Help! Ekos Focus Failure

KenR · August 14, 2024, 12:14am

I’m hoping there is someone here who can help me figure out what’s going on with Ekos Linear 1-pass focusing or Ekos focus logic in general. I have an Arduino MyFocusPro2 updated to the latest firmware. This focuser works perfectly in SGP or just using it manually. I have backlash compensation disabled in the Indi driver.

My problem is Ekos L1P focusing seems to lose it’s mind and comes up with random solutions. It’s trying to take multiple exposures at the same time and seems to move it the wrong direction from time to time. Almost like there are multiple threads running in parallel both trying to focus. Also, the 2-pass linear focus has similar issues but it seems Ekos is able to recover reasonable well and doesn’t leave the focus point in a completely wrong location. I would really like to get the L1P process working. I’m sure it’s some sort of incompatibility between the MyFocusPro2 and Ekos. I would love to correct it but I’m not sure what’s triggering Ekos to go off the rails in a big way.

I have attached a log file of a typical L1P run. All my parameter settings are at the top of the log file. To give you an idea, I’ll paste just the start of the file below and add some annotations. Please review the full log and give me your insight.

Thanks!
Ken

<< Initial position [2024-08-12T21:10:30.099 Atlantic Daylight Time DEBG ][ << Ekos says initial [2024-08-12T21:10:30.099 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:30.099 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:30.100 Atlantic Daylight Time INFO ][ [2024-08-12T21:10:30.211 Atlantic Daylight Time INFO ][ << Ok 100 for overscan [2024-08-12T21:10:30.252 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:30.252 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:31.399 Atlantic Daylight Time DEBG ][ << It moved to 1601 [2024-08-12T21:10:31.400 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:31.400 Atlantic Daylight Time INFO ][ [2024-08-12T21:10:31.403 Atlantic Daylight Time INFO ][ [2024-08-12T21:10:31.457 Atlantic Daylight Time INFO ][ << Now take the 100 [2024-08-12T21:10:31.499 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:31.499 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:32.456 Atlantic Daylight Time DEBG ][ << It’s moving [2024-08-12T21:10:32.457 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:32.458 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:33.605 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:33.606 Atlantic Daylight Time DEBG ][ << Ah ha! Something [2024-08-12T21:10:33.653 Atlantic Daylight Time INFO ][ << Focuser has reached [2024-08-12T21:10:33.791 Atlantic Daylight Time INFO ][ [2024-08-12T21:10:34.595 Atlantic Daylight Time INFO ][ [2024-08-12T21:10:34.620 Atlantic Daylight Time INFO ][ [2024-08-12T21:10:34.705 Atlantic Daylight Time DEBG ][ << Now it finally [2024-08-12T21:10:34.706 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:34.706 Atlantic Daylight Time DEBG ][ [2024-08-12T21:10:34.706 Atlantic Daylight Time WARN ][ << Finally, it again 1600. I know this for a fact as this is my normal starting point very close to focus >>
org.kde.kstars.ekos.focus] - “Linear: Travel 500 initStep 20 pos 1600 min 0 max 15000 maxIters 51 tolerance 0.05 minlimit 1100 maxlimit 2100 init#steps 5 algo 3 backlash 0 curvefit 1 useweights 0”
position is 1700? Ok? >>
org.kde.kstars.ekos.focus] - “Linear: initialPosition 1700 sized 20”
org.kde.kstars.ekos.focus] - “Linear: extending outward movement by overscan 100”
org.kde.kstars.ekos.focus] - “Focusing outward by 200 steps…”
org.kde.kstars.indi] - MyFocuserPro2 : "[INFO] Focuser is moving to position 1800 "
(not sure where to set this) and 100 for 5 * 20 steps is 200. Move to out to 1800. Perfect. >>
org.kde.kstars.ekos.focus] - Restarting focus motion timer, state Busy
org.kde.kstars.ekos.focus] - “autoFocusProcessPositionChange called with state 2 (Busy), focuserAdditionalMovement=100, inAutoFocus=1, m_captureInProgress=0, currentPosition=1600”
org.kde.kstars.ekos.focus] - Abs Focuser position changed to 1601 State: Ok
and stopped? State = Ok. It should be moving to 1800 as is logged above. This seems to be the major issue. It knows it’s looking for 1800, why stop just 1 count from the start? Right here Ekos could have detected the error but it’s downhill from here. >>
org.kde.kstars.ekos.focus] - “Undoing overscan extension. Moving back in by 100 ticks in 0s”
org.kde.kstars.ekos.focus] - “Focusing inward by 100 steps…”
org.kde.kstars.indi] - MyFocuserPro2 : "[INFO] Focuser reached requested position. "
org.kde.kstars.indi] - MyFocuserPro2 : "[INFO] Focuser is moving to position 1501 "
overscan away and we’re at 1501. It should be looking for 1700 here. 1600 + 200 - 100 >>
org.kde.kstars.ekos.focus] - Restarting focus motion timer, state Busy
org.kde.kstars.ekos.focus] - “autoFocusProcessPositionChange called with state 2 (Busy), focuserAdditionalMovement=0, inAutoFocus=1, m_captureInProgress=0, currentPosition=1601”
org.kde.kstars.ekos.focus] - Abs Focuser position changed to 1565 State: Busy
in to 1501 apparently. Here it’s at 1565 and still busy >>
org.kde.kstars.ekos.focus] - Restarting focus motion timer, state Busy
org.kde.kstars.ekos.focus] - “autoFocusProcessPositionChange called with state 2 (Busy), focuserAdditionalMovement=0, inAutoFocus=1, m_captureInProgress=0, currentPosition=1565”
org.kde.kstars.ekos.focus] - “Focus positioning error: requested position 1700, current position 1565”
org.kde.kstars.ekos.focus] - “Focus position reached at 1565, starting capture in 1 seconds.”
knows it should be at 1700 but it reached 1565 and is still in busy state so it’s still on it’s way to 1501 . Regardless, it decides to take an image. >>
org.kde.kstars.indi] - MyFocuserPro2 : "[INFO] Focuser reached requested position. "
requested position. I assume this is 1501? Not sure why it thinks that’s the requested position. Should be 1700. >>
org.kde.kstars.indi] - SBIG CCD : "[INFO] Temperature set to +4.4C "
org.kde.kstars.ekos.focus] - “Capturing image…”
org.kde.kstars.indi] - SBIG CCD : "[INFO] Taking 5.00s exposure on main camera… "
org.kde.kstars.ekos.focus] - Abs Focuser position changed to 1501 State: Ok
reaches 1501 for some reason and it’s already started taking a photo. >>
org.kde.kstars.ekos.focus] - “Focus positioning error: requested position 1700, current position 1501”
org.kde.kstars.ekos.focus] - “Focus position reached at 1501, starting capture in 1 seconds.”
org.kde.kstars.ekos.focus] - Capture called while already in progress. Capture is ignored.
says it’s not at 1700 but has reached 1501 and tries to take another photo while the first one is still in progress. >>

