Re: New Setup

Thanks for the advice guys,

It makes me glad I asked this here to get a bit a "sanity check"
early on before I ended up going on a huge mission. I am a
professional programmer, so I tend to be a bit head-strong on the
coding side of things - but you bring up a very valid and practical
point - Make the thing work first, and then add the bells and
whistles. Programmers are optimists by design.

The other issue is that I cannot find cheap Nema 23 Servo motors. I
think I shall trust steppers to begin with and decide how I feel
about them (albeit, starting with some high-torque ones). To be
honest I was very worried that they would lose steps (bad experiences
with stepper motors when making my own table-sized plotter a year
ago). However, it seems many people are using them and not running
into any problems as long as they don't push the speed too high.

I think I'll go with a simple L297/L298 board to start with,
connected via parallel interface. If I'm up to the challenge, I'll
get a PIC to delegate the stepping operations for me later on - but
of course it all makes sense to get the thing running quickly rather
than pulling hair out at 4AM. Also, with the mill up and running, I'd
be able to do things like make custom motor/encoder mounts if I feel
like plugging in servo motors later on. :)

Thanks for the advice
and probably saving me from having a mill with a logisitcal computer,
sattelite tracking and it's own power-generator attached to it.

:)

Abby

--In CAD_CAM_EDM_DRO@y..., "echnidna" <echnidna@y...> wrote:

> Hi,
> I think most of us on this list would have to agree with James.
> There are so many different aspects of CNC that its more practical

to get something going then tinker with it after its actually working.

> Just the actual mechanical works can be very time consuming.
> As could the electronics if thats your passion
> or just learning and using various cad\cam software packages
> or doing the programming yourself
>
> So make it work firstly, then try to make it go better
> And have fun doing it.
> Regards
> Bob Thomas
>
> --- In CAD_CAM_EDM_DRO@y..., "James Owens" <wotisname@l...> wrote:
> > Hi,
> >
> > I have a question for you, why do you want to reinvent the wheel

with a toy machine? This is a rhetorical question and not meant to
cause offence to you or other Sherline users.

> >
> > The internet and this list has all of the answers you need to

achieve a CNC solution on this machine without you going to the
lengths of designing your own electronics and software. I can
understand you wanting to save money but most of what you require is
free or for a small sum as shareware. Let me start at the beginning.

> >
> > You have your light-weight table-top machine. It has very little

bulk and if you cough in its direction the head will flap about and
does not come with ball-screws to drive the slides. This maybe an
exaggeration but you get my drift. The first thing to look at is the
way the slides are being moved and to remove as much backlash as
possible within your budget. Either throw away the slide nut and make
a new one from something like Moglice or fit ball-screws. Next there
is the question of driving the leadscrews, servos or steppers.

> >
> > To keep it simple you are looking at steppers with a chopper

driver board. A stepper of 200 steps per revolution on a 14
threads/inch leadscrew will give you steps of 0.00035", if you half
step the motor this goes down to 0.00017" which is pretty small. A
chopper driver board was published on this list a few months ago. Dan
Mauch @ Camtronics does one which will drive all the axis and then
there is the Gecko's. I have found a number of circuits using various
chips on the internet. No encoder feed-back is required, your driver
boards and steppers should be man enough not to lose steps unless you
zap them with a pulse train too fast for them to keep up. For
software look at TurboCNC, Kellyware or any number of free/shareware
programs. All of these drive the chopper boards directly from the
parallel port with no extra programming require except setting up.
Finally there is the biggy EMC which is free as well. CAM is the next
thing in the line, but you haven't asked this question.

> >
> > Servos are a different kind of animal and if you don't have the

knowledge now to design a driver board I don't advise you go there.
There are many flavours of servo driver, some expensive, some
complicated and some that work almost out of the box. A servo works
in a closed loop system with an encoder telling the motor how far it
has gone. In basic terms the motor moves, the encoder confirms the
move and the electronics correct any over or under shoot. Jon Elson
of www.pico-systems.com as I understand these drives they only work
under software like EMC or other such software designed to run servo
drivers. Then there is the Gecko's which will run with any of the
software that will run steppers because they accept step and
direction pulses. You will need encoder feed back with servo motors
which should be mounted on the motor, however some put the encoder on
the leadscrew to try and eliminate backlash. I feel that this
practice makes the motor hunt all the time if there is any backlash.

> >
> > An edge finder, either electronic or manual, will take care of

setting up a job on the machine.

> >
> > There is no need to write drivers, design driver boards or

developing digitising probes............... unless you think you can
do better.

> >
> > Regards,
> >
> > Terry
> >
> >
> >
> >
> > Hi Folks,
> >
> > I'm a newbie to the list, and have just ordered a Sherline CNC

Ready

> > Mill, I live in London, England, so I'm expecting it to arrive

here

> > within the next month or so. In the meantime, I've been working

on

> > various plans and ideas for the controller part of the system -

which

> > I shall be building myself. When I see the huge prices

companies and

> > websites are demanding for simple parallel port drivers, it

just

> > doesn't make sense to buy from them when I know I could make
> > something far superior for a much lower price.
> >
> > I have one main question right now before I commit: Steppers or
> > Servos? So far I've been leaning towards steppers, driving them

with

> > the L297/L298 and using a PIC uController to control them and

take

> > input from digitizing probes etc (it'd be a serial device with

an

> > error control protocol, and the PIC would take care of ramping
> > up/down speed and providing steps instead of the PC, so I could

write

> > Windows driver software and performance would not be an issue)
> >
> > I'm just wondering if anyone knows where cheap servo motors

could be

> > found? Specifically, ones that are NEMA 23 compatible. If I

could get

> > inexpensive enough servos, I'd most definately go with this

route

> > rather than trying to hack mouse-encoders onto my finished

stepper

> > setup at some distant point in the future.
> >
> > Any advice/tips etc before I release the wispy spirit from

power

> > drivers?
> >
> >
> >
> > [Non-text portions of this message have been removed]