Zener Diode not Zenering at all ;-( ...

I am designing a circuit with some expensive components ($60.0 plus) and I wanted to protect them from voltage spikes with Zener diode. So I took this Vishay BZX55C4V3 Zener Diode and connected it to 120 Ohm resistor and connected it to a power supply and oscilloscope.

First, as I understand, this Zener is supposed to keep voltage across itself fixed at 4.3V. But if I rump up the power supply to 10 or 11V I can push Zener to about 4.8V. So, no protection there.

Second, I have a neon lamp (one with magnifying glass) directly above my desk and above the circuit itself. Maybe 10"(25cm). If I turn on the lamp, I can see on the oscilloscope that voltage spikes up, across the diode itself, to 5.6V. Granted, this can be induction in the oscilloscope cable. Again, no protection from inductive spikes.

So what it is wrong here? It seems that these Zener diodes are neither regulating voltage from power supply, neither protecting from EMF in the environment.

How can I properly protect the circuit from voltage spikes? Is there a rock solid way to get specified voltage protection?

The simplest way to protect an input against overvoltage is a simple series resistor. The best value for the resistor depends on what voltage you need to protect against, and the function of the input.

Hmnm, adding the resistor in series is only going to dampen the spike, not remove it. I tried, just now, covering the cables with aluminum foil. No avail either.

So, no protection there.

You are expecting the zener to do something it isn't designed to do.

The issue isn't the zener. It is your expectation of the zener.

DROBNJAK:
Hmnm, adding the resistor in series is only going to dampen the spike, not remove it. I tried, just now, covering the cables with aluminum foil. No avail either.

What the resistor does is to limit the current into the pin to a safe value that the internal pin protection diode can safely cope with.

OK, but is there a way to just 100% solid limit the voltage to a desired value? So whenever a spike happens, the component just shaves it off clean.

I mean, all the tutorials on the web are talking about Zener diodes regulating voltage. When in reality they are quite loose. On this one that was rated 4.3V, it can go all the way to 5.0V and even higher.

DROBNJAK:
OK, but is there a way to just 100% solid limit the voltage to a desired value? So whenever a spike happens, the component just shaves it off clean.

Yes, use a silicon diode or Schottky diode between the signal and a power supply with a voltage just below the desired limit voltage, with a capacitor across the supply to absorb the energy. That is how the pin protection diodes work.

dc42:
Yes, use a silicon diode or Schottky diode between the signal and a power supply with a voltage just below the desired limit voltage, with a capacitor across the supply to absorb the energy. That is how the pin protection diodes work.

Many thanks. I'll try that. I've just tried the above lamp trick on L78M09CV 9.0V Voltage Regulator and it didn't glitch at all. No spikes whatsoever. No wander one always see these things on all the Arduino breakout boards ;-). So that's the way to go. Maybe I'll use Zeners for something else.

What is really the real use for Zener diodes, if they are quite loose as voltage regulators?

I tried, just now, covering the cables with aluminum foil.

As you have discovered it is a wast of time doing that. Even proper shielded cable doesn't act like you expect.
What is wrong is your expectations and supposed requirements for protection.
There is a balance between low impedance and effective suppression. The better the suppression the higher is the input impedance.
The ultimate is to add a buffer on the front end, rather than just putting a signal into an expensive chip.

Grumpy_Mike:
The ultimate is to add a buffer on the front end, rather than just putting a signal into an expensive chip.

@Grumpy_Mike
Yeah, that was what I thought Zener will do. Sorry for my newbie's ignorance, what would be a good buffer in this case?

Well you haven't said what the signal or the component is but if it is a logic level signal then something like a 74LS04 or even a simple transistor will isolate anything picked up on the line from the chip.

but is there a way to just 100% solid limit the voltage to a desired value? So whenever a spike happens, the component just shaves it off clean.

Depends on your definition of "clean". Most diodes or zvs or zeners allow some degree of variations above their rated levels, more or less.

I mean, all the tutorials on the web are talking about Zener diodes regulating voltage. When in reality they are quite loose. On this one that was rated 4.3V, it can go all the way to 5.0V and even higher.

That's how they are.

A "cleaner" solution would be a voltage reference, like tl431. However, those devices have limited ability to absorb shocks (fast / large transients).

I mean, all the tutorials on the web are talking about Zener diodes regulating voltage. When in reality they are quite loose. On this one that was rated 4.3V, it can go all the way to 5.0V and even higher.

I had a similar issue and found that zeners require a certain minimum current flowing thru them to function as a voltage limiter (check the data sheet for your zener). You might try six .7v voltage drop diodes in series to limit the board voltage to 4.2v.

Protection with diodes can be a complex matter. There is a good data sheet about the TPD2EUSB30 that goes into many of the issues.

But if I rump up the power supply to 10 or 11V I can push Zener to about 4.8V. So, no protection there.

Look at the datasheet again.... and don't overlook the definition of minimum, nominal, and maximum.
And don't forget how temperature may affect that too. look at the graphs.

You want something that regulates, put a proper 3-terminal voltage regulator there, and adjust it to whatever value you want.

I mean, all the tutorials on the web are talking about Zener diodes regulating voltage.

It's more of a voltage clamping device, not a "proper" voltage regulator.

Just a question but how do you know that the signal you see when you turn on the lamp isn't due to a ground 'artifact' or false indication. Frequently when I need to see anything like that I need an isolation transformer to verify that it (the spike) is from my measurements and not from the measuring equipment.
You mentioned that a zener diode was used in a test that failed...
Yes?
Did you measure the device within it's specifications?
I think Not.
If you compare your measurements against the data sheet you will find the part is OK and your understanding of how it "should" work needs some revision.
If you are asking about a clamp for a supply line or source or for an input then you will likely fare better by asking Mr Google about a "Voltage Clamp".
I did and this was my first observation for the search for "Voltage Clamp".
"About 3,240,000 results (0.23 seconds)"
and this:
Redirect Notice was the first return from the search.
Wiki is generally the expert, having been written by the experts for the general public. Try reading it first. It helps a great deal to know which questions to ask before trying to address your concerns.

Bob

There is this tutorial by Grumpy Mike:

http://www.thebox.myzen.co.uk/Tutorial/Protection.html

I don't know if his previous comments detract from this?

I didn't see anyone mentioning Tranzorbs, they are specifically made for such pin protection,
and very common on industrial embedded controllers. Kind of a low-inductance, fast-acting
zener, with large power dissipation capability.

On my own boards, I've been using simple 330 ohm [or so] series-Rs for years, and had
pretty good luck along the way.

How do you use these Tranzorbs? I've just had a look on Farnell.uk and all the listed ones have voltage protection in 30-60V range. Not exactly the voltage one uses with Arduino components.

You wire them across the supply. However they are not for the sort of spikes you have, they are for mich bigger ones with more area under the curve.