Re: Electronic Lead Screw & Power Cross Feed | Poor CNC ?
Posted by
John Stevenson
on 2005-12-06 12:31:28 UTC
Go and get a coffee, this could be long,..................... we'll wait.
Ready ?
I got involved with this about three years ago as I could see a product at the end of it.
The Frog is able to do this but I didn't like it as being too crude, no screen and relying on croaks to tell you where you were.
My idea was to split the drive train of a manual lathe. Top part of the existing train drove a shaft encoder, bottom part drove the existing leadscrew via a stepper motor.
By using the halfnuts as normal you could spin between a manaul lathe and a power feed / screw lathe.
Stumbling blocks I came across, One I knew sod all about electronics that little in fact that when I had a brain wave the light didn't come on <g>,
no problem, adopt the method called chequebook engineering and pay someone to do this bit.
Next problem was operation.
In 'my' method you had the lathe running, half nuts engaged, small amount of feed applied to cross slide [ X ] and you pressed start.
The preprogrammed move then waited for the encoder signal and drove the stepper motor vis a buch of magic spiky things
in a set relationship, a set distance and stopped the screw at this point which should be in a run off groove.
You now wind out the X, manually, press a button, being return or start again if the program is clever enough and it goes back to the start point.
Next step is to put the next feed on and press start again etc. etc. until full depth is reached.
Now the program running this can be anything from simple two way movement where all
control like depths of cut is via the operator or more technical where the program works the lot out and just requires a start signal at various points thru the program until it stop at the finish.
The actual programming shouldn't be too hard as it's a division program just like my gear hobber. An example:-
Leadscrew of 8 tpi, 200 step motor driving screw at 1:1 and a 500 count encoder.
To cut 8tpi the leadscrew has to move 1" for every 8 revs of the spindle, 8 revs equals 4,000 pulses of the spindle. The leadscrew also has to do 8 revs so the motor will be expecting 1600 pulses
So to gear 1:1 for 8 tpi we need a divide by chip that can output 4000 / 1600 = 5/2 so for every 5 pulses it receives it sends out 2.
16 tpi will be 5/1
32 tpi will be 10/1 etc
These are not very good figures and a bit of time spent working out encoder counts will be able to get a better set of ratios, this just lays out the idea.
Now taking this one stage further it's a single axis CNC program, Aaagghhhh shouts of dismay, we don't need no steenking CNC's I hear
But it is.
Now the opposition comes into play, too expensive, we want a little box about $150 and no CNC.
So lets look at a cost breakdown
We will gladly pay $150 for this but what is 'this'
First we need the $150 black box, then a stepper motor and a driver and a shaft encoder and a few mechanical bits. OK with me so far?
Now we look at a single axis CNC route.
No black box but an old PC and a software licence, then a stepper motor and a driver and a shaft encoder and a few mechanical bits. OK with me so far?
Hang on, what's changed? we now have an old PC and these are usually free or very cheap, OK for arguments sake we say $30 and a license at $40 - $70?? and we were happy with $150??
So we now have what we were looking for , no development, no waiting and we save money ?
No, just re read all I've typed and I can't see where I have gone wrong, why have I talked myself out of reinventing the wheel.
It's not full CNC, it only drives the leadscrew, when the 1/2 nuts are disengaged it's a manual machine.
I need to lie down in a dark room with a damp copy of Exchange and Mart over my eyes...........bye.
John S.
This e-mail was scanned for viruses using BitDefender
Ready ?
I got involved with this about three years ago as I could see a product at the end of it.
The Frog is able to do this but I didn't like it as being too crude, no screen and relying on croaks to tell you where you were.
My idea was to split the drive train of a manual lathe. Top part of the existing train drove a shaft encoder, bottom part drove the existing leadscrew via a stepper motor.
By using the halfnuts as normal you could spin between a manaul lathe and a power feed / screw lathe.
Stumbling blocks I came across, One I knew sod all about electronics that little in fact that when I had a brain wave the light didn't come on <g>,
no problem, adopt the method called chequebook engineering and pay someone to do this bit.
Next problem was operation.
In 'my' method you had the lathe running, half nuts engaged, small amount of feed applied to cross slide [ X ] and you pressed start.
The preprogrammed move then waited for the encoder signal and drove the stepper motor vis a buch of magic spiky things
in a set relationship, a set distance and stopped the screw at this point which should be in a run off groove.
You now wind out the X, manually, press a button, being return or start again if the program is clever enough and it goes back to the start point.
Next step is to put the next feed on and press start again etc. etc. until full depth is reached.
Now the program running this can be anything from simple two way movement where all
control like depths of cut is via the operator or more technical where the program works the lot out and just requires a start signal at various points thru the program until it stop at the finish.
The actual programming shouldn't be too hard as it's a division program just like my gear hobber. An example:-
Leadscrew of 8 tpi, 200 step motor driving screw at 1:1 and a 500 count encoder.
To cut 8tpi the leadscrew has to move 1" for every 8 revs of the spindle, 8 revs equals 4,000 pulses of the spindle. The leadscrew also has to do 8 revs so the motor will be expecting 1600 pulses
So to gear 1:1 for 8 tpi we need a divide by chip that can output 4000 / 1600 = 5/2 so for every 5 pulses it receives it sends out 2.
16 tpi will be 5/1
32 tpi will be 10/1 etc
These are not very good figures and a bit of time spent working out encoder counts will be able to get a better set of ratios, this just lays out the idea.
Now taking this one stage further it's a single axis CNC program, Aaagghhhh shouts of dismay, we don't need no steenking CNC's I hear
But it is.
Now the opposition comes into play, too expensive, we want a little box about $150 and no CNC.
So lets look at a cost breakdown
We will gladly pay $150 for this but what is 'this'
First we need the $150 black box, then a stepper motor and a driver and a shaft encoder and a few mechanical bits. OK with me so far?
Now we look at a single axis CNC route.
No black box but an old PC and a software licence, then a stepper motor and a driver and a shaft encoder and a few mechanical bits. OK with me so far?
Hang on, what's changed? we now have an old PC and these are usually free or very cheap, OK for arguments sake we say $30 and a license at $40 - $70?? and we were happy with $150??
So we now have what we were looking for , no development, no waiting and we save money ?
No, just re read all I've typed and I can't see where I have gone wrong, why have I talked myself out of reinventing the wheel.
It's not full CNC, it only drives the leadscrew, when the 1/2 nuts are disengaged it's a manual machine.
I need to lie down in a dark room with a damp copy of Exchange and Mart over my eyes...........bye.
John S.
This e-mail was scanned for viruses using BitDefender