Si-diodes in Second World War radar & Communication equipment

I am on a temporary setup now that does not have Acrobat, but I somewhat remember Vishay-IR STPS1L30UPBF or 1N5818 dropping maybe .35 volt at 1 amp. These are 30 volt 1 amp Schottky rectifiers.

- Don Klipstein ( snipped-for-privacy@misty.com)

Here are some curves from the RadLab book:

ftp://66.117.156.8/RadLabDiodes.JPG

ftp://66.117.156.8/RadDiode2.JPG

Your data point is dead on the point-contact Silicon diode curve.

John

A couple of fascinating books are Buderi's "The Invention That Changed The World" and Conant's "Tuxedo Park", both about the history of microwave radar, and where the money came from.

Also Bowen's "Radar Days" and the Brit story of HF radar, "Three Steps To Victory" by Sir Robert Watson-Watt, who probably saved England from the Luftwaffe.

John

One of the MIT books says that "a semiconductor triode should be possible." But that wasn't their mandate. The RadLab was disbanded in late 1945.

John

This is a silicon point-contact diode, essentially the same as the WWII parts, expect that they get to use modern, very pure silicon:

Skyworks makes some very low capacitance (below 0.5 pF) schottkies that are similar.

This is 300 mV *max* at 100 mA, so should be down there. I think the schottky curve is sorta similar to the silicon PN curve, which is 60 mV per decade of current.

I posted some WWII diode curves elsewhere, well under 200 mV at 1 mA.

Gee.

John

"JosephKKK Lunatic & Congenital LIAR "

** Examples tested:

BAT46 = 0.261 V @ 1mA

MBR745 = 0.194 V @ 1mA

For comparison

AAZ15 (Ge) = 0.230 V @ 1 mA

The 1N23 ( Silicon point contact) is 0.25 V @ 1mA

...... Phil

Rich Grise inscribed thus:

Yes ! I have some devices that were made in the mid to late 40's.

Also if I can find them I have some pre war point contact detectors that have screw terminals on the ends.

Baron.

--
I just pulled a random 1N5817 out of stock, put 1.000 milliamps
through it and measured 0.1383 volts across it.

JF

Sounds about right.

I have a book here about the develpoment of active devices for radar and it's quite unambiguous about silicon being used for microwave diodes. The early work was actually done by GEC and BTH of the UK in conjuction with military R&D.

As ever the Americans refined the manufacturing process. The early ones were virtually 'hand made'.

Graham

--
Just to make sure it wasn\'t an anomaly, I measured 10 more and here\'s
what I got:

   If      Vf
   mA      V
-------+-------+
 1.000  0.1495
 1.000  0.1350 
 1.000  0.1525
 1.000  0.1344
 1.000  0.1495
 1.000  0.1355
 1.000  0.1510
 1.000  0.1532
 1.000  0.1496
 1.000  0.1370


The equipment was set up like this:



     +-------[WAVETEK 27XT]---[10k]---+----------+
     |+                               |A         |+ 
[HP 6216A]                          [DUT]  [FLUKE 8060A]
     |-                               |          |-
     +--------------------------------+----------+


The 8060A draws 25µA on the 2 volt range, so the current out of the
6216A was set to 1.025mA for every reading in order to force 1.000mA
through the 1N5817s.

Turns out the power supply was impossible to adjust spot on, so I put
the 10k resistor in there to give me fewer µA per degree of rotation
of the knob.  Worked great.

JF

Most DVM's seem to output 1 mA on the diode-test range. I don't know how much of a convention that is. They do seem to disagree on how much voltage they'll indicate: some display the Vf of an LED, some say open or overload or whatever.

John

"John Fields"

** But you well knew that Motorola describe them as having " Extremely low Vf " - now didn't you ??

BTW:

how hot did you make it get first ?

.... Phil

Thanks for the references, John. Sounds like Buderi, who certainly had access to all that, might have put it in somewhat different light than he did, perhaps something along the lines of, "though much theoretical work had been done by 1940, it remained to discover how to apply it in practice." He does make it clear that researchers all the way up through development of the transistor didn't fully appreciate what they could do with potential barrier diagrams and Fermi levels and the like. As I scan through the book, I see multiple references to events over several years where there was clear puzzlement, limited understanding, and/or disagreement about what was going on in observed effects around semiconductors. The serious search for a solid-state amplifier (based on semiconductor materials) was started apparently at least by 1936 at Bell Labs, and I suppose it was there and at a very small number of universities where much of the published work you cite was begun or carried out. Too bad that it's a bit late to be asking the people actually involved in the work! (Wish I'd had the foresight and time to ask my uncle more about the work he did at RadLab. :-( )

Cheers, Tom

Central CMMSH1-20 is a really tiny, about 1206 size, 1 amp 20 volt schottky, great for small buck switchers; measures 201 mV at 1 mA. But it's 280 pF!

John

I started working on military surplus radios and consumer broadcast radios in 1954, went to US Navy ET school in 1956, and spent a few years reparing communications and radar equipment manufactured in the 1947 - 1960 era. Other than the 1N21 and 1N23 diodes used in radar receivers, which I believe were point-contact germanium devices, the first signal diodes I saw were the 1N34 types used in an IFF decoder. They were relatively large axial packages with a hexagonal body shape. All other places where a diode was necessary, vacuum tubes were used for small signals and Selemium-oxide plate rectifiers were used in power supplies. By 1959, when I went to school on the AN/URC-32 SSB Transceiver, both germanium and silicon diodes and transistors were in wide use in new military radio circuits. From my experience, no silicon diodes were used at any time before 1950 in USN military equipment. The first consumer transistorized radios I remember were sometime around 1960. I wish I still had my Allied Radio Catalogs from 1954 and later. You could buy a Raytheon CK721 transistor for something like $3. I think the CK722 cost more. By 1965, you could buy grab bags of transistors and diodes at Radio Shack for $3.

Good Memories, Ken Fowler, KO6NO

I think if you do a Google search for "zero bias diode" you will find things a lot more similar to 1N23 in electrical characteristics.

--
Regards,

John Popelish

The radlab boys observed a number of interesting things that they didn't have the time to pursue. One was negative resistance in diodes, and another was diode mixers that had signal power gain, the precursor to the parametric amplifier. Tunneling was known, too, a long time before Esaki discovered the tunnel diode.

All that is in one book out of 27. In about 5 years, these guys invented modern electronics.

John

"Back diode" is interesting, too. They are, to my knowledge, the only germanium diodes made using an ic-type mask process, and about the only Ge diodes still made at all, except for photodiodes of course. They are still the best microwave detectors.

John

--
Into a short.