Slant-Bed CNC Lathe Build

[barnbwt]

Maybe a swappable gang plate would be a compromise? I read that regulated power supplies have a harder time handling back EMF, but as small as these motors are I can't imagine that's a huge issue here.

 

[Karl_T]

Totally unrelated but EMF noise can/will drive you nucking futs.

My lathe X axis would go unstable on occasion. Chased it for hours and hours. I ended up going to differential encoders to solve it but the root cause was EMF noise.

Two lessons learned:
1) be serious about electrical noise sources and grounding
2) never cheap out on single ended encoders

 

[barnbwt]

Exciting developments (I guess);
I got confirmation the 18-36V (or 36-72V) input range of my main board voltage converters does NOT also apply to the 18-60V range of the motor-power circuits on the daughter boards. Makes sense in retrospect, but their manual does not imply they are independent. So I can use motors to the full 200W level, and my gimpy "24V" (19.8V measured) power supply should still be sufficient at least for testing.

My confusion had stemmed from not realizing there are two levels of amplification here, not one. The 5V source logic circuit regulates the analog and 'line level' (for lack of a better term) digital outputs from the main board, which are supplied by the 12V source. The 12V outputs in turn regulate the 18-60V motor current source on the daughterboard to spin the rotors.

So the sub boards are run directly by those 12V outputs and 5V bus voltage through their connection to the main board, hence there only being external connections for motor power. The converters on the main board let you deliver the 5V and 12V through a single connector at 18-24V (and presumably cleaner power going into the logic circuits). Very convenient.

At 60V is interference really such a certainty? Friends have had stepper motor issues running at 24V that vanished when raised to 48V since the signal was essentially stronger (that was my understanding at least)

This 60V Absopulse is supposed to be a fairly nice industrial power supply so you'd think EM emissions would be minimal, and the power output in the wires pretty clean.

 

[JimDawson]

Friends have had stepper motor issues running at 24V that vanished when raised to 48V since the signal was essentially stronger (that was my understanding at least)

Regarding steppers, at 24V they were way under driving the steppers. The best way is to max out the voltage to whatever the driver is rated at.

In the case of brushed DC servos, you can run them at lower voltage and still get good performance. On my mill, the motors are rated at 140VDC and I run them at 70VDC. They won't reach rated max RPM, but I don't need 600 IPM rapids, and they have plenty of torque for anything that I am doing.

 

[barnbwt]

My target is 100IPM, though

Now, Galil does make a BLDC driver board with 600W per axis (there's a cheap 4-axis unit on ebay) which would be great for a slightly more powerful rig. Maybe next time.

Assuming my axis motors & encoders are in good shape, I think things may fall into place nicely. I suppose limit switches are next on the list; are there any downsides to the capacitive types versus the mechanical ones?

 

[JimDawson]

I suppose limit switches are next on the list; are there any downsides to the capacitive types versus the mechanical ones?

Inductive prox or mechanical is good. I kind of lean towards slow acting mechanical, much more accurate and repeatable than snap acting.

 

[barnbwt]

Slow acting, so it's like a small-travel rheostat or something that crosses a pre-programmed electrical threshold?

 

[JimDawson]

No, it's just a contact closure, but it doesn't snap over like a regular switch. In other words, it doesn't go ''click''

here is an example https://www.automationdirect.com/ad...tal_Plunger_with_Roller_Actuator/AEM2G14X11-3

 

[barnbwt]

Oh; I'd seen those type, but just figured they were beefier coolant proof versions of the cheap micro button/lever switches.

 

[Boswell]

Inductive prox or mechanical is good. I kind of lean towards slow acting mechanical, much more accurate and repeatable than snap acting.

Jim, that is interesting. Intuitively I would have expected the Inductive Proximity type to be the most repeatable and of course with no moving parts, I would also expect them to be more reliable.