ShapeOko serial number 399 says, "Hello ... nobody :("

[tioJim]

Ok. Houston we have a problem. ShapeOko missing in action. Man down. Etc etc.

I won't go into detail yet but likely culprits are

1. GRBL settings
2. My homebrew shield

Main symptoms so far are

1. When a stepper motor does respond it just goes back and forth a single step.
2. Think I've killed two A4988's Although I've been super careful with them.

So ..

To eliminate as many variables as possible I'm going to go back to basics and breadboard up a single driver to a single motor/axis free of any load (including belt and gantry)

Main question for now is ... is this a valid approach? I'm worried that there's voodoo going on like the software or electronics freaks out if there's only one axis or correct operation is dependent on load, other electronic properties of the circuit and so on.

If my Arduino flashed with GRBL can make a stepper do Y20 then Y-20 or whatever via gcodesender then I can progress from there.

So, breadboarded driver, unloaded single axis test, good way to proceed yeah?

[edwardrford]

Judging by your symptoms, I would say the culprit is not grbl settings. Although if you posted your settings here in a code block it would be helpful.

From the sounds of it, you either do not have the stepper motor wired correctly to the driver or more than likely there is an issue with your home brew shield.

Pictures would be helpful as well.

Let's figure this out!

-Edward

[Will Winder]

Did you fix your stepper motor connectors from this thread? viewtopic.php?f=4&t=579

I made the same mistake and saw the same behavior, motors *sound* like they are moving but actually only jitter back and forth.

Based on your earlier description they should be hooked up like:
BLK A = 1A
GRN A_ = 1B
RED B = 2A
BLU B_ = 2B

(you might need to swap 2A/2B on some axis later on to get it to spin the right direction)

[tioJim]

Ok, progress!

Yes I still had my stepper wiring mixed up.
Yes I have bricked two A4988 boards (d'oh, it's ok they died for science, a moment's silence is in order ... thank you, more coming in the post tomorrow)

As for my home-brew board, good and bad. The Z axis seems dead but the X and Y axis live although there seems to be excessive heat on the chips. A short somewhere? I think maybe something's wrong with the enable pin, R10K, VDD wiring also.

Anyway, I did build a single axis controller on a breadboard and that's what I've been testing with and hey presto X50 sends the motor spinning CW and X0 sends it back again CCW from whence it came.

Previously they wouldn't turn, they'd just jog and never finish a command. Also now I can't turn them with my fingers when they're powered up, before I could.

But I'm not happy yet as ... with the belt on nothing moves . It just whines. Is this what is known as stalling?

I need to see a gantry moving before I can sleep!

So I guess you need to know my setup ...

Turns out I wasn't using the revision 3 schematic for my home-brew of the Buildlog.net stepper shield I was using revision 2 so I had to modify and recompile GRBL to account for the different pin config. I've followed the same pin config for my breadboard circuit so I can leave GRBL alone for now. As I can get a stepper to stop/start, go CW and CCW I figure my compiled version of GRBL is Ok. I used the mods to Config.h and Stepper.c detailed in the WiKi somewhere - edited in Xcode and compiled using avr-gcc. Anyway ...

GRBL 0.7d with pin config modified to match earlier versions of Buildlog.net shield. Set as follows ... (again they're got from the Wiki somewhere! You know how easy it is to lose track of the decisions you make/info you use when tinkering!)

Code: Select all

$0=87.489
$1=87.489
$2=320
$3=30
$4=750
$5=1100
$6=0.1
$7=0
$8=9.8
$9=0.05


I tried all permutations of the micro-stepping jumpers. Some were Ok, some killed it, some caused juddering. I'm not sure on the right setting here. Here are my stepper motors http://www.technobotsonline.com/stepper ... s-rev.html. My drivers are the Pololu A4988 breakout boards. My belts on the mechanicals have plenty of play and the gantry moves freely and smoothly.

Things I will do next or haven't done yet

Set current limit on A4988 boards (looks fiddly!)
Used a GRBL calculator for my setup to get GRBL settings
Shelve my home-brew board for now and extend breadboard controller to other 2 axis
As I'm shelving home-brew shield I could update pin config to standard GRBL and use latest build too which IIRC is 0.8?

Where should I go first? Running out of allotted time for this this week, would love to see a gantry move before I must put things away for a few days!

Thanks
Jim

[tioJim]

Well I got a gantry moving by tweaking the pot on my A4988 driver. I guess it gave the motor more juice ...

... but it's operation seems pretty random! There doesn't seem to be any rhyme or reason to which direction it goes or for how long it travels after issuing a command. For example, X5 can send it left. Follow with an X10 and it goes right! Or sometimes left ... What's more X5 might make it travel a tiny bit. Send X50 and whoa! It goes bananas. If it hits the end of it's travel it sometimes bounces back and start going in the opposite direction!

That doesn't seem right to me.

On top of that I flashed GRBL 0.8a, I'm using the correct pins and I used the GRBL calculator and get the proper values for my config and ... it doesn't do jack. It just judders like a damn freak!

Code: Select all

$0=87.489
$1=87.489
$2=320
$3=30
$4=750
$5=1100
$6=0.1
$7=0
$8=9.8
$9=0.05


'spose I better get a commercially made shield to eliminate my DIY tomfoolery but they're harder to source in the UK ...

Despondency, sadness, despair setting in ...

[Will Winder]

Its possible that the movement is correct but gcode just doesn't act like you (or many people) expect it to.

Think of it like cartesian coordinates back in geometry class. G0 X5 doesn't mean go right 5, it means go to coordinate X5. If you are at X0 it should move right, but if you are at X10 it needs to go left. This also means if you enter the same command twice the second time shouldn't move anything.

[tioJim]

Its possible that the movement is correct but gcode just doesn't act like you (or many people) expect it to.

Think of it like cartesian coordinates back in geometry class. G0 X5 doesn't mean go right 5, it means go to coordinate X5. If you are at X0 it should move right, but if you are at X10 it needs to go left. This also means if you enter the same command twice the second time shouldn't move anything.

No I figured it that way Will but thanks anyway. I think there's some other kind of madness going on! I'm sure I'll get there

[forgottenCNC]

I would suggest getting the full step mode working first before you play around with mirco steps as mirco steps are “fake” poles on the stator (coils). You will also need to set your current limit on the A4988 because with a low current limit set on the A4988 there may not be enough current thought the stepper motor to move the motor, a very basic rule of thumb is more current “flowing/alternating” thought a coil the stronger the magnetic field (flux).

This video may help your understanding of stepper motors ( http://youtu.be/bngx2dKl5jU )

I've attached my working complied hex code for Arduino Uno with GRBL shield V2 and the pin outs are as flowed. you will need to change the extension txt to hex

2 - X_STEP
3 - Y_STEP
4 - Z_STEP
5 - X_DIRECTION
6 - Y_DIRECTION
7 - Z_DIRECTION
8 - STEPPERS_DISABLE
9 - X_LIMIT
10 - Y_LIMIT
11 - Z_LIMIT
12 - SPINDLE_DIRECTION
13 - SPINDLE_ENABLE

[tioJim]

Thank you I'll try your GRBL bin. Also after more Googling I've realised setting the current limit accurately is pretty crucial as it looks like not only can it cause the motors to be too weak but also cause them to act sporadically which is what I think I have right now.