RotarySMP wrote:

Any suggestions how to easily make the carousel.comp step through till the desired pin goes high also appreciated.
Cheers,
Mark

Remove the stationary magnet from the cover thing, and replace it with a spring-loaded magnet which remains in contact with the halls as the upper section rises and falls.  The halls will now get positional feedback when the turret is up or down, and carousel will keep the motor-fwd signal high until the appropriate hall sensor is triggered.

Doesn't look like there's any reason you couldn't just 3d print the top cover with appropriate mods for a captive, sliding, spring-loaded magnet.  I'd probably make an aluminum or brass 'carrier' tube for the magnet that presses in to the printed part so the magnet isn't sloppy but also doesn't get hung up on print layer lines.

I don't think I remember seeing another sensor to indicate the turret is down and locked, but as soon as carousel.comp turns off the motor-fwd output it should drop.  You will likely need to play with the hall alignment (or the top hat alignment, to be precise) to make sure carousel.comp doesn't drop the motor signal too quickly or too late to engage the alignment pin(s).

Above should let you eliminate the 'tp' component in your HAL example and just let carousel.comp treat your pneumatic system like a unidirectional motor.

Trylobite sent me a mail which really helped me understand the difference between the original Schaublin turret control, as mine still is, and the mods he made.He replaced the hall sensors with microswitches, which activate once the turret rotates into position. He uses that to signal closing the pneumatic tool change valve, dropping the turret to lock.

Since my Hall sensors do not read high until after the turret drops, bringing the magnet into proximity, I can't use the switches for valve control.

I initially thought it would simply step through, one pocket at time, till the correct feedback sensor goes high. I started writing the attached draft HAL to include a second pause for the pressure to bleed off, dropping the turret onto the hall sensor. But carousel.comp step through pockets like that does it?

Using Carousel.comp, how do make it cycle three times, if it is in pocket #2, to get to pocket #1, verses only moving once to get to #3?

Trilobyte also pointed out that the carousel will keep cycling from one position to the next continually, as long as the pneumatic valve is held open. There is a delay of a few seconds at each position, which appears to be controlled by adjustable restricters in the valving. With those pauses, controlling  the pneumatic valve open loop with simple timing seems feasible. 

I guess I can have it calculate the modula number of pockets to the destination when a change is commanded.
If if measure the time for lift and turn, and the time of pauses, and setp the three delay times whether 1, 2 or 3 pocket moves are needed.
Then by holding the valve open for the appropriate time, it will turn the correct number of pockets.

Any suggestions how to easily make the carousel.comp step through till the desired pin goes high also appreciated.
Cheers,
Mark

Oops, sorry SMC, deleted.

RotarySMP wrote:

Any suggestions how to easily make the carousel.comp step through till the desired pin goes high also appreciated.
Cheers,
Mark

Yes, but I'll wait until you've started your own thread image missing /media/kunena/emoticons/wink.png

RotarySMP wrote:

How do those lathes deal with the order of motion? A Z clearance move first followed my X makes sense for boring tools, with the opposite for turning tools.

Probably time to start your own tool-changer thread rather than blowing up @SMC's project.

Having said that, I'll violate unwritten courtesy and ramble on...

I don't think there's a hard-coded tool change position on those machines... just the expectation that it will be programmed to go to 'home' position.  For Fanuc 21T "C" mode, G28 G0 U0 W0 would move the carriage/slide/turret to the reference position.  If the programmer (or post) calls it on two lines you'd get a sequential X-move then Z-move.  Or W0 then U0 would do Z first, then X.

If you ensure that the tool path before a tool change includes a clearance move of some sort (Z for centerline tools, X for OD and internal boring), then your turret should use a Z-first, X-second sequence.  Your lathe is entirely capable of being used like a gang-tool lathe, with tools on the front and back of the part, turning in both M3 and M4.  You really don't want a boring bar working the rear OD to rapid home in X without a Z-clearance move first.

