AN920

He talkes about a biased Class-B but says this is not a Class-AB. The distortion caused by gain-doubling in Class-AB seems to be more of a problem to him than to deal with the crossover distortion in Class-B amplifiers. He also states that most amplifier designs suffer from so much distortion in the small signal amplifier stages that it swamps the distortion caused by the output stage.

a) C1 helps to improve the impedance match to the tuned circuit. b) Antennas of a certain length have a typical impedance at a certain frequency. The explanation given by the writer is too simplistic. It is a bit more complicated than that. I suggest you read up on type of antennas, impedance and matching. It is too much to explain here in detail. Here is a quick start http://www.borg.com/~warrend/guru.html c) Antennas can be inductive or capacitive. d) In the FM detector the turning on of the transistor places the resistor in parallel with the tuned circuit, lowering the Q, allowing a wider bandwidth to be detected. The transistor will have low impedance between collector and emitter when turned on by base-emitter current. The 100nF cap serves as DC isolation to the internal Gilbert cell quad detector. Another way that is often used in receiver design to obtain wider bandwidth but retaining overall rejection outside the passband it so use cascaded filters like the top-c coupled filter in the diagram. We get wider bandwith but the side skirts remain steep. I added R1,2 to give the inductors a Q value like a real life part.

I don't think "Self" is calling it Class-B by mistake as he states in the foreword of his book The problem comes that most people see a biased Class-B amplifier and think it is a Class-AB. "Self" sees the same amplifier from his Engineering analysis as a Class-B. Also from the book:

I once used this loading effect to design a switchable narrow-wide FM detector in a measuring instrument.

Fundamental mode crystals can typically be pulled by 0.2% and overtone crystals much less than that.

This is technically not correct. Reference: Douglas Self - Audio Power Amplifier Design Handbook, 3rd edition

Last diagram

1) Are u mean the series impedance of the antenna? And if so, What can the antenna be considered from this point? The capacitor add resistance in series so that the low characteristic impedance of the antenna (300, 75 or 50 Ohm) does not load the tuned circuit too much. (This will become more clear when you study the plots below) 2)Are u mean that with C1 it is will be like connecting a high value resistance in parallel with the tuned circuit. Yes, with C1 in circuit the total resistance loading the tuned circuit will be (Z- antenna + Xc) apart from the coupling to the rest of the receiver. 3)Why u consider the antenna as alow resistance? I have heard that it is 50, 75 300 ohm, why antennas are low value resistance? It has been determined that the various antenna types will have that resistance, and can be assumed to be a load of that resistance when connected to a RF circuit. You read up on this using Google. 4)Why antenna is considered to be // to the tuned circuit, why not inseries? The antenna acts as a load between point of connection and ground of the circuit. I have added the diagrams as help. First diagram shows tuned circuit with low loading from antenna (sine source 300 Ohm) and rest of receiver circuit (50k). Note the the dB magnitude. Second diagram shows the same circuit but wil larger coupling capacitors. Note that the selectivity got worse but the dB level to our receiver is better. Third diagram shows a large coupling capacitor that will in effect directly couple the antenna to the tuned circuit. Note total loss of selectivity and severe loss of signal level. Last diagram shows the effect of placing a load R2 (300 Ohm) in parallel with the tuned circuit. It is very similar to directly connecting the antenna with even more signal loss.

Receiver input design is a balance between losses and selectivity in the tuned circuit. More selectivity requires less loading of the tuned circuit. Less loading results in more signal loss, from the signal coupled to the tuned circuit. Connecting the antenna directly will be like connecting a low value resistance in parallel with the tuned circuit, killing the selectivity. The capacitor increase the series impedance Xc to minimize loading while still passing enough signal.

Theory tells us that to reproduce a signal waveform we have to sample at least 2X the bandwidth of the signal we want to capture. In practice we need more for accurate reproduction. For instance in digital oscilloscopes the sample ratio may be from 2.5 to 4+ You pay a price for faster sampling in scopes as you need more and faster memory, as you will be filling up more memory much faster per display capture. Also read http://en.wikipedia.org/wiki/Oversampling

I am not an expert with this program as I hardly ever use it. I find the user interface too clumsy and drawing large schematics a pain. There should be some option to display the node numbers on the diagram. You can go "view spice netlist" as another option. The dots indicate the phasing of the coupled inductors. From help file

If you wanted the circuits to be isolated, you may do something like this. Spice needs to see some path to circuit GND otherwise you will get an error. You may make the resistor much higher in value.

Hope this helps. * L1 N003 0 1mH V1 N003 0 SINE(0 100 1000) AC 10 0 Rser=10 L2 N001 0 100

Here is osc startup with SWcad. Probe at emitter. Also current into emitter and in collector circuit

In help files of SWcad

The program is not free. They used to have a student or demo version. I am not sure if that is still the case. This link http://www.campustech.com/c/campust/title.html?id=gDHxDQRf&mv_arg=NINT100&mv_pc=65 you can buy it for$35 You should be able to use Linear's switchercad also which is free Another free spice http://www.5spice.com/download.htm Here is the same result using SWcad

Program is Multisim from National Instruments. I indicated a total phase-shift of 360 degee which bring you back to 0 degree, correct. In oscillator design the term loop-gain is used to indicate the total gain in the loop (gain of amplfier - any losses from tuned circuit and other elements). Initial design starts with an open-loop analysis and ends with closed-loop analysis. Phase shift caused by tuned circuit (C1//L1+L2) is 180 degree, and amplifier adds another 180 degrees for a total of 360 degrees. I have added the last diagram (wthout the tuned circuit) to show that any change in current into the emitter terminal will cause an output waveform shifted 180 degrees at the collector. I have over driven the amplifier so that the phase shift will be clearly noticeable.

With all basic elements connected we can see that circuit starts to oscillate when power is applied.

It is important to note what happens to phase at the point of resonance. Look at the attached picture of gain(blue) vs. phase(red) and it will become clear. There is no active gain due to absence of the amplifier. Now if you add the amplifier, the additional phase shift will result in a total phase shift near 360 deg with gain above 0dB. Now all oscillation criteria are met and circuit will oscillate Last picture shows the additional signal change through the amplifier circuit. It is clear that the amplifier circuit adds more phase-shift of its own and a total loop gain >1 (~ 5dB) and circuit can now oscillate.

http://www.electronicsforu.com/EFYLinux/circuit/may2003/ci4-clap.pdf

Employ a current limit circuit and power your voice circuit before the current limiter.

I think the text says it all that it is a coupling capacitor to couple back the signal for the circuit to oscillate. You don't need high gain for the oscillator. The total loop gain only need to be >1 to overcome circuit losses and enough phase shift to feed back the signal in-phase. The value of the coupling capacitor is not very critical as long as it has low enough reactance (Xc ~ 10 Ohm) at the frequency of interest. Oscillators with more than 6-8dB of loop gain is normally not a good thing, as it produces many other unwanted products. Think of an oscillator as an amplifier with A>1 with in-phase feedback.

http://www.electronics-tutorials.com/receivers/am-radio-receivers.htm

You may find this handy http://www.kpsec.freeuk.com/components/led.htm

A led will emit enough light for your purpose (to locate the switch in the dark) at low current levels <5mA or so. Depending which battery you use, it may last for days or weeks. If you want to make the led to flicker that can be done with a simple circuit or use a flickering led from your local electronics store. How do you want to use the delay timer and clap switch? Is this part of the project?