[SOLVED] measuring wrong voltage on zener diode

[pwshaver]

Hi all,

I'm having trouble getting a zener diode to give me the expected behaviour and I was wondering if anyone here can shed some light on the issue.
My understanding of zener diodes is that they have a precise breakdown voltage Vz. No current flows until the voltage accross the zener overcomes Vz, beyond which point the voltage accross the zener remains precisely at Vz within the power dissipation limits of the unit.


With this in mind, I assembled the simple circuit shown in the diagram above with the zener in series with a 100 Ohm current limiting resistor. I tested three diodes of two different models. In each test case, the current was within the power limits of the zener. These are the voltages I measured accross the test zener for the two different models:

Model A, part number=1N5221B, Vz=2.40+/-0.12V, Pmax=0.5W, datasheet=**broken link removed**
Model B, part number=BZX79-3V3, Vz=3.30+/-0.07V, Pmax=0.5W, datasheet=**broken link removed**

Model	measured voltage (v)
A	2.92
A	2.82
A	2.88
B	4.18
B	4.19
B	4.19

Then, using a potentiometer (total resistance between the high and low pin = 3.65 kOhm) in the circuit shown above, I've determined the minimum voltage required to observe any current (~0.01A) through the zener:

model	min voltage needed to observe current (V)
A	0.87
A	0.79
A	0.84
B	1.52
B	1.73
B	1.60

I don't know how to explain these observations. In all the cases above, I have the cathode high with respect to the anode, which is at ground. I think this is the correct orientation. Is my understanding of zener diodes fundamentally wrong? How likely is it that all six of the tested zener diodes are faulty or that there's something wrong with my multimeter or protoboard? Is there someting wrong with the circuits I'm using to test the zeners? Is there something else I'm missing in all of this? I would really apreciate it if anyone could shed some light on this issue for me.

Thank you,

Paul

 

[D.A.(Tony)Stewart]

All diodes including Zeners have a threshold voltage and a rated voltage with a specified current where an ESR can be used to predict the VI variations around that operating point.


In your case you ignore this operating point and merely applied a 100 Ohm series R to 12V.

Zener Voltage Range
. . . . . . Nom (V) Min (V) Max (V) mA Ω . Ω . . µA
1N5221B 2.4 . . .2.28 . . . 2.52 . . 20 30 1200 0.25mA

Projecting 2.4V with 20mA @30 Ω would get 2.6-600mV = 2.0V @0mA assuming ESR stays at 30Ω but at 0.25mA the ESR is 1200Ω.

With your 100 Ω 12-2V would drive 10V/100Ω= 100mA which would raise the Vf well above nominal voltage.

If you choose the nominal operating current, then you can use the ESR to see what happens to Vf when you deviate from that current by 50%. this is a guestimate of the range for ESR to change only 5%.

 

[Audioguru]

The datasheet for the 1N5221B shows that when its current is 20mA then its voltage is 2.4V +/- 5%. Your current is (12V - 2.87V)/100 ohms = 91mA so of course your tested voltage is too high.
The datasheet for the BZX79-3V3 shows that when its current is 5mA then its voltage is 3.3V +/- 2.5%. Your current is WAY higher so of course .....

Your idea "No current flows until the voltage across the zener overcomes Vz, beyond which point the voltage across the zener remains precisely at Vz" is completely wrong when you look at a graph of the voltage different zener diodes at different currents. The low voltage ones like you are testing have awful voltage regulation when the current changes (the voltage changes even when their temperature changes). The 2.4V zener diode is shown conducting 0.1mA with a voltage of only 1V! Extend the diagonal line (good voltage regulation would show a vertical line) to see how its voltage is higher than 2.8V at your very high current.
Here is the graph:

 

[BradtheRad]

Notice tolerance on zener diodes is commonly 5 percent.
It can amount to as much as +- 1V for a 20V zener.

If necessary you can obtain them with tighter tolerances.

 

[D.A.(Tony)Stewart]

To get the expected voltage from 12V, one would start with the rated current of 20mA for the 2.4V Zener then use Ohm's Law.
(12-2.4)/20 [V/mA] = 480Ω

Checking the thermal rise @300 °C/W for 48mW gives a rise of 14.4 °C with a tempco of -0.085 %/°C or a drop of 1.2% with a tolerance +/-5% or +4/-6% roughly.

The 480Ω resistor meanwhile dissipates I^2R = 0.2W which would get hot in 1/4W part. Better choices exist for new design of a 2.4V regulator.

Read and understand the Zzt specs for Zeners. Although they are not always given for LED's and other semi's, you can read them from the VI saturation curves and we usually call it the ESR or is reported as RdsOn (FET) or Rce(sat) (BJT).

 

[pwshaver]

Thanks for your replies. I can see that I was expecting far too much from these zener diodes. My application was actually in ESD protection, which was why I was using such a high voltage but I'll reconsider the design now.

 

[Audioguru]

ESD is tens of thousands of volts, not 2.4V, 3.3V or 12V.

 

[D.A.(Tony)Stewart]

ESD protection comes in many forms. (Schottky%1 and ) Silicon diode string, zeners, MOVs, SCRs etc.
For series Led strings , LED protectors or PLEDs also have SCR circuits inside to conduct when an LED goes open.
https://www.littelfuse.com/~/media/...led_protector_pled5ht_sot89_datasheet.pdf.pdf

%1 this is the preferred ESD protection for CMOS, wirh current limits of 5~10mA, in very small junctions, these Schottky diodes must conduct faster than the CMOS jcns. they are protecting.

 

[asdf44]

Notice tolerance on zener diodes is commonly 5 percent.
It can amount to as much as +- 1V for a 20V zener.

If necessary you can obtain them with tighter tolerances.

And it's far worse than that because the zener voltage is calculated at a particular current. At that specific current it's guaranteed to 5% but zeners, especially low voltage ones vary wildly, as the OP discovered, at different currents.

If you actually need accuracy in a low voltage zener you should look at the TL431 and variants.


Also, as others have said, a regular old zener may seem logical for ESD protection but it's (usually) not optimized for that role. Get something dedicated to that task like what SunnySkyguy suggests.