PICTURES -- Full circle milled MANUALLY with "Egyptian Pyramid CNC"

Well, then, you shoulda run the spindle offa a foot treadle. :)

Maybe I need to get some slaves...

That's slick.

-Carl

Thanks... Besides, this method lets me make any curve, not just a circle, a circle is just the simplest illustration.

I was going to say with dro backlash comp isn't needed but I just remembered you don't have a dro yet.

Backlash on a bp screw tends to be highest in the center of each table's range. My X screw is pretty nice but my Y screw is pretty nackered from a oiling issue that previous owner didn't notice or correct. Not a real big issue since I have a dro and don't count turns.

Wes

Very good. But just how long did it take to do that one circle. :-) ...lew...

That's right, I do not have a DRO yet. Also, since I am reading numbers off a laptop, backlash comp is handy as it obviates the need to start with a new set of numbers once the direction is reversed.

I have 0.05" backlash on the X screw. It can be adjusted, but the manual says that it is a "normal" value.

They're always on back order these days.

Approximately 20 minutes. If I could get a helper to read the numbers aloud, it would probably go down to 10 minutes. I may change the script to read the numbers to me using GNU Festival. It is not really difficult, I use it for my talking artificial intelligence robot Splotchy:

No, the problem is that it takes some 9 months + a decade of training to get them to be useful and after all that you only get perhaps 5 years of useful work out of them. Cheaper to just hire illegal day laborers...

Now you need to write a script that will read CAD/CAM generated G code and give you handwheel movements out! :-)

I think that it is, seriously, a great idea. It is the sort of a thing that I want to try out before considering any "CNC conversions".

My own feeling is that for most of my personal needs, I will not need this "milling to a script" type of approach very often, and a rotary table along with various positioning tools ought to serve me fine. But for those less frequent instances, when something complicated needs to be done, "milling by script" might work. Plus, it is fun, as it lets me do what I like, which is computer programming.

For a fun thing to do after you get your dro, zero readout with handle at 0, crank handles .200" and look at dro. If the dro steps .200 in lockstep you got a good screw.

Two things matter with screws, backlash and pitch accuracy. Also bearing end play. Make that three. No make that two, end play shows up in backlash.

Wes

Come back if you manually milled a thread (helix-path). LOL!

Nick

I do not need a DRO for it, I can do it now with a dial indicator.

Go away.

Originaly I was just joking around as some generated G-codes I've done for simple things get to 40,000 lines of G code. But there is times, when 95% of a job can be done with one manual setup and the handwheel movement scripts could save the trouble of setting up a rotary table or programming a CNC. Also, you could have the scripts output in Digital Readout coordinates, the backlash problem would be gone. For the handwheel movements like you're doing, you could have a backlash variable where the user could use a dial indicator and handwheel dial to get their machines actual backlash and enter in the script. Every time there is a direction change, your script could add the backlash amount in the movment. It took you 20 minutes to interpolate a hole, but it that saves you more than 20 minutes on a setup change....

And then you could take the distances in handwheel movements and convert them to step and direction signals, sent them to the printer port connected to stepper motors to turn the handwheels at a certain feed rate,... You've invented CNC! :-)

And 180 passes as you told somewhere else. 6.6 seconds per pass. Tell that shit someone else!

Nick