Topics

MKS Gen-L V2.0 on G11 first light and HELP NEEDED

Jamie Flinn
 

Hi All…

 

Things went “ok” last night but had a number of problems that are a little deeper than the basic config:

 

  1. Setup config using the calculator and the basic G11 setting  (360 …1 in reduction = 12800 steps per rev using 32 microsteps and 16 for goto) – I was able to successfully align according the phone app but every goto on alignment and subsequent object gotos was out by about a telrad – not in the field of view (C11 + 0.5 reducer + 82 degree 24mm eyepice, so a lot of realestate in the view)
  2. Sidreal rate was consitantly fast and driving the stars steadily out – turn the unit OFF and normal drift happens the opposite direction – Lunar rate was closer but the hz faster/slower buttons in the app did not appear to do anything – is there somewhere in the code I can set the rate?...or a better way to fine tue it?

 

I suspect that microsteps are involved here – HOWARD….can you share to working config and the values you set……

 

Anyone with a G11 with Nema 17 0.96 direct I woild LOVE to see what you have

 

Thanks in advance

Jamie

Howard Dutton
 

On Thu, May 21, 2020 at 08:46 AM, Jamie Flinn wrote:
I suspect that microsteps are involved here – HOWARD….can you share to working config and the values you set……

Howard Dutton
 

On Thu, May 21, 2020 at 08:46 AM, Jamie Flinn wrote:
Sidreal rate was consitantly fast and driving the stars steadily out – turn the unit OFF and normal drift happens the opposite direction – Lunar rate was closer but the hz faster/slower buttons in the app did not appear to do anything – is there somewhere in the code I can set the rate?...or a better way to fine tue it?
Those buttons in the app make fine adjustments to the tracking rate... reset any adjustments to zero.  When setup properly you will most likely not need to adjust those at all.

Howard Dutton
 

Better to show me your Config.h file.

Jamie Flinn
 

Interesting - you are using SQUARE rather than PULSE - any reason?

On 2020-05-21 12:04 p.m., Howard Dutton wrote:
On Thu, May 21, 2020 at 08:46 AM, Jamie Flinn wrote:
I suspect that microsteps are involved here – HOWARD….can you share to working config and the values you set……

Howard Dutton
 
Edited

I don't turn PULSE on unless I need it.  For a Mega2560 you basically always need it for the 1.6x speed up, for a Teensy often you do not.

Jamie Flinn
 

Thanks for checking Howard...I also found my "observing site" in the phone app was setting to +79 rather than -79 when using the "use local" option - perhaps that is why I was out about a telrad distance...anyway, here is my config attached



---------- Original Message ----------
From: Howard Dutton <hjd1964@...>
Date: May 21, 2020 at 12:10 PM

Better to show me your Config.h file.


 

Howard Dutton
 

On Thu, May 21, 2020 at 09:36 AM, Jamie Flinn wrote:
I also found my "observing site" in the phone app was setting to +79 rather than -79
The usual "why doesn't it point right" stuff from the Wiki FAQ:
https://onstep.groups.io/g/main/wiki/7118

Howard Dutton
 

On Thu, May 21, 2020 at 09:36 AM, Jamie Flinn wrote:
Thanks for checking Howard...I also found my "observing site" in the phone app was setting to +79 rather than -79 when using the "use local" option - perhaps that is why I was out about a telrad distance...anyway, here is my config attached
I don't use _VQUIET since the motors have more trouble spinning fast in stealthChop mode.  We normally use _QUIET and yes that makes a bit more noise during the slews.

Nothing else jumps out at me as being wrong, but be aware the master branch doesn't enjoy the proven time and time again reliability of release-3.16.  In-fact the master has recently taken on a more experimental nature so I would suggest you continue the testing with release-3.16 for now to rule software out as a possible problem.

If the phone has GPS you can use that to fill the site info. in my App and thus avoid some mistakes.

