Re: Recent s.s.t.c work

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "K. C. Herrick" <kchdlh@xxxxxxx>

Steve Ward (& all)-

I'll lay 2 more photos on Terry:

http://www.hot-streamer.com/temp/KCH_TCH8.jpg

http://www.hot-streamer.com/temp/KCH_TCH9.jpg

I've made the change I mention below & also removed the 1.1 ohm 
series resistors in the gate lines.  The -8 image shows IGBT turn-off 
~0.6 us after primary current z.c. for one pair and ~0.8 us for the 
other pair.  The -9 image shows the 2 "lower" IGBT gate signals with 
no mains v. applied.

In -9 you can see that one of the gate voltages is still slowing down 
before the 0V level is reached.  II think that is due to Miller 
effect in the IGBT,  when the "top" IGBT on that side of the bridge 
turns off, pulling too much current, thru the gate:drain capacitance, 
from the gating transistor, which does not have enough base drive at 
that instant or has lower gain.  I'll need to work on 
that.  Otherwise, the drive doesn't seem too bad, do you think, Steve?

You can see that the supply-current swings are unequal, apparently 
because the intervals between switchings are unequal.  That may be 
due to the asymmetry in the gate drives..

Is 600 & 800 ns for switching after primary-current z.c. pretty slow 
compared to your experience?

Other comments welcome!

KCH

N.B.  All of these are photos made with my Canon A80 mounted on the 
Tek C53 housing, and embellished in Photoshop.  The scope sweep is 
expanded and repetitive.  Not too bad, when I'm careful.

Tesla list wrote:
> Original poster: "K. C. Herrick" <mailto:kchdlh@xxxxxxx><kchdlh@xxxxxxx>
>
> Steve (& all)-
>
> Yes, it's a little slow; and at 100 KHz "a little" is too much.  I 
> measure about 1 us between the fall at one IGBT gate (crossing 0) 
> and the beginning of turn-on at its opposite.  The way I have the 
> drives configured now, the gating transistor doesn't start to turn 
> on at all until the driving wave goes above the IGBT's 
> source-voltage level.  I'm going to change that so that that turn-on 
> starts as soon as the driving wave begins to rise from the -25 or 
> so.  That will give more drive to the gating transistor by the time 
> the driving wave gets above the source-ref. level, & hopefully not 
> cause overlap.
>
> All the drivers are working OK; all the same.  And I have currently 
> only 1.1 ohms in series with the gates--& will try getting rid of 
> that once I improve the delay.
>
> KCH
>
> Tesla list wrote:
> > Original poster: Steve Ward 
> > <mailto:steve.ward@xxxxxxxxx><steve.ward@xxxxxxxxx>
> >
> > Ken, what is the impedance of your gate driver?  I have never seen 
> > waveforms that slow and with that peculiar rising edge.  The 
> > falling edge looks normal.  Is your gate drive transistor (the one 
> > that pulls positive) falling out of saturation perhaps???  The turn 
> > on is usually faster than turn off...
> >
> > The really odd thing is that your gate voltage slows down before 
> > you even hit the 0V line!!!  That is *not* right.  Have you checked 
> > each driver?  Maybe this one driver is damaged and also causing 
> > your imbalance in primary current from hafl-cycle to half-cycle?
> >
> > Good luck... in any case, i think that those large turn on delays 
> > are fix-able.  My gate drivers can supply LOTS of current, as the 
> > mosfets i use in it onl have a 55m ohm Rds.  I dont use any 
> > resistance on the IGBT gates, so i switch at the maximum speed possible.
> >
> > Steve