Re: ND6T AGC implementation for uBIT-X

Jerry Gaffke
 

I'm not so sure that a high SWR is more likely to kill an IRF510.
What characteristic of the RD16HH1 would make it a better survivor?
If we are talking a missing antenna, it's voltage spikes that would kill the final.
An IRF510 is good for double the voltages that an RD16HH1 is.

I haven't thought very much about what happens when an antenna is shorted to ground.

Here's an old thread about voltage spikes at the final:
    https://groups.io/g/BITX20/message/41057
I'm not very handy with magnetics, but my conclusion there is that T11 on our push-pull
uBitx final is a forward transformer, behaves like any filament transformer.  Energy is
coupled through the transformer, not stored in the core.  The center tap of the primary
is at 12v, when one side of the primary is shorted to ground by a FET the other side
will bounce up to 24v.  And one or the other FET is always on.
 
On the Bitx40 with its single IRF510 FET and inductor up to VCC, we are alternately storing 
energy in the core when the FET is on, then releasing that energy when the FET is off.
This is a flyback transformer, the configuration is that of a boost mode switching power supply.
With a high impedance antenna there's nowhere for that energy to go, we can expect some very high
voltage spikes.  But my simulations were showing that at 7mhz and no antenna those spikes
weren't all that bad.  As the drain voltage rises it drags up the gate (miller capacitance)
and keeps the FET turned on.  If you bring the frequency down far enough from 7mhz
the FET does eventually turn off, and you start seeing kilovolt+ spikes
at the Bitx40  FET drain in the simulation. 

Is Gary, N3GO still around?
I wonder how that TR scheme of his worked out.

Jerry, KE7ER


On Sat, Apr 28, 2018 at 06:45 pm, Kees T wrote:
Believe you me, a high BITX antenna SWR will take out an IRF510 in a heartbeat .....not so much for a RD16HHF1 RF MOSFET. Sounds like another ~ 1 sqin board. 

Join BITX20@groups.io to automatically receive all group messages.