How come its OK to mix impedances in a radio system??

Reality? Nah, it'll never sell.

-- "Electricity is of two kinds, positive and negative. The difference is, I presume, that one comes a little more expensive, but is more durable; the other is a cheaper thing, but the moths get into it." (Stephen Leacock)

Hello John,

Yes. I meant the more wideband versions.

I only tried the 99c FM versions from Walmart. All I can say is that the RF performance was absolutely deplorable. Not useful to me without major mods.

To save some cost you only get a button to scan the band. It's hit and miss if you are trying to listen to your favorite station. Maybe they come out with a contractor-special AM version some day that has a Limbaugh speed button ;-)

Regards, Joerg

I don't know about the am/fm ones.

But I took apart a $1.99 "scanning" FM radio, and it uses a variant of the old TDA7000, which converted the signal down to the tens of KHz range, where an active filter could provide selectivity.

The IC used, a 7088 if I remember properly (and I've posted in various newsgroups about the things, when the material was fresh in my mind) adds some circuitry for the tuning. Press a button, and it starts a ramp that feeds the varactor(s). When a suitably strong signal is tuned, there is detector circuitry to notice this, and stop the ramp. The tuning voltage at that point is kept, obviously. Press the tune button again, and it continues from there. Press the reset button, and the voltage starts at the beginning again.

As I said, I haven't looked inside the AM/FM ones. I suspect they are a TRF design, based on cost. Plus, the 7088 datasheet I looked at showed something like that for an am/fm version. It might have even been a descendent of the old ZK414, which after all has to have seen some commercial use over the years in order to be such a long-lived part.

Michael

On Sun, 17 Jul 2005 11:18:22 -0700, John Larkin wroth:

Classic loopsticks for the AM band almost always have a coupling coil wound around the cold end of the loopstick in order to better match the somewhat low input impedance of the following converter/amp stage.

Jim

Hi Joerg and Fred,

I can see you are both on the right track.

It's surprising how many engineers and radio amateurs assume, after tuning up for maximum output, that a conjugate match exists between the power amplifying device and its external load.

Actually the circuit designer doesn't need to know what the internal impedance of a high power HF amplifier actually is. The device manufacturer doesn't even bother to mention it in the data book. Nobody wants to know.

You are wrong on one account, Joerg. The internal impedance of a power amplifier is hardly ever lower than its correct load regardless of whether it is operating under Clsss A, B or C conditions. It is usually much higher than the external load and there is a 'poor' impedance mismatch.

Both tubes and transistors are operated as more like high impedance current sources than very low impedance voltage generators like power stations.

Typical RF PA devices are Beam Tetrodes which may have an internal anode impedance as high as several hundred thousand ohms. It's a long way down to the usual RF load of 50 ohms. With audio ampifiers the loud speaker load impedance can be 8 ohms or less.

At the lower frequencies transistors can have internal resistances approaching a megohm.

Actually, the output transformer turns ratio depends on peak voltage and peak current ratings of the device, on the DC supply volts, and on the power to be developed in a load of given resistance. The internal resistance of the device has nothing to do with it.

Circumstances may occur at extremely high frequencies in which knowledge of a transistor's internal impedance is useful to obtain optimum operation. There is often feedback via the relatively low internal impedance. Circuit designers then take the easy way out and copy the complete amplifier circuits conveniently provided by manufacturers following a lot of experimentation. But they still don't know what the internal resistance is.

Thanks for allowing me in.

If your argument is that the internal plate resistance is lower when you run it in class C service, I disagree. After all, the tube is cut off most of the time. What is its resistance then?

John

Hello Reg,

Got to disagree there, Reg. The biggest honking transmitter I did had two QB5/1750 (the size of a pickle glass, each). Let's just take that one as an example:

These are tubes operated at 5000V on the plates and typically around

If they were current sources they would melt down despite rated 500W plate dissipation.

I don't know where this tube would be on the web but it's smaller sister is. Look at page 14:

With all due respect that would be a recipe for disaster. As a designer you have to know what the internal resistance is at any given drive level. It's all in the data sheet, for transistors as well as for tubes. Else you would not be able to properly calculate the power dissipation, among other things.

Regards, Joerg

Joerg,

I'm afraid you have entirely the wrong idea on how to estimate plate internal resistance.

You need a set of characteristic curves for the tube (or transistor). The graph will show how plate current varies with plate voltage for different values of grid voltage.

Plate internal resistance equals -

(Change in plate voltage) / (Change in plate current)

for constant grid voltage.

It is the SLOPE OF THE GRAPH which determines the plate impedance.

I looked up the graph for the QB4-1100 tetrode.

For a change in plate voltage of 1000 volts, the plate current changes by 100 milliamps for zero change in grid volts. This corresponds to a plate resistance of 10,000 ohms. Which is somewhat lower than I would expect for a tube of that type but nevertherless quite high.

If you have an actual amplifier circuit which shows how much the output tank circuit transforms down the voltage to a 50-ohm load, you will find that the impedance looking back from the load into the amplifier is very much higher than 50 ohms.

Why not design a circuit of a simple Class-A audio amplifier, including the turns ratio of the output transformer. Do you need to know the plate impedance?

The higher powered sister will have a very similar characteristic. Difficult to tell the difference without looking at the figures.

---- Reg.

====================================

regardless

is

impedance

power

had

that

to

them

that

circuit

sister

designer

tubes.

dissipation,

Hello John,

(

The usual way a final amp tube is run is to keep the DC supply constant, or as constant as the power supply affords. Vg1 is modulated which then controls the tube current. Yes, you can get higher impedance values with less Vg1 as can be seen in the graph. But in an efficient transmitter application that isn't done. They want to transit that region as quickly as possible.

