[AMPS] AB1 vs AB2

[Ian White, G3SEK]

KM1H wrote:
>I have been following this discussion with extreme interest and wish to
>thank all the contributors. 
>
>What still has me confused is the low impedence vs high impedence of the
>power supplies. 

Well, perhaps we should concentrate on the output impedance in ohms, and
then whether it's high or low enough for any specific application.

As a reminder, the impedance is the slope of the V/I curve, measured at
the operating point. Also known as dV/dI or dynamic resistance. To
measure it, you set the device up in its normal operating conditions,
then change the load current (dI) and measure the change in output
voltage (dV). The answer is dV divided by dI.

I tend to prefer the word "impedance" rather than "resistance", just to
remind myself that the answer may be frequency-dependent. Voltage
regulators for SSB amplifiers need to maintain a reasonably low
impedance all the way from DC up through at least 3kHz. 

It's relatively easy to make a regulator that works under DC (unvarying)
conditions. Also a nice big electrolytic across the output will stablize
against high-frequency variations, down to a few tens of Hz if the time
constant is long enough. Where most regulators fall over and go high-
impedance is at the repetition frequencies of words and syllables in
speech, ie frequencies of a few Hz.  

>Is a gas regulator tube high or low?  

A few hundred ohms per tube, maybe. (It would be great if someone could
measure one - I don't think I even own one any more!) For a string of
tubes in series, eg for a screen grid, the impedances of individual VR
tubes simply add together. 

There are also serious noise problems with VR tubes (known over here as
the DIY Aurora Effect), and there's a risk of accidentally building a
relaxation oscillator if you simply bypass the tubes with capacitance.

Zeners are a little better, but the dynamic resistances of higher-
voltage zeners are not as low as the more familiar 6-12V diodes. 
You'd be very lucky to get below 200 ohms for a 350V string of zeners.

>What about a series
>regulated tube supply such as is in handbooks for years....using a triode
>for the pass tube?
>
Better... but clumsy because of the heater transformer requirements.
Also, screen grids and control grids can both produce negative current
(due to secondary emission electrodes leaving the grid) and it's vital
to maintain regulation under those conditions too. A series regulator
won't do that unless you provide a shunt resistor to sink the negative
current away to ground. Also, shunt regulators are easier to implement
because most of the "works" can be down near ground potential, rather
than floating up near the output rail.

>In a bias supply how about a gas tube regulator feeding a bias adjust
>pot; the value to keep ~20ma flowing ?
>
It's one answer, but these days you can easily build a solid-state
feedback shunt regulator with an output impedance of an ohm or less.
This is obviously far better than the tube actually needs for good
linearity, but you certainly don't have to worry about regulation any
more! (In his recent articles in QEX, KN5S/K5AM wrote of problems in
stabilizing a feedback regulator, which made him retreat to using a
circuit with rather low loop gain. I hope this doesn't give feedback
regulators a bad name, because they're actually quite easy to stablize.) 

For a screen supply, one advantage of very tight regulation is that it
allows you to use older tubes whose screen emission is higher than
ordinary non-feedback regulators can handle. Also, some tetrodes produce
measurably better IMD performance with very tight screen regulation.
(Please note that I'm not claiming this for all tetrodes, but the
4CX250-350 series definitely do.) GW4FRX did some  interesting
experiments, start with an extremely well-regulated screen supply and
deliberately introducing resistance in series with the output. The IMD
with speech modulation gets worse as the resistance is increased.

Unfortunately this isn't the right medium to discuss circuits in great
detail. There's some information on 4CX250s and 350s in the VHF/UHF DX
Book, and I'm working on a follow-up article on screen supplies for a US
magazine.


>I finally dug out a full RCA spec sheet for the 8122 tetrode as used in
>the NCL-2000 and was amazed to see that they list  AB1 and AB2 operation
>with virtually no change in IMD. Since the Svetlana 4CX400A is a virtual
>clone (comparing E and I ratings, etc) this might warrant some further
>discussion.
>
Although the 4CX400A occupies the same "400W ceramic tetrode" niche as
the extinct 8122, let's not take the word "clone" too literally. 
It certainly isn't an 8122 copy or a beefed-up 8930 - it seems to have
some quite new features, including a 30-second warmup time.


73 from Ian G3SEK          Editor, 'The VHF/UHF DX Book'
                          'In Practice' columnist for RadCom (RSGB)

--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm