Heterodyne:  BJT AM receiver

heard so much Super-heterodyne,  i remember soldered one in university, but it did not work. I had bunch of 2sc9014, it's cheap, ¥0.04, and it's design for low noise amplifier and oscillator.  it's used in every stage, Despite it's not the best choice [ 2sc9018 maybe better]. 

the dream come to true from today.

It's definitely should use separated LO instead of LO-mixer in one stage,  almost all old traditional AM receiver use one transistor as oscillator and mixer, transistor is not expensive as it had been anymore. 

VCC: 12.8 V.

this oscillator is famous nosy oscillator introduce by <EMRFD>, i had built several  ones.  refer to VCO:negtive 1mhz-30mzh  for detailed information. but this is still a amazing  oscillator for it's sure kick start for any not-so-bad Inductor, any frequencies, from 40 khz to hundreds Mhz, it just work fine.

There is a buffer for the output to Mixer, T5.   R7 plus a 47k POT to adjust the output power, we need around 200 mVpp output driving mixer.

i'm going to add some sort of tracking between the LO and the input RF tank.  there is  92 uH oscillator coil in my hand.  I connect 2 short Ferrite rod antenna as a long one, the inductance range in 104uH to 210uH depend on the position of the coil on the Ferrite rod.

base on Robert Weaver's  Band Spread designer.  the cicuits parameter i'm going to use list in below figures.

if choose 92 uH inductor as the oscillator coil, the IF frequency should use 465 khz to get better tracking.  for 455 khz IF frequency, max tracking error acceptable, it's 4%.

LO rebuild  note@2012/11/29

------------------------------------------

the varactor 1sv149, capacitance range 15p to 500pF while bias from 1V to 8V.  frequency could cover 950 khz to 2.5Mhz.  output signal level in 100mV to 200mV range. top band signal level seems always be ~2 times  as the low frequency. buffer the LO is good choice.

if use 1X probe of oscilloscope the output signal would be severely distorted, oscillator is heavily loaded by the probe.

This is very old emitter LO inject BJT mixer, obviously never take it as consideration for a long time.  the MOST important  thing is BJT's bias. Keep Ic as low as 0.3 mA--0.5 mA, otherwise it's might could not  function as a mixer, or bad performance.   recommend LO driven amplitude is 100 mVpp to 200 mVpp. 

@2012/12/12

use 9018 instead of 9014. get more stable mixer. refer to  IF.B0: Gain Vs Stable,

use another 9014, with a little difference bias compared to Mixer.

IF stage Self-oscillating almost the biggest enemy for each builder. the methods to prevent it, listed here[ might not be very reasonable, but these suggestion is everywhere. ]:

0. adjust the IF transformer for better match.

1. use a low β transistor , 60--90 is prefer. and not use too high GBW transistor. [9014 had so high β, it's bad ]

2. LO frequency should far away from IF frequency.  

3. improved the stage impedance match

4. add 50R--200R resistor to base,   problem solved after series the base with 100R resistor, IF stage became quiet at high gain.

5. import negative feedback from tank to base use a small capacitor( xpF--XXpF)

6. Isolation and shield.  layout and line route is also important. 

7. not recommended: reduce IF transformer Q: add xxK resistor to tank reduce it's Q.

8. use tapped IF transformer [ read the following red title section ]

REMEMBER: any self oscillator is because too much positive feed back at some where!!

base on calculation in IF.B0: Gain Vs Stable,      the most better choice for cure oscillation:

The 9014 had very high Ro, 9014 Early voltage is roughly 123V, output impedance Ro~=266k (@0.5mA)

The 9018's early voltage is about 74V,  output impedance 166k(@0.5mA). 

9018 will present a lower impedance to tank then damp it a little bit. [ we did not count this, tank impedance ~= 90k]

a simple detector used, diode and a capacitor. The diode , i use 1N60, not sure if it's point contact (1N34 is point contact). 

Audio gain is 40dB.   this might be critical, cause mixer and IF stage can not provide sufficient gain{around 20dB+30dB}.

it seems hard to get it work,  after put all modules together, I even not  receive any am broadcast signal!

there is some new mistake this time:

1) the couple from tank to mixer , the C18, use 33p. this likely too small use 103 is better.

2) screw  the IF transformer slug seem no effect at all, this finally found out cause by IF transistor broken. the transistor tested on a multimeter, the Hfe  seems fine but it was broken, the IF stage does not have voltage gain at all.

3) my 2 Ferrite Rod antenna screw me: should put them in correct phase to get signal received. [put 2  coil  in correct direction on 2 Ferrite rod ]

1)  the IF transformer is random pickup and without properly design, make it  easy oscillating.  and very narrow band pass, plus no fine tune,  locate  a station almost impossible.

2)  one stage IF seems lack of gain, this time i does not use the LM386 as amplifer(40dB Gain), instead a 30dB gain amplifer(both in voltage gain) . total gain around 20dB(mixer) + 30dB(IF)+ audio(30dB) ~= 80dB. which does not  sufficiently driven the detector(at least 50mVp IF output) when receive the weak station.

3) 2 of 1sv148 is can not tracking each other,  cause: a) no voltage regulator  b) tracing can and should be resolved in a better way.

1. Redesign IF stage : so change T2 to 9018,  R1= 470k, R2=1k. make it stable and have enough BW. now the IF stage can be used as a repeatable receiver module.

2. couple from tank to mixer , the C18, use 103, with simple tracking.

3. suggest high audio gain: 40dB at least

4. tracking still should be redesign. here is the final circuit:

I found that all MW broadcast crowd into the mid 1/3 of the tune Knob, make it even hard to tune.  the reason is 1SV149 really had very high capacitance Ratio: C1v/C8V>=15, even to 25. there must be something to  tame this. So an additional D7 is added limit tune voltage  in 3V-7.5V.

1. it's fine to receive strong station, the broadcast in your city or nearby.

2. cause of no AGC and in-sufficient gain, some place it so loud, and some place you never got any voice. i found under a big bridge , the broadcast station's signal very strong!    without AGC, then this is a Wave Meter!!

3. the voice of BC is very clear,  good.

4. for improve sensitive, add a whip antenna about 1meter is very helpful. 

listen it's voice:(12/13/2012, WangJing)