Modern AM transmitters (if there still is such a thing...) or industrial RF power generators work hard between the extremes. Full on and full off into the tank circuit. The modulation happens via plate PWM or grid PWM. That way you avoid the linear region where the tube dissipates a lot. Tubes and FETs are pretty close relatives in this respect. Even the old AM transmitters drove the RF final hard. Of course, when you have to do a linear such as for SSB or TV transmission that is another story.

However, even an SSB amp can be operated with the tube in "digital mode". A friend in Germany did that. They used to have a rule for ham radio that the highest license class could use tubes up to 150W total plate dissipation rating. The official data sheet was what counted. So, a lot of people used five color TV flyback tubes of the type PL509, rated 30W each. Then one guy figured out an elaborate PWM scheme and cranked several kilowatts out of those five little tubes. This wasn't just about bragging, it was witnessed on a Bird watt meter. That wasn't possible in class AB where the maximum was around 800W and you had to limit it to a minute or the plates would go from dark red to orange and then white.

IIRC the PWM scheme wouldn't work well in pulsed mode (CW). But then again the logic chips he had available in the late 70's weren't nearly as capable as today's, except maybe for the very expensive ECL chips. AFAIK nobody builds these type of amps anymore becuse the rules there changed. Now it's 750W output no matter which tubes you use.

It is surprising that the audio world learned about the PWM concept much later. It wasn't until the mid 90's that class D amps became popular.

Regards, Joerg

Hello Reg,

That is right if you operate the tube in its linear range. Besides single sideband and video transmission that is usually not done. Else you would not be able to squeeze about 90% efficiency out of a large AM transmitter final amp. That is about the efficiency range they offer these days and the one I have seen was actually about 25 years old.

Running an AM or PWM transmitter final in the linear region of the tube would be like stepping on the brake of a car while not releasing the accelerator. Can be done but long term some smoke would appear.

Absolutely. Class A is a fully linear (and pretty inefficient) amplifier topology. Same goes for class AB in push-pull except that it doesn't burn as much quiescent power. But when you pulse the tubes you will get a lot more power out of it. The turns ratio would still matter but with it you merely set how much power you want from the amp or how hard you feel comfortable pushing the tubes when they are on.

It is pretty similar.

Anyway, even modulators use PWM techniques these days and it has made it into the ham radio community. A little info about that can be found here:

Regards, Joerg

Hello John,

Pretty much infinite. But that's exactly what I am trying to say. For maximum efficiency you want it to ping-pong between cut off and full bore. You can't do that in class AB because full bore means heavily into the non-linear range. Same for cut off.

Regards, Joerg

Joerg,

Obviously, as the operating angle decreases, ie., as the tube or transistor goes further into Class-C, the effective internal resistance of the device increases.

BUT YOU STILL DON'T NEED TO KNOW THE INTERNAL RESISTANCE TO DESIGN THE TRANSFORMING NETWORK BETWEEN DEVICE AND LOAD. or for anything else.

Programs TETRODE1 and TRIODE1 take you through the design procedure for Class A, B or C power amplifiers. They don't need the value of the internal tube impedance. One of them calculates it out of curiosity after the design is complete. It is not equal to the load resistance.

Download these 2 programs from website below and amuse youself.

---- ........................................................... Regards from Reg, G4FGQ For Free Radio Design Software go to

Just another comment about maximum power transfer. I don't know if someone else mentioned this.

Given a source, with output impedance ZS, the load impedance that attains maximum power transfer is ZL=ZS* (complex conjugate).

However, given a load, with impedance ZL, the source impedance that attains maximum power transfer is ZS=0 (zero), not ZS=ZL* !!

Curiously, I've known several people who had problems understanding this.

I think the confusion comes from...

Given: SOURCE Impedance = Zo, maximum power transfer to LOAD occurs when LOAD Impedance = Zo

But if LOAD Impedance is _pre-specified_, maximum power transfer to LOAD occurs at SOURCE Impedance = 0 (zero)

...Jim Thompson

Hello Reg,

Thanks. Maybe I'll do that some day but right now I am busy with another design.

My tube designs were all done by hand and I simply looked at them the same way I look at FETs. You can use them as gate (grid) controlled resistors or as switches. The latter is preferred in high efficiency RF amps, for example. Not in linears though unless you do PWM.

At the beginning of the design I just picked the "stress parameters". How much drain (plate) current can I safely demand at which voltage? With tubes I sometimes stretched that a bit since it was for hobby purposes. As long as they didn't get past cherry red ;-)

That was all RF. But I did an audio amp in the same way. There was no turns ratio since I disassembled the coil of a (huge) speaker and rewound it, tapped and all. Just set the max current and then wound it.

Regards, Joerg

Hello Fred,

They usually either know pretty well how to design a power stage or they follow a plan such as a DYI article in a ham radio publication.

Depends. At least in the ham radio community you often have to do trade-offs. In some European countries the legal max used to be given as max plate dissipation. That inevitabley resulted in some radio amateurs walking that fine line between a plate that glows red and one that glows white. Been there.

Regards, Joerg

Anybody who thinks that *isn't* an engineer!

I wouldn't know about radio amateurs.

"Tuning up for maximum output" is fraught with hidden dangers. Depending on circuit design, it is possible to get the loading wrong on many transistor final amplifiers, and get more (sometimes considerably more) than rated power out, usually at the expense of reduced reliability and increased spurious products. Maximum output may not indicate correct adjustment. Clean output at rated power is what to go for. It might just save you from that 3 AM trip up a dirt road halfway up a mountain in the middle of winter.

Maximum power transfer in a system *always* depends on matched impedances. The point is that maximum power transfer is seldom either necessary or desirable, as your power plant example demonstrates.

I was speaking "resistively" ;-)

...Jim Thompson