AstroEQ Discussion Board

AstroEQ R&D => Beta Firmware Releases => Topic started by: TCWORLD on October 16, 2017, 22:53:15

Title: [Completed] AstroEQ 8.11(beta) Release
Post by: TCWORLD on October 16, 2017, 22:53:15
This has now been released. Available from the downloads page.



Hi,

I'm preparing a new release of the AstroEQ firmware. Changes as follows:

Firmware:

Config Utility:


The beta version is attached to this post - both for Windows (AstroEQUploader3.8(beta).zip) and for Debian Linux (astroequploader-3.8.0.zip).

If you are not in the middle of Astronomy stuff and could spare a moment to test that the new firmware works correctly with your mount, that would be great. If all goes according to plan you shouldn't notice any difference with your AstroEQ.



Update (19/10/2017)

I've uploaded a new version (attached). Found a major mistake when tracking down issues with INDILib.



Note for downgrading

The V8.10 firmware uses a slightly modified EEPROM structure. The changes have been designed to be as minimal as possible (only to header). As such downgrading is possible.
When you downgrade with the current release version of the config utility (3.7.3), upload the firmware and then it will detect the EEPROM as corrupt. Continue to the next step which will attempt to repair it. Once this step is complete and you get to the configuration screen, close the configuration utility. The EEPROM structure will now be back to how it was as functional.

Title: Re: AstroEQ 8.10(beta) Release
Post by: aletel on October 18, 2017, 21:52:56
Hello.
I have tried to upload the beta version 8.10 with AstroEQ-ConfigUtility-LINUX and it seems that everything is fine except the one thing. The odd behavior of the motor in Declination. Changes of different speed modes (1x, 2x, 4x, ... 700x) are OK in different micro step setups (8/16/32), but during start of the pointing calibration, the DEC reacts to a short pulse by switching to full speed, so that the star jumps out of picture after the first iteration.
Similar to it the ST4 pulses for DEC goes at full speed. For this reason, I can not use correction for the DEC axis during photo session. This behavior is identical and tested on 3pcs of the Arduino Mega 2560 even on the recently purchased AstroEQ PCB + IC, DRV8825 (tested modified for smoother operation, even without modification). I notice this problem from the version 8.x. In previous versions of 7.x everything was fine and it worked entire season. I use AstroEQ for remote control of homemade mount in KStars configuration (INDI, EKOS, EQMOD on Linux, and Raspberry PI3) and I can not currently realize where the error is.

Thanks, Aletel
Title: Re: AstroEQ 8.10(beta) Release
Post by: TCWORLD on October 18, 2017, 22:50:08
@Aletel,

What happens if you try to slew DEC at a speed < 1x in EQMOD?
Title: Re: AstroEQ 8.10(beta) Release
Post by: aletel on October 18, 2017, 23:28:38
The RA engine at W / E Rate <1 (0.1, 0.25, 0.5, 0.75) responds correctly, DE motor N / S Rate runs briefly at full speed.

I am sorry for my English....
Title: Re: AstroEQ 8.10(beta) Release
Post by: TCWORLD on October 19, 2017, 09:38:01
Could you save your config from the utility and add here as an attachment to a post.

I can't seem to reproduce the problem with EQMOD.

Although given you mentioned Linux I presume you are trying with INDI? I'll have to set up something to test that.
Title: Re: AstroEQ 8.11(beta) Release
Post by: TCWORLD on October 19, 2017, 14:05:38
Right, so I've been trying to track down where the issues are arising with INDILib.

I believe I have found the culprit and corrected it. I've uploaded a new version to the first post of this thread.


For those that were interested, basically this has been an issue for 5 years. Yes, 5 years. Since AstroEQ first came in to existence. The information I was using to determine the correct response to the motor status report (:f command) was flawed. This meant that two of the status flags were swapped and inverted.

Interestingly because EQMOD doesn't take full notice of these flags, and the fact that the flags had inverted meaning it never caused a problem. INDILib on the other hand has over time become more responsive to these flags, and the mistake now causes it a world of confusion.

