High-power varactors

[Andre Kesteloot]

The idea of using the drain-gate capacitance of biased-off power MOSFETs 
as high power varactors is interesting. I attended the conference at 
University of Bath where the paper mentioned in Technical Topics was 
presented, and WA1WLW's talk was very good. (There seemed to be a lot of 
amateurs about - a number of other subscribers to this reflector were 
there; I presented a paper about LF antenna measurements). The Q of 
these MOSFET-varactors was about 30-100 at HF. In the paper, they were 
used as tunable elements in a HF PA matching network with a fairly low 
loaded Q, so a high Q tuning element was not essential. For LF 
transmitting antenna matching, a rather higher Q would be desirable.  
For a small-signal application (like a preselector or a VCO), it should 
be possible to use lower voltage power MOSFETs. As an experiment, I did 
some measurements with a back-back pair of STW34NB20 MOSFETs (200V, 
34A). With a bias voltage of 5 - 200V, capacitance range was 750p - 
140p, with most of the variation at the lower voltage end, as you would 
expect with varicaps. Rough measurements of Q (using a 3.6mH pot core 
inductor over a frequency range of 97kHz - 230kHz) gave values between 
about 500 and 1500, with higher values for higher capacitances. So this 
is obviously a useable technique for LF - the high tuning voltage is a 
bit awkward to produce, but would result in good linearity. Looking at 
the data sheets, similar capacitance variation can be achieved with 
lower tuning voltages by using MOSFETs with lower breakdown Vdss. I 
guess higher BVdss MOSFETs could be used for LF antenna tuning, 
especially if used as a "fine tuning" device making up a fraction of the 
total capacitance.
 
The idea of applying a magnetic bias to a ferrite core to vary the 
incremental permeability and so produce a tunable inductor has been 
around for quite a while - apart from the previous articles cited in 
Technical Topics. In the old RSGB "LF experimenter's sourcebook" there 
is an article about a tunable LF converter using mechanically moveable 
permanent magnets to bias a toroid core, which is apparently a re-print 
of an article from "Ham Radio" by OH2KT from 1974. I guess the problem 
with such a device is finding a ferrite material that gives at the same 
time a reasonably high Q, and a usefully large swing of inductance. To 
get a large inductance swing requires a ferrite with a high initial 
permeability - such ferrites unfortunately tend also to have high losses 
. Lower permeability ferrite has higher Q, but would need higher DC 
control current to get the same bias magnetic flux, and also would give 
less change in inductance. An air-gapped high permeability core like a 
pot core wouldn't really work because the reluctance is mostly 
determined by the air gap rather than the ferrite. The Technical Topics 
article is apparently aimed at MF/HF, but I think the idea would work 
better at VLF/LF, because high-permeability ferrites have lower loss in 
that range.
 
Cheers, Jim Moritz
73 de M0BMU
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://amrad.org/pipermail/tacos/attachments/20031129/32adcf8f/attachment.html