[Amps] Zener Bias Replacement

[Larry Benko]

Tom W8JI wrote:

>That's why I always pick 5v Zeners for voltage references 
>when possible.
>
>There's another issue with a string of diodes besides 
>temperature. They change voltage with forward current.
>
>I'm sure diodes are all different, but I measured a string 
>of ten 1N4005 diodes in series last night.
>
>Room temp:
>100mA 7.59v
>500mA 8.21 volts
>
>After running 500mA through them for 2 minutes:
>
>100mA 7.07 volts
>500 mA 7.64 volts
>
>It doesn't appear to be that wonderful of an idea at first 
>glance.
>
>Anyone else have any actual measurements? Has anyone 
>actually checked the bias stability of various diodes? 100mA 
>7.07v up to 500mA 8.21v doesn't seem so wonderful. That's 
>with NO external heating.
>
>73 Tom 
>
>  
>
Tom,

I measured some 1N4005 diodes this morning and got numbers similar to 
yours.  The current to the diodes was applied for only a few seconds so 
heating effects from the diode current were not fully counted.  The 
voltages are the average of 3 diodes measured.

60 deg F (15.6 deg C)

50mA  0.746V
100mA  0.776V
200mA  0.802V
400mA  0.835V

And at 125 deg F (51.7 deg C)

 50mA  0.717V
100mA  0.746V
200mA  0.777V
400mA  0.812V

Therefore a 65 deg F (36 deg C) increase in temp dropped the voltage per 
diode between 22 and 30mV.  This is considerably better than 2mV/deg.C 
but still pretty crummy.  As an aside, many years ago I measured the 
temp. coeff of many diodes in the 1N52XX family for a design I was doing 
and found that the most temperature stable diodes were either the 6.8V 
or 7.5V ones.  This is slightly higher than what Peter suggested but 
difference may have been affected by the actual zener current used for 
the measurements as the higher voltage zeners have a sharper knee.

No one has mentioned this but if the goal is to keep the tube current 
biased at a constant current then the tube itself might have some temp. 
coefficient which would need to be considered before making a perfect zener.

Larry, W0QE