# Zener diode confusion

#### roombacurious

Joined Sep 25, 2007
4
Hi,

I'm trying to understand how Zener diodes exactly behave but I'm failing so far... I'm making a battery charger for a tool which uses a voltage comparator to stop the charging once the battery is full. I'm using a Zener diode to generate a reference voltage, connected in series to a resistor.

My problem is that the voltage across the Zener diode is NOT stable but seems to greatly depend on the value of the resistor (i.e. the current). I've found that all Zeners seem to produce the specified voltage when the current is 5mA (as in the specs sheet) but any variation to the current (within their dissipation specs of course) will also change the voltage, sometimes pretty significantly.

I.e. I'm not getting a stable voltage irrespective of current which I though was the function of a Zener...

What am I doing wrong???

Thanks,
George

#### Eduard Munteanu

Joined Sep 1, 2007
86
Shoudn't you ground the Zener diode directly? It's supposed to sink as much current as necessary to provide its rated voltage drop.

#### bloguetronica

Joined Apr 27, 2007
1,453
Indeed, a real zener diode is not stable. The breakdown voltage increases with the current biasing the diode. A real zener diode acts like an ideal zener diode and a small resistor in series (resistance Rz). Most datasheets predict the value Rz. So, if you want to stay close to the zener voltage, use a resistor with a big resistance, in order to have a small current biasing the zener. And of course, count with the input bias current of the comparator you are planning to use.

#### roombacurious

Joined Sep 25, 2007
4

Indeed the Zener is directly grounded (i.e. + -> Resistor -> Zener -> - ).

I was not aware of that Rz issue! In all the tutorials I've read Zeners are presented as stable but what you're saying makes much more sense. On the other hand it does limit their use quite significantly...

So is there a better way to produce a reference voltage of one's choice, besides using an IC?

#### GS3

Joined Sep 21, 2007
408
I am looking at the datasheet for the fairchild BZX79C2V4 - BZX79C56 Zener Diodes. For a voltage of, say, 9.1 V @ 5 mA, the differential resistance is specified as 15 ohm which means an increase of 10 mA will produce an increase in voltage of 150 mV. If you need more stability than that then you can follow the zener diode with an amplifier so that the current through the zener is more constant. You will also want to look at voltage drift with temperature. If you want further precision you can look at the LM431
Adjustable Precision Zener Shunt Regulator.

Also bear in mind that individual zener diodes have wide tolerance and you will need to design a circuit where you can have a zenner as a voltage reference but the set point voltage must have some means of being adjustable.

We would need more information about the actual circuit in order to provide useful advice. How is the diode connected? What is the variation in current?

If this is for charging batteries you should also take into account that the temperature of the battery also should be taken into account.

The type of battery is very important. Terminating the charge on reaching a certain voltage can be used in lead-acid batteries but is not recommended for NiMH.

#### roombacurious

Joined Sep 25, 2007
4
It is indeed NiCd and NiMh battery packs that I'll be mostly charging. I know voltage comparators are not the ideal way to control charging for these types of batteries (not least because the internal resistance varies, not to talk about thermal damage) but it is an easy way and will probably be ok for now. I may however add a thermistor to the circuit once I get the Zener issue solved!

The circuit is pretty simple:

(G)----- Battery to charge --.......................................................(+)
|......................................|.......................................................|
|......................................|....... |---- Led ------- Resistor A-------|
|......................................|........|.............................................|
|......................................|........|---------- Charge resistor -------|
|......................................|E..... |C............................................|
|.................................Power Transistor.......................................|
|............................................|B...............................................|
|............................................|.................................................|
-------- Zener --------------------------- Resistor B-----------------

(This ASCII stuff is difficult to draw! Ignore the dots - used for spacing)

The idea is that the Zener creates a reference voltage - if it's higher than the battery one it allows current from the charge resistor to flow to the battery. It also lights the led up. The currents I've been using are about 200mA (charge resistor) + 20 mA from the Led resistor and 5 mA or so from the Zener resistor. The idea is that the battery self-regulates what sort of charge it is happy to accept (by means of changing it's internal resistance). It probably also allows for a sort of trickle charge at the end of the main charge if the Zener is chosen appropriately. Voltage between B-E can range (ideally) between 0.3v (full battery) and 2.0v (empty battery). Practically the upper limit is up to 0.5v higher due to battery internal resistance variations.

