Sine wave generator - problem

[Matthew266]

Hello,
I'm trying to make sine wave generator for an older VFD which requires 2 sine waves with a 120' phase shift to work properly. I tested the VFD with a commercial function generator and it's in working condition. I then designed my of PCB and tried using that, but unfortunetly it doesn't seem to work. I use a STMF103 to generate a PWM output whitch is then smoothed by a RC low pass filter, and scaled
appropriately by a amplifier in the range of V_MOTOR+, V_MOTOR-.
The commercial unit is a NIU fy3200s.

Here is the schematic of my circuit (in attachment): schematic
Generator output shows the output of my circuit, and comparison the commercial one and mine.
I tried to compare the 2 PCB's but the only difference that i could easly find was that the bought one had 50 Ohm output resistors (2x 100 Ohm in parallel). I tried adding them too but it didnt't change anything. i tried  also swapping the phase shift , and changing them to 240' but that didnt work either.
The schematic of the vfd is in the attachements. I attach my sine waves to pins 8,10 and the GND to 7,9 on page 3.

If anybody has any suggestions what could be the problem I would highly appreciate it.

[Benta]

Why does your "comparison" file show three sine waves?

[Matthew266]

If You look closely there are 4 but the osciloscope uses 2 similar colors for channel 2 and 4 so it's a bit hard to see.
In the attachment You can see where the signals from the 2 channels meet.

(edit) The signals from the two different generators have also different Vpp because in My version I implemented a U/f (voltage/ frequency) characteristic for the motor. While testing with the commercial function generator I used a potentiometer to adjust the voltage.

[Benta]

OK, so why are you generating/showing four sinewaves when your drive only needs two?
You've lost me completely.

[Matthew266]

I added this so maybe someone spots some differences in the outputs of these two which I haven't noticed, and to show that the output of my version really is a sine wave.

[Benta]

Well, your trace #1 and #2 seem to be 120 degrees apart. Everything fine, no? No idea what you want with #3 and #4.

[Matthew266]

Yes they are, as well as trace #3, #4, so why is it that when i use My version ( traces #1, #2 ) the VFD isn't working, and when I use the commercial generator (traces #3, #4) it works.

[WatchfulEye]

Why is the amplitude of waveform 2 is much smaller than waveform 3 and 4? The scale for waveform 2 is labelled as 200 mV/div, but 3 & 4 are labelled 2 V/div

[Jeff eelcr]

Also first picture has 3 and 4 at a different voltage also.
Jeff

[Matthew266]

Why is the amplitude of waveform 2 is much smaller than waveform 3 and 4? The scale for waveform 2 is labelled as 200 mV/div, but 3 & 4 are labelled 2 V/div

As qouted, I implemented a U/f characteristic here and thats why they have different scales.

If You look closely there are 4 but the osciloscope uses 2 similar colors for channel 2 and 4 so it's a bit hard to see.
In the attachment You can see where the signals from the 2 channels meet.

(edit) The signals from the two different generators have also different Vpp because in My version I implemented a U/f (voltage/ frequency) characteristic for the motor. While testing with the commercial function generator I used a potentiometer to adjust the voltage.

[Matthew266]

Also first picture has 3 and 4 at a different voltage also.
Jeff

If You referring to the picture with only the 2 traces, they weren't used as You can see in the picture.

[jwet]

Can you remove your V/F correction, its in software, correct? This has to be the problem.  Your signals look great but the fact that the amplitudes are out of whack is the only difference that I can see.

[Matthew266]

Yes it's in software so I could but it doesn't work even without, and also I has to be implemented so the motor doesnt stall, and if I would set a voltage to high it would fry the motor. And as I said before:

If You look closely there are 4 but the osciloscope uses 2 similar colors for channel 2 and 4 so it's a bit hard to see.
In the attachment You can see where the signals from the 2 channels meet.

(edit) The signals from the two different generators have also different Vpp because in My version I implemented a U/f (voltage/ frequency) characteristic for the motor. While testing with the commercial function generator I used a potentiometer to adjust the voltage.

so I dont think it because of the characteristic. Also the characteristic changes the voltage only when I change the frequency, but it doesnt even work with the default frequency that I set.