Pointing with my G11 and even a single star align is within a wide field eyepiece provided it is well polar aligned (which it always is.)  More stars in the align improve things from there.  For getting started I suggest using a one star align since it's possible to misidentify alignment stars and corrupt an alignment.  Get the simple stuff working and build from there.

Jamie Flinn
 

Thanks....I will do that as my working (excellent) photo rig on EQ6 is using 3.16E and never liked 4.x....

Will set up and retest and report back finding :-)

Cheers

---------- Original Message ----------
From: Howard Dutton <hjd1964@...>
Date: May 21, 2020 at 1:02 PM

On Thu, May 21, 2020 at 09:36 AM, Jamie Flinn wrote:
Thanks for checking Howard...I also found my "observing site" in the phone app was setting to +79 rather than -79 when using the "use local" option - perhaps that is why I was out about a telrad distance...anyway, here is my config attached
I don't use _VQUIET since the motors have more trouble spinning fast in stealthChop mode.  We normally use _QUIET and yes that makes a bit more noise during the slews.

Nothing else jumps out at me as being wrong, but be aware the master branch doesn't enjoy the proven time and time again reliability of release-3.16.  In-fact the master has recently taken on a more experimental nature so I would suggest you continue the testing with release-3.16 for now to rule software out as a possible problem.

If the phone has GPS you can use that to fill the site info. in my App and thus avoid some mistakes.

Pointing with my G11 and even a single star align is within a wide field eyepiece provided it is well polar aligned (which it always is.)  More stars in the align improve things from there.  For getting started I suggest using a one star align since it's possible to misidentify alignment stars and corrupt an alignment.  Get the simple stuff working and build from there.


 

Howard Dutton
 

Wondering if maybe those fast (~1MHz) SPI conversations with the TMC2130 etc. aren't reliable in version 4.x... the slew kicks into 16x mode during the goto but never switches back to 32x at the end and hence your tracking is 2x too fast.

In version 4.x adding the following to Config.h would revert to using slow SPI similar to what is done in version 3.16 if you'd like to test that theory:
#define MODE_SWITCH_BEFORE_SLEW ON

At 2X too fast the tracking error is easy to see anytime, the G11's worm will rotate in 2 minutes instead of 4 minutes.

Jamie Flinn
 

I'll check it out and let you know...sounds reasonable

On 2020-05-21 2:06 p.m., Howard Dutton wrote:
Wondering if maybe those fast (~1MHz) SPI conversations with the TMC2130 etc. aren't reliable in version 4.x... the slew kicks into 16x mode during the goto but never switches back to 32x at the end and hence your tracking is 2x too fast.

In version 4.x adding the following to Config.h would revert to using slow SPI similar to what is done in version 3.16 if you'd like to test that theory:
#define MODE_SWITCH_BEFORE_SLEW ON

At 2X too fast the tracking error is easy to see anytime, the G11's worm will rotate in 2 minutes instead of 4 minutes.

Jamie Flinn
 

Hi Howard...

Better results with 3.16 and 3.15 e but still all versions are driving he  mount too fast:

1) tested _quiet, _vquiet, and just TMC2130 = same result

2) tested your settings from your config = slew from home was double or more out of position (vega is not in the south :-) ) - kept on PULSE however

3) tested weight shifts = same

4) tested power on but no align, just manually reefing the mount over to Vega = same rate of westward movement

5) no binding or skipping I can see

6) slew rate fast or slow made no difference - I had the mount slewing so fast it could take your arm off so power is not a problem

7) tested wall-wart 12v 3A  vs LifePo4 battery to eliminate noise and "sour power" = same result

The 12800 with 32 microsteps and 16 for slew did get me very close to alignment stars and I did do goto successfully to M13 and M92 - a few times the goto was in the field or the innner circle of telrad...then worked with the slow and fast button for tracking rate and nothing helped

I did not try using SQUARE....This is a head scratcher since 3.15e and 3.16 on my EQ6 can do a slew from home, then a platesolve once and have the object in the frame....on this G11 rig even the 3rd alignment star is not that close (I assume the process for a platesolve and an align are the same)

but no matter what I do the drive to the west...

