Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Myscope4960 pec editor for LX85

Equipment Software
  • Please log in to reply
54 replies to this topic

#1 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 28 March 2021 - 08:29 AM

Hi All,

 

I've been trying to check out my PEC curves after performing a few PEC training runs on a LX85, which is has been starpatched to A4S4 firmware. Is there a way to view this in the Myscope4960 app created by Andrew Johansen? I just get an error "Raw PEC contains possibly woofy data". Everything else sometime work fine apart from the PEC bit.



#2 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 28 March 2021 - 05:05 PM

Gday Damo

You cant use 4960 with A4S4 as Meade totally rewrote the way PEC works and is stored.

I have a beta version that will work with the new firmwares so will load that later today.

 

Andrew Johansen Melbourne Australia



#3 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 28 March 2021 - 06:23 PM

Gday Damo

New beta version up on my site that should work with your LX85.

Let me know how it goes.

Andrew Johansen Melbourne Australia



#4 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 29 March 2021 - 10:35 AM

Thanks Andrew,

 

I gave myscope4993 ago unfortunately I get an error when i try to run the "read data from 497 HBx" in the PEC utils. A popup box appears with "Invalid argument to date encode". 

 

I did semi brick this audiostar a while back using the Meade ASU and managed to sort it using your AudioStarFixer tool so i don't know if the Meade ASU could have screwed something else up.

 

Damo



#5 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 29 March 2021 - 05:11 PM

Gday Damo

OK, when it barfs, can you then go to the logging tab, do a refresh, then save the log.

That might give me a hint, as i did test it with A4S4 but not all the possible combinations.

 

Andrew Johansen Melbourne Australia



#6 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 30 March 2021 - 01:34 AM

Gday Damo

OK, found the bug. I changed the file save format but didnt retest it in all modules.

Try again and see what happens

 

Andrew Johansen Melbourne Australia



#7 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 30 March 2021 - 02:31 AM

Cheers Andrew

 

Looks like everything is working now not seeing any error messages :) and it also generates a SCOPEDIRECT.rom file which didn't happen before, Interestingly the PEC data points are all zero which probably means my PEC training sessions haven't worked. So I'll give it another go tonight (weather permitting) and see if I can get any data to populate the PEC table.  

 

Now at least I can see what's going on :)

 

Thanks again for your help, much appreciated 



#8 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 30 March 2021 - 02:44 AM

Gday Damo

So I'll give it another go tonight (weather permitting) and see if I can get any data to populate the PEC table.

Just do a test indoors in front of the telly.

Ie start it tracking like normal, then start a PEC train and use the hbx keys

( or a scripted feed ) to blindly fake a PEC train, then read back the data.

I am not sure how it will really work, as the LX85 is the first Audiostar type mount

that has a PEC sensor to keep the model synchronised across boots.

IIRC some earlier firmwares simply deleted the PEC model on booting ( vs unparking )

to prevent a desynchronised model playing back, but i havent looked at the A4S4 code in detail there.

as I cant test it properly as i dont have that sensor on any of my setups

and i dont have an LX85 board to use as a testbench.

Andrew Johansen Melbourne Australia



#9 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 30 March 2021 - 08:35 AM

I couldn't wait until tonight so I gave it another ago at lunch using the keypad during a PEC train as suggested and it works!!waytogo.gif. Still a mystery why the previous training session with the auto guider didn't populate the table, now at least I have the means to trouble shoot.

 

Also, I did a reboot and the PEC data is retained so it must maintain the index using the PEC sensor scan during start-up.

 

Thanks again for your help

 

Damo



#10 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 30 March 2021 - 03:49 PM

I've managed to get a few training runs in before the clouds set in with some interesting results. So if I run a training session using the Autoguider (PHD), nothing shows up in the pec table just zeros. I've also had a ago doing it the old school way manually correcting a guide star using the key pad with the same result. 

 

It seems I can only get numbers generated in the PEC table if I make very large adjustments using the keypad. Could this be just be a number scaling issue in the table? 

 

I can confirm the PEC does get applied, as once I switch on PEC with my keypad mashing PEC table the mount jumps around accordingly smile.gif.

