robgreaves]

Another focuser I have a lot of experience with is the Optec TCF-S. A superb bit of kit. Better than the robofocus in terms of accuracy, repeatability and rigididy, and also that it's its own Crayford focuser that is built like a tank. But - it of course adds back focus, has limited focus range, and also comes with a proprietary thread that you have to use adaptors (sometimes home made) to add to your system.

Again, it's temperature compensated and will self-adjust once it learns your coefficients of expansion (or should I say contraction).

Also, I find the autofocus routines only really work on good steady nights. Trying to use them in the UK was sometimes impossible. It'd finally move to the final calculated position, the seeing had changed, and the stars were well off focus :(

Regards,
Rob.

[jase]

For the benefit of others… As previously mentioned, the RoboFocus controller counts microsteps internally, but you can set how many microsteps equals one control program count or step. By changing the microsteps per count setting, you can change the “fineness” of the focus positioning.

As a general rule, if you are using a Rack and Pinion where the focus changes quickly as the shaft turns, you would might use 1-10 microsteps per count, while a much slower focus movement (such as an SCT course focus knob) would benefit from 16-32 or more microsteps per count. With a count maximum of 64,000, a larger microsteps/count also allows a longer travel without resetting the position count. If you find that you need a finer focus adjustment, simply decrease the number of microsteps per count.

If you are not running the RFCP (software), you can connect to the Robofocus controller over a serial connection. To change the step size (and other configuration components), use the following; FC[DEF]. Commands RoboFocus to set configuration. If DEF=0 then RoboFocus responds with current configuration settings. If non-zero, configurations be set as follows:

D-Duty Cycle. This character has an ASCI value of 0-250 which corresponds to a duty cycle for the stepper of 0-100%.
E-Step Delay. This character has an allowed ASCI value of 1-64 approximately equal to the time delay in ms per microstep of the stepper.
F-Step Size. This character has an allowed ASCI value of 1-64 equal to the number of microsteps needed to produce one “step” (count) of the focus position. High numbers are coarse steps.



Five mintues Paul! Wow, after five minutes I’d want to be well into my first sub exposure by then! Typically hit the critical focus zone in ~70 seconds. Though keep in mind, I’m plate solving with FocusMax to find the most optimal star to focus on i.e 4-7 magnitude. So the total process is a little longer. Below is a snip of an imaging run of the Seagull nebula (http://www.cosmicphotos.com/gallery/image.php?fld_image_id=130&fld_album_id=11) I did recently. I’ve highlighted the main focus run in bold (excludes the plate solve FocusMax AcquireStar function).

11:00:42 Start slew to SeagullNeb...
11:00:44 (wait for slew to complete)
11:01:37 (slew complete)
11:01:37 (request plan-start pointing update)
11:01:37 Ready for SeagullNeb (# 1 of 1 in set 1)
11:01:37 Updating pointing...
11:01:38 (taking 10 sec. exposure, Clear filter, binning = 1)
11:02:15 Plate-solve pointing image.
11:02:16 1359 image stars found
11:02:16 706 catalog stars found
11:02:16 Solved! 228 stars matched.
11:02:16 Average residual is 0.35 arcsec.
11:02:16 Pointing error is 1.741 arcmin @ angle 309.63
11:02:16 Image FWHM is 5.5 arcsec (1.57 pixels)
11:02:16 True focal length is 53.2 cm.
11:02:16 True image center (J2000): 07h 06m 57.5s -10 53' 39.38"
11:02:16 Imager sky position angle is 242.7 deg.
11:02:16 Image FWHM is 5.5 arcsec (1.57 pixels)
11:02:16 (avg FWHM = 8.65 arcsec)
11:02:18 [sync] pointing model updated
11:02:18 Internal guider, always re-slew
11:02:18 Re-slew to target.
11:02:18 Start slew to SeagullNeb...
11:02:21 (wait for slew to complete)
11:02:32 (slew complete)
11:02:32 Target is now centered.
11:02:32 SeagullNeb specifies 3 sets of images.
11:02:32 This is filter group 1 of 3.
11:02:32 Using Red filter for imaging
11:02:44 Do requested auto-focus...
11:02:47 Starting FocusMax AcquireStar autofocus...