This leave me with:

1) the TCM2130 themselves are weird

2) the steps per rotation is actually not 12800 for these motors (a match of your set in specs but not manufacturer)

3) perhaps this G11 does not actually have a 360 tooth gear????

Can you let me know where in the code the base rates are coming from - I will try some kludging to see if I can find

Help Obi-Wan...you are my only hope  (and get that image of me in a Princes Leia outfit out of your mind now!)

Thanks

Jamie



On 2020-05-21 1:02 p.m., Howard Dutton wrote:
On Thu, May 21, 2020 at 09:36 AM, Jamie Flinn wrote:
Thanks for checking Howard...I also found my "observing site" in the phone app was setting to +79 rather than -79 when using the "use local" option - perhaps that is why I was out about a telrad distance...anyway, here is my config attached
I don't use _VQUIET since the motors have more trouble spinning fast in stealthChop mode.  We normally use _QUIET and yes that makes a bit more noise during the slews.

Nothing else jumps out at me as being wrong, but be aware the master branch doesn't enjoy the proven time and time again reliability of release-3.16.  In-fact the master has recently taken on a more experimental nature so I would suggest you continue the testing with release-3.16 for now to rule software out as a possible problem.

If the phone has GPS you can use that to fill the site info. in my App and thus avoid some mistakes.

Pointing with my G11 and even a single star align is within a wide field eyepiece provided it is well polar aligned (which it always is.)  More stars in the align improve things from there.  For getting started I suggest using a one star align since it's possible to misidentify alignment stars and corrupt an alignment.  Get the simple stuff working and build from there.

Jamie Flinn
 

Addendum:

More tests this morning using the SUN as a target and it appears I have found the correct setting

1) I timed worm rotation/motor rotation - at 12800 the rotation is 2 min - at 6400 it is 4 minutes (correct for sidreal)

2) AXIS1_DRIVER_MICROSTEPS 16

3) AXIS2_DRIVER_MICROSTEPS 16

4) AXISx_DRIVER_MODEL  TMC2130_QUIET

With this combo I can go from HOME to the sun and (by using paralax of my finder rings) know I am dead on without even having to do align....I then timed the worm again and it is 4 minute (and about 10 seconds....rough marks with a sharpie as a guide so close enough.

Slews using 3.0 are crazy fast....had to tone it down to 1.5degrees/sec...This should be excellent for yet another clear night testing.

I do have one question I do not understand - microsteps are essentially a slice of the 360 degrees....why would they slow things down or result in slews that are off (more microsteps = more off) - is it not just a rapid fire energizing of the winding at a faster rate? 

Thanks





On 2020-05-22 6:49 a.m., Jamie Flinn wrote:

Hi Howard...

Better results with 3.16 and 3.15 e but still all versions are driving he  mount too fast:

1) tested _quiet, _vquiet, and just TMC2130 = same result

2) tested your settings from your config = slew from home was double or more out of position (vega is not in the south :-) ) - kept on PULSE however

3) tested weight shifts = same

4) tested power on but no align, just manually reefing the mount over to Vega = same rate of westward movement

5) no binding or skipping I can see

6) slew rate fast or slow made no difference - I had the mount slewing so fast it could take your arm off so power is not a problem

7) tested wall-wart 12v 3A  vs LifePo4 battery to eliminate noise and "sour power" = same result

The 12800 with 32 microsteps and 16 for slew did get me very close to alignment stars and I did do goto successfully to M13 and M92 - a few times the goto was in the field or the innner circle of telrad...then worked with the slow and fast button for tracking rate and nothing helped

I did not try using SQUARE....This is a head scratcher since 3.15e and 3.16 on my EQ6 can do a slew from home, then a platesolve once and have the object in the frame....on this G11 rig even the 3rd alignment star is not that close (I assume the process for a platesolve and an align are the same)

but no matter what I do the drive to the west...

