Vintage transistors and tin whiskers

[Cursitor Doom]

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

[DecadentLinux...@decadence.org]

Cursitor Doom <c...@not4mail.com> wrote in
news:r9rgu3$h74$1...@dont-email.me:

Put it in your ten below freezer for an hour and let the cold
disintegrate the non alloyed tin.

Then give it a nice hard rap to break it all loosed like we used to
do to the old tube type TVs to get them going again.

Kidding of course... except for the TV thing. That worked. So
did taking the back off it and re-seating all the socketed tubes and
components. Worked every time. It is almost always a mechanical
connection.

[Pimpom]

I don't know about the whiskers but Ge transistors have high
leakage and no sharp knee at low currents. I wonder if 10 ohms is
too low depending on the voltage applied by your meter. I have a
few of the later AC188 types. I'll dig them up tomorrow and
measure their resistances.

For quick go-no go checks of transistors, I prefer analog MMs.
The static forward resistance depends highly on the current
applied through the probes.

I used to have some OC71s and OC72s a long time ago. They made
neat phototransistors with the paint scraped off from their glass
enclosures.

[Pimpom]

OK, I got the AC188s. With a DMM they measure 25-50 ohms forward
and a few Ks in the reverse direction.

[Cursitor Doom]

On Sun, 17 May 2020 21:06:23 +0530, Pimpom wrote:

>> I used to have some OC71s and OC72s a long time ago. They made neat
>> phototransistors with the paint scraped off from their glass
>> enclosures.
>>
> OK, I got the AC188s. With a DMM they measure 25-50 ohms forward and a
> few Ks in the reverse direction.

I think you've misunderstood the problem here, Pimpom. The leakage in
question is between the case ("screen") and the die and significantly pre-
date the black-painted phototransistors we all fondly recall from 50
years ago. :-)

[Pimpom]

Ah, sorry. Then those low-ohm figures can't be normal.

My dad bought a Sony transistor radio in the early 60s. I
remember that it used Japanese Ge transistors like 2SA12, 2SA15
for RF and 2SB75 and 2SB77 for audio. One of my regrets in life
is that I never got around to restoring it when he was still
alive. This is the model -
https://priceguide.thecollector.com.au/wp-content/gallery/auction-928/1960s-Sony-TR-812-allwave-transistor-radio-Sold-for-56-2019.jpg

[Adrian Tuddenham]

I have had some success with an Eddystone EB35 by replacing them with
silicon transistors and adjusting the base bias to get approximately the
same collector current.

(I was a tester on the production line at Eddystones when EB35s were
going through, so I am the one to blame if they don't meet spec.)


--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk

[Michael Terrell]

If the Tin Man in the Wizzard of Oz tried to grow a beard, they would be really big tin whiskers. ;-)

[abrsvc]

On Sunday, May 17, 2020 at 10:15:34 AM UTC-4, Cursitor Doom wrote:

I have 4 ECG160 transistors (2 that are unused and 2 that look like they may have been used to test something). Send me your address and I'll send them along if you think that these will resolve your problem. I doubt that I will ever use these.

Dan

[Cursitor Doom]

On Mon, 18 May 2020 05:43:18 -0700, abrsvc wrote:

> I have 4 ECG160 transistors (2 that are unused and 2 that look like they
> may have been used to test something). Send me your address and I'll
> send them along if you think that these will resolve your problem. I
> doubt that I will ever use these.

Thanks for the advice and offers of parts, guys; much appreciated. On
reflection I decided to replace with silicon. The purists would go to the
trouble of hollowing out the old transitors and fitting the silicons
inside them. Fortunately, I'm no purist. :-D

[Brian Gregory]

On 18/05/2020 12:59, Adrian Tuddenham wrote:
> I have had some success with an Eddystone EB35 by replacing them with
> silicon transistors and adjusting the base bias to get approximately the
> same collector current.

And if something starts oscillating because you have too much gain with
the silicon replacements then adding a low value R in series with their
emitters will usually fix.

--
Brian Gregory (in England).

[abrsvc]

No problems here. If you want to use the ECGs to verify that everything works prior to "upgrading" to Si, the parts are available. These are small enough that they will fit into a small padded envelope, so shipping will be next to nothing.

Let me know if you want them (no charge).
Dan

[Cursitor Doom]

On Tue, 19 May 2020 17:23:37 -0700, abrsvc wrote:

> No problems here. If you want to use the ECGs to verify that everything
> works prior to "upgrading" to Si, the parts are available. These are
> small enough that they will fit into a small padded envelope, so
> shipping will be next to nothing.
>
> Let me know if you want them (no charge).
> Dan

Thanks Dan, much appreciated.

[Adrian Tuddenham]

There are already emitter resistors, but they are bypassed, so adding
resistance in series with the bypass capacitors (or removing them
altogether) would be the way to go.

I used fairly mundane silicon transistors with a lowish cutoff frequency
and didn't have any trouble.

