Re: [CAD_CAM_EDM_DRO] Bridgeport CNC Mill???

"hanermo" - CNC 6-axis Designs wrote:

> However..
> EMC is limited in terms of hw speed, unless you go into the expensive
> stuff, just as you said.
>

Gee, I don't know about that. Usually I use off-lease Dell computers
from eBay that I
get for about $70 delivered. Stuart Stevenson at MPM in Wichita has
been heard to
crow that his 5-Axis Cincinnatti is now the most accurate machine in
Wichita, after
he solved all the kinematics corrections for machine inaccuracies. (You
can't do
that in Mach, by the way, it is cartesian in = cartesian out.) He has 4
machines now
that have been converted to EMC2 using my hardware, and he is quite
happy with it.
He does parts for Boeing, Cessna and military contractors that he
couldn't name.

> The mentioned servo loop update time, 20 khz max, is low for high
> resolution stuff.
>

Well, the OP is a home shop guy, he is not making hydraulic valves for
aircraft with
nanometer precision. I have yet to talk to anybody working under 400
IPM who has
even MENTIONED the servo update rate as a problem. Now, one place it does
crop up is if you are using the encoder as a tachometer, and extracting
velocity for
a software velocity loop from the encoder counts. But, we have finessed
that one, too.
Many of the servo interfaces, including my Universal PWM Controller, use
encoder
time stamps to compute a much better velocity that could be calculated
from the
number of counts/servo cycle only. But, in a system with velocity servo
amps and
a DC tachometer, this is really not a concern, as the servo amp is
handling the
velocity.

Note that old GE CNC controls had a 60 Hz servo loop, the Allen-Bradley
7320 had
100 Hz, and his existing control was almost certainly also no higher
than 100 Hz.

> The high end stuff is not well documented, and making any changes into
> the screen stuff is hard, practically impossible, unilke with mach3
> screen designer.
>

I have never understood the fascination with the screen designer stuff
for Mach. I used to use the EMC1
TkEMC GUI, which was pretty spartan. Now, I use the Axis interface,
with a little
side panel that shows spindle RPM. This is sure good enough for me, I
see no reason
to change it.

I have also used the EMC2 Touchy interface, it is fine for running the
machine if you have a
touch screen, but it lacks the part preview. I find that real handy to
make sure I have loaded
the right part program, and check the limits of travel and the location
of the (0,0) coordinates.
So, I am not going to give up Axis, at least for now.

> EMC installed base for higher end stuff is very, very low.
>

There are thousands of happy EMC users out there. Sherline sold over
1000 systems
alone, although it was hard to get a ballpark number out of them.
Smithy also sells
their machines with factiry-supported EMC. I know several scandinavian shop
operators who have converted their whole shops to EMC2.

> Plugins are few and far between.
>

What plug-ins do you need? 5-axis? hexapods? All included in EMC2.
There is also the
German Mocca GUI, although I don't see any advantage over Axis.

> Apparently the MESA stuff, some is implemented and some is not. Some is
> very expensive.
> Similar with Mach3 and MESA or Galil stuff.
>

Galil certainly is expensive, I don't think the Mesa or my stuff is
really that expensive, if
you are talking about bench top or Bridgeport-sized machines. You can
definitely go
cheaper for the Sherline and Taig-sized machines with cheap stepper
drives and small
steppers, driven directly from the parallel port by software step
generation.

> As an example, I am using 100-200 kHz hw update rate and will be using 1
> Mh to 2 Mhz, in the near future.
> On a lathe, with a Centipede, and mach3.
>

You will have a 1 - 2 MHz servo update rate? I find that a bit hard to
believe. First, it is
not necessary, unless you need a servo frequency response of many tens
of KHz, at least.
Certainly, the table of a machine tool can't respond to frequencies like
that. Second, when
reading the typical encoder, which may give about 10,000 counts per inch
of movement,
and at typical home-shop speeds of 60 IPM or 1 IPS, then you only get
10,000 encoder
counts/second. If the servo loop is updated at 1 MHz, then it only
receives an encoder
count every 100 servo cycles. The other 99 cycles are essentially
open-loop, where the
only thing the servo code knows is that a wild divergence from commanded
velocity has
not occurred, but it has no idea how close or far it is from the desired
velocity. Then, suddenly,
another count comes in and it has to respond smoothly to that new
information.

