In article ,
Ian Field wrote:

Best one I ever used was the tube version of the Millen.
I switched to the solid state version and regretted it.
Much less sensitive and more fiddly than the tube version.
Tunnel dipper was the worst for sensitivity.

A while back I acquired a small Tektronix component envelope containing very
tiny "top hat" style diodes that there's a faint possibility they could be
TDs.

Copying the Heathkit TD dipper had crossed my mind - now I probably
won't.

By all accounts I've ever read, the Heathkit TD dipper isn't a good
performer. I've played around with one briefly and it had a very
difficult time sensing a resonance even with a large high-Q air-core
inductor and an air-variable cap.

It was actually worse than the "grid dip meter" accessory for an
MFJ-259/269 antenna analyzer, and that's saying something... something
profoundly bad :-(

I compared the MFJ to a more recent Heathkit solid-state dipper (an
HD-1250 if I recall correctly), to a B&W, and to a Measurements 59.
Sensitivity was in that order... the Heathkit was better than the MFJ,
the B&W next, and the Measurement 59 was the stand-out winner. I've
never had a chance to test a Millen; from what I've heard they'd
probably come in between the B&W and the 59.

There are a couple of projects out on the net which show how to make a
GD meter using a (composite) lambda diode. These are
negative-resistance devices which behave a bit like a tunnel
diode. These can be made from two JFETs (one N-channel, one P-channel)
or an N-JFET and a bipolar transistor.

http://users.tpg.com.au/ldbutler/NegResDipMeter.htm

Lambda diodes typically require more voltage and use more power than a
tunnel diode - the negative-resistance part of their curve covers
several volts. Strong oscillators... the dip-meter design requires a
band-specific resistor to tame the strength of the oscillator so that
the dip can be seen.