Here is an example. Let us assume we have 100 volt primary and 100 volt
secondary, and that the normal rectifier has a 10 volt voltage drop. We
replace the tube with 2 series diodes, which have a 1 volt total drop and
then add 9 volts of zeners. That gives us a 10 volt drop, equal to the
tube. Now, assume there is a 10 volt drop in primary line voltage.
Secondary voltage is 90, less 10 volts, which is 80 volts. The output
voltage has dropped by 11%, while the input voltage has dropped 10%. If we
still had the original tube (or a diode and resistor), then the 10 volt
primary drop will result in 90 volts on the secondary - but the voltage drop
will be fairly linear across the tube (and resistor), so will only be
probably 9 volts. Resulting voltage out of the power supply is 81 volts
rather than the 80 volts using the zener diode. In that case, primary
voltage drops 10% and seconddary voltage drops 10%.

Contrary to regulated power supplies, the regulation performance of a non-regulated power supply is defined with regard to load variations, not to mains variations. As a matter of fact a non-regulated supply is inherently unable to tolerate mains variation, whilst it can tolerate load variations if properly designed (think of input-choke power supplies). Regulation (against load variations) greatly benefits from having a series zener in place of a series resistor.

It is however true that, for a mains voltage drop, using a zener diode will cause an extra small output voltage drop.

I use a series zener diode in my car to drop the battery voltage down to about 9V, as required by my VHF handy-talkie. If I would have used a resistor instead, voltage would significantly vary from reception to transmission.

Clearly there is no need at all to use an expensive power zener diode, as a plain power transistor with a small zener connected between collector and base does precisely the same job.

73

Tony, I0JX