lathe/threading question

[Phooey138]

hey, im totally new here, though i retrofitted my Bridgeport series II a couple years back and am quite happy using it with mach3 so far.

I recently picked up a Mori SL1 and am hoping that just sticking with mach3 is an option. What im wondering is this: how can the synchronization of the spindle with the other axes possibly be accurate using a VFD where the steps are converted to a 0-10v analog, it just seems that theres no real position control using this method. id like to use a smoothsttepper (+ any simple step to analog converter board) and a sensorless vector drive or something along these lines, but cant find anyone producing threads with this method and reporting meeting ANSI specs with the results- they just look good on a webcam, which doesn't really tell me if they are good threads. if anyone is making good threads with mach and not installing a servo for the spindle let me know how... please? the machine has an encoder on the spindle already, and wouldn't it just be a matter of a single setting in mach to use this instead of the single pulse per rev method ive read about? would this help, or is the control loop so slow anyway that it wouldn't make any difference?

one final, possibly silly question- a VFD 'steps' through a finite number of digital states when producing a wave anyway, so why cant an actual step/dir signal be used to 'step' it through these states, bypassing the idea of frequency (in the ac sense) altogether. it is a digital device, so why the analog input- should there not be an inexpensive VFD or vector drive which take step/dir commands that make a 3 phase motor behave essentially like a stepper- or even take pulses from an encoder and perform as a servo? i just don't see the intrinsic difference between a real servo and what id imagine a vfd and 3 phase motor could approximate given tiny changes which would negligibly effect the cost of the hardware itself and pretty much just be a firmware change- apart from things like low near zero rpm torque etc- not talking about a real c axis, just good threading. thanks to anyone who even read this long post, and my apologies for my ignorance of the basics of spindle synchronization and threading using mach3!

[RICH]

There is a write up called Threading on the Lathe Mach3 which can found in the members Docs which you should read.

If you use a VFD ( my lathe does not have one ) i would recommend setting / fixing the rpm at which you will do the thread.
During the threading cycle Mach is controlling the feedrate based on it's reading of the real time rpm.  Mach will adjust the feedrate should
 there be a spindle slow down , but, it will not adjust the feedrate upwards above your set rpm. Should your VFD input sense that the spindle
 rpm requires adjustment, and then adjusts the rpm, that would be fighting the built in adjustment that the threading cycle is doing, thus you will have
threading problems.

-how can the synchronization of the spindle with the other axes possibly be accurate using a VFD

I would say it is more accurate based the on  testing we did. The single index is very accurate and there is no advantage to multi slotted disc's
and that's why it no longer used and shouldn't be used. Remmeber that the thread cutting will only be as accurate as YOUR "lathe system".

-theres no real position control ..........meeting ANSI specs  

In the write you will find a table which defines the basis that Mach threading  needed to achieve.
How good your lathe system will do can be tested , just read the write up.

- if anyone is making good threads with mach and not installing a servo for the spindle let me know how 

I have. Just remember that the physical end result is affected by manny things and i believe  most if not all are covered  in the write up.

RICH

[NosmoKing]

What im wondering is this: how can the synchronization of the spindle with the other axes possibly be accurate using a VFD where the steps are converted to a 0-10v analog,

one final, possibly silly question- a VFD 'steps' through a finite number of digital states when producing a wave anyway, so why cant an actual step/dir signal be used to 'step' it through these states, bypassing the idea of frequency (in the ac sense) altogether. it is a digital device, so why the analog input- should there not be an inexpensive VFD or vector drive which take step/dir commands that make a 3 phase motor behave essentially like a stepper- or even take pulses from an encoder and perform as a servo? i just don't see the intrinsic difference between a real servo and what id imagine a vfd and 3 phase motor could approximate given tiny changes which would negligibly effect the cost of the hardware itself and pretty much just be a firmware change- apart from things like low near zero rpm torque etc- not talking about a real c axis, just good threading. thanks to anyone who even read this long post, and my apologies for my ignorance of the basics of spindle synchronization and threading using mach3!

It sounds like you are referring to using a VFD and 3 phase induction motor as a synchronous motor, which is what a servo and a stepper is.
An induction motor can never run as a synchronous motor due to the requirement need for a slip frequency, i.e. a difference between supply frequency and rotor induced frequency, this difference will vary and increase with load.
The closest you can come to that is a motor pulse generator fed to the VFD instead of Electronic Vector detection.
Your original Lathe control could have had a 3 phase induction motor for a spindle, and the threading would have been done by synchronizing or gearing the Z axis to the spindle shaft encoder, which would have been fed back to the CNC control and not the spindle controller, this would take care of any variation in rpm of the actual spindle shaft due to load etc.
 Nosmo

[Phooey138]

Ohhhh... I think i see whats going on then- if mach waits for a pulse to start threading then it DOESNT MATTER how much lag there is, as long as its the SAME every time! you can still cut multiple spring passes and will come in at the same angle with any arbitrary lag - the difference between each pass will be the distance moved (on the 'C' axis) in the first lag time MINUS the lag the second time, not the total lag for both, or even one. and the lag will be consistent to a much smaller time-frame than the duration of the lag itself. im not sure why i had such little confidence in this before you said it was accurate. I was stuck thinking about some kind of real time synch rather than simply repeatable, though unknown, starting conditions and then jamming through it predictably. sounds good now. If i'm misunderstanding though stop me!

As for the VFD, thanks for explaining, i should know this!! while i was posting at 4AM i think i was just tired enough not to care that it might be silly- its good to learn what is silly about it!

Between the two issues ill just get a decent VFD and move right along. this wont be much harder than the mill project. thanks guys!

PS: any luck with tapping, where cutting force is less consistent through the depth?

[Hood]

If your spindle was a stepper or servo then it is possible (by using SwapAxis() ) to do rigid tapping. As you dont have a stepper/servo spindle then afraid a floating holder is really your only option for  tapping.

Hood

[Phooey138]

thanks, a floating holder shouldn't be a problem- as long as it makes threads

[RICH]

Threading will not start until Mach see's  the index pulse and depending on debounce setting a number of times to be valid.
Then the Z axis movement must start from a dead stop and accelerate over some distance, that's why you should be around at least 3
to 5 pitch lengths away from the end, so the axis can arrive at the exact same spot each time. How good your lathe system can do all that  will show up when you try multi start threading or test. Once it starts cutting threads the Z movement is basicaly a G01 move at a set feedrate with monitoring of the rpm to adjust the next cycle.

When i said rpm in real time, i was meaning that the input from the index to Mach is not maniuplated as it was in the past ( what windows does  / kernal speed ect. and thus manipulated by the pc before use).

How threading works along with relevant tests are all explained in the write up which is much better than i can reply in this post.
In the future maybe there will be slaving of the Z to the spindle ( call it electronic gearing or whatever ), or use of an encoder, but for now it is no longer broken and hope it stays that way! 

RICH

[Phooey138]

is there any word on when slaving of the Z to the spindle might be implemented using encoders- or should that be a question in the smoothstepper section, as the Z pulses would actually be re-timed in the SS itself? also- yes, "Rich's Guide to Mach3 Threading" is good stuff!- assuming thats the document people are referring to..?

[RICH]

is there any word on

Your jumping to a conclusion, lets just leave it at "wishfull thinking and may be possible someday" as there would need to be a program setup and defined and a lot of testing.

RICH