Problems that arise include:

Title: Re: AstroEQ 8.11(beta) Release
Post by: aletel on October 19, 2017, 17:01:26
Hi Tom,
(sorry translator)
I uploaded a new version of 8.11 (beta).

There is a change in behaviour, but there is still a problem with the DEC engine in star simulation calibration mode - setting the 5x iterative RA W / E perfectly. Attempting to calibrate the DE S / N, stepper motor is only one repetition (S) and rotates 360 quickly, then turn direction (N) again by 360 . In the simulator, the star returns to the point and the calibration is fine, but then the reaction star, the engine responds with too much correction (again the star outside the image).

I attach my screen with settings and .config.

Thank you Tom = Sherlock Holmes - for search, dedicated time and revealed case with INDI project.
Title: Re: AstroEQ 8.11(beta) Release
Post by: gehelem on October 19, 2017, 20:44:27
Hi

so i've made a quit test with my TMC2100
(OMG everybody should try them ! they are so smooth !!)
Uploader gives a few bugs (linux version)
when clicking on "update", error message :
Code: [Select]
------ DC Axis Update ------
Step Angle: 1.8
Motor Gear Ratio: 2.666666667
Worm Gear Ratio: 144
Oct 19, 2017 9:35:15 PM controlP5.ControlBroadcaster printMethodError
SEVERE: An error occured while forwarding a Controller event, please check your code at controlEvent
java.lang.reflect.InvocationTargetException
at sun.reflect.GeneratedMethodAccessor1.invoke(Unknown Source)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at controlP5.ControlBroadcaster.invokeMethod(Unknown Source)
at controlP5.ControlBroadcaster.broadcast(Unknown Source)
at controlP5.Controller.broadcast(Unknown Source)
at controlP5.Button.setValue(Unknown Source)
at controlP5.Button.activate(Unknown Source)
at controlP5.Button.mouseReleased(Unknown Source)
at controlP5.Controller.setMousePressed(Unknown Source)
at controlP5.ControllerGroup.setMousePressed(Unknown Source)
at controlP5.ControllerGroup.setMousePressed(Unknown Source)
at controlP5.ControlWindow.mouseReleasedEvent(Unknown Source)
at controlP5.ControlWindow.mouseEvent(Unknown Source)
at sun.reflect.GeneratedMethodAccessor2.invoke(Unknown Source)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at processing.core.PApplet$RegisteredMethods.handle(PApplet.java:1408)
at processing.core.PApplet.handleMethods(PApplet.java:1603)
at processing.core.PApplet.handleMouseEvent(PApplet.java:2695)
at processing.core.PApplet.dequeueEvents(PApplet.java:2618)
at processing.core.PApplet.handleDraw(PApplet.java:2429)
at processing.awt.PSurfaceAWT$12.callDraw(PSurfaceAWT.java:1540)
at processing.core.PSurfaceNone$AnimationThread.run(PSurfaceNone.java:316)
Caused by: java.lang.IndexOutOfBoundsException: Index: 4, Size: 4
at java.util.ArrayList.rangeCheck(ArrayList.java:653)
at java.util.ArrayList.get(ArrayList.java:429)
at controlP5.ScrollableList.getItem(Unknown Source)
at AstroEQUploader$GUIScreen.getScrollableListItem(AstroEQUploader.java:1737)
at AstroEQUploader$GUIScreen.getScrollableListItem(AstroEQUploader.java:1726)
at AstroEQUploader$GUIScreen.getScrollableListItem(AstroEQUploader.java:1721)
at AstroEQUploader$UploaderGUI.getMicrostepMode(AstroEQUploader.java:2263)
at AstroEQUploader$UploaderGUI.controlEvent(AstroEQUploader.java:2825)
at AstroEQUploader.controlEvent(AstroEQUploader.java:561)
... 24 more

I've tried without gearchanging option, it looks to work quite well
With option it's KO : motors are too slow...
Need more testing to give more feedback, of course

but thank you again for the good job

Gilles.
Title: Re: AstroEQ 8.11(beta) Release
Post by: gehelem on October 19, 2017, 20:56:00
...
i think my error on uploader was because of loading old config files...
With "legacy" ones, there is no such problem.
Title: Re: AstroEQ 8.11(beta) Release
Post by: TCWORLD on October 19, 2017, 21:05:23
Is there a specific value of microstep mode that causes the config utility to give that error?

