Amps
[Top] [All Lists]

Re: [Amps] input/output transformer switching

To: <amps@contesting.com>
Subject: Re: [Amps] input/output transformer switching
From: "Han Higasa" <higasa@plum.ocn.ne.jp>
Date: Sat, 22 Dec 2007 17:10:15 +0900
List-post: <mailto:amps@contesting.com>
Paul,

Even in gemini-package each FET devices are not exactly matched.
So I think you need an independent gate bias circuits to enable
adjustment of Iq of each FETs in a package.

Set Iq of each FETs to 150-200 mA (not 250 mA) respectively,
to avoid a instability by heat-up.

Use a dummy-load and with adequate output level (200 W), 
slowly and carefully adjust one FET's gate bias potentiometer
while monitoring the level of 2nd harmonic.
This will improve 8 - 20 dB of even-order harmonics compared
to DC Iq balance only.

The high bandwidth device such as 141G/151G may generate 
very high levels of harmonics. Some AR303 experimenters report
only -16dBc of 3rd harmonics at 250 W output -- 6.25W!
LPF is essential.

I think your idea of using relays will work fine, but gate will be 
non-connected when relays are change over -- it is quite dangerous.
If you will not use a separate gate bias you would better use 2.2k ohm
termination resisters from each gate to the ground to ensure stability. 

I recommend to terminate each gate to the ground via 1k or 2.2k ohm
1W resisters, then use adequate RF chokes (cheap 47 uH or 22 uH
inductor and ferrite beads will work fine) to feed gate bias from
1k ohm potentiometer.

Use a 5 V three-terminal regulator to feed power supply of these 
bias potentiometers. Cut power to the regulator when receiving to
avoid heat-up.

Remove R1 and R2, add 0.1uF ceramic caps between gates and
relay to provide DC blocking of both gates and transformers.

Do not forget to connect ground planes of input and output PCB
each other by means of two cupper ribbons, bridge on or near both
flange fix bolts.
Otherwise FETs will be likely to crash by high frequency oscillation.

GL, de Han JE1BMJ

_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps

<Prev in Thread] Current Thread [Next in Thread>