Re: emc conversion of the tape-O-matic

Brad Heuver wrote:

> As I continue to educate myself on what its going to take to complete
> this conversion, I want to see how much I've absorbed.
> I've got the PC with emc running. The tape-O-matic has servo's for the
> X, Y, and Z (quill driven through a fine pitched rack) The spindle is
> varaible speed through the motor, and the head is moved up and down with
> a regular induction motor (1/4 hp). If I get a servo-to-go card, how
> many axis would I need? I know that it would be best to prepare for a
> rotary axis at some time, but doesn't that still only equal 4 axis? Does
> the STG card control the spindle speed, or other functions as additional
> axis?

Functionally, the STG card could control spindle speed (assuming you
get the 8-axis card, which I'd recommend). The code has never been
written into EMC, from some conversations I've had with Fred Proctor.
But, it really wouldn't be hard to have plain RPM control put in.
The only tricky thing about spindle speed is that lathes can have the
speed varied by the radius of the spot they are turning right at the moment.
But, we don't need anything that fancy for milling. Anyway, getting
just an RPM setting put through the interpreter in the S word, and fed
out to a DAC channel as a percentage of full speed, sounds pretty
easy.

> That said, if I can't use the amplifiers on the mill, I'll need 3 of
> those. I may need a new spindle speed drive. Is there anything else
> thats a big item?Power supplies I know of, and the cabling and
> connectors. The mill has a mist coolant setup that I might be able to
> use.

EMC controls some parallel port bits that can be used to drive
opto-relays to operate the coolant pump or coolant solenoid valve.

> I've gotten a bit better at the scanning, and made my way through most
> of the literature I got with the mill. I posted 3 more schematics to the
> RCM dropbox. They are all related to the servo motor drives. The first
> shows the basic setup and power supply, called tapemot.jpg. The second
> shows the Schmitt trigger circuits used. (I can make out the part
> numbers on the originals if that will help anyone make sense of them,
> but they aren't readable on the jpg) that file is called tapetrig.jpg.
> The last file is the shunt amplifier PC board (on the top right of the
> dwg) and the amplifer & null PCC (?) in the lower left. This file is
> tapenull.jpg It also has the output waveforms and timing graphed on it.
> This is all mostly over my head, except the power supply portion, so any
> help from the group as to the suitibility of these amps for potential
> reuse is appreciated.

Ugh! These are, indeed, 3-phase, SCR-controlled servo amps. Anyway,
using 60-Hz power, and 3 phases, it is pretty hard to have very high
bandwidth on these servo amps. I would guess bandwidth would be
somewhere around 10 Hz. Although it is possible to do better, the
circuitry looks VERY unsophisticated. Apparently, each servo amp
is several PC boards full of stuff. I have bandwidth of about 600 Hz on
my system. I could have gone higher, but I had to stay below some
mechanical resonances, like lead screw torsional response. With a
1 KHz servo update rate in the CNC computer (adjustable in the
EMC .ini file) I should get about 300 Hz or so bandwidth on the
positioning loop. Seeing the difference in machine performance
with this servo system, as opposed to the same system but with the
Allen-Bradley CNC control I started with, the higher bandwidth
definitely means tighter control on machine position.

> Also, I noticed that emc is written to support the STG card version 1
> only, but on the STG site, they recommend using their newer card,
> version 2. Is this likely to be an issue? Will a driver for version 2 be
> written for emc?
> Lastly, will the limit switches get hooked directly to the card, or is
> something else needed inbetween? (I think this is just the tip of the
> iceberg on the questions...)

You could connect them directly, but an opto-isolator might be a good
idea.

Well, actually, you will be amazed at how simple it all will be with EMC
and the STG card. It really pulls it all together. I have an 8-channel
Opto-22 compatible board for the auxilliary I/O, the STG card, a cable
to the parallel port on the CPU motherboard, and a terminal board for the
encoder connections. I made up a board to tie all the loose ends together,
and it has all the emergency-stop circuitry on it to watch the e-stop chain
and the STG card's watchdog, and cause an e-stop if anything goes wrong.
Finally, there is a rack of servo amps, which is a pretty small package.
The picture on my web page (picture 6 from the top is a close up)
http://www.artsci.wustl.edu/~jmelson/CNCconv.html
and a wider view in http://www.artsci.wustl.edu/~jmelson/CNCconv.html
showing the whole rack. The top box is a customizable enclosure (Zero,
I think) the servo amp rack is in the middle. Power supplies sit on the bottom
of the rack. Some pictures below show the junction board for the
shaft encoders, and the wire-wrap prototype of the e-stop and interconnect
board.

So, all this stuff fits in a 4' high rack. I even had enough room to make
a drawer that fits in on top of the computer, and holds all the commonly used
tooling, calipers, indi-cal, drill bits, etc.

Jon