[bobeson]

I wanted to share my experiences with tramming my PCNC1100 back into position following a pretty bad crash I had a few months ago. The crash happened during a rapid z down, while the spindle was not rotating. This caused my 1.5" ktool insert mill body to strike a glancing blow off the top left side of my vise, and then continued downward after getting forced to the side of the vise, until the right side spindle shoulder bottomed out on the vise and stalled the Z axis motor. Ouch. This was thanks to a missing decimal point during MDI movement. D'oh!

Anyhow, the first thing I noticed after cleaning everything up and restarting was that I had a terrible surface finish problem. Eventually I realized this was happening because my machine had gottten thrown pretty badly out of tram by the crash.

At first, I tried the method advocated by Tormach where you simply place shims under the machine base, between the machine and the stand. This worked pretty well to put the machine in tram, or so it seemed, but I still had bad surface finish. When I went to re-measure the tram, I realized that I had tweaked my machine base into a helix, because I had pretty large shims under the machine. The result was that the machine was in tram when the table was in the middle of the Y and X travels, but then went out of tram when Y traversed, and also somewhat when X traversed. I think shimming under the machine base is a good bet for small errors, but it seems that the magnitude of my shimming was too much, and I was getting a twisted machine base as a result.

So, I switched to shimming under the column. I started by using a 4x12" cylinder square mounted on the y-axis ways, and using an indicator on the spindle to determine the angle of lean of the column wrt to the y-axis ways. Once I had measurements of the X and Y axis lean angles, I designed two shims, one for each side of the column base. Then I unbolted the column from the base, and used an engine hoist to lift the column while I put a scissor jack under the head near the spindle, and coordinated the lift of the two to make sure I pulled straight up. There are two locating dowels that connect the column to the base. I was careful not to pull the column off the locating dowels; rather I lifted it just enough to slip the shims underneath, and then it went back down easily.

Unfortunately, the results weren't so great. I was still out of tram. Then I realized that referencing the Y-axis ways was a mistake, and the actual table plane is all that matters W.R.T. the column angle. So, I pulled out the cylinder square again and measured the lean angle of the column with respect to the table for both X and Y axis directions. Then I made new shims to compensate for these angles, and installed them under the column after removing the first shims.

This time, the results were fantastic. My table is now square to the column within close tolerances for both X and Y, and my surface finishes have become better than when I first got the machine.

In summary, if you need to tram your PCNC1100, and the errors are relatively small, I would start by using small shims under the machine base front pads, as described by Tormach. If you need more than ~0.050" of shim, you're probably distorting your base excessively and you oughta try shimming the column instead. I'd be happy to walk anybody through it if they need help, or share photos I took of the process if anybody is interested.

[keithorr]

This was thanks to a missing decimal point during MDI movement. D'oh!

would soft limits have prevented the crash?

[LENS_CAD]

I have had a few minor crashes during MDI rapid moves. I know how it's easy to make a typing mistake.

I usually jog rapidly near the desired location with the jog/shuttle control and then use MDI with G01 to get the precise location.

Len <><

[bobeson]

would soft limits have prevented the crash?

Yes, I believe they would have, since the value entered was -03 instead of -0.3, which would have put the spindle through the table if it had completed.

[twocik]

"would soft limits have prevented the crash?"


How would this be setup or does Tormach sell this ?

[philbur]

Google - "soft limits" mach3 -

Phil

"would soft limits have prevented the crash?"


How would this be setup or does Tormach sell this ?

[keithorr]

"would soft limits have prevented the crash?"
How would this be setup or does Tormach sell this ?

Soft limits would have helped. I don't know what Tormach offers. I (and others) no longer use the locked down Tormach version of Mach. We use Mach3 with Tormach screens to get access to all the Mach features available. Soft limits being one of them. No warranties implied or otherwise.

[zephyr9900]

I don't see how soft limits could have prevented the crash. The vise is within the machining envelope. The crash was the toolholder, and ultimately the spindle nose, against the vise, not the machine running out of travel.

Randy

[mactec54]

No soft limits would not of stopped the crash, unless you had set it for that tool

You could set your Z axes soft limit so that the tool could not hit the vice, But then you would have to set the Z axes soft limit every time you changed the length of of your tool.

The soft limits are normally set just in front of the hard limits

[keithorr]

When I typed G1 X-50 F1 into the MDI box, I got a message on the message line (not a popop box) that said "Softlimits System Movement Aborted" so the softlimit does proactively protect even MDI commands.

Thanks zephyr for checking that out. I was operating on a distant memory; hadn't had the situation for a long time.

No soft limits would not of stopped the crash, unless you had set it for that tool

If mach knew the tool was loaded and the tool table was current and the soft limits were set to protect the table, then the crash would have been avoided. The OP stated the move would have put the tool through the table, which would have been set at the lower z boundry. Yes I agree soft limits won't protect the vise while running g code if the vise is inside the soft limit envelope.

[mactec54]

Soft limits are active as soon as a machine is homed, IF they have been SET UP, It does not matter if you are in MDI, AUTO Etc, They are always active,

[mactec54]

keithorr most users are not using the tool table so Mach would not know if the number GCode you put in your program is outside the boundry, the soft limit would not help unless it had been set so the Zaxes can only go down so far & not crash the tool into the table it would of still hit the vice

[zephyr9900]

keithorr most users are not using the tool table

?!? I would think that the vast majority of Tormach users are using the tooltable, Mactec54. I don't think I've ever seen a picture of a Tormach in operation that isn't using TTS holders...