I can work out where in the code it is crashing from the trace, but can't seem to figure out both why it would be doing it, nor how to replicate it.

I'll look in to the gearchanging thing again - seems to be something there that I can't spot. By "motors are too slow", how slow is too slow (roughly), and under what conditions? (high speed movements? low speed movements? go-tos?).
Title: Re: AstroEQ 8.11(beta) Release
Post by: gehelem on October 19, 2017, 21:31:54
uploader : no more errors since i'm using new config files
usteps = always 1/16
motors were too slow when doing a fast goto
i'm sure it's because my conf was not loaded properly
Let me do more tests

Gilles
Title: Re: AstroEQ 8.11(beta) Release
Post by: gehelem on October 19, 2017, 21:48:35
more details :
when i use TMC2100 / 16usteps with gear changing i think movements are 8x too small (might be 10x or 6x...)
Title: Re: AstroEQ 8.11(beta) Release
Post by: TCWORLD on October 19, 2017, 22:24:00
Do you have a multimeter? If so, could you measure the voltage on the mode select pins after performing a high speed movement. I want to make sure the mode select pins were set to the correct values.
Title: Re: AstroEQ 8.11(beta) Release
Post by: gehelem on October 20, 2017, 07:38:06
Sorry !
I've just realized i've removed config pins of drivers to act in shopper mode...
I will put them back tomorrow and try again
Gilles
Title: Re: AstroEQ 8.11(beta) Release
Post by: gehelem on October 20, 2017, 18:11:47
well :
config pins are back on the board (sorry again for that)
i still have this problem of "not enough move" when i set gear changing in eeprom

i went back to "no gear change"
movements seem ok
but i see no difference with my multimeter on config pins if motor are stopped or if they are x800
I read only a few mV

Title: Re: AstroEQ 8.11(beta) Release
Post by: gehelem on October 20, 2017, 18:49:17
new test :
gear changing is activated x16 / x2
During x800 goto, reading conf pins on TMC2100 gives :
CF1 ~ 1v
CF2 ~ 0v (need to confirm this one)
CF3 ~ 1v
During sideral tracking, i read
CF1 ~ 0v
CF2 ~ 300mv > ?? is this what they call "floating" ??
CF3 ~ 0v

Title: Re: AstroEQ 8.11(beta) Release
Post by: gehelem on October 20, 2017, 20:35:56
Annnnnnnnd ! Card burned...
i've made a shorcut between GND and motor output.
i think one of my driver is dead :( (and my arduino, too)
will have to wait a few days....

Title: Re: AstroEQ 8.11(beta) Release
Post by: TCWORLD on October 25, 2017, 20:17:28
The TMC's I ordered have just arrived and I see the problem you were having with microstep modes.

Haven't yet worked out what is happening though.

I'll keep searching.
Title: Re: AstroEQ 8.11(beta) Release
Post by: TCWORLD on October 25, 2017, 20:46:18
Well, found the problem.

For me it wasn't a firmware issue, but rather an issue with the driver board. It seems there are a million different clones of the TMC2100 and some (maybe many) of them didn't follow the original design.

There are two very small solder jumpers on the back of the driver boards labelled CFG1 and CFG2. These are the mode select pins. On the drivers I got, these two solder jumpers aren't shorted which means there is no connection from the mode select pins to the actual chip. Once I bridged the jumpers it, mode select appears to work.
Title: Re: AstroEQ 8.11(beta) Release
Post by: gehelem on October 29, 2017, 22:45:04
Hi,

this was really a good point to check :)
thank you for searching...
At the moment i'm still waiting my new TCM2100, they are somewhere on a boat, between here and there...

