From: Chuck Simmons
Organization: You jest.
X-Mailer: Mozilla 4.61 [en] (X11; U; Linux 2.0.33 i586)
Subject: Re: 1.6V zener diodes?
References: <email@example.com> <firstname.lastname@example.org> <email@example.com> <3DAB0AE5.754C8988@webaccess.net> <firstname.lastname@example.org>
Date: Mon, 14 Oct 2002 22:48:56 GMT
NNTP-Posting-Date: Mon, 14 Oct 2002 15:48:56 PDT
Roger Johansson wrote:
> Chuck Simmons wrote:
> >> >>A led is a better zener than a real zener, and is available in
> >> >>voltages close to 1.6V. The exact voltage depends on the current.
> >> >
> >> >So a LED is better than a zener (main function: almost the same
> >> >volatge over some range of current),
> >> Yes, the voltage is less dependent upon the current than in a zener.
> >> Or you could say that a led has a sharper knee.
> >> (But the knee is actually an illusion built upon the choice of scales
> >> in a diagram, the response is purely logaritmic.)
> >I don't think this is quite right. As I understand it, below about 5
> >volts, zener diodes operate according to the zener effect which is
> >quantum tunneling in the depletion region. The kind of statistics
> >involved in tunneling are not likely to give a characteristic similar to
> >diode forward characteristics. Above about 5 volts, the mechanism
> >changes to avalanche across the depletion region. Avalanche is very
> >abrupt and does not fit a logarithmic curve. In transistors, avalanche
> >can produce a negative voltage-current slope. Moreover, the forward
> >voltage drop of a silicon diode has a 2.2 millivolt/degree C temperature
> >coefficient. Reverse breakdown does not behave that way. Also noteworthy
> >is that I rather think the forward drop of an LED has an even higher
> >temperature coefficient.
> I did a series of measurements years ago, to find the best reference
> voltage at low voltages, trying a zener, and red, yellow and green
> leds, and two 4148 in series.
> The zener had a voltage of 2.1V at 1 mA, so I guess it was a 2.1V
> I used a logaritmic/lin diagram to plot the values in, to get straight
> lines. I took 3 or more measurements of each device, and drew a
> straight lines through the values for each component.
> The values allowed me to do that, they were in a straight line.
> I used currents between 5uA and 2 mA.
> The lines for the leds were more horizontal than the line for the
> zener, and I can see on that old diagram 3 or 4 plotting points for
> the zener, so its response looks logaritmic just like the lines for
> the leds.
> The leds came out as the best, the two diodes second, and the zener
> the worst. It had a more current dependable voltage than the others.
> So the leds were the most constant voltage references.
> I have no idea about temperature dependance effects, I made all
> measurements at room temperature.
> I have also heard that zeners below 5 Volt act differently physically
> than above 5 Volt, but in any case, this 2.1 zener gave a logaritmic
> Maybe they work differently but give a logaritmic response in both
> Roger J.
Just so you know, silicon zeners around 6 volts have nearly zero
temperature coefficient. Two silicon diodes in series are about -4.4
mV/degree C. An 8 volt zener might be 4 mV/degree C. In the zener
mechanism range the TC is about -3.3 mV/degree C. I haven't found it for
gallium arsenide diodes yet but I expect it to be very like a 20 volt
zener (maybe 15 mV/degree C). I'm still dubious because published zener
data says that diodes in the avalanche region are resistive and have a
very abrupt breakdown unlike forward conduction which is not abrupt in
... The times have been,
That, when the brains were out,
the man would die. ... Macbeth
Chuck Simmons email@example.com