I grabbed a couple of the small form factor PC’s from the hackerspace last night and ran a few latency tests. One of the machines (an HP DX2710) gives excellent jitter times (under 20,000ns) so it’s now been set up as the controller PC running LinuxCNC.
I just need to work out the correct parameters to enter into stepconf before I give it a run. If anyone has a linuxcnc configuration file for the X-Carve, I’d appreciate a copy.
The Gecko G540 based controller is assembled in a temporary box and ready to test the motors.
It works! Well the motors move as expected which is a good start. 
I entered a basic configuration into LinuxCNC, initial speed is quite conservative just to be safe.
I managed to scrape off the labels from the stepper motor cables when I was threading them though the cable chain so I took a guess as to which ones were which. Of course I got every one wrong.
Wiring for the home switches is OK but Linuxcnc is not seeing them activate. A problem for tomorrow.
Before I can try to cut something, I have to machine a plate to hold the spindle bracket.
I should apologise for being a bit slack in getting my X-Carve finished, I can only blame Spring, so much to do outside!
Geoff
2 Likes
Disaster! Powered it up to play around with motor speed/acceleration and one of the Y axis motors didn’t move. I swapped over the connections on the G540 and the fault follows so it’s a driver that’s the problem and not the motor.
Gecko used to do a free repair on the first failure of a G540, I hope they still do that otherwise it’s a new driver board at around $A100.
Are you sure you have the Y axis and the A axis slaved together in your config. I believe if they are not sharing Hal pin in the config and one is not inverted that you will see this problem.
Dave
@DavidSohlstrom It worked OK last night (see video) and the configuration hasn’t been changed.
The LinuxCNC computer is not on the network so I can’t (easily) grab an image of the configuration file or stepconf.
I’ve had this G540 for several years and even though it’s not been used until now, it’s well and truly out of warranty.
PANIC OVER!
I was testing the speed via stepconf, not through LinuxCNC. If I fire up LinuxCNC, both motors move OK! It seems that stepconf does not slave the two outputs together when testing. A bit of reading is in order to see if this is documented.
Yes I seen to remember some one talking about that problem on another forum. good to hear the 540 is OK would be a long wait for you to get it repaired.
Dave
I posted a question to the linuxcnc forums and someone confirmed your suspicion, any slaving of axes is not used for testing.
The G540 is a beast with the X-Carve. I can make the X and Y axis move at frightening speeds! Must get those stop switches sorted out before I break something…
I neglected to add captive nuts for the home switches when I assembled the X-Carve so I’m going pull things apart to do that this evening. (I don’t have any post assembly nuts in case anyone was wondering why I’m doing this.)
Plan is to insert 2 nuts into both the rear X MS and the left Y MS. Second nut is in case I decide to add limit switches as some future point.
Does the MS orientation on the Y axis make any difference in how the home switch stop screw contacts the home switch? As the moment, the single t-slot on the out side of the MS is at the top. I remember seeing something about making sure that this slot was upper most on the X axis but can’t find anything about it being necessary on the Y axis.
Hope that last paragraph makes sense.
Looking through the assembly instructions and a few users’ build photos, it seems that the single slot is at the top for the X axis and at the bottom for the Y axis. I have the Y axis upside down 
Just as well I have to take the damn thing apart anyway!
I’m going to add a couple of captive nuts so I can add true limit switches as well. (see this thread Where to locate limit switches?)
Well removing/reversing the Y and X Makerslides wasn’t that bad at all. I’ve added two captive nuts so I can add a limit switch as some point. I was planning to remove the Y plate and X carriage so I could drill it for an additional microsiwtch but as my stepper motor wiring is soldered to plugs for the Gecko G540, it all got too hard. Perhaps when I pull it apart again to do the X axis stiffening mod.
A quick test under Linuxcnc shows that the home switches are now recognised.
No I’m ready to do some alignment and tuning. If I get the z axis plate machined tomorrow, I’ll soon be making sawdust!
No machining today, I decided to finish off the stand for X-Carve and do the machining tomorrow after I clean up the sawdust.
Frame is 90x35 MGP10 pine studs, the legs are 90x45. There’s another frame that will go at the bottom of the legs and have castors mounted on it. There will be a shelf on the bottom frame.
The torsion box top is what the frame/legs are sitting on, it’s a bit more visible ion the second shot.
Here’s a shot of a finished bench for the 6040 CNC router. It also has a torsion box top.
1 Like
A couple more images of the finished bench. Castors added although I doubt I’ll be moving it around much.
I gave the stand and the base of the torsion box, a couple of coats of polyurethane. The top of the box is still the wax that I coated it with when making it (stops the glue from sticking). I don’t think it’s worth trying to remove this.
I still have to cut a shelf for the lower rail but I’ve had enough woodworking for now.
1 Like
The router mount is almost finished. Plate is a standard wheel carriage plate from Inventables (25200-01), cut to length with slots milled for the bolts from the router mount. I still need to make a 3mm spacer to clear the screws that mount the plate to the delrin nut. Getting closer though.
Here’s the parts.
Here they are fitted to the Z axis.
1 Like
Success at last!
After some head scratching over the homing configuration for Linuxcnc, I managed to get some values that are accepted and was able to successfully perform a homing sequence - well partly successfully, the cheap micro-switch I used on the Z axis managed to fall apart (the actuator arm came off) and the carriage crushed the switch. A few minutes with a soldering iron fixed that.
I still need to sort out some acceptable feeds, which I’m happy to do now I can be sure it will stop when the home switches activate. There’s an annoying buzzing from the Y axis which sound like resonance rather than motor noise but that’s a job for tomorrow. I’ve also found the the plate I used for the Z axis is too short. If I enter in 65mm as the travel distance, the bit is too far from the table. I can put an actuator extension so the carriage can sit lower down but still activate the micro-switch at the top of the travel or make a new plate, another tomorrow job.
If it wasn’t so cold, I’d shout myself a beer. Have to settle for coffee instead - but with a chocolate biscuit.
1 Like
TimTam Slam!
Too cold for beer? Does not compute 
2 Likes
After spending what seemed like hours making the spoil board, it turned out it wasn’t too cold for a beer after all. Drilling, countersinking and threading 144 little holes, I deserved a treat!
I wish there was some tool to do this sort of boring work for you, say some sort of computer controlled routing device … 
!
After a bit of tuning, here’s a video of rapid movement on the XYZ-Carve.
4 Likes
Looking good!
Very nice indeed.