CNC drilling machine finally completed

Cute - I take it the Z azis feed is via the special-built pinion gear shown in one of the photos? Looks like you built it around a surplus drill press, what kind was it?

Jim

Sort of....I pressed a precision shaft into the existing pinion gear shown below before the mod;

It's a Cameron MD-70 (aka deep throat), don't think you'll find that surplus.

Direct couplings between the motors and leadscrews? What takes up the misalignment?

The bearings of the shafts involved, leading to wear until the misaligment is less than the potential runout of the shafts. That isn't the case here though, misalignment is already less than the runout of the shafts based on the precision of the couplers involved and mounting methods. I added three more shots showing the motor mounts but there is nothing to show the precision of the couplers, you'll have to take my word on that.

If you say so. But it kinda makes you wonder why machines far more precise than this consider flexible coupling essential.

In my several years of CNC service nearly every machine I worked on had ridgid couplings from the servos to the ballscrews. Certainly the Kitamuras did with big Fanuc 10M servos. There was never an alignment issue with those.'

Pete C.

Theoretically, for a stepper setup or servo setup with encoder on the motor shaft, a flexible coupling is another source of errors due to possible coupling backlash and angular misalignment IMO. However, no doubt there are designs that warrant their use but some people seem to develop a "pet peeve" in this area one way or another. The other extreme of Richard's view came from another fellow, claiming to be a CNC buff, who insisted that the only proper way was to use belt and pulleys and have the encoder on the driven shaft....Anything else was worthless due to misalignment. While standards and the experience of others are invaluable when designing something, being able to think for oneself is also priceless.

Physical reality is not a peeve. Flexibility must be satisified somewhere. Ball bearings with runout in the tenths of thousandths are not going to get along with drive trains that accumulate at least several thousandths of runout, flex, angle, and offset. You can put the flexibility into belts, or helical couplers, or just wrench the bearings. housings, and shafts, but the demand will be satisfied somewhere.

Only if IYO you omit the helical type.

But peeves, irrational convictions and misguided energy are a reality in just about all walks of life. Some scenarios are so convoluted that people end up doing the right things for the wrong reasons.

Not this system....Remember, you were alluding to a design defect associated with this system. Is this all about more precise and sophisticated systems that you have designed by any chance?

It's the magnitude of the demand and the ramifications associated with any "dissatisfaction" that the designer should take into consideration. Rigid couplers are not manufactured and sold as table ornaments. There are systems where they are applicable and used. Mine is one such system.

MMC sells helical couplers. One suited for Y and Z axes in my system goes for $20.00 a pop. That's $5.00 less than the DC gearhead motors I'm using. The relevant load bearing sleeve for them costs 33 cents or I could turn/bore one in less than an hour. But I'll probably need new brushes before new bearings.

There is no flexible equivalent of the X axis coupler. It has three internal diameters, .25", 8mm and 10mm with the 8mm section tapped. Even with all that, there are no misalignment issues.

Didn't know about the gears. So you don't really couple them directly.

I sorta figured that from the fact that he had the same motor driving the spindle downfeed via the pinion gear.

Jim

The fact that there's a gearhead on the motor doesn't make any difference re the sort of coupling required, except perhaps for the fact that any effects of misalignment occur at a lower speed.

Ned Simmons

Ugh. Helical beam couplings are suitable for low torque applications, such as driving encoders, and best avoided for power transmission, particularly in motion control apps. They're relatively expensive, not torsionally stiff, large relative to torque capacity, and can be quite fragile. In places where a helical coupling will work there's usually a less expensive and more robust alternative.

Ned Simmons

Yep, all three motors are the same. I have seven of them, paid $25.00 each. Adding the second shaft for the encoder was a nice exercise in precision lathe work. Second shaft uses spring loaded ball bearings. BTW, the Z axis maximum feed rate would be over 700 IPM (too fast to drill at), X and Y axes are a measly 20 IPM. Still, the 228 holes in the 6"X4" accuracy sample were drilled in 15 minutes for about 15 holes per minute. Certainly beats doing it by hand. Furthermore, the drill motor is controlled by a SSR so I can walk away once the drilling starts and it shuts off when finished (M05). Pretty standard I imagine.

Gears can take some misalignment by their nature, and because the shafts are shorter and the bearings looser.

What alternative would you suggest? Timing belts were dismissed out of hand.

There are probably several. Precise rigid couplers and a "flexible" mount if necessary is one. Each scenario may call for a different type of mount. Note the flat and angle alu setup in the shot below. An ammeter was used for proper alignment.