Repeated FocusMax Errors = Star is too bright

[Eric Coles]

Bob,

It looks like the autofocus routine repeated multiple times and between exposures sometimes as well. Not sure why this happened or what exactly to do. The log is attached.

Eric

[Bob Denny]

Good Morning Eric -- Looking at the log, I can see that ACP kept calling for autofocus a LOT:

05:01:06 Starting Autofocus...
05:03:48 Starting Autofocus...
05:19:47 Starting Autofocus...
05:37:20 Starting Autofocus...
05:54:33 Starting Autofocus...
06:11:40 Starting Autofocus...
06:28:59 Starting Autofocus...
06:45:17 Starting Autofocus...
07:01:51 Starting Autofocus...
07:18:39 Starting Autofocus...
07:35:28 Starting Autofocus...
07:52:40 Starting Autofocus...
08:08:48 Starting Autofocus...
08:25:52 Starting Autofocus...
08:42:21 Starting Autofocus...
09:00:35 Starting Autofocus...
09:16:49 Starting Autofocus...
09:34:14 Starting Autofocus...
09:51:18 Starting Autofocus...
10:08:43 Starting Autofocus...
10:26:37 Starting Autofocus...
10:44:09 Starting Autofocus...
11:00:53 Starting Autofocus...
11:18:11 Starting Autofocus...
11:34:07 Starting Autofocus...
11:51:02 Starting Autofocus...

Each one is the same story:

07:35:28 Starting Autofocus...

FocusMax runs for 3-ish minutes and it's looking good (the second set of [local] times indicates that these log lines are coming from FocusMax):

07:38:24 00:38:24 Focus has converged
07:38:25 00:38:24 Focus position: 69999

But then FocusMax bails out with an error:

07:38:28 00:38:28 Taking images at focus position
07:38:29 00:38:28 Adjusting exposure based on flux target
07:38:29 00:38:29 ** Star is too bright **
07:38:29 00:38:29 Required exposure 0.000 is < 0.10 minimum exposure setting
07:38:29 00:38:29 ...using minimum exposure: 0.10
07:38:34 00:38:33 Aborting - star is too bright
07:38:34 00:38:34 Moving to previous focus position of 70223
07:38:35 **AutoFocus failed. Check FocusMax log for details.

If this just started happening, I'm guessing that a change was made to the FocusMax settings. I don't know how to fix "Star is too bright" at the end of what otherwise looks like a successful focus run. Maybe you remember what was changed. You can get help on the CCDWare FocusMax forum though, or check the troubleshooting section of the FocusMax documentation.

PS - I noticed that the FocusMax Configuration you're running is the one I saved for you the last time I was on your system:

05:01:29 22:01:29 System: Saved by Bob Denny

Since that is the "current" one I learned something new... FocusMax doesn't save off a snapshot of the settings, when you save them it starts using the new freshly saved settings file. So in order to actually take a snapshot of the FocusMax settings (for safekeeping e.g.) you'd have to save the Configuration, then Save it again or switch back to the old Configuration. Then the first one you just saved is the one that will remain for safekeeping as the "last known good" configuration. I thought I was doing that for you. Sorry!!

Since it looks like you are otherwise running great guns, Please take a few moments to save the ACP Profile and the MaxIm Configuration so you can always return to the last known successful setups. Those do save off a snapshot of all settings.

[Eric Coles]

Bob,

This is confusing. How could it go through a successful focus run and then say the star is too bright. What about just forcing per-filter focusing? Otherwise not sure where to go with this issue. Have not seen it before.

Eric

[Bob Denny]

The ACP per filter vs scheduled autofocus mode is irrelevant. FocusMax failed every time it ran, and after seemingly having completed a successful focus. You need to make whatever change is needed to prevent that. I can only guess since FocusMax is a separate module in this modular system, and I am no expert on it. The key seems to be this FocusMax trace line:

Required exposure 0.000 is < 0.10 minimum exposure setting

Guessing here: The star that FocusMax picked is so bright that it couldn't do a final quality check on it. Even at a 100 millisecond exposure it saturated/flattopped. I doubt that the solution is to allow it to take an even shorter exposure... instead I would look at the FocusMax AcquireStar settings and see of you can somehow prevent it from picking such a bright star in the first place.