It just makes no sense to have the servo loop sample position far faster
than the fastest
encoder count rate.

> A user can be up and running very easily and quite cheaply with mach3,
> with about 50.000 commercial customers, and perhaps 2-3 times more hobby
> users, and a vast, responsive, growing community of enthusiastic supporters.
>
>

We have a good community, too, and often when bugs are discovered in the
development
version of EMC, somebody has them fixed within DAYS! And, if somebody
needs an
odd setup, like for adjustable-power laser cutters, tangential knife
machines, etc. it has
been implemented within a day or so.

> If I myself needed a basic, cheap, servo 3D milling machine, I would
> probably use your servo controllers - despite the fact that I have very
> machiliar with mach3, for 8 years, and feel it is excellent for what it
> does.
> The biggest drawback on mach3 is it´s hidden, dependency mechanism for
> all advanced higher end stuff.
>

Yes, and EMC2 allows you to easily create your own real-time components
if you need to
so some odd conversion, sense a switch and trip a relay, or whatever.
Also, for more
complex jobs like an automatic tool changer, you can create the logic in
Classic Ladder,
and have software relays, timers, and all sorts of logic functions to
sequence the ATC
mechanism. Also, it can take over the machine so as to orient the
spindle, move axes
to the right alignment for the tool change, and then restore the machine
before going
back to the G-code work.

And, of course, you can do rigid tapping with EMC2! That has become a
significant
capability to me.

> No-one who makes money with their machine (this was not the goal of the
> Original Poster) minds buying a set of x, say 3 servo drivers, at 1000$
> each, when this raises machine productivity, and results in paying for
> them in a few days.
>

But, my brushless servo drives are $150 each, and work with a wide range
of motors, up
to giant Fanuc 40-Lb monsters.

> However, when such products then wont work with (real random examples)
> dual-gantries, homing, threading, dual home switches, rear tool post,
> dwell, torch height control, cutter radius compensation etc etc everyone
> minds very much.
>

We definitely do gantry machines with two motors and two home switches.
We do a wide
range of homing, from single switches, to homing to the switch first,
then to the index pulse
of the encoder. We do rigid tapping without a servo spindle. I see no
reason a rear
tool post can't work. There are people doing torch height control on
EMC2. I have used
cutter radius compensation since 1998 with the original EMC1.

> The problem is not one of reliability or functionality - it is that you
> cannot know, in advance.
> Only Very experienced people will even suspect that these problems exist.
>

We try to make sure everybody knows all the limitations in advance, and
not let there be any
surprises. If somebody is doing something odd, all the need to do is
ask "has anybody done
this before?"

> Having a basic cnc milling machine is typically about 2-3000$ in parts,
> when all is said and done.
>

I know I could bootstrap a complete desktop system for well under
$2000. So, start
with the $500 minimill, I think that is an X-2. Get 3 motors from
Keling ($60 ea.), and add
encoders from US Digital or other source. Add my PWM controller ($250) and
3 PWM servo amps ($125 ea.)

Machine $500
motors 180
encoders 120
UPC 250
servo amp 375
belts & pulleys 150
power supply 150
total 1725

The motor brackets can be machined on the manual mill before
conversion. The only thing that
might be difficult is machining the pulleys to fit on the handwheels,
that needs a lathe.
If you have any good junk in your junkbox, that can cut the cost. And,
of course,
this INCLUDES the cost of the machine for conversion. (I know about
this one,
I bought a 150 Lb minimill for demo purposes at shows, and have been
hauling it around
for some years.)

Jon