Dibloff's Build Log

Topics Related to the ORD Bot Printer

Re: Dibloff's Build Log

Postby dibloff » Sat Nov 24, 2012 4:43 am

One day after thanksgiving. 11.12 pm. X and Z axes are riding towards the home direction (X rides to the left, and Z rides down). When they hit the end switch they'll stop accordingly. When Y axis runs to home direction - away from the user - it'll hit the end switch but it'll not stop. If I send M119 it'll say y_min: triggered, although that end switch suppose to be Y_max. So by default this is the home position for the Printer:
axes.jpg

If I trigger Z home it'll activate the Z_minus switch:
z_minus_activated.jpg

If I trigger X home it'll trigger the x_minus switch:
x_minus_activated.jpg

But If I trigger Y home it'll run towards Y_plus, but it'll trigger Y_minus switch instead.
Y_plus_activated.jpg

Does anyone know how to swap those end switches in the Marlin firmware?
Thank you!
dibloff
 
Posts: 144
Joined: Sat Sep 29, 2012 2:04 am

Re: Dibloff's Build Log

Postby orcinus » Sat Nov 24, 2012 3:43 pm

If homing moves your Y towards +, then it's not the switch that needs "reversing", it's the Y-stepper direction.
Unless you've changed the way marlin homes, that is.

Check the following parts of Configuration.h:

Code: Select all
// ENDSTOP SETTINGS:
// Sets direction of endstops when homing; 1=MAX, -1=MIN
#define X_HOME_DIR -1
#define Y_HOME_DIR -1
#define Z_HOME_DIR 1


In my case, Z homes towards the max, not min (because i have a switch at the top and set the offset from bed in software instead of fiddling with screws). If all your axes are supposed to home in the - (minus) direction, all three lines should say _HOME_DIR -1. If they are still not moving in the correct directions, check this part:

Code: Select all
#define INVERT_X_DIR false    // for Mendel set to false, for Orca set to true
#define INVERT_Y_DIR true    // for Mendel set to true, for Orca set to false
#define INVERT_Z_DIR false     // for Mendel set to false, for Orca set to true


Fiddle with the settings until axes move in the expected directions (when looking at the front of your ORD bot, -X is left, -Y is away from you, -Z is down). Finally, check how you've hooked up your limit switches to the board. There's two limit switch headers per axis, one is for the negative direction homing switch (min), the other is for the positive direction homing switch (max).
orcinus
 
Posts: 720
Joined: Tue Feb 14, 2012 4:03 am

Re: Dibloff's Build Log

Postby dibloff » Sun Nov 25, 2012 4:39 pm

I managed to machine the parts to mount my digital caliper for travel verification. Since the printer is not officially operational yet I could not print the parts. Flying chips and burnt machine oil again. Drug for my soul…
Anyways. Here are some views of the caliper mounted on the Hadron frame.
X axis:
Resize of DSC01314.JPG

Y axis:
Resize of DSC01312.JPG

Z axis:
Resize of DSC01310.JPG

It does not bind at al, the previous fixture I've made using a file was not so much precise. The max travel I can do is ~110 mm. I think this must be the cheapest method of travel verification if you print the brackets. I've got the files let me know if someone needs them. The 6” caliper sells for $15 @ harbor freight. It is precise to the 0.01 mm.
dibloff
 
Posts: 144
Joined: Sat Sep 29, 2012 2:04 am

Re: Dibloff's Build Log

Postby dibloff » Sun Nov 25, 2012 4:53 pm

New day. New challenge. Hopefully one of the last ones. During the verification of the travel I realized when the Z axis moves up it’ll accumulate about 0.15 mm deviation in the first move, then this will be a constant deviation. So I’m making it move 10 mm but it only moves 9.85 mm. Then 19.85, 29.85 and so on. Here is a video where it clearly shows the deviation: http://youtu.be/wbXK3GG57qI
I started to investigate what might have caused it. I focused my attention on the couplers that join the stepper motor axis and the threaded rod. I did not like the idea of putting the entire gantry weight on it. Initially both the motor and threaded rods were inserted into the coupler ~6.5 mm, leaving space between them. Here is an analytical calculation for the gantry mass. Since my machine is assembled I did not want to take it apart just to weigh the gantry assembly. I only included the main components.
Mass.jpg

As you can see the entire X gantry weighs 1.67 kg or 3.67 lbs. This will translate to 16.4 N force. This force acts on two couplers, so one coupler will see approximately half of this only - depending on the extruder position.
From engineering perspective the coupler is a helical spring with rectangular cross section. It’s probably made out of 6061 aluminum, and has very low stiffness. Also this coupler is not designed for axial loads.
I’ve made some calculation for a helical spring with rectangular cross section with the force shown above applied to it. As you can see the deflection is turned out to be 0.126 mm. This is not far from the indicated 0.15 mm.
deflection.jpg
deflection.jpg (30.21 KiB) Viewed 315226 times

I started to look for solutions to prevent the compression of the coupler. Initially I wanted to have brackets machined/printed and bolted to the makerslides. These would house bearings and provide support for the lead screws.
brackets.jpg

Then I had a better idea, I simply dropped a ¼” ball into the coupler to act as a spacer between the motor shaft and lead screw.
coupler.jpg

Here is a video after its fixed:
http://youtu.be/jDOvTZG_zrE]
I'm pretty sure some people just shoved the axes together in the coupler until they're solidly stacked, I beleive this will make the coupler more rigid. The ball I used is 1/4" and I beleive it was in a bicycle front wheel cup and cone bearing.
dibloff
 
