G540 > 0-10v > VFD > Spindle Speed Wobbles

[PEU]

I have a G540 all wired up and working connected to a VFD, PWM freq set to 50hz and 5% and mach3 calibrated to issue the correct speed via M3 commands.

Voltage is feed to the G540 using a 7812 regulator (user manual states 12v is allowed) and as expected input reference voltage is stable.


The problem I´m seeing is that the motor speed display at the VFD wobbles about 100RPM and I see this replicated at the true speed from the index pulse in mach3.

Is there a way to smooth this?

When I did my own analog PWM to 0-10v converter I just increased the frequency and the wobbling was gone or was less than 5rpm.

Any idea would be appreciated. Thanks!


Pablo

[morestuff]

The 50 Hz update rate is not cast in stone. The G540 uses
a low pass filter to smooth the PWM, so a faster PWM
frequency could be used. It is likely that 100 Hz will result
in less wobble at the expense of slightly poorer linearity.

Steve Stallings

[PEU]

Thanks for the reply Steve.

Tried 100hz at mach3 and the wobbling remains, I wonder if there anything else I can do.
Since the machine is a small lathe I don't have problem with a little nonlinearity, but I do need steady speed for any 0-10v that goes to the VFD.

When I used your pmdx126+pmdx107+geckos 203v it worked like a charm, but for this machine, this combo is out of budget.

Thanks!


Pablo

[morestuff]

Pablo,

Is the 12 volts being used to provide the reference to
the G540 totally isolated from the main power supply for
the G540 and from the power mains? Normally the G540
gets its reference from the VFD and this is fully isolated
because the signal paths in the G540 are optoisolated.

Can you use a battery powered (totally isolated) voltmeter
to measure the analog voltage going to the VFD?

Is it stable?

What percentage of the target RPM is the 100 RPM wobble?

Regards,
Steve Stallings

[PEU]

The 12 V comes from a completelly separate winding in the custom made transformer for the driver, the 7812 heatsink is electrically isolated from the cabinet ground.

I had to use this supply because the VFD does not provide 10V reference, only 24V.

Yesterday I checked the voltage going to the VFD and it wobbled too, did not write it down, I recall about 0.3~0.5V but to be sure will recheck tomorrow.

I will also check with a couple of AA bateries to feed voltage to the VFD to rule out problems there (I highly doubt there are any problem at the vfd)

Thanks for the debug session Steve!

[PEU]

Last year I abandoned the idea and used a pot to solve the problem, now I´m back at it with a lot more debugging done:

To Steve Suggestions:

Now I set up a 10v reference power supply, its only purpose is to provide a reference voltage. Simple circuit: separate winding from power transformer, rectifier bridge, cap,7810,cap output.
Also tested the VFD with a "variable" (***) power supply and it does not wobble at all when connected this way.

I think I pinpointed the problem:

If I disconnect the g540 output going to the VFD and put a multimeter there, the reading is what you should expect, mach3 top speed is 1500 rpm and output is 10v, speed is 750rpm and output is 5v, so there is a direct correlation as one would expect.

Problem arises when I connect the output to the VFD, now 1500rpm (should be 10v) becomes 6.8ish volts and its linear all the way down. I suspect the VFD load (impedance=10K according to manufacturer local representative) cannot be handled by the g540 output.

Can this be the problem? I thought of a voltage follower to solve this, but I´m fully open to suggestions.

but (***)

The 0-10v where made possible with just a 10k linear potentiometer working as a voltage divider from a fixed 12v supply, this kinda confirms that the VFD needs very low current to work, otherwise the pot race would melt in an instant.


Thanks!!


Pablo

[john_100]

Hi Pablo,

I think I'd first try powering the G540 analogue VFD circuit from a
9V pp3 battery and only connect the G540's terminals 7 & 8 to the VFD

if that works try powering the G540 from the VFD
either using a dropper resistor and 12V zenner diode or a 12V regulator IC to
reduce the VFD 24V output to the 12V needed

may be the problem is leakage currents flowing from the VFD via capacitive
coupling between your transformer secondary winding and the earthed core or the primary winding

John

[PEU]

Originally Posted by john_100 Hi Pablo, I think I'd first try powering the G540 analogue VFD circuit from a 9V pp3 battery and only connect the G540's terminals 7 & 8 to the VFD if that works try powering the G540 from the VFD either using a dropper resistor and 12V zenner diode or a 12V regulator IC to reduce the VFD 24V output to the 12V needed may be the problem is leakage currents flowing from the VFD via capacitive coupling between your transformer secondary winding and the earthed core or the primary winding John

I will not have time with the setup until tomorrow, will check with a 9v battery if that solves something, it should

At some point in the dubugging of this setup I tought of using the regulator on the VFD 24v output but it slipped away in my memory, thanks for suggesting it, this should work since there would be complete isolation from the gecko supply. Im sure I have a buck converter sample somewhere to do this more efficiently, but if not a 7810 and little heatsink will do.


Pablo

[john_100]

Hi Pablo ,

I've had a look a the G540 manual but could not find
the current thats required to power the VFD dive circuit
but I imagine its only a few mA

John

ps don't forget to insulate the test battery as I expect
the VFD terminals are live

[PEU]

Originally Posted by john_100 Hi Pablo , I've had a look a the G540 manual but could not find the current thats required to power the VFD dive circuit but I imagine its only a few mA John ps don't forget to insulate the test battery as I expect the VFD terminals are live

The VFD requires very little current, I used a 10K pot alone as a divider for a 12v PSU and it worked fine for many minutes that lasted the test.

[john_100]

Hi Pablo ,

its the supply current for the G540's frequency to voltage converter I was looking for

It will depend on how simple the converter is a simple op-amp based circuit or an IC like the LM2917

the control current into the motors VFD is probably 100uA to 1mA ??

John

[PEU]

I just finished routing/drilling/soldering a small pcb for a TI PTN78000w point of load step down regulator, these are really nice devices, add a Vset resistor and a cap and you are good to go!

Tomorrow I will test it and post results, also will test the 9v idea you proposed.


Pablo