[Mr.Chips]

Mach3, Vcarve Pro.
I was using a pocketing routine and it ran through the Gcode perfectly cutting four levels of 0.100" at each depth on three pieces.
When I started the fourth piece it cut to -0.300" just fine but when the Z was to move to the -0.400" level it didn't stop going down until I hit the E Stop button.
On looking around I saw that the display read -0.400" and the Z was probably at the -0.600 level.

This has happened a couple of times before (holes in the table as proof, LOL), but I failed to notice the display reading so I cant say what was displayed

Any ideas as to what happened?

Thanks

[louieatienza]

Mach3, Vcarve Pro.
I was using a pocketing routine and it ran through the Gcode perfectly cutting four levels of 0.100" at each depth on three pieces.
When I started the fourth piece it cut to -0.300" just fine but when the Z was to move to the -0.400" level it didn't stop going down until I hit the E Stop button.
On looking around I saw that the display read -0.400" and the Z was probably at the -0.600 level.

This has happened a couple of times before (holes in the table as proof, LOL), but I failed to notice the display reading so I cant say what was displayed

Any ideas as to what happened?

Thanks

Coupler loose? Bit slippage? Lost position from binding in linear bearings or other things? Wiring issue? Leadnut or leadnut carrier loose? Spindle slipping from holder?

[Mr.Chips]

I ran another piece after the incident and it ran fine, so I don't think it is a mechanical issue, or I would have seen some erratic Z cutting, is this a good assumption?

Could component overheating cause this problem, I had just ran three pockets prior to the problem?

Has anyone else experienced this, and what was the problem?

Thanks

[pippin88]

Likely software / computer issue. I had a box running mach 3 that would intermittently cause axes to just keep going. It seemed to be related to moving all axes at once but was not really reproducible.
I ended up buying a new computer and using linuxcnc and haven't had any problems.

[LeeWay]

I think you can rule out mechanical failure or slippage.
When that occurs, it won't travel as far rather than too far.
Maybe start trouble shooting for electrical noise and temperature.

[fixtureman1]

Lost comm between computer and controller.

[zool]

Three items:

1. It just takes a tiny slight bump of a wire from controller to stepper motor to throw the cutting off. This is in same vein as fixtureman's suggestion, which I think is check all connections.

2. RE: The cabling from the controller to the stepper motors: Is it shielded? Is ground connected only to the controller end?

3. Do you have an uninterpretable power supply [UPS] on the computer that runs MACH3?

The power that comes out of the outlet is not always 110-115 volts nor can it always be a nice clean sine wave, and it sometimes brings radio frequency interference [RFI] with it.

A UPS that can power the computer and the controller and controller power supply is ideal. The inexpensive ones in big box office supply store I would avoid. The ones they are selling now do not seem to last long under continuous use, even in light duty uses. I do not know if this is because the power supplied is "dirtier" or the UPS from the big box office supply stores are now lower quality.

[Mr.Chips]

Thanks Guys,

Yes the stepper wires are shielded.
No I don't have a UPS.
The previous homeowner had a welder in the garage so I tapped into those fat #6 lines splitting them in to two lines one for the router and the other for the PC and control box,a friend checked them and there were extremely good electrically.
What could I do about the RFI interference?
My PC for the CNC is dedicated XP and never on line, to try and keep it virus free.

What does he mean "lost common between PC and controller", electrically challenged here. Sounds like if it was lost it would be lost all the time?

This has always happened after a long cutting period, so I'm leaning toward a temp buildup in my controller cabnet. I just sent a email to CandCNC this morning to help me identify a 5V DC connection on their power board so I can hook up a fan and get better cooling to the components rather than relying totally on the fan that's built into the PS.

Ill make some long cuts and check component temp before and after the new fan install to so I can see what kind of improvement in cooling I get.

Thanks for your to the point comments, appreciate this.
HH

[fixtureman1]

Static build up can have this effect also.

[Mr.Chips]

Static build up can have this effect also.

How would I prevent static build up?
HH

[fixtureman1]

Get a humidifier or water in front of the heaters.should help.

[LeeWay]

Grounding. Ground the machine. Static dissipates to ground.

[Mr.Chips]

Grounding. Ground the machine. Static dissipates to ground.

