Axes run rough in one direction

[tmead]

Hi ! We've just had our gantry mill delivered from BZT Germany. It uses Leadshine 2278 stepper drives and has been delivered with Mach3. We've switched it on and started running through basic checks. It's not going well - all the axes seem to be really smooth in one direction but noisy and rough in the other.
The Y is even more bizarre, the roughness is such that it loses position and the axes reverses direction. This happens with the motor driving the axis or on its own. The reversal affects all movements after that.
Swapping the individual drives inside the case has not made any difference.

I'm new to all this, so please excuse any lack of background info !

Tim

[Torchhead]

Who furnished the PC and Loaded MACH? Rough running can often be a ragged pulse stream from the PC. Some PC don't play well with MACH3 (especially older versions).
Is it running XP as the OS?

Knowing nothing about the stepper drives or interface I can't comment other than to say that the machine should have been shipped with a MACH profile (XML) that is setup for that machine.

There is a test program in the MACH folder called DriverTest.exe. Run it with MACH not loaded. If the graph is rough (like piano keys) or there are intense bursts of spikes all grouped together you will need to try the optimization steps listed on the MACH Support site.

I don't know how they do it in the EU but we typically look to the company that took our money to provide level one support. They should be able to sort the problems out since the machine should have been tested at the factory.

TOM CAUDLE
www.CandCNC.com

[tmead]

Thanks for your thoughts Torchhead. I'll try to answer as best I can -
We provided a PC, and after several seaches around the web for advice settled on a stable PC (Compaq pentium 4) with a complete cleandown and fresh XP re-install. It's not running anything else nor connected to the network. I've also tried a much more modern Dell 620 laptop with docking station - same results
We are running the XML file that arrived with the machine, I haven't reviewed the contents, just installed it.
Drivertest gave a pretty flat line with occasional very small spikes. I don't know how to interpret the graph/readout any further than that.

I've put a call in to the tech support guys at the manufacturers, but wanted to try all avenues for support, and was worried it might be something to do with our PC as the symptom of rough running in one direction but smooth in the other even with the motors unloaded was odd.

Tim

[Torchhead]

Well, typically that is caused by the motors not being connected correctly (one phase not connected on one side or winding pairs cross connected). The reason I went to the PC is I made the assumption the motors were wired at the factory and worked in test. It's a wiring error on steppers. It can also happen if 8 wire motors are wired wrong. In the drivers you have to get the Step Active Low correct polarity and more important that the drive Common polarity correct (+5 or PCGND). Once again these parameters are for the drives we use like the Gecko 200 series.

If the unit does not have a buffered Breakout board (interface between the PC and the Motor Drives the lower drive current and voltage of new computers could cause funky motion programs. If they are using MACH and the PC for control I am sure they realize this and are using the proper interface components.

As far as software there is not a setting other than those listed in MACH that could cause the problems you are seeing.

TOM CAUDLE
www.CandCNC.com

[tmead]

Thanks Tom, I'll check through this.
The motor wiring is a little difficult to trace, but it looks like this:
the drive has four outputs, labelled A+, A-, B+, B- (no surprises here)
These are connected through to four pins on the motor connector. The motor is an eight wire unit, and the wires are connected in four pairs-

black / green + white -> A+
black + white / green -> A-
red / blue + white -> B+
blue / red + white -> B-

Seemed reasonable to me ?

I'll dig around with the pin config a little.

So far the manufacturer has only managed to send across an updated parameter file with lower velocity + acceleration. Not hugely helpful.

Tim

[PTtec]

I don't realy know much about this control or machine, but normaly this appen in machines with linear scale that have some backlash in the mecanic, try wen moving in the direction that give you the problem push the mecanic in conter direction of movement, our simply disable the linear scales for experience (if tey are availeble of course)
Bruno

[tmead]

The machine is driven with steppers, and no feedback, so I don't think that backlash can be a factor. The issue is apparent with and without loads applied to the motors.

Tim

[PTtec]

other possibility can be hi-frequence in input of the servo that can provide from noise! if you have a osciloscope you can checK the pulses! but normaly the problem would be for the 2 ways of movement!

[tmead]

OK - more info ! The issue is definitely related to the PC end of things. I've had a chance to get a scope on the port and the pulses are all over the place. It seems as if there are some (around ever third/fourth that are pretty stable, but in between these the pulses are unreliable - sometimes missing and sometimes early/late.
I've been into the ACPI and set to standard PC - no difference. The Compaq EVo P4 that I'm running does have onboard video. Could this make a difference ? I'll hunt around for a video card that might fit, not sure if it has an AGP slot.

Interesting thing is that driver test gives a nice straight line with no spikes ?

The issue is exactly the same with the Dell 620 laptop/docking station that I've tried.

I can't believe that this is an isolated case - the hardware is very common and my install is basic. I must be missing something ?

Tim

[ger21]

Have you tried increasing the pulse widths?

[M250cnc]

Another simple thing to try is with the/a motor off the machine.

You can then feel the movement on the table by turning manually and also running a motor again while off the machine as it will then not be influenced by the mechanics of the machine.

HTH
Phil_H

[Torsten]

Originally Posted by tmead Hi ! We've just had our gantry mill delivered from BZT Germany. It uses Leadshine 2278 stepper drives and has been delivered with Mach3. We've switched it on and started running through basic checks. It's not going well - all the axes seem to be really smooth in one direction but noisy and rough in the other. The Y is even more bizarre, the roughness is such that it loses position and the axes reverses direction. This happens with the motor driving the axis or on its own. The reversal affects all movements after that. Swapping the individual drives inside the case has not made any difference. I'm new to all this, so please excuse any lack of background info ! Tim

If i where to guess at this, I would check if your slides have tapered gibb adjusters and if they have any play in them.
What can happen with those is that they have some play in the adjusting mechanism and when the slide is moved one way the tapered gib adjuster slides a small distance along opening up the gib and allow it to slide easier, when moveing the slide the other way they tighten up the gib and make it harder to move.
You could test this by moving the table manually by hand to see if the resistance changes depending on direction.
Another test could be to reverse the Axes in Mach 3's setup (X+ to X-) to see if it will run rough the other direction now.
Good Luck