AF or Bahtinov?

[Dazzyt66]

I've been considering building my own RPI based autofocuser next month (when I get my next 'spending allowance' 😂). This will initially go on my ST80 (permanently setup). As per usual I keep doing my research and I came across this thread:

So now I'm torn. I can get the AF bits and a new case for the RPI which'll cost about £60 or I can do the 10p mod that @ollypenriceuses (and do it today) and then buy a Bahtinov mask for about £15.

My question is: I want to be able to do my imaging mostly using Ekos scheduling. At the mo I either check the focus before I start, using focus assist in Ekos (manually until I have the best number) and then lock the focus on the ST80. For subsequent nights I just quickly check the focus numbers in Ekos are roughly what they were previously and leave it to go. Does the focus change overnight due to seeing etc (or if I move to a new target) if doing long sessions? Or once its setup generally is that it for the night? If its the latter then 10p and a Bahtinov seems the way to go given that the ST80 is as basic as it gets (although I do love it!!).

Cheers

Daz

[Clarkey]

I have an auto-focuser and I would always focus at the start of a session and on changing filters using offsets (so not applicable with a DSLR). I do have the set up set to re-focus at 10% deviation of FWHM, but it very rarely needs to re-focus during a session. This is even with a FL of 1600mm. At very long FL and detailed imaging you might need to make changes during the session as the effect of temperature is likely to be exaggerated. The main cause for needing refocus would be large changes in temperature altering the optical train. In Cumbria it is always cold so this is not an issue🤣. I think at the FL of the ST80 I would be surprised if you really need to re-focus, but you could possibly get a marginal benefit. I generally do not need to refocus my ED80 - in fact I don't think I ever have. Change of target would not make any difference - infinity focus is always the same.

You could save even more money and make your own Bahtinov mask. I made my 'temporary' one six months ago and I have not felt the need to buy a new one. Eventually I will probably leave it outside and the dew will get at it as it is made from cardboard. But for now I'll save my £30 for some other expensive bit of astro kit!

[fwm891]

I have an autofocus unit (sestoSenso 2) which I use on a WO ZS 73. Whether I use it depends on which camera I attach.

If I attach my ASI 533 then I will use the AF unit. However, if I attach my Nikon 800E the files take so long to download and assess focus if becomes a slog - then a visual on the magnified camera screen is much faster.

The other issue is subject brightness, if your FoV has few faint stars tthen an AF unit will be the better option otherwise your moving to a nearby brighter star each time you change a filter. Using a mask or a magnified screen view with faint stars is almost impossible.

Edited April 22, 2021 by fwm891

[Clarkey]

8 minutes ago, fwm891 said:

The other issue is subject brightness, if your FoV has few faint stars tthen an AF unit will be the better option otherwise your moving to a nearby brighter star each time you change a filter.

Very good point I forgot to mention. This can be a right pain when using a mask. Having said that, you can pre-focus with the mask then run the AF routine to check the focus has not changed. Interestingly, when I use a Bahtinov mask and the AF routine there is always a slight difference. Not sure why. It would not be noticeable visually so not an issue really a problem.

[jacko61]

33 minutes ago, fwm891 said:

. However, if I attach my Nikon 800E the files take so long to download and assess focus if becomes a slog - then a visual on the magnified camera screen is much faster.

 

I had this issue with my Canon 70d until recently. A 3 minute exposure would take 3 minutes to save.  Some kind soul told me to turn off noise reduction in the camera and long exposures now take about a second to save. No idea if Nikon have the same setting but it would be worth a look. 

Graeme

[tomato]

1 hour ago, Clarkey said:

I have an auto-focuser and I would always focus at the start of a session and on changing filters using offsets (so not applicable with a DSLR). I do have the set up set to re-focus at 10% deviation of FWHM, but it very rarely needs to re-focus during a session. This is even with a FL of 1600mm. At very long FL and detailed imaging you might need to make changes during the session as the effect of temperature is likely to be exaggerated. The main cause for needing refocus would be large changes in temperature altering the optical train. In Cumbria it is always cold so this is not an issue🤣. I think at the FL of the ST80 I would be surprised if you really need to re-focus, but you could possibly get a marginal benefit. I generally do not need to refocus my ED80 - in fact I don't think I ever have. Change of target would not make any difference - infinity focus is always the same.

You could save even more money and make your own Bahtinov mask. I made my 'temporary' one six months ago and I have not felt the need to buy a new one. Eventually I will probably leave it outside and the dew will get at it as it is made from cardboard. But for now I'll save my £30 for some other expensive bit of astro kit!

Well, this is an instance where my experience has a different perspective. I am currently imaging at 1050mm FL at 0.94 arcsec per pixel. I monitor my subs through NINA and sometimes (rarely) I can take 20+ 3 min subs with the HFR values staying steady, but other times it can drop off straight away, and gets worse with every sub, even though the OAG performance is steady. Visually the subs don't look much different to my eye but they are firmly at the bottom of the list after any star analysis in APP. Focus stability on my rig does depend on orientation, elevation on the sky and temperature, stability improves after midnight when presumably everything has thermally settled down. I have a dome so the kit is permanently set up but this time of year I can see daytime temperatures of 30 degress inside so it definitely isn't a constant temperature environment. It is worth noting that my rig is both big and heavy, as this I think is a contributing factor to focus stability.

