Diode and very small amplitude high frequencies signals

[Roy Lewallen]

In other words, this data is just a plot of a diode's DC I vs V
characteristic, right?

What is of more interest is the slope at a given DC operating point. If
we pick 0V, for example, the above data (within the limits of its
precision) gives a flat line around that point (+5mV 21 Mohms, -5mV 21
Mohms). With a 100 uV signal, you might as well throw a 21 M ohm
resistor in there instead.

Exactly. At very small currents, the diode is just a resistor (shunted
by a capacitance). At slightly small currents, it's a very poor diode,
with reverse current almost equal to the forward current, so on each
negative half cycle you suck out nearly all the charge you delivered
during the positive half cycle.

Roy Lewallen, W7EL

 

[Roy Lewallen]

No, it means its a better diode at low currents. See my curves again,
http://www.picovolt.com/win/elec/comp/diode/diode-curves.html Note the
1n458 and the JFET diodes, which follow the theoretical 60mV/decade rule
down to very low currents. As for Roy Lewallen's "ratio of reverse to
forward current" argument, there is no reverse current for these fine
fellows, at least for DC and reasonably low frequencies.

Sure there is. All diodes have reverse current.

It's the very

crummy gold-doped 1n4148 that falls over. Awwkk!

The gold doping is done to dramatically reduce charge storage time.
Without it, the voltage across a diode continues to be in the forward
direction for some time after you reverse the current through it. While
a non-gold-doped diode might look good in DC tests, it makes a lousy
rectifier of RF. In the extreme case, it acts like a PIN diode (which is
simply a diode designed intentionally to have a long charge storage, or
reverse recovery, time).

Alas, life is full of tradeoffs.

Roy Lewallen, W7EL

 

[Roy Lewallen]

I need to clarify this. My comments apply only to junction diodes, which
virtually all silicon diodes are. Schottky diodes don't exhibit this
charge storage effect. That's one reason they're often used in high
frequency switching supplies. Their leakage current is, however, much
greater than silicon diodes.

Roy Lewallen, W7EL

 

[Steven Swift]

When I needed diode clamps that would work with +/-100mV with less
than 10pA of leakage, I always used selected Jfets with some back
bias. The DPAD10, etc. series of diodes that were designed for this
were too expensive. But that was when I was at Fluke and I could
specify exactly what I wanted and the vendors would come knocking.
Most of the parts were specially selected from National process 50,
which was used to make 2N4416 RF transistors.

The diode equation works over all ranges of the diode. We also used
2n2484 over 9 decades of current (pA to mA) of DC and tested them to
be very log-linear. At AC, as already stated, you have to include the
capacitance in the model. A finger print on the transistor body would
ruin the response.

Steve.

 

[Winfield Hill]

Steven Swift wrote...

When I needed diode clamps that would work with +/-100mV with less
than 10pA of leakage, I always used selected Jfets with some back
bias. The DPAD10, etc. series of diodes that were designed for this
were too expensive. But that was when I was at Fluke and I could
specify exactly what I wanted and the vendors would come knocking.
Most of the parts were specially selected from National process 50,
which was used to make 2N4416 RF transistors.

Yes, the PAD-1 and pn4117 JFET curves on my plot show your point,
http://www.picovolt.com/win/elec/comp/diode/diode-curves.html
with about 150 to 200mV forward voltage at 1pA. In that setup,
the current for under 100mV was well below what I could detect.

The diode equation works over all ranges of the diode. We also used
2n2484 over 9 decades of current (pA to mA) of DC and tested them to
be very log-linear. At AC, as already stated, you have to include the
capacitance in the model. A finger print on the transistor body would
ruin the response.

Not to mention the long reverse-recovery time of those parts. I put
up my graph to counter Dave's claim of a 600mV "barrier potential",
not to argue that these diodes make good low-level RF rectifiers. :>)

 

[lemonjuice]

Don't people still use 2N2369As, or at least the plastic version?
If not, what do they use instead?

(Does gold doping work for PNP transistors? I don't see why it wouldn't,
but I've never seen a specific reference to a gold-doped PNP.)

Without going into the details of Quantum chemistry and electronic
configurations (which is certainly more challenging and exciting) you
can easily see it from a simplified perspective of the energy bond
model of semiconductors.
For a npn transistor doped with Gold the equilibrium density (ed) for
impurity electrons in Silicon doped with gold is given by the well
known Fermi-Dirac statistics equation.
ed = B*Nd*exp(-Edonors/kT) E donors is the energy required for Golds
electron to freely dwell in the Silicon lattice ... a value much
smaller then the E gap energy value for Si to create an electron hole
pair.
Smaller values of Edonor imply larger values free donor electrons.

If a pnp transistor was to be formed from Si doped with Gold the
equilibrium density (ph) for impurity holes in Silicon doped with gold
is
ph = A*Na*exp(-Eacceptor/kT) Eacceptor is the energy required for a
electron to be stripped from the Silicon atom by the Gold atom.

As Eacceptor is greater then Edonor only a npn transistor can be formed
from Si doped with Gold.

Its even more exciting to look at the Physics of Gold plated
transistors. Very intriguing .. just like the MOSFET transistor!