TRF : the origin of Receiver

Feb/9 2012

get confidence, intuition, build a TRF receiver, you got to hear the voice.


Crystal Set: the state of art

schematic showing in left is a fundamentals crystal set, and the right one is a practical one with improvement, such as use taped tank connect to detector.

* a high Z headphones is very important for a crystal set

* the Antenna is critical, you definitely hard to hear voice without a big antenna

* even with all above, you still might hear no signal if you are far away from broadcast station. Not to mention the indoor sets.

* audio volume usually small. unless you nearby the station or you have an art-of-state antenna.

* selectivity is poor, normally.

Crystal set is not the receiver easily to get a acceptable voice volume. that's might be the reason why so many people like it and built numberless remarkable machines. for example ,Geoff's crystal set :

other guys


TRF is an advanced version of crystal set. TRF, i mean 'pure' TRF, no regen, no reflex. TRF is popular around 1920s.

the right schematic is a basic improvement for antenna couple method, the transformation: light couple to input tank, preserve LC tank Q value. with a resonated antenna, performance could be much better.

* if just attach a Audio amp to crystal radio, you won't get any sensitive improvement. more loudly, but still not enough. selective, sensitive just keep be a crystal radio.

* adding more and high gain Audio amp stage does not make crystal radio sensitive, and high gain audio amp might pick up so many noise. this is not the correct way to make a better TRF.

* Improve sensitive: the only way to improve sensitivity for crystal radio is add RF amp before detector stage. the reason is not only a low noise high gain audio amplifier is hard to design. here is why:

  • diode detector for the crystal radio actually work as a square law detector, that is, output voltage is proportional to input power , for more mathematically , proportional to square of input voltage.

Vo = αVi2

α value:, if a 20mV input give an output about 0.2 mV, then α=0.0005. test condition: load of detector is around 1.5 MΩ.

Measurement on several schottky diodes, shows detection in the square law region takes place at input voltages below 200 mVpp.

  • for a typical configuration, a load resistor 50KΩ, or the detector's RD might also 50K(no bias). the -60dBm(50R system) signal source generate an output voltage about 0.1uV, such a low level signal will disappear in the noise of the RD or load resistor. because 50K resistor's thermal noise at room temperature will also around 0.1uV.
  • an usable output must have a acceptable SNR, at least 8dB, required input power 50 dBm. and most of detector diode data-sheet 'out-voltage vs input power' start from -30dBm or -40dBm.

great information about diode detector:


ferrite loopstick antenna dominate portable AM receiver market. But I've built a regen and a reflex receiver with ferrite loopstick antenna. both of them seems hard to receive any signal indoor. the only way to test them is put them near the window. the broadcast signal is so weak in door. it's super uncomfortable while you can not receive any signal on workbench.

look at this amazing crystal radio antenna(left) ,it's hard to make but beautiful.

i built the right one, a loop antenna use PVC tube , M3 screw, pieces of PCB board. cross conner length is 30cm. it is much better than ferrite antenna, even not resonated. I also test it use AM superhet receiver without input tune capacitor, these is more stations (not count the images,-:). use LOOP ANTENNA for your TRF, except there is strong signal around you.

Practical TRF

here is an popular simple TRF

* 3DG6 is an popular china made old transistor back into my childhood. it's hard to find it. but 2sc9011 might be a good replacement.

* the ANT should couple to tank through a small capacitor

* Q1,Q2 form 2 stage DC couple amplifier. DC coupled amplifier need to calculate it's bias, R1 might vary form xxxKohm to xMohm.

* Possible improvement is get signal from a tap of L1, to match low input impedance of BJT and preserved the tune tank Q.

Now show you my version, similar to TechLib's all band receiver.

  • VCC 7.2V-8.4V
  • loop antenna, the ferrite stick antenna is not enough for weak signal especially indoor.
  • but transistor BE junction as detector.
  • transistor's quiescent current 1mA, voltage of transistor's collector around 3V
  • A high-Z ear phone might work, but i use an Audio Amplifier for playing aroud, gain x10.
  • C4, C3 bypass the RF signal
  • if omit the variable capacitor, it works as an all band receiver, of course, bad sensitive, that's TechLib's all band receiver.
  • stage one couple to stage 2 without considering the impedance match. stage one output impedance is about several kilo ohm, input impedance of stage 2 around xxx Ohm, transformer coupled method is better, but it not very convenient to make a audio couple transformer.
  • detector bias is not in it's best performance. both stage work as amplifier actually. while Ic=1mA , Vbe=0.65 volt, will be a bad detector, even though it works. the first stage should adjust the Vbe to around 0.2V - 0.3V, that will make a better detector stage.
  • add a detector before the stage one, receiver became a crystal radio with audio amplifier. add a detector stage after stage one will transfer it to be a receiver with RF amplifier.

how about the performance?

  • Pick up local strong station only near window. large single loop (3mx3m square) does not help improve the sensitivity but loudly.
  • Noisy. i guess the indoor antenna contribute the most of noise, not sure about this.
  • Bad selectivity.

TRF problems and improvement

* bad selective

AM band station is crowd. the single tuned TRF, ie, 12 kHz bandwidth at 1 MHz equates to a LC tank with it's Q is 83. this Q is not very easy to achieve without pay attention to antenna coil. with a BJT as the first stage amplifier, low input impedance make Q even worse. i hear 2 stations voice same time, some time a strong station occupy almost all tune range, it's awful. cause it's loaded Q value is so bad.

* selectivity of a TRF receiver was not constant in tune range.

the frequency increased, the selectivity became worse. The AM broadcast band extended from 540 KHz to 1500 KHz during the 1920's. If the tuned circuits in the receiver were designed for the proper bandwidth (10 KHz for an AM broadcast signal) at 540 KHz, the bandwidth at 1500 KHz was almost 30 KHz, which occupy 3 channels on the AM band.

after 1920's, more and AM stations began to transmitting signal, the AM band became crowd, people need a new receiver design, that was the super-heterodyne.

* multi tune TRF tracking problem

multi tune TRF get better selectivity but tracking is the problem, hard to keep track, make it hard to alignment.

* noise

My TRF hear so much noise, wide bandwidth, not fine tune the amplifier , make it so noise.

* The gain of the amplifiers generally decreased as the frequency increased

* self oscillate

if the RF stage have too much gain, it will became a oscillator, one solution is multi stage, keep each stage's gain low.

* high input impedance, light couple to tune circuit.

* antenna couple method

the following 3 method widely used in AM receiver. couple by transformer, capacitor, and a combine of them. they are not for loop antenna but worth to know.

* (a) transformer couple

depend on AT(antenna) resonance frequency(AT inductance and AT capacitance), when frequency raise, it show 3 type voltage transfer curve: trace a, resonance frequency below the fmin, trace b resonance frequency high than fmax, trace c resonance frequency in the receive band.

* (b) capacitor couple

transfer more voltage when frequency raise.

* (c) compilation

trace a, the Capacitor transfer function,trace b, the transform couple transfer function, trace c the combine. NOTE: AT resonant frequency lower than the band.

* TRF improve step 1, tap couple

* tap exploit a impedance match, and light couple to anntna tank.

* much big audio volume

* selective is improved, but still not acceptable

* TRF improve step 2, high input impedance

* the very high input impedance ( Mega ohm), via the bootstrap capacitor C1, C1 work as a current source for AC perspective.

* selective is much better. strong stations does not occupy all the tune range now.

* TRF improve step 3, high input impedance combine tap copule [ Feb/22/2011]

selectivity is dramatically better. station only occupy a narrow tune range, feel like a superhet receiver.

Final Project

  • using tap combine the high z input RF stage dramatically improve the selectivity, it's acceptable now.
  • add diode detector, D1, 1N60 without bias. for better performance, use FET high-Z stage couple to the detector.
  • now it receive more stations. and low noise. the tap provide most contribution for reduce the noise.