Electronics > Projects, Designs, and Technical Stuff

Single phase variable frequency drive

(1/2) > >>

gztproject:
Hi,

I'm trying to build a single phase VFD for controlling the speed of a 150W fan but somehow it seems I've made an IGBT toaster  :palm:.
I tested the attached schematic on a breadboard using a separation transformer on the input mainly so that I could use scope more freely, but it also limited the current to about 5A. I first tested everything with a lightbulb to represent a resistive load and all the waveforms were looking OK. Switched to the actual motor and it spinned as expected, I was able to control the speed. A thing to note here is that the input voltage at that point dropped to about 100V due to (undersized) separation transformer on the input. There were some spikes on the output from the inductive nature of the motor but the protection diodes (D3 and D4) took care of that.

The problems started when I switched to 230V directly. After the series current limiting resistor R13 turned glowing hot, I quickly disconnected everything only to discover the negative IGBT Q2 was toast. I changed it for a new one and repeated the whole process. Toasting included  |O

After quite some head scratching I didn't get far.

My current working hypothesys is that the parasite inductance of the wound(!) series resistor prevents the line to absorb the spikes from the motor and in turn raises the voltage over the second IGBT, toasting in the process.
I am thinking of putting ~100n capacitors and 400V zenner (or better, tunnel) diodes across +IGBT and MotN (-IGBT and MotN on lower side). Also an 8uF cap in parallel with the output.

Any thoughts would be appreciated :)

The KiCad project is available on Github (along with all the code for the Arduino): https://github.com/gztproject/VFD/tree/master/PCB

Thanks, Gapi

gztproject:
So, there's an update:
After adding all the bypass stuff the results were still pretty much the same, so after some more head-scratching I went over the basics again:

* Conclusion 1: The IGBTs I chose originally (STGD5NB120SZ) are quite specific, they seem to lack the reverse diode and have really high gate resistance (4kohm). Also their max reverse voltage is 20V. Switched to a more conventional IGBTs (IXA12IF1200PB).
* Conclusion 2: I did not take into account the relatively high current spikes when charging and discharging the IGBT gate capacitance. After some research, switched from driving them manually to a dedicated Infineon IGBT driver (1EDI05I12AHXUMA1).
At the moment I'm waiting for the new boards to arrive, will report how it turns out. All the updates are available on github.

-Gapi

NiHaoMike:
Most likely you'll find that changing the frequency to a single phase motor doesn't give very much dynamic range and it would work a lot better if you remove the capacitors from the motor and drive it as the 2 phase motor it really is.

BTW, once you're ready to build a (much) bigger VFD, used Prius inverters can be found for $90 or so. Probably overkill for anything you're likely to do with it at home but it's hard to beat the low cost.

rvalente:
Check these microchip ans:

AN967

AN889

Zero999:
This is not something I've done before.

The schematic originally posted looks wrong to me. It has a voltage doubler on the input, which will give 2×230√2 = 650V, which will toast 400V IGBTs.

The motor should be AC coupled to a half bridge, to avoid DC in the winding which would cook it.

Get rid of the voltage doubler circuit. Use a full bridge rectifier and a full H-brige driver for the output stage.

Is a VFD actually needed? It's just a fan so phase control will probably work, since torque drops dramatically, with speed.

Navigation

[0] Message Index

[#] Next page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod