Hi,
Suddenly AF is no longer working for me. AF consistently fails, ending most times by displaying the Auto Focus Error message: “Auto Focus received a bad HFR value 2 times in a row! Something maybe wrong!”
Im using the Atik 16200 camera and Baader Steel drive focuser with a C14 SCT. AF was successfully working about 2 weeks back but not since I downloaded the latest few SGP updates (dont know if this is just a coincidence).
I have tried alot of different combinations of values for Exp time, Bin, Step size, minimum star diameter and crop value put the problem remains.
Any idea how to correct this?
Thanks,
Anders
Link to Logs
Useful Info
OS: Microsoft Windows 10 Pro Ver: 3.1.0.434 .NET: 4.8
You have a pretty massive step size set (400, 500 and 600 steps … NOT microns btw). I suspect you have an HFR of 0 because the first image in AF is so badly out of focus that there is not data to be found.
Are you starting AF while near focus? How did you decide on those values? For your focuser that’s moving about 1360 microns per data point and the initial move to the first data point is moving 4760 microns. Unsure if those numbers sound right to you or not.
You can fine tune these numbers when you start getting success, but for now, move to 9 data points (up from 7) and then, to determine your step size, get close to focus (eyeballing it is ok). Start looping Frame and Focus with Image History on. You will see HFR appear in the upper left corner. Note the HFR and current position of the focuser and slowly move focus out until your HFR is ~2.5 to 3 times the starting value. Note change in steps. Multiply this number by 2 and then divide the product by 9. This is your step size.
Thanks for the input. It was very helpful to understand that the error message in this case was triggered by HFR being 0. The AF routine is working perfectly for me again. The problem I had was more to do with the light from the full moon together with quite a few hot pixels of my new camera which I’m still getting acquainted with. The solution was to create a AF dark library (which I had not used before) and then try again with the slightly darker sky early on tonight with the following result:
The step size of 400 in my case I believe is fine and is acually based on the formula you mentioned.
Btw, I really like the latest enhancements of the AF function.
Hi Ken,
I have been using autofocus for years now and it has worked like a champ. I have always obtained perfect v-curves. However I am having the same issues as the OP with the new auto focus routine.
I have always used a number 5 times out from actual focus. Is this too much for the new routine? Also, you state to divide by the actual number of data points. This in contrast to the help file formula which I have always used which subtracted 1 from the data points as in my case it would be 9-1=8.
I am getting some filters to focus but have not successfully established focus with my Baader blue filter or my Astrodon 3nm O3 getting the same error as the OP.
No logs here as I have only tried it in 1 outing and have been experiencing horrible weather ever since.
This will cause pretty steep curve slopes and we are still getting reports in about the new quadratic fit routine. As such, we don’t have much data about how the those kind of “tall V” shapes get fit. We did A LOT of testing around 3x though… Give that a try. If it doesn’t work for you, just send us logs and we’ll take a look. So far, there hasn’t been a single person we were not able to get going with some settings adjustments.
Thanks Ken for the quick reply. Major snow storm here in PA so hopefully mid to late next week for a test. One other thing is my original comment was the step size. Should I divide by the number of data points or the number of data points -1?
Ultimately, it doesn’t matter… it’s just a rule of thumb intended to get you to a reasonable HFR on the outer bounds of the AF run. If you already know your step size and data points required to yield 5x, then you are probably safe trying a step size value that is ~ 0.6 * [current step size]