11:02:47 ** Starting Acquire Star Sequence **
11:02:48 Catalog: GSC-1.1
11:02:49 Telescope ACP->Gemini is connected
11:02:52 Binning = 2
11:02:52 Taking 5 sec pointing exposure
11:03:10 14303 catalog stars found
11:03:12 1305 plate image stars found
11:03:14 ** Plate solved **
11:03:14 Plate(J2000) RA : 07:07:01.6 Dec: -10 54' 29.7" Roll ang.= -62.715188
11:03:14 Plate(Topo) RA : 07:07:26.1 Dec: -10 55' 11.6"
11:03:15 Current pointing: RA 07:07:26.1 Dec -10 55' 11.6"
11:03:15
11:03:15 ** Getting catalog stars between mag 4 & 7 **
11:03:16 Looking for target star > 70 deg altitude <<<-- Don't focus on a star lower than 70 degrees! Seeing will impact result!
11:03:16
11:03:16 Field search = 2.0 x 2.0 (degrees)
11:03:16 1 potential target stars found
11:03:16 Current target stars = 1
11:03:17
11:03:17 Field search = 4.0 x 4.0 (degrees)
11:03:18 6 potential target stars found
11:03:19 Target stars = 3
11:03:19 Identified target stars:
11:03:19 Star 1 mag 6.50 RA 05:08:33.9 Dec -35 42' 29.2" (Topo)
11:03:19 Star 2 mag 4.59 RA 05:04:43.6 Dec -35 28' 19.3" (Topo)
11:03:19 Star 3 mag 6.30 RA 05:04:45.3 Dec -35 41' 39.5" (Topo)
11:03:20 Using 6.50 mag star at RA 05:08:33.9 Dec -35 42' 29.2" (Topo) <<<--star to be used to focus
11:03:20 Slewing 36.5823 degrees to target star
11:03:48 No star centering desired
11:03:49
11:03:49 System: FSQ
11:03:49 ** Beginning Focus run **
11:03:49 LS: -0.343528 RS: 0.344151 PID: 0.85 NFHFD: 10
11:03:50 Move direction: In
11:03:50 Temperature = 617.0
11:03:50 Filter number = 0
11:03:50 Current position = 769
11:03:50 Focus Start HFD: 20
11:03:50 Exposing Central Region = 50% of CCD width
11:03:58 Target star found at X = 1692, Y = 975
11:03:58 Focus exposures will be 1.00 sec
11:04:03 Position , HFD , Mean Best Focus , X , Y , Flux
11:04:03 769 , 4.08 , 0 , 1642 , 996 , 141472
11:04:10 815 , 18.30 , 0 , 1641 , 997 , 483033
11:04:10 On correct side of focus
11:04:10 Moving to start HFD
11:04:10 Move to Near Focus HFD
11:04:16 791 , 8.44 , 0 , 1641 , 996 , 464858
11:04:16 ** Starting Near Focus **
11:04:16 Position , HFD , Mean Best Focus , X , Y , Flux
11:04:16 791 , 8.44 , 767 , 1641 , 996 , 464858
11:04:20 791 , 8.57 , 767 , 1641 , 996 , 460714
11:04:25 791 , 8.52 , 767 , 1641 , 996 , 460177
11:04:30 791 , 8.59 , 767 , 1641 , 996 , 495194
11:04:34 791 , 8.72 , 767 , 1641 , 996 , 477172
11:04:35 Best Focus is: 767
11:04:43 Star is too close to frame edge or is too weak
11:04:43 Increasing frame width
11:04:51 767 , 3.99 , 767 , 1641 , 997 , 411660
11:04:53 Focusing Completed
11:04:53 Focus time = 64 sec
11:04:54 Performing return slew of 36.5834 degrees
11:04:54 Slewing to: 07:07:26.1 Dec: -10 55' 11.6"
11:05:17 Acquire Star completed
11:05:18 FocusMax auto-focus successful!
11:05:18 HFD = 3.99
11:05:18 Focus position = 767
11:05:18 Autofocus finished.



