Zener diodes in series as a voltage regulator

[JC]

First, if you haven't already run across this, it may help:

http://www.allaboutcircuits.com/vol_3/chpt_3/11.html

Second, as a practical matter, as usual Paul coincides pretty much with every thing I've ever heard about using Zeners: The Wattage rating is very generous in the real world, so de-rating it to about 20% or less of the stated rating is prudent.

Third, rather than give up on the idea, I would suggest devising a stack of zeners that gets you closer to your target output Voltage.  You may not hit 283V exactly, but you should be able to get closer to your target.  A stack of 5 56V zeners would give you 280V +or- tolerance, for instance.  Then, size that first dropping resistor to consume most of the difference between your expected "raw" Voltage and your target Voltage. Not necessarily all, but certainly most.

After the zener stack, use a small cap to ground, like .1uF or smaller.  Remember, at this point you shouldn't be needing any filtering or additional reservoir, you are concerned with bypassing whatever noise you suspect the zeners might produce.  Feed this to your circuit without any additional resistance in series or filters in parallel.

The reason I recommend you persevere is that I have often found shunt regulators, where possible, to improve the sound of audio amps.  This certainly seems like a case where one might be used.

[Paul Joppa]

I've been avoiding comment on the oscillation problem, but you can use big caps with Zeners. It's the gas regulators that can oscillate.

[JC]

I guess my point is, Paul, I don't see any particular reason for using a big cap in that position.  Presumably, prior filtering and reservoir are taking care of their jobs, and the stack of Zeners should be opposing any ripple that may be remaining.  Why filter something that isn't there?   Is it your experience that a higher capacity is needed to filter out Zener noise?

[Paul Joppa]

Yes, Zeners are quite noisy - even the special "quiet" ones. The affordable ones are much worse!

When we built the test box to select a shunt regulator design, we included a Zener string, bypassed with a 10uF capacitor. That was to roll off the noise above a couple hundred Hz.

If you are following the shunt reg with current source plate loads, that will reduce the noise substantially and you might not need a bypass. But a cathode follower has not power supply noise rejection to speak of and you do have to be careful there.

[dbishopbliss]

First, if you haven't already run across this, it may help:

http://www.allaboutcircuits.com/vol_3/chpt_3/11.html

Second, as a practical matter, as usual Paul coincides pretty much with every thing I've ever heard about using Zeners: The Wattage rating is very generous in the real world, so de-rating it to about 20% or less of the stated rating is prudent.

Third, rather than give up on the idea, I would suggest devising a stack of zeners that gets you closer to your target output Voltage.  You may not hit 283V exactly, but you should be able to get closer to your target.  A stack of 5 56V zeners would give you 280V +or- tolerance, for instance.  Then, size that first dropping resistor to consume most of the difference between your expected "raw" Voltage and your target Voltage. Not necessarily all, but certainly most.

After the zener stack, use a small cap to ground, like .1uF or smaller.  Remember, at this point you shouldn't be needing any filtering or additional reservoir, you are concerned with bypassing whatever noise you suspect the zeners might produce.  Feed this to your circuit without any additional resistance in series or filters in parallel.

The reason I recommend you persevere is that I have often found shunt regulators, where possible, to improve the sound of audio amps.  This certainly seems like a case where one might be used.

Thanks for the link.  I've read it a couple of times but I think there may be a "chicken and the egg" issue that I'm not understanding. 

The link shows a battery source so the voltage is constant.  On the other hand, I am using a transformer, so the voltage varies depending upon the current (not to mention the power company, etc. but I'm going to assume my power is fairly stable for this question).


???V ---+--2.5K--+ 280V
        |        Z 
        |        Z 
        =        Z
        |        Z 
        |        Z 
 GND ---+--------+


So, if I take a swag based on values from PSUD using a current of 20mA and 40mA, then the rectified voltage will be between 340 and 380V.  I'm not saying that the current going across the 2.5K resistor is 20 or 40mA, I'm just using those values as a way to guess what my starting voltage will be in PSUD.

Using the high value of 380V, then the 2.5K resistor will be dropping 100V with a current of 40mA (100/2500=0.04).  Therefore the 2.5K resistor would need to be at least 4 watts (10 watts for safety).

Now, if my circuit is drawing 20mA, then there will be 20mA of current through the zener diodes, so they will be dissipating close to Paul's recommended 20% of their rating -  1.12W (56*0.02).


380V ---+--2.5K--+--280V--+
        |        Z        |
        |        Z        |
        =        Z 20mA   8 20mA
        |        Z        |
        |        Z        | 
 GND ---+--------+--------+


If the voltage is on the low side, then the 2.5K resistor will be dropping 60V with a current of 24mA.  Will the diodes continue to regulate as long as the circuit doesn't draw more than 24mA?  In other words, is there some minimum current that the diodes need to draw to work?  How do I know this value?

Or... am I completely off?  This is why I wish I could take a class on this stuff.

[JC]

As far as I understand it, you will get Voltage regulation as long as the Voltage at the top of the Zener stack exceeds the Zener breakdown Voltage.  So, the trick is to size the dropping resistor so that the Voltage at the top of the stack exceeds the Zener Voltage under the widest set of circumstances regarding input Voltage and current draw through the resistor that you can.  As with most shunt regulators, that means that the range of current you expect your load circuit to draw is limited, and, of course, you would like for your supply Voltage to stay within a narrow range as well.

There is, as you say, a minimum current through the Zener in order for it to be in it's regulation zone, but the only way I know of determining that is by looking at the characteristic curve for the Zener you propose to use.  Of course, you know the maximum current already.

Yes, there is sort of a chicken v. egg thing going on here!  You need to be able to predict the top and bottom limits of current you expect your load to draw at the regulated Voltage, and you need to predict how much you expect your supply Voltage may vary depending on the factors that may cause it to vary.

[Paul Joppa]

You can download data sheets on Zener diodes from several manufacturers, and there are different data sheets for different power ratings as well. Get a few of them - any one of them tends to be confusing, but if you have a few sometimes one will explain something that the other leaves out.

You will be startled at how far from perfect these devices are - but you will have a much deeper understanding, as well.