Serious guessing here:

05:01:09 22:01:09 ** Beginning Find Star **
05:01:09 22:01:09 ...Filter 1: R (slot 2)
05:01:09 22:01:09 ...Flux target: 300000K
05:01:09 22:01:09 ...CCD central region: 75%
05:01:09 22:01:09 ...Star subframe width: 100
05:01:09 22:01:09 ...Base exposure: 0.50 sec
05:01:09 22:01:09 ...Min/Max exposure: 0.10 / 5.00 sec
05:01:09 22:01:09 ...Target star binning: 2
05:01:09 22:01:09 ...CCD central region: 75%
05:01:09 22:01:09 Taking initial image
05:01:19 22:01:19 ...Looking for star
05:01:22 22:01:22 Target star found
05:01:23 22:01:22 ...Exp: 0.50 Bin: 1 HFD: 10.74 Flux: 2173956
05:01:23 22:01:23 ...Unbinned X: 3418 Y: 3178
05:01:23 22:01:23 Setting subframe parameters
05:01:23 22:01:23 ...Focus binning: 1
05:01:23 22:01:23 ...Frame width: 100
05:01:23 22:01:23 ...Exposure: 0.50
05:01:23 22:01:23 Taking subframe image
05:01:28 22:01:28 Star X: 3418 Y: 3179 HFD: 7.76 Flux: 6001621
05:01:28 22:01:28 ** Star is too bright **
05:01:28 22:01:28 Required exposure 0.020 is < 0.10 minimum exposure setting[/high]
05:01:28 22:01:28 ...using minimum exposure: 0.10
05:01:28 22:01:28 ...Exp: 0.10 Bin: 1 HFD: 7.76 Flux: 6001621
05:01:28 22:01:28 ...Unbinned X: 3418 Y: 3179
05:01:28 22:01:28 Target star found
05:01:28 22:01:28 ** End Find Star **

So FocusMax AcquireStar is picking a star with a flux of 200000-ish, but then it shifts gears to the subframe and suddenly the star has a flux of 600000-ish! What happened? And that flux is double the 300000 flux target setting in FocusMax AcquireStar. I don't know why this is. I apologize but I'm just not that deep of a FocusMax expert.

[Dick Berg]

Hi Eric,

I notice in your log that at the start the Flux Target is set to 300000K, (not 300K) - mine at this point says 120K for comparison:

05:01:09 22:01:09 ** Beginning Find Star **
05:01:09 22:01:09 ...Filter 1: R (slot 2)
05:01:09 22:01:09 ...Flux target: 300000K
05:01:09 22:01:09 ...CCD central region: 75%
05:01:09 22:01:09 ...Star subframe width: 100

and again here on the second try:

05:03:52 22:03:52 ...Filter 1: R (slot 2)
05:03:52 22:03:52 ...Flux target: 300000K
05:03:52 22:03:52 ...CCD central region: 75%

I'm surprised about 300 million. Where would it have come from? Is there a setting in FMx where you might have put in 300,000, thinking the units were single ADUs instead of thousands of ADUs? It may mean nothing, but perhaps it's something you should look into. You'd need a long exposure of a bright start to get 300 million ADU. I don't know how AcquireStar works in FMx (I don't use that feature), but it may be what governs selecting the magnitude of the focus star. The 300 million ADU is set that way in the "Find Star" portion of FMx. For you at 7th mag it's clearly too bright.
A few seconds later, that value is restated as 300K:

05:01:30 22:01:30 Filter: R (slot 2)
05:01:30 22:01:30 ...Flux target: 300K
05:01:30 22:01:30 ...Focus method: Standard

and again:

05:04:11 22:04:11 Filter: R (slot 2)
05:04:11 22:04:11 ...Flux target: 300K
05:04:11 22:04:11 ...Focus method: Standard

Bob, the fluxes Eric measures are in the millions, not tens or hundreds of thousands. FMx picks a star in "Find Star" based on 300 million ADU. I suspect FMx plods on in any case, having picked the wrong star and sets the minimum exposure to the user's minimum time. It only fails the whole process at the end, when what I would call the confirming exposures are too short. Does this make sense to you?

