[Amps] RE: 4CX250B Screen Supplies

[Ian White, G3SEK]

Will Matney wrote:
>Bill,
>The only problem is that the test tone signal is a constant peak to 
>peak voltage no matter what the frequency. On the human voice, the peak 
>to peak voltage would vary thus varying the current. Now if the test 
>tones voltage would be made to vary up and down, it may be made to be 
>close but most test equipment I've seen use a constant voltage output. 
>The peak to peak voltage changes with the volume in a voice modulated 
>signal too. Needless to say, when using the voice, a mere varying of 
>the mic to and from the mouth would change the voltage along with the 
>differences in the modulated signal of the voice itself. All this being 
>said, the signal voltage and current would swing widely as compared to 
>a two tone test. This can be verified by watching the peak to peak 
>voltage on a scope using both voice and the test tone. The regulator 
>must be able to deal with this as you know. Adding capacitance will 
>help the response also as you mentioned.

I'm jumping in here after a few days away, but all previous comments 
have been read and noted too.

Several people have pointed out that a two-tone test can give better IMD 
results than you'd observe in a practical voice test. This has been 
agreed many times over in the list archives.

The real point is that even under these favorable conditions, a 
zener-regulated supply (with a capacitor) *still* produces notably more 
IMD than an electronically regulated constant voltage supply. This is 
not surprising, because a well designed electronically regulated supply 
will have a much lower output impedance than a string of zeners.

The output impedance of a string of zeners is the Zzt value of the 
individual diode (from the data sheet) multiplied by the number of 
diodes in the string. For a typical 350V screen supply, many designers 
use something like the 1N5369B (51V) in a string of 7 - but each diode 
has a Zzt of 27 ohms, so that's 189 ohms for the string. You can get 
this down to about 50 ohms using a string of 20 diodes at 15V each... 
but that's about the lower limit for bare zeners. VR tubes fall 
somewhere in between those extremes: better than a bad zener string, but 
not as good as the best.

If you make the step to electronic regulation, the output impedance 
drops straight down to a few ohms or even less. This is probably much 
lower than most tubes really need, but there seems to be no middle 
ground between "very low" with electronically regulated supplies and 
"too high" with zeners or VR tubes.

However, voltage stability is not the only reason for using electronic 
regulation. The other two reasons - equally important - are that you can 
include fast-acting trip protection for both the screen *and* the power 
supply itself, without spoiling the voltage regulation.

It's obviously true that electronic regulation with full protection is 
more complex than a string of zeners. I'd love to find ways to do all 
those things more simply - but I won't compromise on either performance 
or protection.


-- 
73 from Ian G3SEK         'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek