during the 4 months it wouldn't work reliably--often not at all--I "applied my technical skills" to it.
The biggest change, IMO, was rotating the power supply inside the case (barely fits that way), then moving the controller board to the far side of the power supply, away from the router-motor relay. Why? Because in the original configuration they run the router's 120V motor cable directly over the uCtrlr chip. (as in 1/4" directly above). That required modifying the metal of the box so that the connectors could come out that way. It also meant that the DB grounding studs no longer connect to the frame ground...THERE.
I then modified the grounding (in addition to the metal-box change) so that the grounding wires from the stepper motors didn't first connect to the digital ground. Sure, that makes sense when they're a shield, but their bizarre non-grounding of the bed leads to huge static discharges that can be seen on the digital ground, that make their way to the steppers whenever the hapless human touches one of the motor bodies during, say, an HDPE cut. So, I routed the static around the board and to the chassis, then the digital board just got it's normal K-contact.
I added a fan, since the stepper motor drivers get quite hot during normal use. At one point one even went into into some over-temp protection mode, rr something; one motor driver just cut out on a long (LONG) run, and then it came back on just fine after it cooled. Adding the fan, never had a problem since.
I added a circuit to output the router-motor-relay control signal out to my crazy-ass power-control box
. Yeah, it should have been all opto-this and that, but I was lazy that day
.
Because of the static drain wiring changes I made (a tip to the knowledgeable: 410K required bed-to-chassis ground; otherwise it interferes with the z probing). The touch panel couldn't be used directly on the bed, since the polarity has "ground" the spindle, and the "voltage" on the bed...which means that a grounded bed won't work. So, I reversed the polarity, and now I don't even need to use the touch plate; I touch off to the material, with an 0.020" plate to compensate for the Z-height error in the controller.
Of course, I changed the X- and Y-axis cables, but that isn't a "change" inside the box per se, right?
In addition, the laptop I use beside the machine is not actually the computer running the control panel software. That computer is about 30 feet away in another room, away from electrical noise. I use a Silex USB device remote server over wired 1Gb ethernet to connect to the controller, and I use a remote-terminal session from the laptop to the actual controlling computer. That way, the laptop can completely die if it wants (a hand saw fell on it once, DURING a cut!) and the CNC will not be affected.
Okay, I'll stop
.
Regards,
Thom