Posts: 144
Joined: Sat Sep 29, 2012 2:04 am

Re: Dibloff's Build Log

Postby dibloff » Sun Nov 25, 2012 6:04 pm

Here is the travel verification for the X and Y axes:
http://youtu.be/vVzGarRTlVo
As you can see both of them are off for ~ 0.15 - 0.2 mm at the initial step. I'm not sure what is causing this. I suspect it can be attributed to the belt tension, or maybe too much clamp on the excentric nuts. I'll work on it a bit more to make sure the numbers are right.
BTW. Here is the calculation for the X and Y axes: 200 (steps/revolution) * 16( microsteps/step) / 18 (teeth) /2.032 (MXL belt pitch) = 87.4890 - as I mentioned in an earlier post with this number the measurement was off. Considering it's a linear system I simply used a multiplier to find the correct number. I'm using 87.74352.
dibloff
 
Posts: 144
Joined: Sat Sep 29, 2012 2:04 am

Re: Dibloff's Build Log

Postby cozmicray » Mon Nov 26, 2012 12:18 am

Really like your digital caliper mounting.
I have been wrestling with same Ordbot hadron calibrations.
Z-Axis ---- Steps/mm setting in Marlin may be the culprit.

Here are a couple of my spreadsheets showing my efforts with digital dial gauge
The ---Incorrect Z steps per unit --- topic covers it
This seems the default setting
#define DEFAULT_AXIS_STEPS_PER_UNIT {78.7402,78.7402,2519.68,106.707} // default steps per unit for ORD Bot Hadron


Z_axis_steps.jpg
Z axis steps per mm test readings


Z_Axis_Measurement.jpg
Z axis measurements


I'll post my X and Y measurements soon

Tnx
Cozmicray
cozmicray
 
Posts: 168
Joined: Thu Sep 20, 2012 4:32 pm
Location: Blue Bell, PA

Re: Dibloff's Build Log

Postby orcinus » Mon Nov 26, 2012 12:51 am

I'm not sure i understand why so much fuss with the calibration.
Unless you're doing something really weird, the settings from the ORD Bot wiki should pretty much be spot on.

If you're not using the stock threaded rods, all it takes is about two iterations of measuring (i'd suggest measuring as big a span as possible, although about 10cm works fine too) and adjusting the steps/mm. I've marked two spots on the Z rails for that purpose - i set the gantry at the bottom one, then lift Z until i hit the second one (a known distance apart). The issued Z movement command / real length = the factor by which you have to adjust the steps/mm in Configuration.h, that's all there is to it.

(although i really like your caliper mount :D)
orcinus
 
Posts: 720
Joined: Tue Feb 14, 2012 4:03 am

Re: Dibloff's Build Log

Postby dibloff » Mon Nov 26, 2012 1:17 am

I found the problem with the Y axis homing. It was the end switch. I connected it on the Azteeg X3 to the Y_min terminal. It should be @ Y_max (since the Y axis will home @ max position). Everything is set up as far as hardware. I've made the first ghost print. It looks transparent :)
I tried to tweak the XY feedrate - adjusted it to 18000 mm/min (or 300 mm/sec) but it dances with the same rate. I re-sliced etc. Could not make it run faster. I also changed the layer height in pronterface from 0.4 to 0.25 still when the slicing is done I get 50 layers for a 20 mm high part instead of 80 layers. When I click on the Setting->Slicer Settings in pronterface I get the "Entering Skeinforge settings:Slic3r/Slic3r.exe --load Slic3r/config.ini -- ignore nonexistent config" error message is what I get.
My other problem is that the fan on the extruder does not run. It's connected to the 2nd extruder on the Azteeg. Does anyone know how to turn it on?
Thank you!
dibloff
 
Posts: 144
Joined: Sat Sep 29, 2012 2:04 am

Re: Dibloff's Build Log

Postby orcinus » Mon Nov 26, 2012 3:00 pm

M106 = fan on
M107 = fan off
M106 S[0-255] = fan on, set custom speed

I'd strongly suggest two things:
1) don't use slic3r integrated into Pronterface (use it separately), as it's a pain to configure and manage
2) don't aim for insane speeds at first - get your printer working right at low/moderate speeds first

Oh, and BTW, the feedrates are limited by the caps set in Marlin configuration and the acceleration (the acceleration dictates the maximum speed you will be able to attain within a limited distance).
orcinus
 
Posts: 720
Joined: Tue Feb 14, 2012 4:03 am

Re: Dibloff's Build Log

Postby dibloff » Thu Nov 29, 2012 4:10 am

Fan is running. Everything is running. The extruder gets stuck sometimes. Not sure why. Took it apart today, removed the sharp edges from the gear. Readjusted the screw. Got the new borosilica glass plate from McMaster. They shipped it in one day... I just ordered it yesterday morning.
Anyways. She is running fine. I need to do some prints now to see how accurate the machine is.
Also I realized if I print a 0.35 mm wall with a 0.35 mm nozzle it'll be ~ 0.65-0.7 mm thick. I guess it's because the nozzle compresses and flattens the layer.
I've read somewhere that tightening the screw for the qu-bd extruder can make the filament to flatten and stuck into the tube.
Also I'm not sure why, but when the filament comes out from the nozzle it'll curl up and stick to the nozzle's outside tapered surface (when it's not printing) I need to get a small metallic stubby brush to clean that surface.

Again, just wanted to thank to everyone who helped me with the build and setup. I really appreciate your help guys.
dibloff
 
Posts: 144
Joined: Sat Sep 29, 2012 2:04 am

PreviousNext

Return to ORD Bot

Who is online

Users browsing this forum: No registered users and 33 guests

cron