On Tue, 3 May 2005 22:31:29 -0700, "John Smith"
wrote:

Well, it doesn't flash,

Hi Brett,

How could you tell? Everything about this is far too fast for the eye
to perceive. But we let that pass.

and it does seem to work... albeit in a bit of a
"sloppy" fashion (tuning is sloppy as to light changes in the REF led)...
and you are correct, the 100k (and the 1k is there in series too) is barely
enough to keep from burning out the leds and a kludge...

Barely enough? The problem is it is far too much. You would have to
pump 3KW (matched) through this SWR meter to run the forward LED at
rated current.

The eye is a miserable detector of absolutes, and what you perceive as
bright with whatever current is eking its way through the circuit, is
a sad comparison to what the LED could support. The proper comparison
is done by what is called optical pyrometry, but then that is the
wrong kind of balance for this application.

If you HAD to make this circuit work, I am interested, and certainly I think
you of an intellect capable--what changes would you make--keeping as much of
the original circuit as possible?

Research the "Bruene SWR meter" to find out where you went wrong in
the circuit topology. If you are going to substitute LEDs for meters,
take heed that they are power indicators, not current indicators.
This is some advantage in one sense, and a divergence from the
standard presentation in another sense. Not keeping track of the
sense can get your necktie wrapped around the axle. Mixing LED types
like for Green forward power and Red for reverse power (or t'other way
'round) can really gum up the works when it comes to sizing other
component values.

A simple test of how confounding the eye can be is found in the
question:
"How bright can the Red be in proportion to the Green?"
By the simple electronics, this is a slam dunk. By appearance from
those same electronics, you are off by two orders of magnitude. By
conversion efficiencies there are still major errors of appearance.
Once you get all these balanced out, you still have to answer the
question whose psychology will still perturb most users. Try writing
the answer for a User's Manual (the ultimate test of functionality).

73's
Richard Clark, KB7QHC