One reason the tool change position isn't hard-coded (in Fanuc) is that it's much faster to program a specific tool change location than wait for the slide to go all the way back to home.  I've done this on production runs of short parts... back the turret up so the longest tool barely clears the stock/part, and substitute that location for the standard "G28 G0 U0..." move.  Saves a ton of cycle time on long part runs.

So I'd leave out a specific TOOL_CHANGE_POSITION in your INI file and get used to programming a version of G53 G0 Z0 as your 'default' safe position.  You probably aren't going to be using the tailstock and automatic turret commands at the same time.  And you'll get used to drilling with the turret pretty quickly; it's so much nicer than using a tailstock by hand.

Tailstock work:  manual tool indexing
Non-tailstock work: auto tool changes at G53 Z0 X0 (or somewhere)

Hi,
Will this button open a new GUI when BUTTON is pressed? 

Wenn er frei beweglich ist dann passt schon der enable nicht. Nennt sich auch son oder servo on

Hi all,  

please guide as how we can add .py files as codes for user buttons in qtplasmac interface. The code works ok on terminal and base machine in raspi 5.

thanks

That's what I kind of figured. I'll have to run a temp bypass on the SSR for now and also re-think how it's wired to keep the vfd on without the control program running. The old vfd was analog so this never came up before
Thanks
M

I wired it so that if the control was closed or the e-stops tripped, it would cut power to the axis drives and spindle vfd. The enable button on the screen does as you say. 
until now I've not really wanted the vfd powered without the control software running so it wasn't an issue

Wrong thread. Sorry, deleted

The Schaublin tailstock has no sensor, and a very large and long nose. Everything is so large reletive to travel, that you dont even need the tailstock that close to the saddle for a turret tool crash.  I'll take a look at fitting a switch to detect the tail stock installed. It might be best to inhibit the tool changer as long as the tail stock is mounted. At least initially.

How do those lathes deal with the order of motion? A Z clearance move first followed my X makes sense for boring tools, with the opposite for turning tools.

Hallo,

ich versuche gerade den Einstieg in LinuxCNC, um eine kleine Maschine mit zwei Linearachsen (Kreuztisch) zum laufen zu bringen.
Dafür würde ich gerne zwei AC Servos verwenden, komme aber nicht mehr weiter.
A6-Serie 400W RS485 AC-Servomotor-Kit 3000rpm, 1,27Nm mit Bremse, 17-Bit-Absolutwertgeber IP67 - A6-RS400H2B1-M17 | StepperOnline

Ich bin mir nicht ganz sicher ob das Problem an meiner (ersten) Linuxcnc Installation bzw. Konfiguration liegt, oder am Servotreiber.
Hardware ist ein PC mit Mesa 7i96 (ohne s) Karte, die per discovery auch erkannt wurde. 
Wenn ich im Pncconf wizard bei X Motor "Test/Tune Axis" versuche oder über die UI den Motor im Jog Modus bewegen möchte passiert nichts.
Der Motor bleibt auch frei beweglich und hält die Position nicht

Im Gegensatz zur Software von Stepperonline, wo der Motor sich manuell ansteuern lässt.
Hier kann ich aber in der Software auch den enable Schieber aktivieren.
Ein Enable Pin wird aber nicht explizit erwähnt, auch hier im Verkabelungsbeispiel nicht:
A6-Serie 400W RS485 AC-Servomotor-Kit 3000rpm, 1,27Nm mit Bremse, 17-Bit-Absolutwertgeber IP67 - A6-RS400H2B1-M17 | StepperOnline

Ich hätte den Treiber differentiell angeschlossen, wie im Datenblatt des Treibers auf Seite 44 dargestellt.

Bin ich zu voreilig den Motor direkt drehen lassen zu wollen und sollte erst mal tiefer in die Konfiguration meiner Maschine einsteigen, oder übersehe ich beim Servoantrieb selbst etwas?
Leider zwei unbekannt, da ich bisher nur GrblHAL in Verbindung mit CL57Y Treibern verwendet habe.

Viele Grüße,
Martin