 

Below are a few lines from the PEC txt file, definitely not all zero but small never the less

 

 

Recno Plot_0_X, Plot_0_Y, Plot_2_X, Plot_2_Y, Plot_3_X, Plot_3_Y, Plot_4_X, Plot_4_Y, Plot_5_X, Plot_5_Y, Plot_6_X, Plot_6_Y, Plot_7_X, Plot_7_Y, Plot_8_X, Plot_8_Y,
-----------------------------------------------------------------
0000 0  -0.0066 -0.5 0      0  0 -0.5 0 -0.5 0 0  0 -0.5 0 -0.5 0
0001 1  -0.0066 0.5  0.0001 1  0 0.5  0 0.5  0 1  0 0.5  0 0.5  0
0002 2  -0.0066 1.5  0.0002 2  0 1.5  0 1.5  0 2  0 1.5  0 1.5  0
0003 3  -0.0066 2.5  0.0004 3  0 2.5  0 2.5  0 3  0 2.5  0 2.5  0
0004 4  -0.0066 3.5  0.0005 4  0 3.5  0 3.5  0 4  0 3.5  0 3.5  0
0005 5  -0.0066 4.5  0.0006 5  0 4.5  0 4.5  0 5  0 4.5  0 4.5  0
0006 6  -0.0066 5.5  0.0007 6  0 5.5  0 5.5  0 6  0 5.5  0 5.5  0
0007 7  -0.0066 6.5  0.0009 7  0 6.5  0 6.5  0 7  0 6.5  0 6.5  0
0008 8  -0.0066 7.5  0.001  8  0 7.5  0 7.5  0 8  0 7.5  0 7.5  0



#11 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 30 March 2021 - 04:03 PM

Gday Damo

Can you connect to your mount using my app

Once connected, save the log file.

Then zip and post the log + scopedirect.pec file here.

 

It might be a scaling problem, as Meade have massively changed the

way they do PEC in the later Audiostar firmwares,

The old method stored data as 150 single byte values

where #x80 = 0, #81 = +1, #x7F = -1

The new system uses true 2 byte integers,

and uses variable data lengths based on mount type.

I thought i had that covered, but may have missed something

 

Andrew



#12 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 30 March 2021 - 04:43 PM

Gday Damo

I just redid a dummy ( scripted ) PEC test here and it worked happily ( as an LX90 ).

With the new format, the raw data is massive relative to the PEC so what you need to do

to see the "PEC as arcsecs " is to heavily zoom in on the Y axis

or turn off the raw data trace and zoom to fit.

ie in the old system a raw value of +1 might now be +200

 

Only other thing i can think of is you had PHD set to measure but not send guide pulses????

 

Andrew Johansen Melbourne Australia



#13 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 31 March 2021 - 01:24 AM

Gday Damo

It seems I can only get numbers generated in the PEC table if I make very large adjustments

Just for fun i retested with a script that sent pulses every 8.3 seconds

ie should be one adjust per bin.  I used sets of 4 ie E,E,W,W

and I used a range of 50, 100, 150 and 200ms pulses.

All showed up clearly in the data.

I have to say the mount didnt react anywhere near as accurately as i would have expected

to the script, but it did react and store data down to 50ms.

Andrew Johansen Melbourne Australia 



#14 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 31 March 2021 - 09:48 AM

Thanks again Andrew,
 
I've had another look at the plots and I'm still struggling to get it to show me anything. I've attached a log file + ROM and all the pec data I could find.
 
I double checked the PHD setup and it is using pulse guiding to send corrections to the handset, during the PEC training run I was also monitoring the PHD guiding and I could see the corrections being applied and the response of the guide star. So I'm pretty confident that side of things is working ok.
 
I've also run PHD in manual mode and issued large manual corrections (5000ms) during a training run and I still cant get anything to show up in PEC data.
 
I'll give it another ago tonight and see if I can get anything different.  Also do you think it makes a difference if I guide using ST4 or Ascom? so far I've only used Ascom
 
Cheers
 
Damo

Attached Files


Edited by Damo19, 31 March 2021 - 09:49 AM.


#15 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 31 March 2021 - 04:14 PM

Gday Damo

 

I've had another look at the plots and I'm still struggling to get it to show me anything.

Thats cos there is nothing there.

In reality, when testing you only need to use traces 6 and 8

ie Raw and Raw as arcsecs

 

I was also monitoring the PHD guiding and I could see the corrections being applied and the response of the guide star.

OK, in that case, it should be sending real pulses

 

I've also run PHD in manual mode and issued large manual corrections (5000ms) during a training run and I still cant get anything to show up in PEC data.

What speed was the handbox set to????

