Slant-Bed CNC Lathe Build

[barnbwt]

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.

I can see chips, fluid, and just grime messing with inductive/capacitive types, I guess. Do temperature/humidity impact their readings as well (beyond thermal expansion)?

 

[Karl_T]

I've got the inductive proxes on both my CNC lathes, never had a failure. got the mechanical on my CNC mill. Did have one of them fail. Pretty small sample size though.

FWIW, this is something I'd shop for on eBay and buy spares. you need three per axis so the cost adds up. I like to wire them at 24 volt NC. That way a wiring failure causes a machine stop and 24 volt is just more robust for machine logic wiring.

 

[JimDawson]

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.

They are pretty reliable, but more difficult to wire in depending on the application. They come in 4 flavors as far as the wiring goes NPN NC, NPN NO, PNP NC, and PNP NO so you have to make sure you are getting the correct device for the application and the other hardware that you are using. The limit switches have both NO and NC contacts and are compatible with all hardware.

I can see chips, fluid, and just grime messing with inductive/capacitive types, I guess. Do temperature/humidity impact their readings as well (beyond thermal expansion)?

Chips can be a problem depending on the sensitivity, and coolant could affect the capacitive devices. Normally they are rated at IP67. Temperature (within their operating range) and humidity do not affect inductive sensors but humidity could affect a capacitive device. I have never seen a capacitive device used in any machine tool application, they are normally used in applications where something other than metal needs to be sensed, wood for instance.

 

[Karl_T]

Jim is there any application difference between PNP and NPN. I've never worried about it.

Now with a bit of work I probably come up with the Laplace transform describing the electron's potential across the NP junction

 

[JimDawson]

It depends on what you are connecting it to and what the other related hardware looks like. Depending on the Galil model, it will take a NPN or PNP input, or NPN only. The issue is the common to the inputs. LS COM is separate from IN COM so can take different input types. The real problem come in where you should be wiring travel and home limits as NC. That way a wire break or other failure looks like a limit trip, and this is what you want. It's just simpler with mechanical limit switches IMHO. But I'm lazy

 

[barnbwt]

My motors arrived, hopefully I'll be able to jog them in the near future, but they look pretty good so far. I'll have to do some research to determine EXACTLY what winding they have, but it'll be 3-4A at 60VDC for around 5000rpm, and 190-210W --every winding for this series-type falls near this regime.

Three came with circular mounting adapters, slotted for belts to pass through, and 1"/25mm timing belt pulleys.

Three of them have brakes (presumably 12VDC) which I may or may not remove depending on the current requirements. Surprisingly, the brakes are not rigid like I'd expected, and the motor still has about 3 degrees of movement. The one motor with no brake feels nice and smooth/tight in the bearings, just as you'd expect for a PMDC motor.

Three of the motors have Japanese Daido encoders on them (by their blue colors, I'd originally thought they were Tamagawa Seiki's). I still haven't found the exact datasheet with the part number construction for this older series, but they appear to be 2500C/T incremental encoders, with an 8-wire pin-out helpfully provided on the label;

V/gnd, A +/-
B+/-
Z +/- which I presume is an index-pulse channel)
Two wires per channel which I believe makes them double-ended. The current Daido H48 series use a 5V supply and are good up to 6000rpm

Now, two of the encoders have hairline cracks on the plastic cases; I'm not sure how significant this is in reality since I know ABS plastic moldings tend to do that anyway as they age. The motors are very clean and the power wires covered with loose shrink-tube, so I don't think they were getting soaked in coolant or anything like that.

I did also want to mention that these motors were a bit larger than I was expecting (in length). I guess BLDCs are a little more energy-dense since comparable units were around 5" in length, these PMDC motors are about 7" before the encoder is stacked on. I'll find a way to fit them, but folks considering these should know they won't be getting a stepper-sized package like a BLDC provides, but something about 1/3 larger.

 

[JimDawson]

Yup, differential output on the encoders. That is the only way to fly. Yes, the Z is the index pulse.

The only place you might need a brake is on the X axis, mine will drift down without it when not powered.

 

[barnbwt]

Jim is there any application difference between PNP and NPN. I've never worried about it.

Now with a bit of work I probably come up with the Laplace transform describing the electron's potential across the NP junction

That sounds noise sensitive, to put it lightly

I see plunger type and pivot-type limit switches (the big industrial ones). Do the arms give more resolution/accuracy, or is it purely to make adjustment a bit easier once they are installed? Lastly, once you're looking at industrial switches (sealed, large enough for a good bit of voltage & current) is there much of a difference as far as accuracy or longevity going with a Chinese product, vs a Honeywell or Allen Bradley? I am mostly thinking of the durability of the contacts as 12V sparks across them a bunch of times.

Hmm, I probably need to start sketching out how the E-stop will be implemented; I've seen a number of home builds where this was half-assed, and "feed hold" was essentially the kill switch. With steppers I'd be less concerned about crazy runaways, but I know that's what servos love to do when an encoder dies/etc so I'd like to do this one up correctly.

 

[barnbwt]

Yup, differential output on the encoders. That is the only way to fly. Yes, the Z is the index pulse.

The only place you might need a brake is on the X axis, mine will drift down without it when not powered.

Yeah, I'm thinking that, too (would also be nice even on Z as a safety mechanism though).

Not to drift my own thread too terribly much, but I am considering a much more upright slant configuration (60deg vs 45) and the brake really would become a necessity. My discussion about ganged-tooling earlier got me thinking; gangs don't work very well if you have a tailstock supported workpiece (unless I'm missing something) and my 7" max stickout would stand to benefit from that ability. So how to get the tool gang around the workpiece so the desired tool can be selected?

If the headstock itself can move about 2-3" at an angle to the X-axis (across the table toward the operator in this case) the tool gang could slide freely behind the work piece without interference. This would not be a controlled axis but a momentary shift similar to a powered draw-bar. Combine that with a similarly 'dumb' tailstock center which can engage or disengage from the end of the workpiece a short distance. It seems like it'd be a pretty fast way to change tools without adding a ton of complexity or volume (maybe some loss of rigidity?) I'd be able to put right and left hand side tools at opposite corners of the gang positions and cut on the full 24" interior length between the spindle and the far wall. Anyone try something like this before?

 

[JimDawson]

The arms make adjustment easier. Normally for machine limits I use the roller plunger type and make a cam to operate them.

The slow acting switches are more repeatable than the snap acting ones. The industrial switches come in both snap acting and slow acting. Normally you are going to wire these directly to the inputs, so the current is a couple of milliamps, any contact should last in the millions of cycles.