[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: A prospective s.s.t.c. input-synchronizer (fwd)



Original poster: Gerry Reynolds <greynolds@xxxxxxxxxx>



---------- Forwarded message ----------
Date: Thu, 21 Dec 2006 09:37:56 -0800
From: K. C. Herrick <kchdlh@xxxxxxx>
To: Tesla list <tesla@xxxxxxxxxx>
Subject: Re: A prospective s.s.t.c. input-synchronizer (fwd)

Steve (& all)-

No, I meant the reduction from 120 KHz or so to around 106.  But as I 
posted yesterday, I can see now how that comes about; I have to 
reconfigure somewhat my connections between the IGBTs and the 4 
mains-capacitors (the big blue items in the image I posted) so as to 
minimize the area of the magnetic loop.  And also, bring the leads going 
to the primary closer together, to minimize that loop as well.

I surely agree about the difficulty of probing those switching 
transients.  Especially since they originate in the mains-referenced 
circuits: no common ground with the scope.  I'll likely also give up in 
trying to diminish them; with the air-core current transformer, they 
seem to have been diminished somewhat, anyway.

And as to that xfmr and in response to your other posting of today...

At present I rather like the air-cored one since it appears to yield a 
more understandable performance.  But I will check as to whether I'm 
introducing any excess phase-shift.  At first observation, it would seem 
not: I note that the switching transients occur with about the same 
delay from the transformer-output zero-crossings as with the 
ferrite-cored xfmr..

Regarding the loading of the transformer, you'll note that my shorted 
bridge-rectifier acts to load it in essentially the manner you describe: 
open circuit until the clamping levels (in my case, ~ +/- 1.4V) are 
reached.  One surely wants to clamp the output to some reasonable level 
since otherwise excess voltage could appear, leading most likely to ruin 
of the transformer.  And with a ferrite core, would not symmetrical 
clamping be best?  Otherwise, the hysteresis curve could be adversely 
shifted (and here my nomenclature and even my recollection of 
theory--50-odd years on--may be fuzzy).

Ken

Tesla list wrote:
> Original poster: Gerry Reynolds <greynolds@xxxxxxxxxx>
>
>
>
> ---------- Forwarded message ----------
> Date: Wed, 20 Dec 2006 23:06:25 -0600
> From: Steve Ward <steve.ward@xxxxxxxxx>
> To: Tesla list <tesla@xxxxxxxxxx>
> Subject: Re: A prospective s.s.t.c. input-synchronizer (fwd)
>
> By anomalous Fr, do you mean the 14MHz noise?  That is very typical of
> large IGBTs switching quickly, and is caused by the IGBT C-E junction
> capacitance being charged through the stray inductance of the H-bridge
> layout.  I battled that one for awhile, and eventually gave up, since
> the coil still works fine, even with these harsh bursts of HF noise.
> You have to be particularly careful about probing such noise, as many
> times the scope might say they are worse than it really is.
>
> Steve Ward

[snipped]