I thought I vaguely remembered that the bandgap voltage of germanium was about 0.6 V, you calculate with 0.74 V and https://en.m.wikipedia.org/wiki/Band_gap claims it is 0.67 V at 302 K.
There is some small correction term on top of the bandgap voltage. That's why silicon bandgap voltage references usually produce 1.23 V to 1.25 V instead of 1.14 V or 1.15 V.
This is all wrong. The temperature coefficient gets more negative the lower the base-emitter voltage is, like Mark wrote.
Did you take the huge temperature-dependence of Is into account? It looks like you assumed it to be constant over temperature and made a sign error; the temperature coefficient would be positive if Is would be independent of temperature.
Last edited:
The full equations are in the attachment of https://www.diyaudio.com/community/...ith-ic-voltage-regulators.359652/post-7021353 but not in the simplest form.
That is mostly an urban legend: during the manufacturing and encapsulation, Ge transistors are subjected to transient temperatures much higher than the regulatory 75°C: bonding the connecting wires, electrically welding the metal case, sealing the all-glass enveloppe, etc.
To achieve significant unwanted diffusion, such temperatures need to be maintained for very long periods of time.
Anyway, the kovar leads + the thermal shunt of the glass to metal sealing is sufficient to keep the Ge well within the safe zone, even with an overheated soldering iron used for an excessive length of time.
As a teenager, I wanted to test the validity of the legend on glass-encapsulated devices: i didn't manage to make them unworkable, except when the glass was melted.
I had no refined means of measuring Hfe, breakdown voltage or leakage currents, but they still worked in simple circuits
To achieve significant unwanted diffusion, such temperatures need to be maintained for very long periods of time.
Anyway, the kovar leads + the thermal shunt of the glass to metal sealing is sufficient to keep the Ge well within the safe zone, even with an overheated soldering iron used for an excessive length of time.
As a teenager, I wanted to test the validity of the legend on glass-encapsulated devices: i didn't manage to make them unworkable, except when the glass was melted.
I had no refined means of measuring Hfe, breakdown voltage or leakage currents, but they still worked in simple circuits
1/3rd of Ge's melting point isn't much above 100C, so dopants will begin to be mobile perhaps in the 100C--200C range (storage temperature is often quoted as 100C) - but yes pretty slowly. Silicon's melting point is much higher and diffusion even of phosphorus isn't an issue at any reasonable temperature. Creep and diffusion are generally pretty much locked out below ~1/3rd of the melting point for crystalline solids although notable exceptions exist (nickel superalloys).
Ge has very significant leakage currents even from low temperature which is why the operating temperature is very limited, as leakage-runaway is a thing (just like Si Schottky diodes)
Ge has very significant leakage currents even from low temperature which is why the operating temperature is very limited, as leakage-runaway is a thing (just like Si Schottky diodes)
Ge melts at ~900°C, thus ~1200°K, and the primary reason for specifying such low storage temperatures probably had to do with the poor manufacturing methods of the time: alloyed transistors were plunged in a silicone goo, which was probably not very healthy to begin with, but was also spiked with various oxides to improve the thermal conductivity.
That system was already not ideally stable under gentle ambient conditions (look at the number of failed NOS transistors), but elevated temperatures certainly exacerbated the effect of unwanted chemical reactions
That system was already not ideally stable under gentle ambient conditions (look at the number of failed NOS transistors), but elevated temperatures certainly exacerbated the effect of unwanted chemical reactions