Indeed Rob, the TCF-S is a excellent focuser. Though as you point out, not ideal in all situations considering the back focus. i.e you wouldn’t install one on a fast APO. Doing so would sacrifice your f-ratio. This is where RoboFocus comes into its own element (no footprint on imaging train). If you want something better than a RoboFocus and has an extremely low profile, it’s very hard to go past the FLI PDF - http://www.flicamera.com/fli/focusers.asp?product_model=PDF. I purchased one a while back, but have not used it as I’ve found the RoboFocus meets the needs.



This maybe true to an extent, but your seeing conditions would have to change considerably for this to have an impact. I think if the seeing was that woeful, I’d call it quits for the night. You wouldn’t be able to get any high resolution work done. Perhaps just bin the chip and collect RGB data to ensure the night is still productive.

Keep in mind that autofocus routines such as those utilised in FocusMax don’t simply take a quick image of a star to judge focus. By default it takes five exposures of the same star to validate fluctuations in seeing and calculate the average. This is configurable – you can take more or less. (I’ve highlighted these in the above output in red) - You will note the HFD fluctuations range from 8.44 to 8.72. It is good practise to schedule an autofocus run every 30-40 minutes during an imaging run. Should seeing have changed, the focuser will then compensate for it. Temperature shifts are a different topic in which I'm happy to discuss. There is nothing worse spending a night under dark skies to realise the following morning that your five hours of data collected is soft on all sub exposures due to a minor focusing mishap. I write it off as experience.

:)

[RB]

BTW what is the CFZ for the 150 and the 130s guys (with & without F.R), anyone know?


I assume you use it on the 150?
What adaptors do you need for this?

Cheers

[Zuts]

Hi RB,

I got my CFZ from CCDCalc. This allows you to select a scope and FR and then calculates it.

Paul

[jase]

Native - TOA-130 F/7.7 CFZ – 130 micron
Native – TOA-150 F/7.3 CFZ – 117 micron
Reducer - TOA-130 F/5.8 CFZ – 74 micron (this is using Tak part# TKA31580)
Reducer – TOA-150 F/5.8 CFZ – 74 micron (this is using Tak part# TKA32580L)
Note the TOA-35 & 67 (TKA31582 and TKA31583 respectively) are flatteners.

There is no difference in f-ratio between S and F TOA-130 version. For those wondering. - the S comes with a 2.7" while the F has the 4" focuser (monster).

There is a discrepancy in the documentation. The TOA-130 manual indicates the reducer drops the f-ratio to F/5.67, but the reducer documentation indicates its F/5.8. Not a big difference. If there is someone has a CCD camera and plate solving abilities, they could upload their plate solve info with and without the reducer. We will then be able to confirm the f-ratio, but you'd need to describe your imaging train i.e. what adapters you've got in line.

[rally]

The Critical Focus Zone (in a perfect system) can be calculated by

CFZ (microns) = Focal ratio x Focal ratio x 2.2

1 micron = 0.001mm

In the real world it is often the case that the CFZ is up to a third larger than the theoretical figure. - Does anyone know why ?

Stating the obvious - the focal ratio of the system must include all optics in the imaging train.

Cheers Rally

[RB]

Thanks for the info Paul, Jase and rally.

Yeah the 4" focuser is massive Jase, the smaller 2.7" while not flimsy pales in comparison.

[jase]

I'm not sure what you mean Rally. The CFZ is directly related to the f-ratio. The faster the f-ratio, the shorter/smaller the tolerances the CFZ will be.
As an example, the native FSQ-ED is 530mm @ F/5 which equates to a CFZ of 55 microns. With the F/3.6 reducer installed the CFZ drops down relative to the f-ratio, so 382mm @ F/3.6 becomes a 28 micron CFZ.

