DIY VVVF Inverter¶
Introduction¶
So it's no secret that I really enjoy the sounds various inverters make. The only problematic part is that most of these inverters are located in other countries, which makes listening to them in-person slightly challenging.
Therefore, the obvious course of action is to build my own inverter that imitates the sounds of these other inverters.
I'll freely admit that I'm not the first and probably not the last person to do this - if you search for vvvf inverter
on youtube, you'll find plenty of people who have made their own inverters.
Anyway, I set out to create something that would hopefully power a small induction motor that I picked up at my local scrapyard.
Process¶
So, to start I found what parts I had laying around, which included this old 2U security server that I also found at that same scrapyard.
I thought that this would be a good starting point to build my inverter from, so I got to work removing the interior and planning everything out. Don't worry, the machine was really old, so it's not like any good compute power was wasted.
Anyway, after removing all of the old parts inside, the chassis looked like this:
Now, it was time to plan out the important parts - this was the harder part.
Firstly, I had some large IGBT modules laying around from a previous project, so I wanted to design this project around those. I got the modules out of their storage box and measured everything out to plan out how this would all fit.
I also needed to figure out an effective way to snub the inductive kick caused by quickly turning off the IGBTs with a large inductive motor winding connected. In the past, I've accidentally killed one of these modules by improperly snubbing the circuit.
So, I got on mouser and found some capacitors that should be sufficient for my use case. In case you'd like to learn more about snubber capacitors, here's a good resource.
To get AC in and motor connections out, I mounted two terminal blocks in the back of the chassis.
After some initial thoughts, I realized that I had some server fans laying around that I could use to cool the inverter. I was planning to design something likely very overkill, so I wanted to put in adequate cooling.
I planned to get a large aluminum heatsink and attach the IGBTs to the heatsink, then use front to back airflow to pull air through the heatsink and cool the IGBTs.
Unfortunately the fans I had were just a few mm to big to fit in the chassis, so I had to mount them at the back with a small piece of aluminum angle bar stock.
Next it was time to mount the heatsink, but before doing so, I needed to add in the wires to/from the terminal blocks.
This just took a moment and was done with some 12AWG wire I had laying around. It's a bit silly compared to the power capability of the IGBT modules, which are rated at 800A, compared to the 20A rating of the wire, but then again the whole project is overkill.
I also took the time to drill the mounting holes for the heatsink.
Finally, I was able to mount the heatsink, which weighed quite a lot.
I drilled four mounting holes in the chassis and in the heatsink, then threaded in some M5 screws.
Naturally, I removed all the metal shavings prior to energizing anything with a leaf blower.
Before I could mount the IGBTs, I needed to assemble the snubbers onto the modules, which was done with patience and more soldering.
Here are a few pictures showing that painful process.
Truthfully that process wasn't particularly difficult, I'm just not very good at soldering and kept melting off the plastic coverings on the terminal connectors before I figured out you could just remove them. I also forgot to get heatshrink tubing that was the right size, so I just used electrical tape to cover the connection of the snubber capacitors to the wires.
Anyway, it was finally time to mount the IGBT modules, so I unscrewed the snubbers and began to mount them.
I made sure to use thermal paste so that they would conduct the heat away into the heatsink where it would be rejected to the air.
At this point, there were a few things left to do - such as adding the rectifier, filter capacitors, and bus bars to connect everything together. I unfortunately forgot to take a lot of pictures at this stage because I got too excited, so here's a picture of the final result. Let's just say that a lot of time was spent drilling and taping holes in the heatsink, and figuring out the bus bar nonsense.
With that, the main power part of the inverter was done. In the next page I'll go over my trials and tribulations with the gate drivers.