I need a bit of help with Zeners

[tubeswell]

... you can never, ever use a zener without some form of current limiting.

Hi R.G. In the hope of learning something (and acknowledging that I probably have missed something else you said that was relevant, but) amps I have built seem to be able to quite happily take a sole 15V 5W reverse biased zener on the HT CT. (I figured that allowing for a max current draw of 200mA for e.g.; a 5F6A PT, the zener would only ever dissipate a maximum of 3W. None of the situations where I used zeners in this way has blown up yet and I've had these amps running for hours at a time.).

[Steve Conner]

Well, here the rest of the amp is functioning as the "current limiter". It's a fundamentally different use of a zener: knocking 15 volts off instead of regulating something down to 15 volts. Maybe RG can explain better...

[Albert Kreuzer]

HA! I just hand-soldered a TPS61165. Don't mess with me!

I had to look up TPS61165 - wow!

[R.G.]

Remember that a zener's voltage is specified at a given current (Izt). If you force more or less current then the Izt your Vz will vary from specified.

Actually, for the benefit of the beginner (in understanding, at least) that asked the original question, I was simplifying pretty heavily. I didn't want to confuse him with discussions of the zener knee, sharpness thereof, internal bulk resistance, small signal impedance, temperature sensitivity, noise density, internal capacitance, and phase of the moon. 8-)

Yes, real zeners, like real voltage sources, deviate from perfection. The skilled and experienced designer may not know all of the deviations, but *will* know that deviations from ideal exist in every component, and that he should look up the specified deviations on the component datasheet, as well as needing to know what deviations matter for the design at hand, and which ones do not matter. In many cases, knowing which things matter and which don't is the mark of the expert, not encyclopedic knowledge of details.

Precision is a funny thing. People assume that precision is a good thing. However, obtaining too much precision is like pouring good whiskey on the ground - it costs a lot and does you no particular good. One has to know when precision is worth the price.

To the audience: zeners are imperfect. Read the datasheets, and be aware that it is quite difficult to achieve high precision with zeners. Your mileage may vary, and some assembly is required.

[Enzo]

I seem to recall reading somewhere that zeners below some low voltage - 3 or 4 volts maybe - were not actually zeners. I forget the rationale for that assertion though. Implication was that while they were used as zeners, the internal process was somehow different than in a "true" zener. In terms of us guys using them, it made no practical difference apparently.

I don;t claim zener expertise, but also seems to me I recall that for good regulation the current through the zener should be much larger than the circuit draw from atop it. Which is why a couple op amps might run with supplies regulated by a 1 watt zener, and a circuit with a bunch of them would use a 5 watt zener. SInce the draw for the circuit would come through the same resistor limiting current for the zener, the zener current would have to be large compared to the circuit current to keep the operating conditions in the zener more or less constant - and thus the zener voltage stable.

Is that even remotely correct?

[R.G.]

I seem to recall reading somewhere that zeners below some low voltage - 3 or 4 volts maybe - were not actually zeners. I forget the rationale for that assertion though. Implication was that while they were used as zeners, the internal process was somehow different than in a "true" zener. In terms of us guys using them, it made no practical difference apparently.

Close, but actually it's the other way round. True zener action as defined in the semiconductor textbooks happens below 5? 7? volts, something like that. Above that, the mechanism is avalanche breakdown, which is similar but different, a little. Sorry for being fuzzy - it's been over 35 years since that lecture.

And you're right - it makes not much practical difference.

I don;t claim zener expertise, but also seems to me I recall that for good regulation the current through the zener should be much larger than the circuit draw from atop it. Which is why a couple op amps might run with supplies regulated by a 1 watt zener, and a circuit with a bunch of them would use a 5 watt zener. SInce the draw for the circuit would come through the same resistor limiting current for the zener, the zener current would have to be large compared to the circuit current to keep the operating conditions in the zener more or less constant - and thus the zener voltage stable.

Is that even remotely correct?

Yeah - if you want a zener's voltage to be stable, keep it at a constant current. There were lots of circuits driving zeners from constant current sources for precision references, and it still wasn't good enough once we discovered band-gap references.

However, as a practical matter, there are three ways to use a zener - (a) as a rough-and-ready shunt regulator, (b) as a series voltage dropper, and (c ) as a precision reference. This last is very, very difficult to do because of the internal impedance of the zener. No zener has a truly vertical current/voltage curve, so any current change affects the voltage.

In both the series droppers and the rough-and-ready regulators, you have to be not so much concerned with precision, because you can't get it that way.

You are correct in that for a shunt zener regulator to not shift voltages, you have to keep it's current approximately constant. It's not the load current so much as the variation in the load current. In a shunt regulator, the series resistor does the voltage dropping, and the zener manages the voltage approximately constant by eating up the variation in load current. Load current goes down, the zener eats more; load current goes up, the zener lets it have more of the available current. When you get near the zener's knee, regulation goes out the window as you starve it for current. When you get to too high a current, thermal effects change the voltage and either the zener or the dropping resistor burns up. So a shunt regulator has to have a varying current, otherwise you may as well drop the voltage with a resistor, you'd get the same result. But varying currents vary the zener voltage, they have to, or there's no regulation at all. It's a no-win if what you want is precision.

But it *is* cheap.

Case in point - those %^#@&*% silly zener regulators in the Fender Blues Deville and Hot Rod Deville. They were too cheap to put in three terminal regulators, so they used the cheapest possible resistors and zeners. The result is that most of these amps have burned spots under the resistors, and some of them have charred-black spots under the resistors or intermittents from the resistor leads melting solder. It is not easy to design a good, precise, zener-shunt-regulated power supply.

It may be impossible. I've never seen a good one.