7 replies to this topic

[mgermani]

Holy acronyms, Batman!

 

I'm having some difficulty with my new EAF in EKOS. I have a Zenithstar 61 and am using a Canon 600D DSLR. I did get it to work eventually by changing the step size to 10, but the autofocus routine fails as often as it passes. I'm using EKOS with an Astroberry (Raspberry Pi).

 

What I'm wondering is, what settings do people use in the autofocus tab in EKOS when using a DSLR? Exposure/ISO, step size, other settings? Most tutorials I've found online are using dedicated astronomy cameras. Any advice is appreciated!

 

CS,

Mark

[FrostByte]

Are you starting the routine close to focus? Have you measured backlash? Which routine are you using?

 

I use ISO800 with my 600D and a 1-second exposure. Step size 40. I'll have to check my other settings, but I mostly use default.

 

One other thing - make sure you have the EAF selected in the focus module instead of the Canon autofocuser! Ask me how I found out that one....

[mgermani]

Thanks @FrostByte, those are all good questions!

 

I stumbled over starting the routine close to focus - it needs to be REALLY close, as it turns out, or it detects no stars. But yes, I'm within about 25 ticks of focus when I start the routine. I'm using whichever routine is the default. I did measure backlash, and after testing it I found it to be pretty close. in my case, 35 ticks. I was using 3-5 second exposures with an ISO of 800, so maybe I'll step that back. My initial step size was 50 and it was too big - the focus routine wouldn't even start properly. I've got it down to 10 ticks now, and that seems to work better.

 

Should I let it automatically select a star, or should I manually choose my star each time?

 

Thanks again!

[FrostByte]

I let it automatically select the star and use the whole field of view. Are you getting a good v-curve?

 

I use the polynomial routine and SEP for detection. Works well for broadband, but not so much with a narrowband filter. For that I’ll have to use like a 20-second exposure and switch the detection to gradient. 
 

You could try switching the routine to “linear” to see what kind of focus curve you’re getting. It’ll be slower but should be very robust. 

[mgermani]

Thanks for the suggestions! I'll have to try them. I've only tried with a broadband filter (L-Pro) so it'll be interesting to see how things work with my L-eXtreme moving forward. The curve I'm getting is ok, but my seeing isn't great from where I image, so perhaps it's throwing the curve a bit. I manually selected the star last time, but perhaps I'll let it automatically choose one next time and see if that improves things. I'll try "linear" curve too. Lots of variables here!

[FrostByte]

Yeah, there are lots of things to try, but once you get it working, it should be pretty effortless at that point. 
 

If you think seeing might be throwing off the curve, there’s an option to take multiple photos for each point which should average them together and give you a more robust HFR measurement. 

[mgermani]

Question - should guiding be turned off during autofocus? I'm assuming so. I tried both, and it worked both ways, but I'm assuming corrections can really throw off short exposures and the curve.

[FrostByte]

I turn guiding off, but only because the mirror slap gives PHD2 fits. I don’t think it should affect the autofocus routine, but I could be wrong.