Thursday, February 3, 2011

An SCR Audio Frequency Amplifier

I came across a book some years ago in which the author briefly mentioned that it was possible to employ an SCR (silicon-controlled rectifier) as a low frequency linear amplifier. I've never heard of this anywhere else, but the question of whether or it was possible has remained in the back of my mind ever since.

Well, today I decided to investigate. After running down a few blind alleys I eventually hit upon a working circuit.

      
The gate (input) impedance is an inverse function of the anode current. The 10k Ohm resistor in the anode supply line sets the gate input impedance to 50 Ohms. Not having a suitable wide-band matching transformer on hand, I decided to try a narrow-band match using a capacitive divider. The combination of L2 and C4/C5 resonate at 725Hz. The impedance is stepped-down from 99.5k Ohms to match my 600 Ohm headphones. The loaded Q is higher than I would have liked, but aside from that it produces an efficient match. 

The circuit produces just over 21db of power gain with low distortion. I quickly connected this amplifier to the simple receiver that I constructed yesterday

One of the first stations that I heard was W2LB on his 1929-style breadboard transmitter. He built it using a #27 triode for the Bruce Kelly Memorial 1929 AWA QSO Party. Larry has posted some photos of this transmitter on his QRZ.com web page. Hearing this great little rig was the icing on the top of a very FB day in the shack.

Three stages of SCR amplification ought to drive a small loudspeaker handily; at least on 80m. For now, I'm happy with the single stage. It's going to be fun telling my contacts that all my receiver gain is provided by a 1962-vintage silicon-controlled rectifier!

This evening (2/7/11) I ran a simple circuit model through a "Spice" simulator. The SCR was modeled using a 2N3906/2N3904 transistor pair. It was necessary to increase the anode-to-supply resistance to 70k Ohms in order to present a 50 Ohm input impedance. Aside from the fact that my vintage 2N2323 SCR produced somewhat more gain, the model behaved much the same as my breadboard circuit. The circuit model is shown below, along with the quiescent DC node voltages.
 
Notice that transistor Q1 is in saturation. This essentially leaves Q2 acting as a common-base amplifier. However, in this case the collector and emitter terminals are reverse-connected. This, along with the forward-biased diode shunting the amplifier input port accounts for the reduced gain as compared with a conventional common-base amplifier. 

Here's a plot of the amplifier input impedance from 100Hz to 100kHz.

In the image below, a 725Hz input signal is plotted at the top and the amplified signal appearing across RL is plotted beneath. The signals are in-phase; just as one would expect from a common-base amplifier.

2 comments:

  1. Michael,

    Hope winter is treating you well?

    Very interesting, I did not know SCRs were possible to use in linear service. I thought they would saturate on hard because of their 4-layer structure. I must try this soon!

    Regards,
    Alan

    ReplyDelete
  2. Hi Alan,

    Thanks, we're having a great winter. I only wish it weren't flying by so quickly!

    Reverse biasing the anode and cathode throws a monkey wrench into the normal regenerative switching mechanism, so the circuit is quite stable on that account.

    73/72,
    Mike

    ReplyDelete