Sean,

I think you mix up class C with class D. Class D is where the active device
is either fully on or off. This is sometimes referred as 'digital amps'; in
fact it is some form of pulse modulation (pulse width modulation, duty cycle
modulation, whatever). Think of it as a power ADC. You need a filter at the
output to reconstitute the original (analog) signal.

Class C is when the active device conduct for less than 180 degrees of the
signal cycle (assuming a sinewave signal). The output will be the top (or
bottom) of the sinewave, but for less than half a cycle. This is of course a
grossly distorted replica of the input signal. Therefor, a tuned filter
(also called a 'tank', normally a parallel LC circuit) at the signal
frequency is added at the onput. The filter starts to 'oscillate' at the
signal frequency, thus providing an amplified version of the input signal.
One consequence is that class c has very narrow bandwidth, unless the
reconstitution filter is tunable or has low Q; in the latter case the
efficiency suffers. In fact, one can argue that the output filter takes the
power of the output 'pulse' and transforms it into a sinewave. The effective
output power is the area under the output 'pulse' minus the losses in the
output filter.

As for the transistor to use, it must a) be a power device, b) have good
high-frequency specs (Fc, low capacitances, low rise/fall times, the usual
stuff). By nature it does not have to be particularly linear.

Cheers, Jan Didden


----- Original Message -----
From: "Sean Breheny" <shb7@CORNELL.EDU>
To: <PICLIST@MITVMA.MIT.EDU>
Sent: Tuesday, July 10, 2001 6:17 PM
Subject: Re: [EE]: Class C amps


> Hi all,
>
> Hi all,
>
> As usual, thanks for the quick replies.
>
> First of all, let me explain that I am not designing for a specific
> application right now. I am just trying to write a short explanation of
> how class C amplifiers are designed. I want to make sure I explain, in
> this writeup, how to select a transistor for such an amplifier.
> Therefore, I don't want to just say "pick one which says it will work for
> this application", I want to explain what is required for a transistor to
> work well in a class C amplifier.
>
> The whole point of my tutorial is to help make people "RF literate". It
> does this, in part, by explaining how RF circuits work at the transistor
> level. I think it is valuable to know how this works even if all you are
> doing is just using prebuilt modules. In addition, the lessons learned in
> trying to understand such RF circuits are useful in making sure that you
> make intelligent design choices even in using off-the-shelf modules.
> Besides, aren't you curious about how radio works at all levels? I
> wouldn't be in the hobby or in the profession if I weren't.
>
> Secondly, thanks for the book recommendations, but I already have
> Motorola's RF Device Data book and Mini-Circuits RF designer's guide. In
> fact, it was looking through Motorola's book that raised this question in
> my mind to begin with.
>
> After considering your responses, I think the problem is that "class C"
> can mean several different things and the references I have give only one
> meaning of it. They consider class C operation to be where the active
> element (BJT,FET,tube,etc.) is only either completely on or completely
> off. While in the on state, they consider it to act like a small
> resistance. This gives you a behavior which can approach 100% efficiency
> (looks like 85% max in their graph for a practical case) where the output
> amplitude depends only on the conduction angle and supply voltage, not on
> the input amplitude.
>
> Apparently class C can also refer to the case where the transistor or
> tube is not fully on in the conducting portion of the cycle. In this
> case, the output amplitude depends on the input amplitude. Because the
> average DC current can still be much smaller than for class A, it is more
> efficient than class A. So, the a similar efficiency analysis applies as
> applied above, but the circuit's output amplitude is now much less
> dependent on the supply voltage and totally dependent on the input
amplitude.
>
> If you are using a FET or tube, then, AFAIK, there is no speed-reduction
> penalty for going to the completely on (ohmic) region of operation, so
> the "saturated" model of class C operation makes complete sense for FET
> or tube circuits. It also has the advantage that you can do AM modulation
> by changing the supply voltage. I think this is why my references use
> this model.
>
> For BJTs, though, internal charge storage effects create speed penalties
> for going into saturation. In general, these cannot be modeled as
> capacitances because they invovle some pure delays (called Td and Tsd, Td
> being the pure delay to begin to turn on and Tsd being the pure delay to
> begin to turn off after being completely on). These delays depend only on
> the amount of base overdrive, and are pure delays (a sudden change in the
> input causes no change in the output until Td or Tsd time has elapsed),
> so they do not act like capacitances. Therefore, S parameters or other
> small-signal models would not model them. Yes, real switching circuits
> also have capactive delays which would be modeled by S paramters, but
> those are not the whole story for BJTs.
>
> I think this now explains why the datasheets have only S parameters and
> no timing parameters: they are for upper end HF,VHF, and UHF applications
> where it is very difficult to make a BJT come out of saturation quickly
> enough. So, they assume that you are thinking along the lines of the
> "linear for part of the time" class-C model and give you the parameters
> which help you to design such a circuit.
>
> I just wish that the books I have made this distinction. It's because of
> all of these little frustrating, subtle points that I have wanted to
> write this tutorial in the first place. It will contain a lot of
> explanations of such things which can often stump beginners.
>
> Thanks again for your help,
>
> Sean
>
> --
> http://www.piclist.com hint: To leave the PICList
> mailto:piclist-unsubscribe-request@mitvma.mit.edu
>
>

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu