Re: Flyback Converter

From: 	bmack[SMTP:bmack-at-frontiernet-dot-net]
Sent: 	Saturday, January 10, 1998 10:00 PM
To: 	Tesla List
Subject: 	Re: Flyback Converter

You are delving into a very interesting and tricky subject perhaps
( I am refering to the Motorolla quote). Device input conjugates change
frequency.  For certain wide band applications , a compensation network
would be required to insure stable operation.  This is certainly not the
with TC's.  A simple series gate series resistor provides two functions.
Reduction of Q and current limiting to the gate ( check that data book for
Igmax) , especially if you are using a high performance driver.  
You should also follow Allen's suggestions about the diode clamp
circuit-very important.

As far as efficiecy is concerned, I don't think it's even 70% on a flyback
( I'll check) due to the half sinsiod waveform etc.. BTW use a reverse
schotkey diode  from drain to ground. this should reduce the stress on FETs
during the off cycle.

Running sparkless primaries in NY

Jim McVey
> From: Tesla List <tesla-at-pupman-dot-com>
> To: 'Tesla List' <tesla-at-pupman-dot-com>
> Subject: Re: Flyback Converter
> Date: Saturday, January 10, 1998 7:05 AM
> From: 	RODERICK MAXWELL[SMTP:tank-at-mail.magnolia-dot-net]
> Reply To: 	tank-at-mail.magnolia-dot-net
> Sent: 	Friday, January 09, 1998 9:13 PM
> To: 	tesla-at-pupman-dot-com
> Subject: 	Re: Flyback Converter
> > Greetings max
> > 
> > Message text written by INTERNET:tesla-at-pupman-dot-com
> > >   On each gate there will be a 12ohm carbon composition resistor to
> > dampen inductive ringing in the gates. Also each gate lead will be
> > threaded through a ferrite bead to prevent self-induced ocillations.<
> > 
> > The resistor is fine, but I'm not sure about the ferrite beads - you're
> > shorting the leads to reduce inductance but now with the bead
> > you are creating an inductor - it may be counter productive.
> > 
> > I also have two schottky diodes
> > 0v->|---driver output---->|------12v
> > because I'm using a driver chip and I don't want ringing taking the
> > output negative or too high.
> > I'm not sure about calculating the output -
> > You can calculate the switching losses and the snubber losses but
> >  I think most of the losses are in the output transformer - say 80%
> > efficient.
> > 
> > Have fun
> > 
> > Alan Sharp (UK)
>     Thank you for the information! I did wonder about that. I'm getting
> most of these methods from a Motorola Tmos databook. They mention that: 
> "mosfets high input impedence and high freq. reponse may cause parasitic
> oscillations at freq. greater than 100Mhz when all gates are driven from
> a common node. Without individual gate resistances a high-Q network is
> established that may cause the device to oscillate when operating in or
> switching through the active region. The device transconductance,
> gate-to-drain parasitic capacitance, and drain and gate parasitic
> inductances have all been shown to influence the stability of the
> circuit". 
>   "By decoupling the gates of each device with lossy elements such as
> resistors or ferrite beads, the Q of the circuit can be sufficiently
> degraded to the point that oscillations are no longer possible".
>                      Motorola TMOS Power Mosfet Data   page 2-7-11
>                                   Frankensteins Helper
>                                           Max