[AMPS] Shunt regulated high voltage psu

[Ian White, G3SEK]

rich Wrote:

>-  Bargain-price/cheapie/fallout zeners diodes are basically units that
>fail to pass the manufacturer's final test.  I was given a box of them
>which was marked "semiconductor scrap".  Some were usuable.  Some were
>not.  None were first class.

My remark about "cheap zeners"  came from an earlier statement by Rich:

>-  A series of 5W. 20V - 30V zeners are a good choice for such a shunt
>regulator because they can be mounted on a perf-board, they are
>reasonably priced, and the screen V can be adjusted in 20 - 30 V steps.

It had nothing to do with the alleged quality of the semiconductors. Put 
it down to the "common language" barrier.

If the well-known 1N5300 series of 5W zeners are chained together as 
Rich describes, it looks at first sight like they can make a voltage 
regulator with a low dynamic resistance - which is the change of screen 
voltage, for a given change of screen current. The dynamic resistance of 
the diode chain is equal to the dynamic resistance of one diode, 
multiplied by the total number of diodes in the chain. If you use say 12 
* 30V diodes for a 360V regulator, the dynamic resistance of an 
individual diode has to be multiplied by 12.

The data sheet for the 1N5300 series shows that the dynamic resistance 
is a strong function of diode current. It increases sharply when the 
diode current goes down, which happens when the screen current is 
greatest. So the regulation is worst on speech peaks, right at the very 
moment you need it the most. That's my objection to a chain of "unaided" 
zeners for a screen supply.

Using a string of zeners to bias a shunt transistor is not necessarily a 
good way to improve voltage regulation. Typically the diode current is 
no more than required to supply base current to the transistor (a few 
milliamps) but at low currents the dynamic resistance of the zeners 
increases dramatically. The improvement in regulation from adding the 
transistor may be largely lost by having reduced the diode current.

Also, a voltage regulator such as a string of zeners ("amplified" or 
not) makes it very difficult to protect the amplifier against screen 
current going outside the normal range, due to overdrive, incorrect 
loading or a range of other causes. This may or may not damage the tube, 
but it will damage signal quality and it needs to be detected. But any 
kind of current sensor that you put between the zener chain and the 
screen grid will add to the dynamic resistance of the screen supply, and 
spoil the voltage regulation. The main reason why I'm so keen on active 
regulators is that you can put the current sensor *inside* the feedback 
loop, so it has no effect on voltage regulation.

I rate voltage regulation and current-sensing protection as equally 
important. The amplifier needs both. Having tried just about every kind 
of screen supply from VR tubes on upwards, an active regulator is the 
only way I've found that can deliver both.

>-  The (your?) reported IMD difference between using a precision shunt
>screen regulator and a zener-string regulatior was too vast to be true
>with quality semiconductors.
>
The person who made the IMD measurements is more than able to respond to 
that one.

-- 
73 from Ian G3SEK          Editor, 'The VHF/UHF DX Book'
                           'In Practice' columnist for RadCom (RSGB)
                            http://www.ifwtech.com/g3sek

--
FAQ on WWW:               http://www.contesting.com/FAQ/amps
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com