Testing IGBTs?

[BigMax]

Hi,

 

Over the past few years, with reference to 'larger' (ie >30kW) VSDs, I've noted a migration from 'dumb' IGBT modules to 'smart' types and back again!.

 

By dumb, I mean basically 3 connections, Gate, Emitter, Collector with no inbuilt Overcurrent or SOAR protection.

 

By 'smart', I mean with comprehensive inbuilt protection, with two power connections and a multipin 'interface' to the rest of the VSD (or other product).

 

I see each having inherent advantages and disadvantages. The smart types rely on protection designed by the people who probably know best, the IGBT manufacturer themselves, the dumb types rely on external protection circuitry. The smart types cost many $, the dumb types are relatively cheap. Catastrophic failure of the smart types usually leaves the external driver circuitry undamaged, catatrophic failure of the dumb type can cause cascading damage to external circuitry.

 

Which all brings me to my point / question.

 

When faced with IGBT failure on a large and complex drive, consideration needs to be made regarding replacement of all IGBT modules (up to 12 or more!) or risk having to spend time dissassembling and reassembling when static tests didn't reveal all the failed modules. Blindly changing all modules when not all have failed can be expensive.

 

Testing of smart modules external to the VSD has not proved practical (for me anyway), but testing of dumb types is a practical proposition. Now that the dumb types are making a return, here's what I suggest (assuming 'enhancement mode' IGBTs, are there other types?).

 

Connect a 100W lamp in series with the IGBT collector and +ve of a 200VDC power supply. Connect the IGBT Emitter to -ve of the same power supply. Short the IGBT Gate to the Emitter. Apply power, no current should flow, the lamp should be dim. Remove power. Connect another 0-30VDC power supply +ve to gate, -ve to emitter. Set this power supply to 0V. Apply 200VDC power, slowly increase the other power supply Voltage from 0V upwards, a good IGBT should go from 'off' to fully 'on' over a few Volts once a particular Voltage range is reached (anywhere from 6-20V depending on the type, compare to a known good example to be sure).

 

Without being too dramatic, I'm sure other forum members know what they're doing (), but 200VDC can be LETHAL and the IGBT gate can be subject to static discharge damage if care is not taken.

 

ANYWAY, what do other members think? What IGBT test methods do you use?

 

Cheers!

 

BigMax.

[GGOSS]

Great info BigMax,

 

Another source of useful information re the testing and failure of IGBT's is as follows:

 

http://www.ab.com/drives/techpapers/ieee/29_2.pdf

 

Regards,

GGOSS

[karthik_eeeng]

;q.i am beginer in electrical & electronics.so any one pls help me to improve by basics.