[Boatanchors] RE: selenium rectifier... replacement strategies

[Barry L. Ornitz]

Dave Roscoe, w1dwz@rcn.com, wrote:

> However, I take umbrage at him  accusing the ARRL , by 
> referring to the constant 700 mv forward drop of a 
> Silicon diode , as  propagating a "myth".
> 
>  It is , in fact, a FACT !
> 
> If you use the term "relatively"  before the word  
> "constant".
> 
> The slope of the E/I curve is ALMOST vertical throughout 
> its USEABLE range of currents, with a  forward drop of 
> 0.7 volt   ( + or - )  "Close enough for government 
> work" , as the saying goes.

Sorry Dave.  I am quite correct here.  And PLEASE edit your 
posts to avoid quoting entire messages.  My original post 
was quite long and quoting it again and again is not only 
poor manners, it is an extreme waste of the bandwidth of 
this mail reflector.  If everyone would turn off automatic 
quoting in the mail program, the owner of QTH.NET would cut 
down a huge amount of what is actually being sent.

But to get to your statement, look at any data book for a 
diode you are familiar with.  The voltage drop is dependent 
on current.  For example to following table is data for the 
1N4007GP diode from General Semiconductor.

    Diode            Diode
  Voltage, V       Current, A

    0.4             0.00002
    0.45            0.00006
    0.5             0.0002
    0.55            0.0006
    0.6             0.002
    0.65            0.006
    0.7             0.018
    0.75            0.04
    0.8             0.105
    0.85            0.25
    0.9             0.5
    0.95            1
    1               2
    1.05            4
    1.1             6
    1.15            8
    1.2            10
    1.25           12
    1.3            14.5
    1.35           17
    1.4            20
    1.45           22.5
    1.5            25

So what is the typical range of operating current?  This 
diode is rated at one amp average current, with peak 
currents up to 30 amps.  At one amp of forward current, the 
voltage across the diode is 0.95 volts, a full quarter of a 
volt higher than the 0.7 like you say.  In a capacitor 
input filter, the peak currents seen by the diode can 
easily be 10 times the average current, so operation here 
can produce peak voltages across the diode of 1.2 volts.  
Going in the other direction, if you use this diode in a 
low power circuit where it passes only one milliamp, the 
voltage across the diode is slightly less than 0.6 volts.

I would certainly call a usable range of currents for this 
diode to be from microamps to perhaps 10 amps peak.  The 
voltage across the diode for this range of currents varies 
over 0.8 volts.  I would hardly call this constant or even 
close enough for government work.

If you want to learn something about semiconductor 
behavior, plot this data on linear graph paper with voltage 
on the X-axis and current on the Y-axis.  At high currents, 
the line will be pretty linear.  This is the internal 
resistance of the diode you are seeing.  The line can be 
extrapolated back to zero current and it will intersect the 
X-axis at about 0.6 volts.  But the current is not really 
zero here.  It looks that way because the Y-axis scale is 
not fine enough.

Now take the same data and plot it on semilog graph paper.  
Here the voltage is plotted on the linear X-axis, but the 
current is plotted on the logarithmic Y-axis.  The same 
data looks a lot different doesn't it?  For currents below 
an amp or so, the line looks pretty linear.  This shows 
that the current is an exponential function of the voltage 
which is ideal diode behavior.  When the line begins 
turning toward the right is when internal resistance 
becomes significant.

Excel is a pretty lousy program for plotting scientific and 
engineering data. {Ask an engineer when was the last time 
he used a pie chart except to explain something really 
simple to his pointy-haired boss!}  But you can use Excel 
to plot this data.  If you are not comfortable with semilog 
scales, plot the log of the current versus the voltage.

Until you have plotted the data in this fashion, it is 
difficult to understand the explanation from words alone.
The graphs that I referred to yesterday plot the diode 
current versus voltage on both log and linear scales and 
they compare copper oxide,selenium and germanium rectifiers 
to a small signal silicon diode (1N4148) and a medium power 
diode (1N4007).

Those graphs total about 600K in Adobe Acrobat PDF format.
I will be glad to send copies to those that ask.

        73, Dr. Barry L. Ornitz     WA4VZQ     ornitz@tricon.net