What do you think about these driver ? about awesome silence ?
They can be driven in chopper mode : don't know what it means but with that you can achieve really smooth moves
(but config pins must stay float)
Title: Re: AstroEQ 8.11(beta) Release
Post by: Jonesdee on October 31, 2017, 10:41:49
Hi Tom

(moved from 'Loss of COM Port' Thread): With DIY Mega 2560 and DRV8824 stepper drivers, slews initiated from Cartes du Ciel are very slow indeed and the motors are very noisy.  Slewing via EQMOD NESW controls is fine.but Spiral Search is also noisy.

Downgrading to 8.05 cures the problem.

Let me know if you need me to save any files or do any other testing

Cheers

Dave
Title: Re: AstroEQ 8.11(beta) Release
Post by: Jonesdee on November 05, 2017, 11:58:43
Update to my last post on slew speeds :

Using the EQ-5 Custom Pulley config, the Goto Rate multiplier usually appears as 808 in the config setup. When upgrading a Mega 2560 to 8.11 for the first time, opening the custom pulley config the correct multiplier appears.  If microstep gear changing is then disabled, the multiplier changes to 507. re-enabling Gear Changing then sets the multiplier to 504.  Loading the config file again brings 808 back.  This does not happen on susequent config changes (i.e. once 8.11 has been written to the board).

Similarly, when downgrading from 8.11 to 8.05, the first timethe config file is loaded, the goto multiplier shows 101. Loadingthe config file a second time corrects it to 808.  No other parameters change.

I also mentioned in an earlier post that upgrading took a very long time - this is because I was leaving my mouse pointer on the 'Next' button.  It seems that after some operations, the button does not function until the pointer is clicked elsewhere on the screen then returned to the button.  Maybe an issue with the touchpad on my laptop?

I should receive some 0.9 deg stepper motors this week, so will try 8.11 with TMC2100 driver boards.
Title: Re: AstroEQ 8.11(beta) Release
Post by: Jonesdee on November 19, 2017, 17:10:00
Hi Tom, when you have a moment (I know that you are very busy !), could you please let me know the make of your TC2100 drivers (if they are marked)?  I bought a pair which are labelled "BigTree Tech" and each behaves differently.  The first appears to run at the correct sidereal speed and slews quite slowly, while the second runs more than 4 times faster in both sidereal and slew modes and causes the motor to be a lot noisier.  My drivers are mounted in sockets, so are easy to change around and the issue definitely follows the driver board.  I can't see any physical differences between the two units (solder links etc).

Replacing them with 8825s and re configuring 8.11 accordingly fixes the issue and the controller runs normally (with the known skip while tracking).  I have not fitted the diode array to this controller as I was hoping the the TC2100s would fix the micro stepping problem.

I am testing with 10V 0.4A 0.9* resolution motors, running in 16 microstep mode on both drivers.

Cheers

Dave
Title: Re: AstroEQ 8.11(beta) Release
Post by: TCWORLD on November 19, 2017, 20:31:43
Hi Dave,

Mine say Big Tree Tech on them as well.

Make sure that the three red dotted jumpers are shorted (and no others):


Title: Re: AstroEQ 8.11(beta) Release
Post by: Jonesdee on November 19, 2017, 23:02:16
Hi Tom

Thanks very much for responding so quickly!

As far as I can tell (with my 70 year old eyes and a magnifying headset), these jumpers are not shorted on my boards - many thanks for the tip.  Just to be sure, I will get a second opinion in the morning then hit them with the soldering iron if my suspicions are correct ::).

Hope all is going well in the US and that you are enjoying there.

Cheers

Dave
Title: Re: AstroEQ 8.11(beta) Release
Post by: Jonesdee on November 20, 2017, 16:06:58
Hi again Tom

I tried linking the jumpers today, however on each board one solder pin of one of the pairs disappeared when I applied the iron (16 watt pointed tip).  Currently CFG1 remains open circuit on one board and CGF6 remains open on the other.  Strangely, this hasn't impacted the performance of either unit, one seems to run OK and the other still runs fast!  Anyway, I decided to make one last expenditure on this and have ordered a pair of 5v boards from the original supplier (Watterot) and these should arrive in the coming days.

Will let you know how these perform, as I understand that the required jumpers are connected by default on these units.

Rgds

Dave