Optimum Optimill MH50V CNC conversion

More
2 years 10 months ago #287296 by Unlogic
After finishing the control panel I had to do a little side project as the fancy adjustable gas spring equipped monitor mount I had purchased proved to be way to flimsy...



It would hold the weight just fine but if I pressed a button on the control panel the whole thing would flex and shake for a long time afterwards.

So I realized I needed something more heavy duty but I was unable to find anything more sturdy readily available so I decided to build one myself.

Went to the local metal supplier and picked up some rectangular and square tubing with thick walls along with some solid round bars of 30mm and 25mm diameter.



Everything temporarily laid out before machining.





Everything machined ready for welding along with some brass washers that I managed to pick up at Bauhaus of all places.

    

Please Log in or Create an account to join the conversation.

More
2 years 10 months ago #287300 by Unlogic
Things tend to move about during welding so I welded the arms first before doing the final machining and drilling the holes for the hinges in order to make sure everything was straight and aligned.

 

 

 

 

To prevent warping of the base plate during welding I made this little jig, but I'm not sure if it helped much at all because the plate became slightly warped anyway...

 

First test assembly of the finished parts.

 

Time for a some paint.

 

 



 

Please Log in or Create an account to join the conversation.

More
2 years 10 months ago #287302 by Unlogic
In order to be sure the arm was correctly leveled I attached the control panel before mounting the arm to the wall. That way I could set the spirit level on top of the control panel.



The arm attached to the wall next to the cabinet.



This is how it looks all assembled with the screen attached.

The following user(s) said Thank You: besriworld

Please Log in or Create an account to join the conversation.

More
2 years 10 months ago #287304 by Unlogic
As I now have the cabinet mounted on the wall I could do some more wiring work. One thing that was missing was the solenoid for the mist cooling (FogBuster). As I only have one 24v power supply it needed a flyback diode installed, it was a bit fiddly but I managed to fit the diode inside the connector itself.



 

 

Next up I wired the two Noctua 12v fans in series so that I could run them on 24v. Noctua does offer 24v fans but they are much more expensive than the 12v versions for some reason...

 

This is how the cabinet looks now with the servos laying on top waiting on me finish machining the ball screw mounts.

 

The only stuff left to wire now is the limit switches, external e-stop, probe and VFD which will all come later when I'm done with the mechanical conversion of the machine.





 
The following user(s) said Thank You: tommylight, Clive S, rodw

Please Log in or Create an account to join the conversation.

More
2 years 10 months ago #287778 by Unlogic
I have Probe Basic up and running now with all the HAL customizations I had previously added in my Axis configuration.



It wasn't smooth sailing however as I hit a few snags along the way which I had not foreseen in my preparations.

Those who have followed this thread know that I had initially planned to use a HP Elitedesk G2 Mini computer for this build but I abandoned it for an older Core 2 Quad setup instead as the Elitedesk had horrible delay and network problems.

 I knew that Probe Basic required OpenGL 1.50 or later and the Intel specs for the Core 2 Duo motherboard (Intel DQ45EK) I was using as specified the following in the manual for the motherboard.



But in fact Probe Basic doesn't need OpenGL 1.5 it actually needs OpenGL Shading Language (GLSL) which is something different. Now days the OpenGL shader version number always matches the OpenGL version number but this wasn't always the case as OpenGL 3.2 and below used a separate version number for the shader for some reason...

 

This meant that the 3D preview window in Probe Basic wouldn't work on the Core 2 Quad based computer work even though it worked fine in Axis... image missing /media/kunena/emoticons/sad.png

At the same time as this played out I realized thanks to Rodw's post in the thread linked below that there are actually multiple versions of the HP Elitedesk G2 Mini computer.

forum.linuxcnc.org/27-driver-boards/4691...ead?start=300#287022

That explains why my experience installing Debian in legacy mode on that computer differed so much from what RotarySMP described in his thread linked below:

forum.linuxcnc.org/26-turning/41498-scha...fit?start=260#227063

The Elitedesk G2 Mini version I have seems to use a laptop CPU that explains why legacy mode was so broken. Laptops haven't focused on legacy mode for a long time since Windows requires secure boot, TPM etc these days.

Based on this information and together with the need for newer pixel shader version support I decided to give the Elitedesk another try but this time in UEFI mode. The computer fought me every step of the way in legacy mode as I've described previously but in UEFI mode it was the opposite and everything installed without a hitch and it booted just fine from the NVMe drive.

It would also launch Probe Basic just fine as it support a much higher OpenGL and OpenGL Shading Language version. The latency  in UEFI mode is 2-3 times that of the Core 2 Duo at around 35000-45000 but still usable with the Mesa cards. But most importantly the wired network now worked properly which surprised me as it was pretty much unusable in legacy mode.

This meant I had to revise my recently built control panel slightly. As I had built a rail based mounting system on the back of the monitor I had to  build two new rails for the Elitedesk.

 

 

Once Probe Basic was up and running I came across some very strange work offset coordinate errors that I couldn't find anything about online.
Turned out that my e-stop related HAL changes somehow interfered with the initialization of the current offset in Probe Basic.

I didn't fully manage to find the root cause of this but I was able to find a workaround by changing the initialization as described in this thread:

forum.linuxcnc.org/qtpyvcp/50922-configu...ied-properly?start=0

The only thing I have left software wise for now is to modify Probe Basic in some way so that I can control the jog speed of my physical buttons but I'll tackle that a bit later.

The next step is to modify the mill so that I can add central lubrication and ball screws for X and Y. I'll start with the central lubrication as I'm currently waiting for HTD 5M pulleys and polyamide belts from Norelem for the ball screws.
 
The following user(s) said Thank You: besriworld

Please Log in or Create an account to join the conversation.

More
2 years 10 months ago #287787 by PCW
That's interesting about the 800 G2, If you have good long term network
performance on the 35W model (in UEFI mode) I should change my post about
that model.

Please Log in or Create an account to join the conversation.

Time to create page: 0.209 seconds
Powered by Kunena Forum