You are correct in that the imaging train will impact the f-ratio. Folks automatically believe that because they've got the Celestron/Meade F/6.3 reducer installed, that they will be imaging at F/6.3 on a F/10 SCT, but you need to take into consideration spacers and where the reducer is placed in the imaging train relative to the focal plane.

The Lumicon Giant Easy Guider (GEG) has a focal reducing lens that can be moved to provide F/10,F/6.5,F/5.5 and F/4. This is purely what the GEG is capable of, it is not an indication of what you'll be imaging at.

"F Ratio calculations: In the past I have unfairly critiscised Lumicon for quoting optimistic f/ratios for the GEG. In fact this is not true - I was neglecting two facts: the intrinsic f/ratio of my LX200 is actually f/10.7 (rather then the specified f/10), and the GEG itself adds a further 115mm to the focal length, bringing the system f/ratio to f/11.2. If the system f/ratio had been f/10 to start with, then the quoted Lumicon figures would be pretty much correct. F/ratios in all modes were calculated as accurately as possible by measuring the distance between star points at the extreme edges of the 35mm frame." Quoted from http://www.astrocruise.com/geg.htm

So once you know your precise f-ratio, you can successfully calculate the CFZ. Plate solving (used for astrometric calculations/pointing) is an accurate way of determining what your true focal length is (and in turn your f-ratio).

[rally]

Jase,

Re CFZ being potentially larger than the theoretical value :

It came from something (authoritive) that I read. (I can't find it anywhere - possibly in the bookshelf - so I will try to dig it out)
As I recall the article didn't provide any specific detail on that issue.

Added this after posting by editing
-------------------
Found a web reference from the The New Astro - Ron Wodaskis' book "The New CCD Astronomy" Page 39 - but no detail
www.newastro.com/newastro/samples/c2a.pdf

I have read further into the book to see if there is any detail now and cant find anything yet.
--------------------

But in trying to hunt something down about it I did find this interesting site - Wilmslow Astro
http://www.aozc64.dsl.pipex.com/software/formulae.htm

It contains lots of useful astro formulae some of which include calculating CFZ in a number of different ways.
Online Calculators are included for each formula.

I'll leave the experts to figure this !

Cheers

Rally


Critical Focus Zone

Calculate the length of the zone in which the focused image of a star is smaller than the size of its Airy disk.

cfz = 2 x f-ratio x AiryDisk

which simplifies to:

cfz = 4.88 x f-ratio^2 x wavelength

For CCD cameras, if we take a 2x sampling ratio:

ccd fz = f-ratio x pixel size

where:

λ = wavelength of light (Red ~700, Green ~530, Blue ~470)

Note that because at low f/ratios the size of the Airy disk becomes significantly smaller than typical CCD pixels sizes I have introduced a value for the CCD Focus Zone. The value for the CCD focus zone takes the larger value of the CFZ, or where the Airy disk is half the effective pixel size (2x over sampling ratio) the CCD focus zone value defined above. For small focal ratios the CFZ gives a misleadingly small figure for imagers.

[Zuts]

Hi Guys,

Still havnt managed to use it yet as it's raining. When I do, do I need to do two seperate sets of V curves, one with a reducer and one without?

Thanks
Paul

[montewilson]

HI Jase - I have taken on board what you have written and here are my new questions which I expect you will answer without even batting an eyelid!

I agree with you that FocusMax seems to have the runs on the board. I am not using DL so getting it work means I will have to enter into the dark art of ASCOM of which I am rather afraid.

CCDWare have CCDAutoPilot which claims it can run Focus Max as well as Bisque imaging programs. It kind of acts as a front desk for all these programs. I will ask on their forum if I need to have DL installed to use FMax. If not, COOL!

What is the minimum linear movement of the tube that RoboFocus can do on the FSQ? You mentioned some values but I am not sure what they relate to. Should I be able to get within the CFZ with with say a couple of pulses to spare or will I be jumping right over it?