Any thoughts / improvements on this without complicating things too much?

Thanks

#### bloguetronica

Joined Apr 27, 2007
1,453

Indeed the Zener is directly grounded (i.e. + -> Resistor -> Zener -> - ).

I was not aware of that Rz issue! In all the tutorials I've read Zeners are presented as stable but what you're saying makes much more sense. On the other hand it does limit their use quite significantly...

So is there a better way to produce a reference voltage of one's choice, besides using an IC?
No, the zener is OK, as long as you limit the current biasing it. Most comparators only need input currents in the order of 100nA. If you consider a current 20 times greater than that, say 2uA, the voltage drop across Rz (the hypotetic series resistor) will have no significance, sinçe Rz value is in the magnitude of of 10Ohms. Thus, the voltage drop across the Zener will be aproximately equal to the voltage across an ideal zener with the same Vz.

#### GS3

Joined Sep 21, 2007
408
Zeners have a wide tolerance so you either specify a zener diode with especially low tolerance or you provide a means of adjustment.

You also need to limit the I when the battery is very low and would take more than is good for it.

And, as I said, absolute voltage is not recommended as termination for NiMh batteries. Delta V is better and even that has its limitations. Delta V and Delta T are better and, in any case, a timer should provide termination if everything else fails. Overcharging is the main cause of damage to batteries. I recommend you check out some pages about charging NiMH batteries such as http://www.powerstream.com/NiMH.htm

#### Ron H

Joined Apr 14, 2005
7,014
No, the zener is OK, as long as you limit the current biasing it. Most comparators only need input currents in the order of 100nA. If you consider a current 20 times greater than that, say 2uA, the voltage drop across Rz (the hypotetic series resistor) will have no significance, sinçe Rz value is in the magnitude of of 10Ohms. Thus, the voltage drop across the Zener will be aproximately equal to the voltage across an ideal zener with the same Vz.
Rz is not independent of current. It is actually just the slope (di/dv) of the zener's VI curve at any given current. At 2uA, Rz might be many kohms.
I'm not sure why our OP mentioned comparators, since there are none in his schematic.

#### bloguetronica

Joined Apr 27, 2007
1,453
Rz is not independent of current. It is actually just the slope (di/dv) of the zener's VI curve at any given current. At 2uA, Rz might be many kohms.
I'm not sure why our OP mentioned comparators, since there are none in his schematic.
I think you are talking about the overall resistance of the Zener diode. I'm just talking about the hypotetical series resistance that the real zener has.

See this datasheet: http://www.fairchildsemi.com/ds/BZ/BZX79C2V4.pdf

See the value Zz stated at 5mA. It is the value for the hypotetical series resistor I ilustrated. What happens is that at lower currents the ideal zener diode in the equivalent association will keep the same voltage drop, being biased by less current, thus having a greater resistance. When talking about significant currents, it is Rz (the resistor in series) that has a greater effect.

If you see the specs for the 2V4 diode, for example, you will see that Zz is much smaller than the ideal Zener's impedance of 480Ohm (2.4V / 0.005mA). So the equivalent real zener diode would have at least 480Ohm, and it may have 580Ohm since Zz = 100Ohm.

#### Ron H

Joined Apr 14, 2005
7,014
I think you are talking about the overall resistance of the Zener diode. I'm just talking about the hypotetical series resistance that the real zener has.