This isn't a big deal for me as I use the autofocus routine in NINA, but putting a Bahtinov mask on and off every 5 subs would get tedious. 

[Dazzyt66]

2 hours ago, Clarkey said:

You could save even more money and make your own Bahtinov mask. I made my 'temporary' one six months ago and I have not felt the need to buy a new one. Eventually I will probably leave it outside and the dew will get at it as it is made from cardboard. But for now I'll save my £30 for some other expensive bit of astro kit!

I've just knocked up my own cardboard bahtinov mask so fingers crossed will get to try that out tonight - and then (if successful) I can think what to spend my £50 on instead...😂

[fwm891]

56 minutes ago, jacko61 said:

I had this issue with my Canon 70d until recently. A 3 minute exposure would take 3 minutes to save.  Some kind soul told me to turn off noise reduction in the camera and long exposures now take about a second to save. No idea if Nikon have the same setting but it would be worth a look. 

Graeme

Thanks Graeme, No not that sort of issue for me just file size with 1:1 binning. SGP setting will allow me to crop the AF field size so I'll try that next time the 800E goes on the ZS73.

[MercianDabbler]

1 hour ago, Dazzyt66 said:

I've just knocked up my own cardboard bahtinov mask so fingers crossed will get to try that out tonight - and then (if successful) I can think what to spend my £50 on instead...😂

I did that too. Much cheaper I used printer paper and then sprayed it with acrylic after cutting to keep the damp out. Obviously it's not strong but I wanted to see if it helped before committing to anything that was more effort or money.

I used it once with my SP102. My experience was that the diffraction fringes were less evident than the ones in the 'how to' on the Interweb and so it was less easy to tell when they were properly aligned. I didn't notice any improvement in my end result.

Since then I've gone back to manually focussing using the x10 enlargement on the camera screen. I think that with CA in action there is no ideal focus point, just 'least bad'

[Dazzyt66]

21 minutes ago, MercianDabbler said:

I did that too. Much cheaper I used printer paper and then sprayed it with acrylic after cutting to keep the damp out. Obviously it's not strong but I wanted to see if it helped before committing to anything that was more effort or money.

I used it once with my SP102. My experience was that the diffraction fringes were less evident than the ones in the 'how to' on the Interweb and so it was less easy to tell when they were properly aligned. I didn't notice any improvement in my end result.

Since then I've gone back to manually focussing using the x10 enlargement on the camera screen. I think that with CA in action there is no ideal focus point, just 'least bad'

Yeah, I've just been watching YT on how to use one!! 😂 Never been so exited about a bit of cardboard! Interesting about the CA - I know I'm gonna get CA but hopefully the mask will give me a little more focus accuracy than I currently get, then it looks like I can lock the focuser down and forget about it for a while...

[AstroMuni]

3 hours ago, Dazzyt66 said:

I've just knocked up my own cardboard bahtinov mask so fingers crossed will get to try that out tonight - and then (if successful) I can think what to spend my £50 on instead...😂

Do you use/planning to use the Bahtinov mask assistant in Ekos?

Edited April 22, 2021 by AstroMuni

[Dazzyt66]

5 minutes ago, AstroMuni said:

Do you use/planning to use the Bahtinov mask assistant in Ekos?

Yeah, I'm gonna give it a go (I've never used a mask before so never used the assistant) - assuming my mask works of course... 🤞

[Dazzyt66]

Wow! It worked! And the Bahtinov assistant in Ekos is very cool! 👍😀

[rickwayne]

Your estimate for  autofocuser parts is in the ballpark. I started with a Pi running Ekos, added a Waveshare Motor HAT board to it (US$26), got a stepper motor ($30), pulley ($10), and a timing belt ($8).

The big advantage is refocusing as the scope cools down overnight. 1℃ seems enough for my 336 f/5. I never knew when to stop tweaking the Bahtinov mask, though a quantitative figure of merit would have really helped.

[barbulo]

I’m following this thread closely because I have the same doubts as Daz, except that I’d opt for the ZWO EAF instead of the DIY solution. I’ve considered the latter but having the ZWO for €180... (focuser + temp sensor) I wouldn’t take the risk.

My main concern is how EKOS would manage the autofocusing routine with a DSLR. Does it use the live view or takes pictures? Can perform when the scope is pointing at faint objects, or it needs brighter stars?

[Dazzyt66]

3 hours ago, barbulo said:

I’m following this thread closely because I have the same doubts as Daz, except that I’d opt for the ZWO EAF instead of the DIY solution. I’ve considered the latter but having the ZWO for €180... (focuser + temp sensor) I wouldn’t take the risk.