As I have the venerable "106" could I use that focal reducer on my scope? That is such an exciting alternative to buying a pricey Canon or Nikon telephoto lens. Would that mean my FSQ will be a 400mm f3.75?

[jase]

Firstly, Rally, thanks for providing this information. “Note: real world values for the CFZ are approximately 10%-30% greater than the theoretical values...” I have pondered on this. Need more time, but I believe I can understand the reasoning. Assuming that the CFZ shifts based on wavelength (blue short, red long). If you use the calculator you provided, the CFZ does shift in size.
So if we plug some values into the formula - cfz = 4.88 x f-ratio^2 x wavelength
For example if I use the blue (~470nm wavelength), f-ration of F/5 (FSQ) and Pixel size of 9 microns for STL11k (no binning so 1x1) – the CFZ is 57 microns. If I use the same parameters, but change the wavelength to red (~656nm Ha) – the CFZ changes to 80 microns. So based on wavelength alone the CFZ will shift, but I not sure if it is justified as high as 30%.
Perhaps, to circumvent this ever shifting CFZ, calculators/formulas only base the measurement on the start of the visible wavelength being 450nm (blue). Using the above formula with a 450nm value, the FSQ hits the 55 micron CFZ I previously stated. I also plugged in the values for the TOA-130 @ F/7.7 and obtain a CFZ of 130 microns. So in a way, this confirms that the default value for establishing the CFZ is at the 450nm wavelength. Makes sense to an extent, i.e. you wouldn’t try focusing on a star in the NIR or IR wavelengths though it is possible considering they emit over a broad spectrum. On reflection, this would be an interesting test with an achromat. Considering they are not well corrected instruments, focusing all wavelengths simultaneously – compared to an APO that is.
Will give the airy disk some further thought - The CFZ remains the same, but then the CCD Focus Zone (CCDFZ) changes based on pixel bin values. There would have to be more to it than simply the significant smaller size of the Airy disk.
Thanks for formula page too. I’ve been after something like this for sometime. The CCD internal reflections calculator is great. Will be sure to use it in the further to determine what is happening in some of my images.



Yes Paul, you’ll need two v-curve profiles – one native (no reducer) and the other with the reducer. This is because the reducer will shift the focal plane. Make sure you perform more than one v-curve run. The more runs you do, the greater the accuracy (each successful run is automatically added to the model). I usually aim for 8 or more. You can have as many v-curve profiles as you need.



Hi Monte. You don’t need MaximDL to run FocusMax. It can integrate with CCDSoft if that is your acquisition tool of choice. Indeed CCDAutoPilot supports FocusMax. As you mention, CCDAutoPilot is the “front desk” for backend applications. It’s the glue to brings everything together to automate an imaging session. I use ACP which does an identical job – the output I provided above is a snapshot of an ACP log. I have ACP doing everything, plate solving, automatic focusing, filter offsets, automatic guide star selection, telescope point model, automatic acquisition of skyflats and other calibration frames and more. I simply upload my imaging plan that I’ve created from TheSky and watch it do it thing. Both CCDAutoPilot and ACP are very powerful packages. I went for ACP as I’m an ASCOM advocate and perhaps more importantly it has a web user interface (some CCDAutoPilot doesn’t have yet). A thin web interface was a must considering I’m uploading and running sessions over a satellite link to the remote observatory. Satellite is the only form of “broadband” due to the remoteness. Anyway, a slight diversion from your questions…



I have not measured the linear movement of the FSQ draw tube. Each pulse is very small and not easily perceivable. I don’t have any precision tools to make a measurement, but when I’m down at the observatory, I’ll see what I can come up with. As previously mentioned, typically a microstep is 0.0254mm (depends on the gearing used in the particular motor).

