Boxford ATC toolchanger component

Hi All,

Having got to grips with comp and finally learning enough about HAL to make it work, I have written a toolchanger component for Boxford ATCs.

This should work on 240, 160 and 125 and maybe others that have 8 station turrets and
use a 'move past pawl and reverse back to stall to lock' methodology.

This will work without an index signal at each tool station.
The 160 and 125 don't have one ( at least mine didn't) and the ATC proximity switch on my 240 has never worked so I did away with it.

You will need to read the toolchange.comp file, because it contains info on setting up in comments at the header.
You may need to fine tune the parameters using setp on a running system and then hard code them in the .comp file once finalised.

For ease of testing I have made a .tar.gz of my test directory from my ~/emc2/configs directory.
You can simply copy the directory into your ~/emc2/configs and from inside the new directory run sudo comp --install toolchanger.comp
Then run emc, select the Boxford240Test entry from the list and it should run.

There is also a man file, toolchanger.9.gz which can be put in /usr/share/man/9/ if required

I have hacked several settings from my actual machine setup, to make it run in 'simulation-ish'
One of these forces it into an enabled state with 'setp iocontrol.0.emc-enable-in true'

DO NOT use this with a real machine attached, my house computer does not even have a parport.
(It requires EMC2 > 2.4 to run, if you have 2.3.x you will have to uncomment the NML_FILE=emc.nml line
in 240F.ini and insert emc.nml from another set up into the directory.)

Once started, home the axes, when you home 'A', a message will tell you that no tool is currently selected.
Switch to the MDI and enter M6T1 and the tool number will update without any actual rotation.
From menu Machine > Show Hal Configuration start Hal Configuration.
Select pins > toolchanger , open the Watch window and click on each pin in turn until they all display

See screen capture



Then you can make a succession of M6Tx commands in the MDI text box and watch the signals change and the values increment and then roll over at 360 degrees.

As for using for real, just copy the .comp file across and use the info in the header to change your .hal and .ini files
Do not attempt to use the ones in the .tar.gz with a real machine, they would be dangerous.
I would suggest that you copy your existing machine config directory to another name and use that to implement the changes to .hal and .ini files required.

You cannot jog the ATC with this component loaded because it hijacks the ATC stepgen, so I have written a basic jog facility into it, which is accessed via the supplied pyvcp panel.
This is mainly of use when setting up parameters, to sort out any over or under shoot in tool position.

Hope someone finds it of use

regards


PS
The component itself is now available for download from the contributed components wiki
wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Con...oolchanger_component

thanks i'm just about to start to look at controlling my toolchanger , you have posted literaly as i started to look at it ,
so i need to look at adding some grey code decoding for tool positioning , which my toolchanger has , the only other is mine has a dc motor rather than a stepper

so i'm hoping you have given me a big step forward , thanks again

ok anyone can guide me to change this to use 3 opto switches for positional info based on the following truth table , and to use a dc motor instead of a stepper

truth table :

tool 1 2 3 4 5 6 7 8

opto 1 1 1 0 0 1 1 0 0
opto 2 1 1 1 0 0 0 0 1
opto 3 1 0 0 0 0 1 1 1

i know i'm missing the obvious
i hope to just use 2 pins to operate 2 relays for motor on off and direction

Hi

I am just about to go out and will look at this properly later

The first thing that occurs is to increment the position_cmd in the progress_level 1 section, instead of allocating it the required position in level 0 and then waiting for it to get there.

Something like this would detect tool station 1

case 1:
while( !input1 || !input2 || !input3)
{
position_cmd += 1;
break;
}
// then deal with reversing back to pawl after reaching required station

Hope that makes sense

Will give it some more thought later

regards

thanks any help appriciated

Hey that's good, didn't know that was in there.

As I understand it, what John is pointing to is something that does the equivalent to the below.

If you take your truth table and assign values of 1 to opto1, 3 to opto 2 and 6 to opto 3
you will get a unique sum total from each of the lines in the truth table which identifies the tool position ie.

tool 1 2 3 4 5 6 7 8
opto 1 1 1 0 0 1 1 0 0
opto 2 1 1 1 0 0 0 0 1
opto 3 1 0 0 0 0 1 1 1

sum 10 4 3 0 1 7 6 9

Except that it will probably do it at a bit level, the theory is the same.
You could feed your 3 inputs into one component and get a single unique value out the other end as each tool-station passes.

I got your emailed drawings thanks, really need to know how many input pins you have available and how you are thinking of driving the changer.
If the DC motor is geared and slow enough, the 2 relays (on/off and fwd/rev) could be feasible to control it.

If you want to go off - forum until we have the polished final solution for publication, that's fine, you have my email.

cheers

Hi Andy

A friend of mine has been offered a Boxford 240 lathe for £1000.

If it is in good condition, take it, he can always sell on Ebay for more than that

He is dithering because the existing CAD/CAM software is licensed for educational use only.
(He wants to use the lathe as part of a small business he has making security products)

Boxford are notoriously difficult to deal with, they are only interested in supplying packages to educational establishments and will not entertain private individuals.

Does anyone know how easy it is to run this lathe through the serial port without using the Boxford software?

If the lathe is 1980s - early 1990s as it probably is, the software is proprietary, quite restricted and rather old (it may not even run on anything above Windows98 or even just DOS).
The serial connection may require a dongle and to run from something else would require a lot of reverse software engineering, which is probably not the way to go.

The 160TCL I converted had a 25 pin serial connector and I just used it to connect to the BOB!

How difficult is an EMC2 conversion?

There are several variants of the 240, if it is the straight 240 TCL which it almost certainly is because it comes with Boxford software, it is just an inflated version of the 160 TCL.

Therefore it is a stepper powered, probably Parker Digiplan SD2 or SD3 stepper drivers, 2 pickups on the spindle disc for index and pulse, good heavy transformer and rectifier bridge, DC spindle motor with a GEC speed controller and FWD and REV via a reversing contactor.

If he is lucky and it is a new, just pre 250 TCL lathe, it may have a Mitsubitshi inverter and 3 phase spindle motor. Even easier to convert.

In other words a prime candidate for conversion, even just using the original equipment.

There was a Fanuc version which I have, the 240F with AC servos, 3 phase supply and Fanuc OT mate controller.
I have seen a Heidenhain T150 version but don't know if it is original.

With your assistance I am sure conversion is the way to go, just depends if he has the time or simply wants a machine that runs out of the box.

If he decides to convert I am quite happy to assist where I can.

regards