This leave me with:

1) the TCM2130 themselves are weird

2) the steps per rotation is actually not 12800 for these motors (a match of your set in specs but not manufacturer)

3) perhaps this G11 does not actually have a 360 tooth gear????

Can you let me know where in the code the base rates are coming from - I will try some kludging to see if I can find

Help Obi-Wan...you are my only hope  (and get that image of me in a Princes Leia outfit out of your mind now!)

Thanks

Jamie



On 2020-05-21 1:02 p.m., Howard Dutton wrote:
On Thu, May 21, 2020 at 09:36 AM, Jamie Flinn wrote:
Thanks for checking Howard...I also found my "observing site" in the phone app was setting to +79 rather than -79 when using the "use local" option - perhaps that is why I was out about a telrad distance...anyway, here is my config attached
I don't use _VQUIET since the motors have more trouble spinning fast in stealthChop mode.  We normally use _QUIET and yes that makes a bit more noise during the slews.

Nothing else jumps out at me as being wrong, but be aware the master branch doesn't enjoy the proven time and time again reliability of release-3.16.  In-fact the master has recently taken on a more experimental nature so I would suggest you continue the testing with release-3.16 for now to rule software out as a possible problem.

If the phone has GPS you can use that to fill the site info. in my App and thus avoid some mistakes.

Pointing with my G11 and even a single star align is within a wide field eyepiece provided it is well polar aligned (which it always is.)  More stars in the align improve things from there.  For getting started I suggest using a one star align since it's possible to misidentify alignment stars and corrupt an alignment.  Get the simple stuff working and build from there.

Dave Schwartz
 

Good that you found the right parameters. Errors like that (too far, not far enough, too fast, too slow), after verifying correct location and date/time info, are always in the parameters dealing with motor steps, microsteps and gear ratios.

Microstepping is a bit more complex (or maybe I don't understand your understanding). Full stepping is done by just turning the current in a coil full on or full off - a square waveform. Microstepping changes this into a modified sine waveform (like the difference between a true sine wave and modified sine wave inverter). The sine wave is divided into multiple time slices (the number of slices = the microstepping level) and the current is controlled at each time slice so that the current at that point describes (approximately) a 'stepped' sine wave. It is the interplay of the current value, and thus the magnetic field strength, in the two coils that convinces the motor's rotor to be held between the locations it would naturally want to fall into due to the physical geometry and static magnetism of the rotor and stator.

So microstepping is a way of turning a 200 step motor into a 400, 800, 1600, 3200, 6400, etc motor, Although the accuracy of those positions between the 'natural' steps is not as precise, the motion becomes much smoother than the no-microstepping jump from one natural position to the next.

This article https://www.geckodrive.com/support/step-motor-basics.html describes so many of the things we talk about with stepper motors.

On 2020-05-22 10:45 a.m., Jamie Flinn wrote:

Addendum:

More tests this morning using the SUN as a target and it appears I have found the correct setting

1) I timed worm rotation/motor rotation - at 12800 the rotation is 2 min - at 6400 it is 4 minutes (correct for sidreal)

2) AXIS1_DRIVER_MICROSTEPS 16

3) AXIS2_DRIVER_MICROSTEPS 16

4) AXISx_DRIVER_MODEL  TMC2130_QUIET

With this combo I can go from HOME to the sun and (by using paralax of my finder rings) know I am dead on without even having to do align....I then timed the worm again and it is 4 minute (and about 10 seconds....rough marks with a sharpie as a guide so close enough.

Slews using 3.0 are crazy fast....had to tone it down to 1.5degrees/sec...This should be excellent for yet another clear night testing.

I do have one question I do not understand - microsteps are essentially a slice of the 360 degrees....why would they slow things down or result in slews that are off (more microsteps = more off) - is it not just a rapid fire energizing of the winding at a faster rate?

Thanks





On 2020-05-22 6:49 a.m., Jamie Flinn wrote:

Hi Howard...