EkosFocusIssueL1P.txt (47.4 KB)

KenR · August 14, 2024, 12:20am

I should also mention that at the end of the focus run it selected these points for the curve which do not match the actual focuser position. Then settled on 1527 as the focus point which is completely out of focus. It appears to be using “theoretical” positions rather than the actual position reported by the focuser. I’m pretty sure I’m dealing with some sort of race condition but where to start looking?

Linear: points=[(1700, 0.991997, 1), (1680, 0.958589, 1), (1660, 0.890479, 1), (1640, 0.938845, 1), (1620, 1.08277, 1), (1600, 1.15129, 1), (1580, 1.43183, 1), (1560, 1.70612, 1), (1540, 2.05459, 1), (1520, 2.36188, 1)]

Thanks

johneevans · August 14, 2024, 7:00am

Hi Ken,

There have been issues with the MyFocuserPro2 in the past. If you search the forum you’ll see some threads. Here’s one for example:
https://www.indilib.org/forum/general/13289-linear-focusing-strangenss-again.html?start=0#92263

What I suggest is:

Make sure you are on the latest MyFocuserPro2 firmware. This has fixed issues in the past.
If 1. does not fix things then set AF Overscan = 0 and see if that fixes the issue. If it does then we may be able to optimise things a bit better but lets see if that fixes the issue first.

Hope this helps.

KenR · August 14, 2024, 9:20am

Thanks John. I am on the latest firmware but I did not try the AF Overscan setting. I will try that and report back.

Your analysis of the log file in the thread you provided is exactly what I am seeing. Since this spans so many firmware revisions I’m thinking the issue may be in the Indi driver which hasn’t been updated in a very long time - if at all.

Ken

[quote]
Hi Ken,

There have been issues with the MyFocuserPro2 in the past. If you search the forum you’ll see some threads. Here’s one for example:
https://www.indilib.org/forum/general/13289-linear-focusing-strangenss-again.html?start=0#92263

What I suggest is:

Make sure you are on the latest MyFocuserPro2 firmware. This has fixed issues in the past.
If 1. does not fix things then set AF Overscan = 0 and see if that fixes the issue. If it does then we may be able to optimise things a bit better but lets see if that fixes the issue first.

Hope this helps.[/quote]

KenR · August 17, 2024, 7:51pm

I fixed it. See here for details.

myFocuserPro2 Driver Fix for Ekos - INDI Forum (indilib.org)