See this datasheet: http://www.fairchildsemi.com/ds/BZ/BZX79C2V4.pdf

See the value Zz stated at 5mA. It is the value for the hypotetical series resistor I ilustrated. What happens is that at lower currents the ideal zener diode in the equivalent association will keep the same voltage drop, being biased by less current, thus having a greater resistance. When talking about significant currents, it is Rz (the resistor in series) that has a greater effect.

If you see the specs for the 2V4 diode, for example, you will see that Zz is much smaller than the ideal Zener's impedance of 480Ohm (2.4V / 0.005mA). So the equivalent real zener diode would have at least 480Ohm, and it may have 580Ohm since Zz = 100Ohm.
I thought you were proposing that the zener in your comparator example could be biased at 2uA, and would still have the value of Zz specified in the datasheet. What was your point about 2uA?
And I have never heard of anyone refer to an ideal zener's impedance as being V/I (480 ohms in your example). It makes no sense. That's like saying that you are drawing 1 amp off your 12V battery, so the impedance is 12 ohms. Oh, wait. The current changed to 2 amps. Now the impedance is 6 ohms. What the hey???

#### bloguetronica

Joined Apr 27, 2007
1,453
I thought you were proposing that the zener in your comparator example could be biased at 2uA, and would still have the value of Zz specified in the datasheet. What was your point about 2uA?
And I have never heard of anyone refer to an ideal zener's impedance as being V/I (480 ohms in your example). It makes no sense. That's like saying that you are drawing 1 amp off your 12V battery, so the impedance is 12 ohms. Oh, wait. The current changed to 2 amps. Now the impedance is 6 ohms. What the hey???
It was precisely because Rz that I proposed the diode to be biased at about 2uA. Thus you minimize the voltage drop effect across Rz (the hypotetical resistor), and the voltage across the Zener will be aproximately the rated Zener voltage.

Just like I said:
...What happens is that at lower currents the ideal zener diode in the equivalent association will keep the same voltage drop, being biased by less current, thus having a greater resistance. When talking about significant currents, it is Rz (the resistor in series) that has a greater effect....
Thus, if you had it biased at 5mA, you would jave a voltage drop across the zener bigger than expected, due to the effect of Rz.

Remember that a real Zener can be represented by an ideal Zener and a series resistor. That is why they give Rz on the datasheet (as Zz if I recall).

#### skyboltone

Joined Sep 27, 2007
8

#### SgtWookie

Joined Jul 17, 2007
22,221
"Dz whatever you want"

Well, if you grounded your emitter you would certainly put a load on your plate supply I predict a bright flash coupled with a cloud of acrid smoke from the vicinity of your Rs and Zeners.

Do you know what your maximum desired plate current will be, along wiith your desired plate voltage?

Skyboltone,

How about starting a new thread with your topic; that way things won't get all jumbled in with the problems in progress.

Edit your post, copy all of your text out, then delete everything but a short line of dots - save it, and then start a new thread. Then I'll move mine over as well.

#### Ron H

Joined Apr 14, 2005
7,014
It was precisely because Rz that I proposed the diode to be biased at about 2uA. Thus you minimize the voltage drop effect across Rz (the hypotetical resistor), and the voltage across the Zener will be aproximately the rated Zener voltage.

Just like I said:

Thus, if you had it biased at 5mA, you would jave a voltage drop across the zener bigger than expected, due to the effect of Rz.

Remember that a real Zener can be represented by an ideal Zener and a series resistor. That is why they give Rz on the datasheet (as Zz if I recall).
I think you don't fully understand zeners. Below are plots of the I-V curve for a 2.7V zener (the lowest voltage spice model I could find on short notice). Notice that the slope (Zz) of the curve is a function of the zener current. Also notice how low the voltage is at 2uA.
For low voltage zeners, ohmic series resistance is generally insignificant.

#### Attachments

• 19.5 KB Views: 87
• 20.8 KB Views: 86

#### SgtWookie

Joined Jul 17, 2007
22,221
.............