In his booklet, Rundfunkempfang ohne Röhren, dating from 1959, H.G. Mende presented an AM broadcast band regenerative detector circuit using a point-contact transistor (PCT).
I breadboarded this circuit using a Western Electric 2N110 transistor; one of several that were given to me by Jack Ward of the Transistor Museum. My regenerative detector came to life with a minimum of fuss. As a quick test I swapped this circuit in place of the 2N107 alloy-junction based regenerative IF detector in a recently made 40m superhet receiver.
The minimum detectable signal strength with the PCT regenerative stage increased to 1uVrms. As expected, the 2N110 produced conisiderable self-noise. Of the three devices on hand I selected the least noisy unit (date-coded November of 1956). I could still hear "eggs frying" in the phones with the antenna disconnected, but the level was now low enough that I could live with it. I found the regeneration could be smoothly adjusted using the variable resistor located in the emitter-ground return. I only noticed pulling (frequency-injection locking) on the strongest signals.
The DX came rolling in with this setup during the early evening hours on 40m. HB9CVQ and EI7KD both had big signals here. A couple of interesting signals were heard from Brazil. It's not often that you hear an 18 year-old girl working CW on 40m, but Lia, PU2AIL, was heard knocking off the DX in great style. Also heard was PQ8XB operating from the Reserva Biológica do Parazinho.
The results were good enough that I decided to attempt to build a 40m simple superheterodyne receiver exclusively from 2N110 point-contact transistors. However, I was unsure whether or not these devices would oscillate in the vicinity of 7MHz, given their datasheet-listed 3dB-down frequency is only 1.5MHz.
Fortune smiled on me the next evening when I found that one of my two remaining 2N110s would not only oscillate on 40m, it would generate as much as 20mW under crystal-control!
Well, I couldn't resist connecting this oscillator to my Czech military surplus straight key and calling a few CQs. This, despite the fact that changes in the thermal equilibrium were pulling the crystal-locking frequency around more than one could expect to get away with. It amounted to a slow, but very pronounced chirp on the keyed signal. Hearing no replies, I ceased transmitting and returned to thinking about the superheterodyne design.
However...the next morning I was surprised to find an email message from K8GU/3. Ethan is a scientist engaged in studying the ionosphere down in Silver Springs, Maryland.
Good morning, Mike,
I've been spending a lot of time listening with an SDR
(SoftRocks+Rocky) lately and last night I noticed your curious-looking
signal in my waterfall. It took a moment for your signal to come up
on the QSB, but I had pretty good copy on you. I started rewiring the
station so I could give you a call, but my wife handed me our
7-week-old son for a diaper change. The best I could do was hit
"Print Screen" as I whisked him off. So, you'll have to live with a
screenshot for now...
Please click here to see his screen-shot of my 20mW signal. That is so cool...thank you Ethan!
Getting back to the superhet receiver; I decided not to try to place a 2N110 transistor directly into the 7MHz signal path. I opted instead to use a germanium diode single-balanced mixer in conjunction with a 2N110 crystal-controlled oscillator. The variable IF output signal feeds a second 2N110 working as a self-excited, regenerative IF amplifier/detector/audio frequency amplifier. Here's the resulting circuit
Please notice the lack of an explicit RF feedback path in both the oscillator and regenerative detector. I'm making use of a unique characteristic of these early point-contact transistors. Under certain conditions a negative resistance appears at the base of these devices. Unlike a tunnel diode, for example, the value of negative resistance produced can be easily controlled by adjustments made at emitter and the collector. For example, the detector regeneration level is adjusted by RV1 in the circuit shown above.
T1, 2: 3.5 to 5.5uH adjustable, 16-turn primary, 2-turn coupling
T4: ditto above, except 1-turn coupling
T3: 1:4 balun (Mini-Circuits T4-1, etc.)
T5: 455kHz shielded IF can (yellow-slug)
L1: 470uH molded RFC
C1, C2: 114pF
R3: 330 Ohms
D1, 2: germanium point-contact diodes (1N34a, etc.)
CV1: 20 to 420pF air-variable capacitor (main tuning)
X1: 7376.4kHz quartz crystal from junkbox
HP: high-impedance magnetic (mine are 1950's Czech army surplus)
Q1: 2N110 point-contact transistor (date-code March 1957)
Q2: 2N110 point-contact transistor (date-code November 1956)
The total DC current on the -22Vdc supply is -4.87mA (107mW power drain).
As you'd expect, the receiver is somewhat less sensitive on account of the gain-less mixer stage. I can't hear my 1uVrms signal-source. Aside from that and the afore-mentioned weaknesses, the all-PCT receiver appears to function normally. In one night of listening quite a few DX stations were heard along with several domestic QRP stations. Some stronger stations were copied with the headphones lying on the workbench. I did not find it necessary to re-adjust the regeneration setting from one end of the 40m CW band to the other.
By the way, I came across an interesting passage in Chih-Tang Sah's, The Fundamentals of Solid-State Electronics, on the subject of these nearly forgotten PCT relics.
"In this first point contact transistor, the minority carrier pathway through the bulk or base of the single crystal Ge had probably contributed only a small percentage of the total curret. Instead, the surface channel path may have dominated the current making it a p-channel junction-gate field-effect transistor." p. 705
The author proposes that 90% of PCT action was due to field-effect and only the remaining 10% on account of (base-injected minority current) bipolar-transistor action.
The PCT was a manufacturer's nightmare (William Shockley refered to certain steps in their fabrication process as "witchcraft"). As if the wide device-to-device performance variability, dismal production yield and user reliability numbers weren't bad enough, they tended to be horribly noisy. Even the earliest junction-type germanium transistors exhibited a 20dB noise improvement!
"People beyond AT&T were beginning to suspect that all the original enthusiasm over the transistor had been premature. 'Current Bell statements concerning the transistor are far more subdued and give an impression that it is under wraps,' noted a September 1949 article in Consumer Reports. 'Very little is said of immediate practical applications. Such transistor difficulties as high noise level are stressed.'" p191, Crystal Fire, Riordan and Hoddeson
The 2N110's used in my receiver described above first appeared in 1955. But as late as 1952, as this Electronics Magazine editiorial suggests, transistors remained at best, a tantalizing but unfulfilled promise.