Message-ID: <003c01c06205$fadb2bc0$e29c01d4@g4jnt>
From: "Andrew Talbot" <G4JNT@thersgb.net>
To: rsgb_lf_group@blacksheep.org
Subject: LF: Re: Re:amplifiers
Date: Sat, 9 Dec 2000 17:32:17 -0000
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>

About 95% correct I reckon Jim, I think the overlap betwen turn off and on
is a function of  the FET itself, and turn off will always be faster that
turn on even for an infinietly steep drive waveform from zero impedance.
But the choke between the devices does absorb this transient (not usually
done in SMPSUs for some reason) but now makes the extra diodes even more
essential.  At first I was unhappy with the Decca gate drive circuitry - a
square wave direct to the gate seemed wrong and why does every SMPSU Dc
restore the drive signal and often use an extra PNP transistor to get fast
turn off ?  But testing really did prove it works better so it looks as if a
few SMPSU designs could be improved/simplified with direct drive.  In
hindsight, a negative drive voltage will do a very good job at turning the
device off !

Thanks for the pointer about HF/VHF sprogs from the midpoint to tank Johan,
It hadn't really occured to me but a well screened enclosure will help a
lot, then a simple LPF  to complete the filtering.

I hadn't come across the current switching topology, might be time to look
at the 'YXM design in more detail.   I'm radiating abt 600W at the moment, a
BPSK test message on 137790 (will turn off well before the VE beacon comes
on at 2000z).  Tested at 500W earlier this afternoon for just over 1 hour.
After fitting a small 12V fan to the heatsink it now runs COLD. In fact the
only components that do get hot to touch are the output transformer core (50
deg) and the tank coil just above ambient.  Strangely the transformer gets a
fair bit hotter than when testing into a dummy load at the same power - due
to a slightly reactive load I guess.

Andy G4JNT


>From 'MBU
>I think this is the main function of the "transient soak" chokes in the
>Decca design - diverting the current into the fast diodes rather than
>through the Mosfet diodes during the switching transient. There is
>not much chance of cross conduction through the mosfets, since
>the transformer coupled gate circuit inherently prevents the gate
>drive from overlapping.