The Cyber-Spy.Com Usenet Archive Feeds Directly
From The Open And Publicly Available Newsgroup
This Group And Thousands Of Others Are Available
On Most IS NNTP News Servers On Port 119.
Cyber-Spy.Com Is NOT Responsible For Any Topic,
Opinions Or Content Posted To This Or Any Other
Newsgroup. This Web Archive Of The Newsgroup And
Posts Are For Informational Purposes Only.
X-Mailer: Mozilla 4.79 [en] (Win98; U)
Subject: Re: Problems with MOSFET drivers
NNTP-Posting-Date: Sat, 16 Nov 2002 18:24:59 GMT
Organization: AT&T Broadband
Date: Sat, 16 Nov 2002 18:24:59 GMT
Mike Poulton, Winfield Hill wrote:
>> It likely has to do with what's often called shoot-through, the
>> fact that one FET turns on before the other turns off, causing
>> destructive rail-rail supply currents. The effect gets worse
>> with high load capacitance.
> That does explain the sudden and complete device failure! I'm not
> sure why shoot-through within the drivers never occured to me, since
> I was already concerned about it in the big mosfet H-bridge due to
> the lack of dead time.
Normally a well designed driver IC has circuitry to completely prevent
any shoot through at its output stage. The only way this might occur
is if the IC's substrate is flooded with carriers from excessive reverse
currents driven back into its outputs. This can lead to anything from
slow switching performance (and shoot through) to self sustaining areas
of latch up within the IC. Latch ups may cause anything from a little
unexpected extra dissipation to a degradation or blocking of some or all
functionality to a complete melt down of the IC.
> Perhaps there are transients in the feedback signal that I couldn't
> see on my scope which are causing multiple state changes during the
> rise/fall times. They would have to be 300mV transients, but that is
> certainly possible. This would explain why a clean squarewave from a
> function generator doesn't cause this problem.
I think this is a red herring. See my other post on this subject.
>> Again I suggest using a comparator to get fast-edge logic-swing
>> input signals to your driver ICs, thereby allowing the internal
>> circuitry a chance to safely do its tough job.
> Alright, a comparator will be included in the next version -- along
> with quite a few other new parts.
A comparator is good because, if employed properly, it will precisely
define the switching threshold within the tesla coil feedback circuit
and ensure that both driver ICs are always switched at the same time.
>> I've been meaning to point out that a square wave isn't optimal
>> drive for a resonant circuit, it's much better to use a shorter
>> POS and NEG drive, say 20 to 25%, with an intervening zero-volt
>> or floating period,
> A bit of dead time is good to prevent shoot-through, but I was under
> the impression that an essentially square voltage drive *was* optimal
> for a resonant circuit since it provides zero-current switching. It
> also prevents the free-wheeling diodes from ever having to conduct.
> If you do the switching at any point on the primary current waveform
> other than zero crossing, won't the switching losses and component
> stresses go way up?
Yes, one would normally expect square wave drive to minimize I^2*R
losses, but it may not always lead to zero current or zero voltage
switching or minimum switching losses. -- analog
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup