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.
Subject: Re: Problems with MOSFET drivers
Date: Tue, 19 Nov 2002 01:52:43 -0000
References: <email@example.com> <firstname.lastname@example.org> <email@example.com> <firstname.lastname@example.org>
NNTP-Posting-Host: modem-153-11-60-62.vip.uk.com (188.8.131.52)
X-Newsreader: Microsoft Outlook Express 5.50.4522.1200
"Mike Poulton" wrote in message
> On 18 Nov 2002, "markp" said:
> > Actually what you need is an edge detector to determine whether the
> > edge of the drive signal preceeds the edge of the zero cross signal,
> > and decrease the frequency by a small bit if it does. Similarly if the
> > drive signal lags the zero cross signal then the frequency is
> > increased by a small bit. There's no need to measure current, and I
> > don't think you'll be able to do that accurately enough to stay in
> > resonance anyway.
> All of these schemes rely on the feedback winding giving an accurate
> representation of what's happening in the secondary resonant circuit.
> I'm not sure it does.
With good inductive coupling from feedback to secondary it *has* to follow
it, but it relies on good coupling.
> The goings-on of the primary also influence it's
> signal greatly. This makes it much more complicated to deduce what the
> feedback means.
Why? If there is good inductive coupling between the the windings the only
perturbations will be due to things like leakage inductance. With adequate
filtering the feedback winding will represent what is happening on the
output voltage, it *has* to because it is coupled to the same flux as the
output winding. If this were not true by the way your original system of
driving the FETs from this is totally invalid. I'm proposing inserting a PLL
into all this, which is actually a filter to remove such things as leakage
inductance noise spikes that could end up causing wierd oscillations.
> Measuring average current over a fraction of a second
> is easy, and dI/dF near the resonant frequency is quite high. A tiny
> bump of the frequency knob on the function generator drops the current
OK, assuming a high Q on the output this would be true I guess.
> Consequenty, it won't be hard to measure current
> accurately enough. Besides, I want an LCD current meter on this thing
> anyways and the same PIC that does the frequency control can do that,
> > As an aside, I would change the way the transformer is driven. Use a
> > centre tapped primary, and feed this via a constant current source
> > (i.e. large inductor) from the main DC suppy. Then use two fets to
> > ground, one either end (push-pull). Ensure there is about 5us of
> > OVERLAP on the control signals to keep the current flowing. This
> > becomes self limiting for output voltage which is now solely dependent
> > on turns ratio and supply voltage. This might not be what you want for
> > a Tesla coil, but hey..
> Yeah, this isn't really a "transformer" in the normal sense of the word.
> It's a series LC resonant circuit with an inductively coupled driver.
Mmmm, it acts like a mutually coupled inductor, effectively sticking in
energy in phase to keep it oscillating. Nevertheless, it *is* a transformer!
> The turns ratio essentially determines the input impedance, not the
> output voltage.
> The voltage is determined by the Q of the resonator and
> the input power. The physical configuration of the primary is critical,
> and using a split primary will not provide a balanced drive since the
> two halves do not occupy the same space.
Both halves of the primary are equally coupled to the secondary aren't they?
If not, why not?
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup