Reflex AM Detector

Interesting, an extra 12.7pF cap parallel to the coil would better suppress the carrier signal. Keep it up!!!

Thanks! Yes, adding a cap will help to filter the carrier better, but it will also narrower the low-frequency bandwidth of the detector.

No, i m not talking about the lowpass cap. Sure increase it just lowers the corner freq. I mean a bandstop tank made with the existing 2mH plus parallel extra 12.7,pF! This would almost completely block the carrier.

Here you can see the modification of your circuit http://everycircuit.com/circuit/5032795027013632

I got what you meant, hurz. What I mean is that, theoretically, replacing an inductive load by a high-Q LC tank makes an RF amplifier more narrowband. But as I now checked in the simulator there is still pretty much bandwidth in this circuit :) Capacitor added!

It just blocks the carrier and does not make the AM bandwidth smaller. Its still 100kHz. The only thing one could argue, the cap is quite small and the inductor big, so the parasitic capacitance of this coil might already be bigger then 12pF. However this can also be changed by adapting the lowpass impedance. Now the sinewave looks perfect ;-)

Hurz, the LC tank is the load of the RF amplifier. If this tank has (external) Q factor such that Q > 5, then its bandwidth will not be enough to fit the whole AM signal bandwidth (200 kHz in this case). This is what I am talking about.

stanislav_maslovski, the lowpass -3dB corner is already at about 100kHz. So having a notch at 1MHz wont change anything. Anyway Q will be much higher than 5! The notch freq is 10 times higher then the AM bandwidth!!!

Ok, let me try to explain better. This transistor in this circuit is both RF and LF amplifier. At RF the load of this amp is the LC tank (and the input of the detector) , and at LF it is the low-pass filter. As you say correctly, the cutoff frequency of this low-pass is around 100 kHz. Adding the capacitor you suggested does not change this cutoff frequency. However, adding this capacitor changes the load impedance at RF, thus it affects amplification at RF. In other words, it affects the amplitude of the RF signal at the point BEFORE the low-pass (at the collector). Do you get it now?

Now more on Q. In this circuit, the characteristic impedance of the LC tank is about 12.5 kOhm. This tank is shunted by the input impedance of the detector which is relatively low (let us say 10-20 kOhm), and because of that the EXTERNAL (i.e. loaded) Q of this LC tank is low, and its bandwidth is large enough to fit the 200 kHz AM signal band.

To make it short, the cap improves the demodulator output a lot! Now, were exactly do you see a problem?

It improves suppression of the carrier, yes, but depending on the other parameters of the circuit, it can reduce the amplitude of the detected LF signal, and also distort the spectrum of such signal (in a situation with many frequency components in this signal). I do not get why you do not get it :)

The mentioned frequency distortions will appear when the bandwidth of the LC tank 2*df = f_res / Q is smaller than twice the modulating signal bandwidth.

Keep cool, i like what you are publishing on EC. There are not many circuits i pay attention when users are going public with it. But your name is always promissing good circuits ;-) check this and let us continue in discussion http://everycircuit.com/circuit/4557394358042624

I checked your circuit. It surely demonstrates what you want to show, but as I said above myself, in THIS particular circuit, with THESE particular parameters of all the elements, there is no problem of adding the cap: The external Q of the LC tank in this case is below unity, and the respective resonance in transistor's load impedance is very wide around 1 MHz. But imagine another situation: one may think that 12 pF cap is too small in practice for 2 mH coil (as you mentioned yourself!), and thus one may want to split the 2 mH coil in two parts, with the smaller part shunted by a cap to get the same notch effect at 1 MHz. Then, there can appear a problem which is shown and explained here: http://everycircuit.com/circuit/6430306658680832

I wont split the coil in two parts!