Better results with 3.16 and 3.15 e but still all versions are driving he  mount too fast:

1) tested _quiet, _vquiet, and just TMC2130 = same result

2) tested your settings from your config = slew from home was double or more out of position (vega is not in the south :-) ) - kept on PULSE however

3) tested weight shifts = same

4) tested power on but no align, just manually reefing the mount over to Vega = same rate of westward movement

5) no binding or skipping I can see

6) slew rate fast or slow made no difference - I had the mount slewing so fast it could take your arm off so power is not a problem

7) tested wall-wart 12v 3A  vs LifePo4 battery to eliminate noise and "sour power" = same result

The 12800 with 32 microsteps and 16 for slew did get me very close to alignment stars and I did do goto successfully to M13 and M92 - a few times the goto was in the field or the innner circle of telrad...then worked with the slow and fast button for tracking rate and nothing helped

I did not try using SQUARE....This is a head scratcher since 3.15e and 3.16 on my EQ6 can do a slew from home, then a platesolve once and have the object in the frame....on this G11 rig even the 3rd alignment star is not that close (I assume the process for a platesolve and an align are the same)

but no matter what I do the drive to the west...

This leave me with:

1) the TCM2130 themselves are weird

2) the steps per rotation is actually not 12800 for these motors (a match of your set in specs but not manufacturer)

3) perhaps this G11 does not actually have a 360 tooth gear????

Can you let me know where in the code the base rates are coming from - I will try some kludging to see if I can find

Help Obi-Wan...you are my only hope  (and get that image of me in a Princes Leia outfit out of your mind now!)

Thanks

Jamie



On 2020-05-21 1:02 p.m., Howard Dutton wrote:
On Thu, May 21, 2020 at 09:36 AM, Jamie Flinn wrote:

Thanks for checking Howard...I also found my "observing site" in
the phone app was setting to +79 rather than -79 when using the
"use local" option - perhaps that is why I was out about a
telrad distance...anyway, here is my config attached

I don't use _VQUIET since the motors have more trouble spinning fast in stealthChop mode.  We normally use _QUIET and yes that makes a bit more noise during the slews.

Nothing else jumps out at me as being wrong, but be aware the *master* branch doesn't enjoy the proven time and time again reliability of *release-3.16*. In-fact the master has recently taken on a more experimental nature so I would suggest you continue the testing with release-3.16 for now to rule software out as a possible problem.

If the phone has GPS you can use that to fill the site info. in my App and thus avoid some mistakes.

Pointing with my G11 and even a single star align is within a wide field eyepiece provided it is well polar aligned (which it always is.)  More stars in the align improve things from there.  For getting started I suggest using a one star align since it's possible to misidentify alignment stars and corrupt an alignment.  Get the simple stuff working and build from there.

Jamie Flinn
 

Thanks!  good link

J

On 2020-05-22 12:06 p.m., Dave Schwartz wrote:
Good that you found the right parameters. Errors like that (too far, not far enough, too fast, too slow), after verifying correct location and date/time info, are always in the parameters dealing with motor steps, microsteps and gear ratios.

Microstepping is a bit more complex (or maybe I don't understand your understanding). Full stepping is done by just turning the current in a coil full on or full off - a square waveform. Microstepping changes this into a modified sine waveform (like the difference between a true sine wave and modified sine wave inverter). The sine wave is divided into multiple time slices (the number of slices = the microstepping level) and the current is controlled at each time slice so that the current at that point describes (approximately) a 'stepped' sine wave. It is the interplay of the current value, and thus the magnetic field strength, in the two coils that convinces the motor's rotor to be held between the locations it would naturally want to fall into due to the physical geometry and static magnetism of the rotor and stator.

So microstepping is a way of turning a 200 step motor into a 400, 800, 1600, 3200, 6400, etc motor, Although the accuracy of those positions between the 'natural' steps is not as precise, the motion becomes much smoother than the no-microstepping jump from one natural position to the next.

This article https://www.geckodrive.com/support/step-motor-basics.html describes so many of the things we talk about with stepper motors.