The focuser should be able to sit in the CFZ for at least a couple of steps before you notice changes to FWHM/HFD figures. Your microsteps per step are too large if you are jumping right over it. You can test this manually. Simply select a star in CCDSoft and start grabbing subframes for it (so you get rapid updates). Monitor the FWHM/HFD values while you manually push the in or out button on the Robofocus controller. Make sure you push and release, don’t hold the button in. One beep noise is one step (not a microstep). Keep moving through the focus range (from out of focus, in-focus, then out of focus). Make the initial first pass and take note of the focus position reported by the RFCP (software). Then make a second pass, but take a closer look at the FWHM/HFD measurements as you proceed. A previously mentioned I usually get 4 steps before values change considerably. Make sure you sit in the CFZ for a while to monitor the FWHM/HFD measurements as they will fluctuate a little due to seeing. They shouldn’t be varying too much as the 3.5 pixel/arc delivered with the STL11k’s 9 micron pixel size is far from high resolution. Once you know how to reach the CFZ manually, the progress to automation will be natural.

You may have already done this but make sure you tell the Robofocus controller the focuser travel (i.e. the draw tube all the way out to all the way in). The process is called calibration. It is important to set this up correctly as you’ll lose your absolute positioning info if not.
To calibrate the RoboFocus, use the OUT button to run the focuser all the way “out” (or at least as close to all the way as you want). Release the OUT button, and turn off the RoboFocus. After ten seconds, turn on the RoboFocus while pressing the IN button. The RoboFocus will read the IN button, will know that it is in the training mode (it will beep five times). Let up on the button during the beeping. After the beeping ends, the focus will begin running in without holding the button. When the focuser has moved all the way in (or as far as you want it to go which may take several minutes), you may press either button to stop the training run. When the motion stops, the RoboFocus records its position as =2 and will record the total focuser travel as the number of steps run, and calibration is now complete. Thus, a count of 2 is the innermost position. As the RoboFocus moves outward, the position increments by one for each step of the motor. You can read the MaxTravel number of steps from the Configuration screen in the RFCP (you will need to refresh the screen after calibration). Something to be aware of is that you don’t need to let the focuser go through the full travel length of the draw tube. I stopped mine around 50 steps before the focuser was all the way in. It doesn’t make a huge difference. What is important that the CFZ position should reside roughly equal on either side of the focus travel.



Sorry, no can do. The FSQ-106N does not support the new Q reducer (TKA36580 Reducer-QE 0.73x (FSQ-106ED)). It drops the FSQ down to 382mm @ F/3.6. There simply isn’t enough backfocus on the FSQ-106N to support it.

[montewilson]

Ive been chatting with a guy called Frank Barnes on the CCDWare forum and he had this information about RF and the FSQ.

Hi Monte
I use a RoBoFocus Microstep of 1, a MicrostepPause of 10, and a Duty Cycle of 75%. This is the same for both of my FSQs.

@Focus2 does like to start close to focus. It makes a single pass from inside of focus, through focus, to outside of focus to get the best focus setting. I have found that the following parameters work very well for a FSQ/STL11K and the clear filter :

Samples = 20
Averaging = 1
Focus Range = 400
Subframe Box Size = 100

I found that a Mag 5 star exposed for 0.5 sec. is what is needed in my sky conditions. One thing that I found with @Focus2 that can help it be successful is to experiment and find how long an exposure is required to keep the "out of focus" star at the beginning of the focus run above an ADU count of 800. In my experience, when the star was fainter than that value at the start of the run, @Focus2 usually didn't work. Stay above that value, and it works almost everytime.

Both my RoboFocusers are attached directly to the standard shaft of my FSQs and they are both FSQ-106Ns, not the newer ED model. I don't have a 10:1 focuser on either one. HTH ....

[jase]

Sounds like you're all set now Monte. Let see some images.;)

[Zuts]

Hi,

I finally got to use my robofocus the other night. It took me about 15 minutes to do about 5 v-curve runs. After that and on a subsequent night it was able to find focus in about 34 seconds.

This is with Focusmax, integrated with CCDSOFT 5. I can highly recommend Focusmax. I have never tried @focus, and probably never will, given that Focusmax is so easy.

Paul

[montewilson]

OK I agree it's about time but here is the update on what I am doing and why I am not giving results at the moment.

My daughter turned one on ANZAC day and this Sunday is her birthday party. As Lou Reed says, "there are no stars in a New York sky, they're all on the ground" so all the stars I will be seeing will be much closer to me than normal ;)
I knew this break was coming so I decided to have a big overhaul of my equipment and a number of upgrades.

I have a remote guide head for my camera so I no longer need the STV. I will keep it for public open nights, it's cool for video sky displays.

I am making a plate to move the FS-60c forward about 70mm. This will help with Dec balance and clearance between the two cameras.

This plate will also hold my new guide scope. I bought a cool red dot gun sight for $79 from a gun shop in Sydney - www.magnumsports.com.au - the Nikko is not on their site but the owner Andrew knows me well so if anyone is interested tell him it is the one he sold Monte, he knows which one it is. It is like a much attractive Telrad. See picture. You'll have to gerry-rig a holder for it but it's worth it.

I am installing a new bad-ass pier at the ASNSW dark sky site where I normally image. It is part of a galvanized street light pole that I bought at the tip recycling centre for $10. I will have a better connection between the pole and mount. I will also be burying the wiring to the house for power and cat-5 and installing a water system for camera cooling.

I am going to get a copy of CCD AutoPilot and FocusMax.

I will install a second Robofocus on the FS-60c. I have made a USB operated switch that can change between the two stepper motors so I can use the one Robofocus box from a distance.

All the wiring loom is being re-run so it will be lighter and more flexible.

In all, I am making the best use of this two month break.

[Zuts]

Except you never mentioned actually taking a photo :P

Good luck with all your projects, the main one of course being your little girl :)

Paul

[leinad]

Hi Gents,

Loong time no post, and fitting to post in a thread thats so old :P

I've just returned to setting up my system again, and in an attempt to ensure everything is setup correctly and optimized I was hoping you could check my calcs on autofocus..

I havent had any issues as yet with focusmax with 4 microsteps; but without crunching all the numbers just yet; just checking Im understanding this correctly.

I havent as yet measured the exact draw tube distance and counts to work out the step/microns, so Im assuming what others reported on yahoo forums to be similar.

Going by the previous mentions of steps to microsteps; and ensuring the focusroutine(FcsMax) is running correctly; I believe as Jase mentioned for the FSQ at 3 microsteps or even two gives a +- 3 counts for focusing in the CFZ and accounting for change; though focuxmax magic can mean measure the best HFD anyway as long as the counts(steps/micron) dont overlap the CFZ.

Not sure how critical checking this against diff filter(nm), but will work out after for the satisfaction of knowing nayway..

*gulp, my workings below.

So 3 or 4 microsteps should be fine; 2 may be a waste of time(pun).



4000 steps Focuser draw tube distance

using 1:1 step /microstep
------------------------------

55micron CFZ
Tak106
reported as: 6.7microns/step

55 micron / 6.7 micron = approx. 8 steps(counts in CFZ).

2 microsteps x 6.7 = 13.4 micron 55/13.4 = 4.1(4) tolerance steps(count)
3 ...steps x 6.7 = 20.1 micron 55/20.1 = 2.73(3)
4 ...steps x 6.7 = 26.8 micron 55/26.8 = 2.05(2) <--- +- half 1/2 CFZ
5 ...steps x 6.7 = 33.5 micron 55/33.5 = 1.64(1-2) <--- +- almost 2/3 CFZ

[cfranks]

Just 2 quick questions. Those using FocusMax, are you using v4? If not, could you please tell me the actual version number you are using.
Thanks
Charles

[cfranks]

I didn't mean to change the subject, it's just that using TheSkyX @Focus2 with my RoboFocus setup always requires a 4 step manual adjustment. I tried to test FocusMax 3.7.0.51 with MaximDL and couldn't get it to work. Lots of 'internal errors' that I didn't take a note of, of course!

Charles