Re: Glass caps for bypass

Don Pearce wrote

Very abysmal AC coupling capabilities for the
dielectric in a 0.1uF cap because the capacitance is a very strong
function of the applied voltage.

In a coupling situation the applied voltage is constant, so any
non-linearity is irrelevant - as far from abysmal as you could get, in
fact. Coupling caps in general have but one job - to charge up to the
required DC voltage, then stay there. They see virtually nothing by
way of signal; certainly many orders of magnitude less than any
resistor, for example.

Same could be said of decoupling caps. A point worth exploring to the extent
that it is not true.

The difference is that, with decoupling, we tend to compare any AC voltage to
zero, whereas with coupling caps, the error is compared to the signal voltage.
In neither case, if a big enough cap is chosen, is the error very large compared
to the signal, and in neither case is the signal voltage across the cap very
large anyway.

But only if a big enough cap is chosen. What you say is only true within the
open loop bandwidth of the coupling, and only with respect to the steady state
response of the system.

Outside the bandwidth, the reactance of the cap becomes significant, obviously,
and hence so does the signal voltage across it, and so too its linearity with
respect to voltage.

An amp with coupling caps sufficient for a certain low frequency limit, which
has its bandwidth extended beyond that limit by the use of feedback, will place
an increasingly large signal voltage across the coupling caps as the input
frequency falls from the open loop limit to the closed loop limit. Distortion
will result. The nature of the distortion depends very much on the kind of cap,
with intermodulation distortion probably the worst enemy.

Caps all get increasingly non-linear as the bias voltage rises. For valve amp
coupling caps the DC bias is often hundreds of volts. But even where it is
comparatively low, and within the range of an ordinary ceramic cap, that cap may
be well be operating in its most non-linear region.

This of course is just as the feedback is falling away and no longer correcting
the error.

No point me pointing out the problem of transient response, since no-one seems
to know what I'm on about.

Talking of feedback, and polystyrene caps, the only such cap in my old PA amp is
used for compensation in the feedback loop. Perhaps to avoid temperature and age
drift in a situation where tuning is critical?

cheers, Ian


.

Relevant Pages

  • Re: Pentode voltage amp question
    ... to DC stabilize pentode bias by simply taking the screen ... up the bypass cap and off we go. ... screen voltage is the same. ... Except gain is now around 5 times what it was. ...
    (rec.audio.tubes)
  • Re: Power Inverters
    ... The "designer" copied the circuit ... Newer power supplies have built-in power factor correction to make ... the slightly higher voltage won't matter. ... > When the fast rise of the modified sinewave slams into the cap, ...
    (comp.sys.laptops)
  • Re: More cap tuning
    ... the voltage withstand is related to the oxide thickness. ... Is the oxide layer in a 600V cap 100 times the ... "forming", as you say, because the electrolyte in the cap is ...
    (rec.audio.tubes)
  • Op-Amp Design - Stage 2 Current to Voltage
    ... MH> current and converts it a voltage before it's off to the output stage. ... MH> The current input goes to the base of a transistor whose emitter is ... MH> from cap? ... the transistor is now forward biased. ...
    (sci.electronics.repair)
  • Re: Bally Mata Hari Question- Is bigger, better?
    ... If your game has more than two flippers, use a 15,000 mfd cap. ... Electrolytic Capacitor. ... coming directly from the bridge rectifier of the +5 volt supply ... The cap tends to hold up the voltage longer - providing current ...
    (rec.games.pinball)
Loading