PEC will only get registered if the handbox is set to speed 1.

The fact you got a model to work yesterday at lunch indicates it did work

so something else is going on.

When using PHD, were you using "Pulse Guiding" or default guiding

( as they use different commands that behave differently )

What ASCOM driver are you using??? ( as old drivers may not know the LX85 )

If you set serial tracing ON in the driver, you can then view the ASCOM log

to see what commands are getting sent.

Pulseguides will be like :Mgxdddd#   where x = n,s,e,w and dddd = millisecs

Normal "moves" will be :Mx# followed by :Qx#

 

Also do you think it makes a difference if I guide using ST4 or Ascom?

As per above, ASCOM can actually cover 2 modes.

ST4 should always work, but can hog CPU time

 

Just for fun, and as an indoor test

Start your mount and get it dummy aligned

Connect my app.

Go to the "Misc Utils" tab and click the "Slew/Focus Test" option

A floating dialog will appear, and in the middle is a "pulseguide" section.

Set the rate to say 200 first

Start a PEC train on the mount and once running, use the E or W direction keys

to send pulses to the mount as each bin changes

After a few at 200, set it to 100 and repeat then set it to 50 and repeat.

When done, reread the PEC from the mount and see if its there.

Also, save the log at that point.

If there is no model stored, send me the log.

 

Andrew Johansen Melbourne Australia



#16 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 01 April 2021 - 12:22 PM

Hi Andrew,

 

Had ago with the guide pulse test using your App still no luck (logs attached). I also had a try using the ST4 port to see if that would make any difference and unfortunately that turned out the same. In both cases I set the handbox to guide speed (key 1).

 

I'm starting to suspect the data is not being recorded in the Audiostar. So I'm going to run some tests using PHD to measure the PE with and without PEC switched on (after doing a couple of ST4 guided training runs). At least then I'll now for sure if there is a PEC table to read in the first place.

 

Cheers again

 

Damo

 

 

 

 

 

 

Attached Files



#17 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 01 April 2021 - 04:48 PM

Gday Damo

Had ago with the guide pulse test using your App still no luck

You didnt appear to do any West pulses??? but that said, the normalising routine should have fixed that.

 

I'm starting to suspect the data is not being recorded in the Audiostar.

Correct, your download after the key pressing episode didnt show a model being stored

ie No need to run PHD to see if there is model there, its all zero.

 

One last test before i give up and read the code to see if they have done something specific

Tell your mount it is an LXD75, then set it to polar and astronomical.

Clear the PEC model using the Hbx and start a train.

Use the pulseguide panel in my app to send some east and west pulses then see if something gets saved.

 

Andrew



#18 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 01 April 2021 - 10:13 PM

Gday Damo

OK, i gave up and started reading the code.

It is very early days but it looks like they forgot to allow for the LX85 in the training code.

There is a std ( hardcoded ) variable that stores the encoder ticks per bin for various models.

There is a value there that suits the LX85, but it doesnt appear to be allocated.

As such, it looks like the code still does the train, but uses an uninitialised variable on the stack

to convert the raw pulse data to arcsecs :-)

Need to do a LOT more decoding to confirm it, but try the LXD75 mode test.

It should store a model

 

Andrew



#19 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 02 April 2021 - 06:21 AM

Hi Andrew,

 

Looks like you have cracked it! if I run the training session with the AudioStar set to LXD75 the PEC model gets created no problem bow.gif.  

 

So I guess I could still upload a PEC model to the LX85 using your App if I create one using say PEMPro as the issue only seems to be related to recording the PE rather than applying it (hopefully anyway). 

 

Thanks again for all your help with this, its good to finally get to the bottom of what was going on.

 

Damo



#20 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 02 April 2021 - 02:20 PM

Gday Damo

So I guess I could still upload a PEC model to the LX85 using your App

Nahh. Should be easy to patch it for a specific LX85 model

 

its good to finally get to the bottom of what was going on

What amazes me is why no one else has reported it yet.

I knew the PEC "structure" had changed based on a casual code review,

but it worked on my Test Bench (LX90/ETX125), even tho it gave slightly odd results.

What i can now see is it appears they have also changed the training code

to use time per bin ( vs time slew key was pressed ), and that would explain

why i get such irregular results.

More in a day or so till i see what else they have screwed up.

 

Andrew Johansen Melbourne Australia



#21 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 03 April 2021 - 07:59 PM

Gday Damo
 

it appears they have also changed the training code

to use time per bin ( vs time slew key was pressed ), and that would explain

why i get such irregular results.

And it does.

Just finished reading the code and confirmed they have changed to a time based calculation,

vs the old method of using the summary time a key was pressed.

With the Audiostar, they can do this as the encoder reading is much faster

then the motorola based units  ( New units read at up to 30x per second )

That said, in earlier firmware, they also kept a tab of if any guide commands came in within the bin.

If not, they just jump, but now they dont check, they just process.

As such, timing errors in the bin trip are now recorded as PE

I confirmed this by doing a PEC train where i used no inputs, just let it run til finish.

You get a really interesting model that way  :-)

 

If you want to get your mount to do PEC as an LX85,

can you confirm which model  LX85 you are selecting,

( Your log says "LX85-8 SCT" )

If confirmed, i can send you a patchette to test with

( just add it to the existing patch and reload )

 

Andrew Johansen Melbourne Australia



#22 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 04 April 2021 - 02:35 PM

Hi Andrew,

 

Wow that was quick :) sounds like the changes for the LX85 were more than just the addition of a worm encoder, I guess they also wanted to make use of the extra processing power, or not as the case may belol.gif. With regards to the LX85 model the log is correct,  its the Meade 8" SCT. When I was trouble shooting I did go through the other LX85 modes just in case one of them worked and it look like the only thing changing was the focal length but I could be wrong. 

 

If you have a patchette, I can give it ago and report back. If things look promising I'll run a back to back test with and without PEC enabled to compare the RA trace.

 

Damo



#23 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 04 April 2021 - 05:32 PM

Gday Damo

sounds like the changes for the LX85 were more than just the addition of a worm encoder,

Nope. The worm encoder is specific for the LX85, but the other changes now apply to

all models that can do PEC. ( As do the motor tracking changes )

 

I guess they also wanted to make use of the extra processing power,

No, or they would have done it when they brought in the new processor.

Meade changed the method and storage mechanism for PEC when they brought in Starlock

on the LX600s. My guess is whoever did the fix for that simply copied it over into the Audiostars

to ensure both use the same algorithms. Just as per normal, they dont fully understand their code

and forget to do stuff :-)

 

If you have a patchette, I can give it ago and report back.

OK, will knock up a few lines you can paste into your existing patch and try.

More in a few hours.

 

Andrew Johansen Melbourne Australia



#24 OzAndrewJ

OzAndrewJ

    Aurora

  • *****
  • Posts: 4,730
  • Joined: 30 Nov 2010

Posted 04 April 2021 - 10:34 PM

Gday Damo

OK, have attached the patchette.

Just cutnpaste the red bits into the master patch and save

( Its a text file so can open with any normal text editor )

 

I tested on my testbed with it set as an LX85 8"SCT and didnt send any pulses

and got a screen full of sawtooths :-)

See how it goes

Andrew Johansen Melbourne Australia

 

;//================================================================================================
;// Below here are Bug Fixes
[-------------   Bug Fixes   -------------]
S=UnChecked
> ----------------------------------------------------------
;//================================================================================================
[Allow PEC for LX85 8"SCT]
S=Checked
> Allows LX85 model to do PEC training
;// F8-B02E model selection
R 0D6C30  41 56     // change VM1 to LX85-8 SCT
R 0D6C33  7F 26     // change if equal goto xF8-B0B3, to "goto xF8-B05A" and test there jumps on

;//------------------------------------------------------------------------------------------------
[Fix Drive Train / Cal Sensor Resets ]           // *** Tested OK ***



#25 Damo19

Damo19

    Lift Off

  • -----
  • topic starter
  • Posts: 19
  • Joined: 28 Mar 2021

Posted 05 April 2021 - 06:14 AM

Hi Andrew,

 

I've applied the patch and run a dummy RA train with some pulse inputs and its looking good so far! Also the model name entered in Audiostar is "LX85 8 ACF" which is slightly different to what was reported in the log. Either way it seems to work fine from the test run results.

 

I'm all setup now so if I get lucky with the weather I'll do some proper PE training runs using PHD etc. Its going to be really interesting to see how much the PEC straightens things out.

 

Once again thanks for sorting this out bow.gif  I think it would have been a really long wait for Meade to resolve this themselves given how many firmware updates they put out.

 

Cheers

 

Damo




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: Equipment, Software



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics