Hi, mikasi
I have built such device two years ago for experimental work. If you want I could send the schematics and the needed programs but you can get the basic ideas from ANALOG DEVICE'S site or from Texas Instruments.
In my device I use a Mitsubishi IPM and an Analog Device's motor control DSP and it works with motors up to 2,2 kW.
Let me know if you need some additional help.
i'm very interrested too, can you share your work with us ? post your data in FM, it would be very interresting for many peoples here...
Have someone any idea about protecting the igbt module agaisnt short cuts on motor lines ? or an efficient technique to control the igbt desaturation ?
The most simple solution, so far I know, is to use MC3PHAC chip from Motorola. It's a DSP with all the firmware inside, you don't need to write any code. It can be serial conected to PC (or MCU) if you like to integrate in an control loop
Back in the good old days when I designed variable speed drives the HEF 4752V from Philips were used a lot in designs. We later went to more uP controlled stuff. I personally designed drives up to 36kW at 380VAC and 575VAC. I am sure there's a lot of new stuff out, but I am not up to speed with that since my retirement. However take a look at the Philips notes under motor control h**p://www.semiconductors.philips.com/acrobat/applicationnotes/APPCHP3.pdf if you need to understand the basics.
Another good site for intelligent drivers is h**p://semikron.com
i'm very interrested too, can you share your work with us ? post your data in FM, it would be very interresting for many peoples here...
Have someone any idea about protecting the igbt module agaisnt short cuts on motor lines ? or an efficient technique to control the igbt desaturation ?
I used to use the following methods in my designs:
1/ I had a gated clamp (to monitor only when device is on) over the IGBT so that I could measure the saturation voltage over the C-E terminal. The moment that went over a certain threshold I knew that the device was in trouble and took the necessary action. This saved a lot of expensive devices in the development stage and later in the field.
2/ I designed an electronic fuse using a extra IGBT into my HT line. Basically you take 2 current measurements from your current sensor in a small time frame, say 10uS apart. When this dI/dT is above a threshold that you set you can predict that there is a short-circuit or another abnormal condition long before it even reaches a dangerous limit and terminate the
supply. So what you are looking for is when the slope of the current-rise is to high, you know there's a problem.
We used to have a crowbar test. Each design should be able to withstand a "crowbar" short over the output at full power! This was a must as our stuff were used in harsh mining conditions.