My main concern is how EKOS would manage the autofocusing routine with a DSLR. Does it use the live view or takes pictures? Can perform when the scope is pointing at faint objects, or it needs brighter stars?

Well, thanks to the guys here, I can say I’ve abandoned the AF ideas. The Bahtinov did a great job and I can’t see why I would need to refocus again during the night with the ST80, so for 5 mins at the start I’ll use the mask and the mask assistant in Ekos for now.

Having used the focus module in Ekos with my DSLR (albeit manually) I can confirm it takes images (and auto selects a star from that image to focus on - if you have it set that way). It will then take regular images until focus is achieved (or in my case when I manually stop it after reaching the lowest point possible). I believe if you have an auto focuser it takes an image, adjusts, takes an image and so on until it has established the best part of the focus curve (rather than use liveview). The magnitude of the stars it can detect will depend on what exposure time you have put - I usually go 5 - 10 seconds and haven’t had a problem yet wherever I’ve targeted.

HTH

Daz

[barbulo]

Thank you Daz,

I haven’t yet had the the opportunity to test the EKOS’ focus assistant. Before selling my 150P I used the APT’s Bahtinov aid: happy with it for 1-2 hour sessions; used live view so no pictures taken.
Currently I’m moving to a new configuration with Astroberry, environment where APT can’t be used with DSLRs under INDI. So far. 
I don’t mind using the Bahtinov mask and re-focus every 30-60 minutes, but I don’t fancy the idea of using the camera shutter uncountable times. I prefer the way APT works with live view. 
On the other hand, the idea of automating a 2-3 (or plus) hour imaging session with an auto focuser, is quite sexy. 
I’ll give a second thought. 
 

[Buzzard75]

I will say this for a Bahtinov, it will get you close, but it's not always perfect. I have found a number of times that I can nail focus with a Bahtinov and the stars look good with an auto-stretch in APT. However, when I check the same image in Pixinsight, I find I have holes in my stars meaning my focus is off. I don't trust Bahtinov's any more. I will use them to get close and then run an AF routine. It seems to perform better. As for refocusing throughout the night, it's already been mentioned, but temperature and seeing conditions changing throughout the night will affect your focus and it may need to be redone occasionally. It's good to recheck it after a certain change in temperature has been detected and periodically based on time. This is where a good AF routine will be more beneficial than using a Bahtinov mask. It can be automated where using a Bahtinov mask requires user intervention.

[Clarkey]

32 minutes ago, Buzzard75 said:

However, when I check the same image in Pixinsight, I find I have holes in my stars meaning my focus is off. I don't trust Bahtinov's any more. I will use them to get close and then run an AF routine. It seems to perform better.

I have found that the AF routine and the mask give different focus points, but it has always been very close and not enough to see donuts when processing. I would hope if you get a good Bahtinov specifically for your scope it would give you accurate results. (Mine is a cardboard special!)

[BCN_Sean]

I've got both a Baht and a Moonlite protocol AF system which I put together; and after having a look at both (though even with building things, it makes the cloud god angry so not had that much testing time) I'd have to say that for imaging the AF is better but if (when) I get an ST80 for a quick visual it'd probably not be worth spooling up the AF system.

The biggest advantage of the AF system over the Baht I've found at the moment (aside from it being a bit more precise than fingers on a focus ring or stopping ring creep) is I've not got a nag in the back of the head if I've set up and then I've got to go to work about losing x of the sequence due to slippage or temperature changes.  To me, the Baht is good enough for where I am at the moment, but having the AF and sequencer on it just reduces the failure risk a little.

[Dazzyt66]

This is a greatly informative thread. For me, for the moment, I’m gonna stick with Baht only. I’m pretty handy so could do the AF, but I’m not sure with the ST80 it would bring much more benefit - yeah AF would be nice but I think I’ll relegate it to a future project when I’m better at AP 🙂

[BCN_Sean]

1 minute ago, Dazzyt66 said:

I’m pretty handy so could do the AF

What surprised me was how easy (and cheap) it was to put together, if you are a tad handy with a finger burner then the hardest thing is working out a mount.  When I've got things a bit more settled on mine, there's a few bits in the code that I still need to look, I'll be getting it up in DIY in case folk want to crib some ideas.

[barbulo]

1 hour ago, BCN_Sean said:

What surprised me was how easy (and cheap) it was to put together,

What DIY solution do you refer?

[BCN_Sean]

36 minutes ago, barbulo said:

What DIY solution do you refer?

I found a moonlite type code on Github, then it was built up with an Arduino Nano and a small 5v stepper motor.  When I get some time in a bit, I'll make a thread on it in the DIY.

[Dazzyt66]

6 hours ago, barbulo said:

What DIY solution do you refer?

There’s also a non Arduino version code written by a guy on cloudy nights, if you search on the indi forum (if you are using Ekos) there are a few drivers already written for different motor solutions (the cloudy nights version was the one I was going to do)