[slimneill]

Our machine has Kflop sending Step & Direction to 3 Viper drives what are controlling 3 DC Servo Drives for X,Y,Z axis.
We have Kanalog and 1 Konnect board hooked up in the system too.

We have one Kanalog DAC connected to the spindle to control Spindle speed.

We have been experiencing the machine loosing it’s position and not returning to the same place (adding or missing steps). This issue only started happening after we hooked the spindle VFD to the Kanalog DAC. After many weeks of trouble shooting we found that the issue only occurred when the wire coming back from the VFD to DAC Ground was connected. If we disconnected this wire, the repeatability went back to normal. When we hooked back up the DAC Ground and spindle control the positional repeatability went bad. We put a LED in series with the ground wire at the Kanalog terminal and the repeatability issue seems to have gone away.
The ground wire is shielded and connected and grounded as specified in the Hitachi VFD manual.

Is it possible that the ground was picking up interference and somehow affecting the Step/Dir output on the Kflop? Is there a more permanent fix that we should implement? The LED was the only thing lying around that seemed to work.

Thanks,
Dan

[TomKerekes]

Hi Dan,

Yes I think that would indicate noise being introduced into the KFLOP/Kanalog GND and affecting the Step/Dir signals. It sounds there might be a ground loop problem. Grounding is very complex. Its hard to make suggestions without knowing all the details of your system's components and all the Grounds (DC and earth) and what circuits are isolated. Are the Viper Step/Dir inputs isolated? Can you describe the Step/Dir wiring? Shielding? What mode are you using (LVTTL or open collector)?

Kanalog Analog Outputs share a common DC GND with KFLOP (not isolated). KStep has one Analog Output that is isolated. You might consider using that. It isn't as linear or as high resolution as the Kanalog Analog Outputs but works well for basic speed control. So that is an option you may consider.

HTH
Regards

[slimneill]

Hi Dan,

Yes I think that would indicate noise being introduced into the KFLOP/Kanalog GND and affecting the Step/Dir signals. It sounds there might be a ground loop problem. Grounding is very complex. Its hard to make suggestions without knowing all the details of your system's components and all the Grounds (DC and earth) and what circuits are isolated. Are the Viper Step/Dir inputs isolated? Can you describe the Step/Dir wiring? Shielding? What mode are you using (LVTTL or open collector)?

Kanalog Analog Outputs share a common DC GND with KFLOP (not isolated). KStep has one Analog Output that is isolated. You might consider using that. It isn't as linear or as high resolution as the Kanalog Analog Outputs but works well for basic speed control. So that is an option you may consider.

HTH
Regards

Right now: the Vipers recieve Step & Direction from Kflop JP5 Pins 1-8. The cable used is Cat5 Twisted pair cable.
Can Step and Direction be sent to the vipers from the Kanalog? Will it be better islated?

You said above that Kstep has one analog Output that is isolated. We don't have Kstep. Did you mean Konnect?

[gfacer]

One suggestion based on work I'm planning, split the work of the cat5 wire into 2 cat5 cables, 4 internal wires taking data signals and the twisted pair of each of those going to ground on the kflop side. This allows the twisted pair cable to protect the data signal from emf. My understanding anyways, not an expert.

[TomKerekes]

Hi Dan,

The twisted pair doesn't really help if the signals aren't complementary with a differential input.

Kanalog doesn't have any Step/Dir outputs.

Oops I read KStep by mistake. With a circuit you can make two Konnect outputs create an isolated analog signal. But it is a bit complicated. See:
Konnect PWM to Analog Example

Regards

[engnerdan]

I am using a Hitachi WJ200 VFD in analog input mode. A KFLOP and KANALOG with no issues. Here is how i have mine connected

I used a 2 conductor/ shielded (communications wire), it has a 100% foil shield and a drain wire. At both ends of the wire I only have a minimal amount of the 2 conductors exposed (I would guess 1/2"-3/4" at the VFD end). At the KANALOG end I kept the red wire (the analog signal) wire as short as possible and the black wire (analog ground) takes the shortest path to the ground terminals. The drain wire is connected to the chassis ground of my machine at the KANALOG end, as it is advisable to ground shields at the end where the signal/voltage originates. The VFD end does not have the drain wire connected, this prevents a ground loop through the shield.


-Dan

[slimneill]

Thanks,

We reran all the grounds to one common grounding point (we found one power suppply with a floating ground loop) and installed a One to One Isolation Transformer between the Kanalog and VFD. It's called AIM1 One to One Analog Isolation Module.

The previous step issues have all gone away now.

[Jon.N.CNC]

Out of interest, was the vfd screen wire wired at just one end or equipotentially at both ends?

Edit.. Just noticed you already said.

So just to confirm this issue was ac eddy currents from the vfd polluting your earthing network and the isolator effectively filters the ground noise from the other components? Or did I miss something?

[slimneill]

I'm not an electrical guy.

We fixed the grounds and isolated the Kanalog from the VFD and it went away.