DIODE

    Unless their non-critical diodes are not generally used to provide a voltage drop, although they are frequently used for 'diode-O' 2 supplies together, or as mentioned elsewhere, they provide protection for reverse gear. For lower forward voltages, it is more common to use Schottky, which has a forward drop of 0.3 V or so. You usually do not use a high-value R to drop voltage unless you are working in very high voltage / low current applications. for rails below 5 V, a 0.7V drop is generally intolerable; although it can be insignificant in systems that use 12V or high. Again, your specific application. Zener can be used in both derivations and serial topologies, depending on what you try to do, but they are only used in low precision applications. It is quite common today in computer electronics to have secondary 1.2V supplies that can be maintained at a 5mV tolerance or change dynamically in steps of 5mV, such as the SVID rails on a CPU.

Silicon Diode 

     The standard silicon diode should work, as you get closer to the parameters. I found the manual for that radio online: the output transistors were manufactured by CTC, it has not worked for a long time. They are normal silicon transistors, so the base-emitter junction voltage, when polarized forward, is the same as that of an ordinary silicon diode. The junction voltage versus temperature is 100% predictable if you know the materials. In this case, simply silicon. You will want to have polarization diodes that are also pure silicon (doped with P and N bonds). This means a rectifying diode. Do not use Schottky diodes, since their union is N-doped silicon plus metal and the junction voltage as a function of temperature does not follow the same path as a PN silicon junction. The circuit used for the bias is intended to present a very low impedance voltage source for the transistors, which tracks with the temperature. If you draw Q506, the 12V in your collector, and note that the two diodes in question go from the base Q506 to ground, and then observe the two 4.7 ohm resistors in parallel from the Q506 emitter to ground, you will notice this: - Q506, R514, R515, D501 and D502 form a current regulator that will feed a controlled current through the parallel combination of R516 and R520. - D501 and D502 only carry the current flowing through R514 and R515, which will not be more than 2 mA maximum. So, diodes do not necessarily have to be strong. - D501 and D502 do not have the good derivation from them. What this tells me is that D501 and D502 should be silicon diodes and not fast ones. It does not want a "signal diode" since it could rectify the abundant presence of RF and alter the voltage through them. A rectifier diode is slow in comparison and your idea of a 1N4001 is a very good start. I can not see how the D501 and D502 models are, but for this circuit to track the temperature, you want them to be in good thermal contact with the heatsink. Therefore, try to find diodes that, at least, are in the same package as the ones you are replacing, so that the thermal connection remains the same. R515 will be your polarization adjustment potentiometer. I'm sure the same manual tells you how to adjust the bias, and you'll have to do it after replacing the diodes. There is a fan of Atlas radios on the network, he publishes in many places.