[Phil Hobbs]

On 2020-05-20 10:40, Adrian Tuddenham wrote:
> Brian Gregory <void-invalid...@gmail.com> wrote:
>
>> On 18/05/2020 12:59, Adrian Tuddenham wrote:
>>> I have had some success with an Eddystone EB35 by replacing them with
>>> silicon transistors and adjusting the base bias to get approximately the
>>> same collector current.
>>
>> And if something starts oscillating because you have too much gain with
>> the silicon replacements then adding a low value R in series with their
>> emitters will usually fix.
>
> There are already emitter resistors, but they are bypassed, so adding
> resistance in series with the bypass capacitors (or removing them
> altogether) would be the way to go.
>
> I used fairly mundane silicon transistors with a lowish cutoff frequency
> and didn't have any trouble.
>
>

Retrofitting germanium transistor circuits can be tricky, especially the
ones where they rely on the apparent beta being negative, i.e. the CB
leakage being larger than I_C/(true beta).

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

[Adrian Tuddenham]

Phil Hobbs <pcdhSpamM...@electrooptical.net> wrote:

> On 2020-05-20 10:40, Adrian Tuddenham wrote:
> > Brian Gregory <void-invalid...@gmail.com> wrote:
> >
> >> On 18/05/2020 12:59, Adrian Tuddenham wrote:
> >>> I have had some success with an Eddystone EB35 by replacing them with
> >>> silicon transistors and adjusting the base bias to get approximately the
> >>> same collector current.
> >>
> >> And if something starts oscillating because you have too much gain with
> >> the silicon replacements then adding a low value R in series with their
> >> emitters will usually fix.
> >
> > There are already emitter resistors, but they are bypassed, so adding
> > resistance in series with the bypass capacitors (or removing them
> > altogether) would be the way to go.
> >
> > I used fairly mundane silicon transistors with a lowish cutoff frequency
> > and didn't have any trouble.
> >
> >
> Retrofitting germanium transistor circuits can be tricky, especially the
> ones where they rely on the apparent beta being negative, i.e. the CB
> leakage being larger than I_C/(true beta).

These were relatively late germanium technology (0C171) with a high ft
and internal screen, so they didn't need the external neutralising
network of the earlier germanium types. That encouraged me to try a
direct substitution which only needed some adjustment of the base
biasing network.

[John Larkin]

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.

--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

[DecadentLinux...@decadence.org]

John Larkin <jlarkin@highland_atwork_technology.com> wrote in
news:uledcfl5p1mkbbopb...@4ax.com:

> On Sun, 17 May 2020 14:15:31 -0000 (UTC), Cursitor Doom
> <c...@not4mail.com> 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.
>

General Electric? Hehehe.. You mean Ge as in Germanium.

https://spectrum.ieee.org/semiconductors/materials/germanium-can-take-
transistors-where-silicon-cant

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

https://www.americanradiohistory.com/Archive-Catalogs/GE/GE-Transistor-
Manual-No.-5---1960.CV01.pdf

[Phil Hobbs]

On 2020-05-21 13:33, John Larkin wrote:
> On Sun, 17 May 2020 14:15:31 -0000 (UTC), Cursitor Doom
> <c...@not4mail.com> 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.

[DecadentLinux...@decadence.org]

John Larkin <jlarkin@highland_atwork_technology.com> wrote in
news:uledcfl5p1mkbbopb...@4ax.com:

> On Sun, 17 May 2020 14:15:31 -0000 (UTC), Cursitor Doom
> <c...@not4mail.com> 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.
>

Also as a side note to the title of this entire thread...

'Vintage transistors' were NOT 'tin plated' like todays parts are.
They were Nickel Cadmium plated. A far superior plating but with
carcinogenic dangers that caused Cadmium and Cadmium alloys to be
banned. But you can still buy, eat and die from Beryllium parts.

[DecadentLinux...@decadence.org]

Phil Hobbs <pcdhSpamM...@electrooptical.net> wrote in
news:0b74dede-27e7-ca27...@electrooptical.net:

https://iopscience.iop.org/article/10.7567/JJAPS.14S1.57

[John Larkin]

Our FLIR has a germanium lens.

https://www.dropbox.com/s/uda77g9w66x3u9f/Flir_E45_WA_Lens.JPG?raw=1

[Phil Hobbs]

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.

[Phil Hobbs]

Germanium makes really good IR lenses. Besides the built-in filtering
action, it has a refractive index of 4. The optical power of a given
surface is (n-1)/r, where n is the index and r is the radius.

Glass is generally around n = 1.5-1.8, so a Ge lens of a given power has
4-6 times the radius of curvature. Aberrations are much reduced due to
the weaker curvature, so a simpler lens can have better performance.
Also of course the diffraction spot size goes like lambda, so in terms
of the diffraction limit a Ge lens at 10 um is like 100 times easier to
design than a glass lens at 500 nm.

[Phil Hobbs]

Oh, and the dispersion of Ge out in the thermal IR is much less than
glass in the visible, so you don't even need to achromatize it. The
tempcos of index and of optical path length are quite large, so you do
need to athermalize in general, something that's rarely required in the
visible.

[John Larkin]

On Thu, 21 May 2020 17:26:09 -0400, Phil Hobbs

That lens stays in focus up to the point it touches a part. It will
clearly show the hot spot on an 0603 resistor.

Here's a tiny dual transistor,

https://www.dropbox.com/s/dd072w1z2gmfpbt/Dual_NPN.jpg?raw=1

obviously two separate chips inside.