Randy

[keithorr]

keithorr most users are not using the tool table so Mach would not know if the number GCode you put in your program is outside the boundry, the soft limit would not help unless it had been set so the Zaxes can only go down so far & not crash the tool into the table it would of still hit the vice

Even without a tool offset and/or work offset, the system would still protect the machine.
If I had no values entered for tool offset, nothing for work offset and was just using the machine in machine coordinates;

Workpiece is 2 inches off the table, mill in 1 inch in length. I would have to program the G code or MDI to Z3.0 to get the tool to touch the top surface of the workpiece. If the Gcode or MDI had Z-3.0 (3 inches below the table) the softlimits would throw an error and not allow the machine to begin the move.

In this particular instance. The OP entered a value below the table. I don't know if he had offsets or not. Never mentioned. Doesn't matter. If he had softlimits set up and had zeroed the machine at the beginning of the session, the crash would not have occured. Mach would have thrown an error message and not allowed the machine to move at all. No move, no vise crash, no nothing. It looks ahead; does not wait until it is at the limit to stop the machine. Proactive, not reactive. What can you say now, that nobody zeros their machines in the morning? If they don't, they've earned a crash.

And yes to zephyr, the one inch buffer decel zone is only in play during jog. You can mill right up to the soft limit at a rapid after which the limit switch comes into play. (assuming you zero the machine and do all the other things I know you already do)

Happy Holidays to all. I'm going sailing.

[zephyr9900]

Workpiece is 2 inches off the table, mill in 1 inch in length. I would have to program the G code or MDI to Z3.0 to get the tool to touch the top surface of the workpiece. If the Gcode or MDI had Z-3.0 (3 inches below the table) the softlimits would throw an error and not allow the machine to begin the move.

I dispute that, Keith. (I'm going to take the opportunity to pile on you while you're out on your boat. )

On the Tormach I observe that machine coordinates take Z0 to be at the upper limit switch, so in machine coordinates all Z is negative. It is just the magnintude of the negativeness that the softlimit is looking at. And, as I just posted, with Tormach's stock lower Z softlimit the spindle nose could be 3" below the table before the lower Z softlimit was triggered, if it were not for the limit switch.

Randy

[Scott_M]

Doesn't matter. If he had softlimits set up and had zeroed the machine at the beginning of the session, the crash would not have occured. Mach would have thrown an error message and not allowed the machine to move at all. No move, no vise crash, no nothing. It looks ahead; does not wait until it is at the limit to stop the machine. Proactive, not reactive.

Lets pile on Keith while he is out sailing.


I just enabled soft limits in PCNC rev. 3.12b and did a few tests. With the machined homed ( Positive in all 3 axis )After ensuring that the soft limits worked while jogging I typed G1 Y-100.00 F3 in the MDI and Y started moving in the Y- direction. I let it go and it went all the way to the soft limit and stopped.

It did not look ahead and throw an error message. It was reactive not proactive.


Added- I just ran some g-code from file and it did error out and throw a soft limit warning before moving.

Is this how your machine acts Randy ? Or did I set my button up incorrectly ?

Or maybe the Tormach Flavor of Mach3 acts differently ?


Scott

[zephyr9900]

Is this how your machine acts Randy ? Or did I set my button up incorrectly ?

Scott, I don't have a softlimits button on my screenset, so I just used Mach XML Viewer to set SoftLimit to 1 in my XML. My machine respects the softlimits on both directions on all axes for jogging, but I can type ridiculous values into MDI and the machine will move until the limit switch. But I've been messing with reconfiguring soft limits so much that I don't know what I've done... Right now soft limits are only effective for jogging.

Or maybe the Tormach Flavor of Mach3 acts differently ?

As far as I know, I am using the Tormach flavor of Mach3. I'm using PCNC 3.12b, and only changed the stepper and spindle motor tuning for my Old Skool machine (original stepper drivers and VFD). I have a fresh, untweaked install of 3.12b on my desktop PC, and I'll go through the motor tuning tweaks again and generate a fresh XML to try.

Randy

[mactec54]

Scott You have set it correctly, & that is how it is supposed to work. It is ony a digital stop, That you can ajust if you want, to work anywere in the machine work envelope, it is always better to hit a digital stop than a hard stop.

[mactec54]

Randy you want to give your soft limit a least .100 before your hard stops (anolog switches) you should be able to drive the machine at full rapid to the soft limit & it will stop right there, You do not need a deceleration before a soft limit, only the hard stop you can have the deceleration

[zephyr9900]

You do not need a deceleration before a soft limit,

For a controlled stop (which, according to my understanding is the purpose of a soft limit in the first place) I need at least the .125" of deceleration that Tormach has programmed for the axis motors. I agree that 1" is probably overly much, but it does give a good audible warning when approaching the limit, and as I discovered doesn't affect gcode operations in the deceleration zone.

If the soft limits didn't have a controlled stop, they'd be of no more use than the limit switches. If I hit a limit switch, I need to re-reference the machine (and pick up X-Y-Z on my workpiece again) before I can continue, since it is a sudden and uncontrolled stop with probably lost steps on an axis.

What's your rationale for not needing the deceleration zone, Mactec54?

Randy