[Eric Coles]

Dick,

I looked through the preferences in ACP and Focusmax and did not see the 300M setting. The "Use Focusmax Acquire star feature" is unchecked in the ACP Preferences under AutoFocus.

Eric

[Bob Denny]

The "Use Focusmax Acquire star feature" is unchecked in the ACP Preferences under AutoFocus.

Now that you mention that, I can see that it's not selected in your log. Look at 05:00:21 - 05:01:06 ... that's ACP's "dumb" focus star selection process. Some people still like it. AcquireStar is definitely off in ACP. FYI if you select AcquireStar in ACP, then ACP will use the star magnitudes in the FilterInfo list (where the focus offsets are also saved) and send them as "hints" to FocusMax AcquireStar. It's generally a better way to go as far as I am concerned. But I've left the older "dumb" logic in there because I don't want to force long time customers to give up old habits. Now they have a choice.

I looked just now and I can't see any way to force FocusMax to use AcquireStar. YOu have to explicitly clicjk that button in FocusMax, and you have to explicitly call a separate function from any outside program like ACP. But maybe there is and I don't know about it. Even more remote is the possibility that someone patched your AcquireSupport library script to force the call to AcquireStar??? Is that even possible?

What bothers me, given your specific problem, is the shift in flux from 200000-ish to 600000-isn going from full frame to subframe.

05:03:52 22:03:52 ** Beginning Find Star **
05:03:52 22:03:52 ...Filter 1: R (slot 2)
05:03:52 22:03:52 ...Flux target: 300000K
05:03:52 22:03:52 ...CCD central region: 75%
05:03:52 22:03:52 ...Star subframe width: 100
05:03:53 22:03:52 ...Base exposure: 0.50 sec
05:03:53 22:03:53 ...Min/Max exposure: 0.10 / 5.00 sec
05:03:53 22:03:53 ...Target star binning: 2
05:03:53 22:03:53 ...CCD central region: 75%
05:03:53 22:03:53 Taking initial image
05:04:01 22:04:01 ...Looking for star
05:04:02 22:04:02 Target star found
05:04:02 22:04:02 ...Exp: 0.50 Bin: 1 HFD: 10.29 Flux: 2005378
05:04:02 22:04:02 ...Unbinned X: 3420 Y: 3178
05:04:03 22:04:03 Setting subframe parameters
05:04:03 22:04:03 ...Focus binning: 1
05:04:03 22:04:03 ...Frame width: 100
05:04:03 22:04:03 ...Exposure: 0.50
05:04:03 22:04:03 Taking subframe image
05:04:08 22:04:07 Star X: 3420 Y: 3180 HFD: 7.58 Flux: 5705269
05:04:08 22:04:08 ** Star is too bright **



How can this be? Note that this is all contained within FocusMax, ACP has no control or influence at this point, it just told FocusMax "Focus and let me know when you are finished". FocusMax injects its log into ACPs (thank heaven so we can analyze).

I can't read all of your older logs to see if AcquireStar has been used in the past, for how long, who has access to change things on your system and when something might have been changed. So please look at past ACP run logs. Do you see

05:03:52 22:03:52 ** Beginning Find Star **

in any earlier ones? And how far back? If you see it start to be used say a month ago... then I'd go on the humto to find who changed FocusMax settinga a month ago.

[Eric Coles]

Bob,

Everything has worked as it should up to now. I have not made any changes in preferences. How should this be setup? I don't want to make an uninformed random decision that could make things worse. It should be setup the way you intended.

Eric

[Eric Coles]

BTW, I posted this issue on the CCDWARE forum and have not received any reply.

[Eric Dose]

Eric,

In your place, and assuming you do want ACP to use FocusMax's AcquireStar feature (as I do from ACP every night): I would test FocusMax AcquireStar without ACP from several places around the sky. Make sure it's a clear night so that the flux doesn't fluctuate. If you have problems, the FocusMax forum is much more likely to help you diagnose this.

My experience over 10 years is that proper FocusMax (including AcquireStar) settings, a good focuser, and a good sky will yield sharp autofocus every time, with or without ACP. There's not much chance ACP (or any other control program) can apply FocusMax AcquireStar if the latter isn't already very stable when run standalone.

EDIT: FocusMax writes its own log files which are highly detailed and very useful.