Solving Zener Diode Problems: 1N5226BT/R 3.3V, Zz = 28 ohms

[roeb]

I am using 1N5226BT/R zener diodes (3.3 Vz, Zz = 28 ohms) and am having trouble getting a constant voltage source...

Silly me, I designed a circuit assuming zener diodes would act ideally... According to the datasheet the minimum voltage reverse breakdown voltage would be 3.1 V. Here's what I've done so far:

1: I have two zener diodes in series with a ~200 ohm resistor. I get about 2.5 V over each diode.

2: I have two zener diodes in series with a ~120 ohm resistor and I get ~3.11 V over each diode. Great! Turns out I actually need ~9 volt instead of ~6 volts so I decide to try again.

3: I keep the 120 ohm resistor and add a third zener diode in series with the others, now I get 2.4 volts over each diode, ugh!Does anyone know what I am doing incorrectly?

I am using a transformer (wall wart) to supply the initial voltage 12 volt 500 mA (seems to hover at 16-18 volts open circuit).

It seems that supplying more current to the diodes got them in the ideal range. However, I am concerned if I try that again, I will be putting too much current through the diodes (500 mW diodes and 1/4 watt resistors).
3.3 Volts * 18V/120 = .495 Watts --> getting close to the limit...

 

[Pumblechook]

Is the PSU producing well smoothed DC?

Are you sure you are reading the resistor values correctly?

 

[roeb]

Thanks for your reply,

I am measuring the 'exact' values of the resistors by using a multimeter (ideally they are 330 and 220 resistors in parallel).

The PSU is not really regulated at all, however I am using a 330 uF electrolytic cap and a 1 uF polyester cap in parallel with the diodes to smooth the output.

 

[Pumblechook]

I will have to find some zeners and play with them. I have bench PSUs with V and I readout.

 

[berkeman]

Zeners are not very good voltage sources, and certainly are not meant to be used in series to stack up to some voltage. Not to mention their tempco...

A far better solution if you want to make a 9V voltage regulator is to use either a fixed 9V linear regulator, or a variable positivie voltage regulator like the LM317. Check out the LM317 datasheet to see if that wouldn't work better for you.

 

[roeb]

Heh, yeah I am wishing I would have gone that route, but I've got this 100 piece bag of 3.3 zener diodes burning a hole in my pocket.

It seems if I reduce the series resistance to 87 ohms it bumps up to 2.75 volts per diode which is good enough for my purposes, but kind of defeats the purpose of using a voltage reference...

Since this is all being done on an already etched board, I can't modify it at this point but for future reference are there any other good voltage references that can be used excluding lm317/lm7805 and zener's? the voltage regulating chips are expensive and large.

 

[Redbelly98]

While Zeners are not very good for constant voltage, the 3-ish volts (and lower) are particularly bad. If you got an actual 9V Zener it should work better. See Figures 9 and 10 here:
http://www.datasheetcatalog.org/datasheet2/3/074g7qee7jr0tgehc1k7j342htcy.pdf
V varies much more with I for the low voltage types.

But as berkeman said a regulator (9V or variable) would be far better.

 

[berkeman]

Heh, yeah I am wishing I would have gone that route, but I've got this 100 piece bag of 3.3 zener diodes burning a hole in my pocket.

It seems if I reduce the series resistance to 87 ohms it bumps up to 2.75 volts per diode which is good enough for my purposes, but kind of defeats the purpose of using a voltage reference...

Since this is all being done on an already etched board, I can't modify it at this point but for future reference are there any other good voltage references that can be used excluding lm317/lm7805 and zener's? the voltage regulating chips are expensive and large.

A 5.1V zener has the closest to a zero tempco, so if you are trying to do some voltage reference based on zeners, you would use a 5.1V zener biased at a reasonably stable operating point (aka, power hog), and multiply that voltage up or down with suitable mirrors.

Next best choice would be to use the classic shunt regulators, which are a bit more trimmed than just Zener diodes by themselves:

http://www.linear.com/pc/viewCategory.jsp?navId=H0,C1,C1154,C1002

And best is to use a linear voltage regulator like the 7805 or 317, since they use trimmed voltage references (trimmed in production), and bandgap references that are designed for low tempco error.

Are you getting graded on this project, or is it just on your own? If I were grading a project for a linear regulator design, I think you see where I would be counting + and - points...

However, if you were constrained in the graded project to use discrete non-trimmed parts, I'd expect to see a 5.1V zener burning way too much power at the center of it, right?

 

[roeb]

Yep, I'm just making this for fun. It's actually part of a switch mode power supply that I designed (modified, really). I figured since I am plugging this into a wall wart I would be better off having a stable voltage so that once I calibrate it, I would never need to recalibrate it regardless of the wall wart I use.

Unfortunately, the zener diodes seem to have backfired on me and I would have been better off just leaving them off and playing around with the pot each time I need to readjust the voltage.

I really like switch mode power supplies, right now the only 'chip' that I am using on this board is a 555 timer and I think I may be able to redesign this so that I can use a hartley or colpitts oscillator so that I won't be using any IC's. Basically I am seeing if I can design a switch mode power supply that is cheaper than one I can buy... unfortunately I don't know if I trust this power supply that I just made to not melt down if I leave it plugged in for a long time (it boosts up to 150 volts)

Here's a picture of it working:
http://img377.imageshack.us/img377/3600/1020653jn2.th.jpg