On 2020-05-21 14:16, wrote:
Phil Hobbs wrote in
:

On 2020-05-21 13:33, John Larkin wrote:
On Sun, 17 May 2020 14:15:31 -0000 (UTC), Cursitor Doom
wrote:

Gentlemen,

I have some old OC171 transistors from an old Eddystone short
wave radio I'm restoring. The problem is I suspect they've
developed whiskers, as the resistance readings from e,c and b to
the screen connection are all far too low (sub 10 ohms). A
colleague has suggested blasting the whiskers by tying the ecb
leads together and zapping them against the screen with a 500nF
cap charged to 500VDC. Now this seems a bit counter- intuitive
to me and I'd have thought higher current lower voltage would
be safer for these delicate germanium devices, but WTF do I
know? Is it feasible to remove the whiskers by this sort of
method or any other?

Thanks,

CD

Does anyone still make ge transistors? I can't think of any use
for them.

The only production ge devices I know of (excepting SiGe) are
back diodes, which I think are the only germanium parts
fabricated using lithography.


Ge makes good photodiodes for some uses. They're very leaky, but
if you make the epi thin enough, they can cover 350-1800 nm, which
otherwise requires expensive stacked-die devices. Garden-variety
ones are more like ordinary InGaAs, i.e. 800-1800 nm, which is
much less interesting.

Cheers

Phil Hobbs


https://iopscience.iop.org/article/1.../JJAPS.14S1.57


That's a thin-film bolometer, similar in general character to the ones
used in modern uncooled microbolometer cameras. Compared with
visible/NIR photodiodes, they're slow and very insensitive, but they
sure do have a wide wavelength range.

20 years ago I built an interesting system called "Footprints" that I've
discussed here a few times over the years. It used an array of 96
carbon-ink pixels screen-printed on a 9-um thick free-standing film of
PVDF (which is basically fluorinated Saran Wrap). The pixels were 3 x 5
mm on a 6-mm pitch, to leave room for the wiring. The readout
multiplexer was a single red display LED per pixel.

Diodes ideally conduct in only one direction. The particular LEDs I
used leaked less than 50 fA between -5V and +0.5V bias. Interfacing
them to an AC-only sensor such as a pyoelectric requires a bit of bias
current, which in my gizmo was supplied by four green display LEDs,
which allowed the processor to adjust the average bias current between 0
and about 5 pA per LED.

The optical system was a moulded Fresnel lens made out of HDPE
(bleach-bottle plastic).

When it was done, it had very competitive sensitivity: noise equivalent
delta-T of just over 0.1 K, not bad for something so minimalistic.

We tried licensing it, but couldn't because the parts cost was too low.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com