On 2020-05-22 10:45 a.m., Jamie Flinn wrote:

Addendum:

More tests this morning using the SUN as a target and it appears I have found the correct setting

1) I timed worm rotation/motor rotation - at 12800 the rotation is 2 min - at 6400 it is 4 minutes (correct for sidreal)

2) AXIS1_DRIVER_MICROSTEPS 16

3) AXIS2_DRIVER_MICROSTEPS 16

4) AXISx_DRIVER_MODEL  TMC2130_QUIET

With this combo I can go from HOME to the sun and (by using paralax of my finder rings) know I am dead on without even having to do align....I then timed the worm again and it is 4 minute (and about 10 seconds....rough marks with a sharpie as a guide so close enough.

Slews using 3.0 are crazy fast....had to tone it down to 1.5degrees/sec...This should be excellent for yet another clear night testing.

I do have one question I do not understand - microsteps are essentially a slice of the 360 degrees....why would they slow things down or result in slews that are off (more microsteps = more off) - is it not just a rapid fire energizing of the winding at a faster rate?

Thanks





On 2020-05-22 6:49 a.m., Jamie Flinn wrote:

Hi Howard...

Better results with 3.16 and 3.15 e but still all versions are driving he  mount too fast:

1) tested _quiet, _vquiet, and just TMC2130 = same result

2) tested your settings from your config = slew from home was double or more out of position (vega is not in the south :-) ) - kept on PULSE however

3) tested weight shifts = same

4) tested power on but no align, just manually reefing the mount over to Vega = same rate of westward movement

5) no binding or skipping I can see

6) slew rate fast or slow made no difference - I had the mount slewing so fast it could take your arm off so power is not a problem

7) tested wall-wart 12v 3A  vs LifePo4 battery to eliminate noise and "sour power" = same result

The 12800 with 32 microsteps and 16 for slew did get me very close to alignment stars and I did do goto successfully to M13 and M92 - a few times the goto was in the field or the innner circle of telrad...then worked with the slow and fast button for tracking rate and nothing helped

I did not try using SQUARE....This is a head scratcher since 3.15e and 3.16 on my EQ6 can do a slew from home, then a platesolve once and have the object in the frame....on this G11 rig even the 3rd alignment star is not that close (I assume the process for a platesolve and an align are the same)

but no matter what I do the drive to the west...

This leave me with:

1) the TCM2130 themselves are weird

2) the steps per rotation is actually not 12800 for these motors (a match of your set in specs but not manufacturer)

3) perhaps this G11 does not actually have a 360 tooth gear????

Can you let me know where in the code the base rates are coming from - I will try some kludging to see if I can find

Help Obi-Wan...you are my only hope  (and get that image of me in a Princes Leia outfit out of your mind now!)

Thanks

Jamie



On 2020-05-21 1:02 p.m., Howard Dutton wrote:
On Thu, May 21, 2020 at 09:36 AM, Jamie Flinn wrote:

    Thanks for checking Howard...I also found my "observing site" in
    the phone app was setting to +79 rather than -79 when using the
    "use local" option - perhaps that is why I was out about a
    telrad distance...anyway, here is my config attached

I don't use _VQUIET since the motors have more trouble spinning fast in stealthChop mode.  We normally use _QUIET and yes that makes a bit more noise during the slews.

Nothing else jumps out at me as being wrong, but be aware the *master* branch doesn't enjoy the proven time and time again reliability of *release-3.16*. In-fact the master has recently taken on a more experimental nature so I would suggest you continue the testing with release-3.16 for now to rule software out as a possible problem.

If the phone has GPS you can use that to fill the site info. in my App and thus avoid some mistakes.

Pointing with my G11 and even a single star align is within a wide field eyepiece provided it is well polar aligned (which it always is.)  More stars in the align improve things from there.  For getting started I suggest using a one star align since it's possible to misidentify alignment stars and corrupt an alignment.  Get the simple stuff working and build from there.