[beernuts]

I do have a Chinese mill that works very well in general but does have some
noise issues. There are some mistakes made with the wiring, kind of obvious
stuff like non shielded cables for the VFD, high power cables running across
the PP in plug and stuff like that.
I start the G code, the spindle starts and the coolant starts. X/Y move to
the first cutting position but will not reach it precisely as for short
times the motors stall instead of traveling while the spindle rev's up. This
is not influenced by the speed I travel. I can see the stopping/stalling on
the Mach3 screen too.

I manage to get around the problems by starting the spindle and coolant
while in PAUSE after moving to the first X position. This works ok for now
and I will look into the noise issues very soon.

For now I would like to know what settings can influence/enforce the noise
sensitivity. The following points are of interest to me:

1) if the acceleration is set too low, can that influence/lead to motor
stalls?
2) if the timing for step pulse and direction pulse are set wrong, can that
have an influence?
3) the signal settings for Dir low active / Step low active , can they have
an influence? How can I set these right?

Any advice welcome,___

[ger21]

1) Only if your motors have resonance issues at lower speeds. Sometimes faster acceleration can help with resonance. But, if resonance is not an issue, then no, lowering the accel shouldn't cause any problems. However, faster acceleration is usually desirable.
2) Probably not stalls, but possibly. Your steper driver manufacturer should have a minimum recommended pulse width that the drive needs to work properly.
3) Dir low active will just reverse motor direction when changed. Step low active can result in lost steps if not set properly. Again, the driver manufacturer should specify the correct setting.

[beernuts]

thanks a lot for the quick and detailed response, Gerry.

The drives have no names and no brand. The machine supplier provides a XML file with even the most basic settings wrong so I can not trust any of those.
For about 6 months all went quite well besides a few noise issues (eg stopping). I increased the debounce and things calmed down. But suddenly this VFD noise issue appeared with no apparent reason. That's why I looked deeper in the box and discovered that the wiring is lousy.

I did read most of the noise related posts but became confused over time, mainly because of the terminology used. I do understand some of it but would like to get things a bit more to the point.

I understand the advice about a star connector. But what exactly should I connect to it?
Here as I understand it:
1) I connect Earth that comes from the power point
2) I connect all the shields from the motor power cables and the switch cables and so on
3) It is best to connect the star connector to the case

Now I start to get confused. I read about Common of the Power Supplies but am not sure what that is. My power supply is connected to the earth via the power point. Do I make a connection from that earth to the star connector?

Then I read about the common of the PS, is that actually the - (negative)?? Does that mean I connect the PS - (negative) to the star connector?
The PS supplies the motors as well as 5V to the BOB.

I can also see a grounding point on the BOB. Do I connect that to the star connector too?

Any help much appreciated.

[RCaffin]

I understand the advice about a star connector. But what exactly should I connect to it?

Mains Earth (of course), power supplies 0 V, all screens, and yes, attach the star to the case. 'Power supply Common' is usually the same as 0 V, or the more -ve rail.
Do NOT connect the other end of the screens to anything. Screens should be single-ended. If you wish to earth a motor for instance, run a separate earth wire inside the screen.

Earthing is tantamount to a black art. There are theories ... but some hacking is usually also required.

Cheers

[Al_The_Man]

There have been a few papers recently on grounding both ends of the screened cable, this is contrary to what was originally an attempt to avoid induced noise created by the dreaded term, 'Ground Loops'.
It is now recommended by EEC and others that it is Very important that Equi-Potential bonding take place, this normally eliminates the stress on grounding shields at one end only due to this practice eliminating ground loops when carried out properly.
http://www.automation.siemens.com/do.../emv_r.pdf?p=1
Al.

[tvan56]

I have been looking at some drives recently and with
one mfgr. I noticed they recommended grounding both
ends of the shields. I thought that was odd with past
tradition. Must be the new thought.

[Al_The_Man]

Must be the new thought.

Yes, but it must go hand in hand with the equi-potential bonding practice.
See ch6 in the Siemens link I posted.
Al.

[RCaffin]

There have been a few papers recently on grounding both ends of the screened cable, this is contrary to what was originally an attempt to avoid induced noise created by the dreaded term, 'Ground Loops'.
It is now recommended by EEC and others that it is Very important that Equi-Potential bonding take place, this normally eliminates the stress on grounding shields at one end only due to this practice eliminating ground loops when carried out properly.

If there are ground loops potentials and you earth both ends of the screen you may get severe problems. Don't.
But DO run a separate heavy earth wire in parallel to and outside the screen to all motors etc. That is Equi-Potential bonding, and it does take the strain off the screen.

In the mainframe computer area you may find that the metal cabinets are tied together with what is essentially flat battery earth braid. Serious stuff!

Cheers

[Al_The_Man]

If there are ground loops potentials and you earth both ends of the screen you may get severe problems. Don't.

On the other hand it the bonding is done correctly there should be no ground loop potential!.
Al.