Hi,

tried an it made worse an crashed everytime if by filtering you mean through ess on inputs

This is wrong somehow, there is no way increasing a filter time constant on an input should do that.

Have you electrically bonded all the metal parts together?. It sounds to me like you may have a power
supply issue or alternately some leakage current into the frame of the machine.

Try unpluging the spindle speed control from the AC line. If the motor has poor isolation it will
have leakage current into the frame of the machine and will very very easily overwhelm computer level
signals.

Craig

Hi,
this is probably a bit off topic but here goes.....

I have stated that a VFD  often requires line reactors, essentially high inductance filters to accommodate the current pulses that occur in the rectifier/capacitor input circuit.
I have attached two pics, one of the rectifier/capacitor circuit (representative) and its output voltage. Note the ripple on the DC. The second pic is the Fast Fourier Transform (FFT)
of the output voltage waveform.  Note the spectral components at the 2nd, 4th, 8th and 10th harmonics. Low order harmonics like this require a highly inductive filter because of the low
frequencies involved.

I have also stated that a phase controlled bridge such as OP is using is noisy and requires line filters capable of filtering higher frequencies. Accordingly I have attached another two pics,
the first a circuit of a phase controlled bridge rectifier (representative). Note that the gate drive voltage sources are low voltage pulses to turn the SCRs on. The output voltage of the bridge
is also shown. Now look at the second pic being the FFT of the output voltage. Note how the harmonic distortion extends all the way out to over 100kHz.....noisy as hell!

Craig

The last fft certainly gives a graphic indication of how those sharp edges translate into the frequency domain. As you say, it is an aside since the op says it isn't RFI from the spindle drive, but it's interesting to see anyway.

Allan

Hi Allan,
yes its quite dramatic.....when I was cropping the screenshot to post I inadvertanly cut off the frequency scale off the bottom of the FFT pic.
The full picture showed significant harmonic energy to 100kHz and certainly measureable out to 1MHz.

I suspect that OP will find that the speed control is still at the root of his problems, whether he knows it or not. What else can produce the sort of hash this speed control
produces?

One feature of this speed control is that the DC produced is not isolated from electrical earth, at least in New Zealand where we have a MEN electrical system. That means
that the insulation resistance of the motor OP is using is the only thing which prevents line level voltages from getting into the frame of the machine. Additionally there
is a good possibility of the so called, and dreaded 'homopolar' current.

The 5 phase steppers I use on my mill are subject to this phenomenon. They have good insulation to the frame as measured by a Megger.  The drivers are direct feed from the AC
mains to a DC link voltage of 150VDC. The magnetic coupling between the stepper coils and the frame mean there is leakage to the frame, several mA at least. Without earth bonding I have
measured 130V on the frame of the machine......a great surpise to me I can tell you. I have investigated quite thoroughly and it turns out the voltage induced into the frame
is in the kHz range, ie phase switching frequency so the current is induced in the frame rather than resistive leakage.

Homopolar current is not common, there is little mention of it in the literature and that is seems to fly in the face of Kirchoffs Law its little wonder that it gets missed. That's what happened to me
anyway!

Craig