[UKRobotics]

Hi all,

I am now nearing the completion of my bridgeport retrofit project however there are just a few irritating glitches remaining which prevent me from using the machine properly.

I think I have narrowed the fault down to interference from the VFD which actually seems to be getting into the computer and making it unstable.

If I run a g-code program I can runn the machine at full speed (5 metres per minute) and I get a fast smooth response and everything appears to work just nicely. This is without the spindle running.

Anything up to about 50% of my max spindle speed and it still runs nicely but as soon as I run the spindle much above 50% the machine starts to respond in an unusual way.

This can take a number of forms:

1. When doing rapid movments the table will judder whilst trying to reach its target location
2. The spindle will ramp up to full speed and then ramp down a bit and then ramp up again.
3. When the spindle gets close to full speed mach 3 will go into e-stop mode stating a driver watchdog error as the cause.

My setup is as follows:

I have a Huanyang 4hp VFD in the rear cabinet of the machine and it is mounted in the top left corner. The output wires are not shielded but all three phases are twisted together and there is only a short (6 inch) exposed area before they go into earthed metal conduit up to the motor. There is an airgap of at least 1ft between the output wires and the pc or any other circuit. The single phase supply wires are unshielded and do run alongside a lot of other wires however it is my current understanding that the input side should be 'clean' and only the output side is noisy. The control cables (forward, reverse and speed are shielded; earthed at the VFD end. The direction wires are connected though cnc4pc solid state relays which then goes to a cnc4pc breakout board. The speed control (0-10v) wires go to a cnc4pc variable speed board which then goes to the cnc4pc parallel port breakout board.

In my opinion the faults seem to suggest noise from the VFD is actually getting into the pc and causing it to output some strange stuff. This would explain the driver watchdog errors, the juddering axes and ramping up and down of the spindle speed.


Anyyway ... basically I'm really stuck with this. I need to solve this problem as I cant use my machine properly in this state but I don't have a clue what to do next. I would be really greatful if somone could help me out here.


Regards,

Dom

[Al_The_Man]

I seem to remember you had this problem previously, Did you try grounding ALL the power supplies as shown in one of the previous links this solved an OP who posted with a similar problem?
The 0-10v for the VFD should still be isolated.
Although I have found most VFD's use an internal logic supply that is referenced to ground, unlike SCR drives that can be referenced to the incoming mains supply, so they have to be isolated for sure.
Did you take a ground conductor from motor frame to VFD to Earth Plate? or just use the conduit?
Also as pointed out previously, try grounding the PC P.S. common direct without relying on the M.B. screws.
Someone also mentioned that US Digital encoders has cause problems in the past.
Al.

[jeffrey001]

VFD produce lot of noise too, try using sheilded cable from VFD to motor supply, separate power cable from VFD from those carrying data, use sheilded cable everywhere.

[UKRobotics]

Al,

You are right, I did post a similar thread and in hindsight I should have posted this there. The difference is before I was unsure as to the cause of the interference, now I'm near certain its the VFD at fault but i'm still no closer to a solution.

I did have a read through this thread http://www.cnczone.com/forums/showthread.php?t=71923 as it seemed particularly relevant however I'm still a bit confused to be honest.

My understanding here is quite limited so I'd be really greatful if you could explain (in as simple terms as possible) what needs to be grounded to what.

In my cabinet I have a number of power supplies.

1 x 140vdc linear PSU for the servo motors
1 x 24v PSU for the logic side of the servodrives, various relays, drawbar solonoid, etc
1x 12/5v PSU which runs my break out boards, sensors, etc
1 x PC PSU inside the computer running Mach 3


The spindle motor frame is earthed with a cable from the motor to the earth point on the VFD which is connected to the earth plate.

[Al_The_Man]

You may have guessed by now that I subscribe to grounding the common of all power supplies, where ever possible.
Although I do not use Mach at all, I do use PC based controllers, that have 5v encoders and TTL inputs going back to the card and are PC powered.
These systems often have VFD's and other high powered devices and 24vdc I/O.
Unless it is Electronically not acceptable due to design, I ground everything, and so far it has payed off.
You could try one at a time and see if it makes a difference.
To test, if you connect a 100ohm resister from common of the supply to be tested to ground and if there is virtually no voltage across the resistor it should be OK to ground it.
There are also 3 phase inductors or chokes that can be wired on the VFD motor output for noise suppression, many VFD manufacturers advise using them.
Some even suggest on the input as well as output.
Al.

[keebler303]

One more comment/suggestion. From the name, it sounds like you are running one of the cheap Chinese VFD's. I wouldn't be surprised if these cheap drives are really noisy (electrically) and have the potential to cause a lot more problems than a "big name" VFD. With that in mind, I would put the VFD in a grounded metal enclosure separate from the rest of the electronics. If that isn't possible, just try to put some grounded metal shields between the VFD and your low voltage signals. In any case, make sure you still have adequate cooling air to the VFD and try to keep grounded metal shields around your wiring.

Matt

[UKRobotics]

So Al, just to double check before I set fire to anything .. You are reccomending I could connect the 0v from every single power supply in my system to a single point and that single point should also be connected to earth ?

That is; with the exception of the 0v supply from the control connectors of the VFD ?

[Al_The_Man]

So Al, just to double check before I set fire to anything .. You are reccomending I could connect the 0v from every single power supply in my system to a single point and that single point should also be connected to earth ?

That is; with the exception of the 0v supply from the control connectors of the VFD ?

That is the method I subscribe to, isolation on the input to the VFD should be OK, Although as I say, I have found many that have isolated I/O (from the mains) anyway.
Do the 100ohm test on any commons to be on the safe side.
Also if you have not grounded the PC power supply externally, it may just presently be referenced to ground through the MB screws which I consider inadequate for machine control applications.
Al.