Thanks Lee,
I did not have ground wires on the various parts, good thought.
HH

[zool]

Mr. Chips stated: "This has always happened after a long cutting period...."

There is also the stepper motors that get hot. I searched here on CNCZone for: "stepper motor overheating". Lots of threads, but this one seems to be the most analytical: http://www.cnczone.com/forums/steppe...g_problem.html

It may or may not be the problem, but the discussion about how to figure out the problem may help you, if only to eliminated a source of problems.

[Mr.Chips]

Thanks Lee,
I did not have ground wires on the various parts, good thought.
HH

This afternoon I used my VOM and checked between the ground at the electrical outlet and all the linear bearing rails, and the router bit, all had a ground connection, so I can check that off the list.

I used shielded cable on all stepper motors.

Was going to do a long run and check the temp as Zool had suggested but the battery in my infrared temp meter was dead so ill have to do that after I pick up a battery. I'll post how that comes out when I get the temps of the various components.

Thank you all for taking the time and offer suggestions.


HH

[jsheerin]

I had this happen to me a few times. The stepper wouldn't move even though Mach3 was telling it to. It would (for whatever reason) happen on Z+ moves most often, so Mach3 would think the bit was higher than it was. Then when it plunged for the next cut, it would go way too far down, go through my spoil board, carve up the bed rails, etc.

It turned out to be a problem with my parallel port. It wasn't supplying enough current to drive the inputs of my G540 correctly. It was very obvious when looking at the output of the parallel port driving a resistor on an oscilloscope. I have no idea how it ever worked to begin with. I found a good test for this was to command a very slow move, like 0.001 ipm. The counters in Mach3 would move but the motor would not. I installed a pci card and only used certain outputs that I measured to have enough current capability and never had another problem.

[turmite]

I am having this same problem right now and am curious if you can check the pp output without having an oscope, since I don't have one.

Mike

[Mr.Chips]

I had this happen to me a few times. The stepper wouldn't move even though Mach3 was telling it to. It would (for whatever reason) happen on Z+ moves most often, so Mach3 would think the bit was higher than it was. Then when it plunged for the next cut, it would go way too far down, go through my spoil board, carve up the bed rails, etc.

It turned out to be a problem with my parallel port. It wasn't supplying enough current to drive the inputs of my G540 correctly. It was very obvious when looking at the output of the parallel port driving a resistor on an oscilloscope. I have no idea how it ever worked to begin with. I found a good test for this was to command a very slow move, like 0.001 ipm. The counters in Mach3 would move but the motor would not. I installed a pci card and only used certain outputs that I measured to have enough current capability and never had another problem.

That sounds kind of reasonable, I just need to get one of my electrical tech friends to check this out, I'm pretty electrically challanged.

Could you spell out the procedure in simple terms, so I can explain it to a technical person and get mine checked out?

Thanks
Hager

[zool]

try a RS232 mini-tester to determine function of parallel port

RS232 MINI TESTER

I had different parallel port problems, but this helped.

Also, try a different parallel cable.

[jsheerin]

If you have a G540 (other controllers are probably different), you need to make sure the parallel port can supply 5mA or more of current at a voltage greater than what the G540 needs to see a signal at. It's in the manual, but I think 4V-5V should be fine (5V is theoretically what it should be). To test a parallel port, you can buy a db25 connector at Radioshack, solder on some wires, and attach some 630 ohm resistors between all the pins (or just one at a time) and ground. So iirc, pins 1 through 17 are signals and 18-25 are ground, so you would put a resistor between pin 1 and pin 25 for example, turn pin 1 on using Mach3 and measure the DC voltage across the resistor with a multimeter. It should be 4-5V. If it's something like 3V or lower, you might have a problem. Just to clarify, when the output is off you should measure 0V. Do this for each pin (1-17) and see if they supply enough current.

V=I*R, so I=V/R, so if you measure 1V/630ohms, that would mean that pin can only supply 0.0016A, or 1.6mA - not enough to drive the G540. If on the other hand you measure 4.5V/630ohms, that's 7.1mA - more than enough.

To further clarify, you have to take the voltage measurement when the pin is loaded by the resistor. If you measure without a load, no current is flowing and thus the voltage should always be 0 or 5V. If the pin can't supply enough current, the voltage will sag down some amount based on the load you have applied.

Hope this helps.