From wklero–(at)–sn.net Sat Sep 28 21:15:56 CDT 1996
Article: 15221 of rec.audio.tubes
Path: geraldo.cc.utexas.edu!cs.utexas.edu!howland.erols.net!newsfeed.internetmci.com!ncar!csn!nntp-xfer-1.csn.net!news-2.csn.net!usenet
From: wklero–(at)–sn.net (Bill Kleronomos)
Newsgroups: rec.audio.tubes
Subject: Re: Triode Mode and Plate Imp
Date: Sun, 29 Sep 1996 00:03:42 GMT
Organization: Minimal, at best.
Lines: 137
Message-ID: <52keah$ou--(at)--ews-2.csn.net>
References:
NNTP-Posting-Host: 199.117.100.11
X-Newsreader: Forte Free Agent 1.0.82

kwarde–(at)–yrell.net (V. Keith Warden) wrote:

should I take up a new hobby;)?

: Keith – kwarde–(at)–yrell.net

————————————————————————–
Keith:

You don’t need a new hobby! There is a reason that engineers designing
tube-type electronics were/are? paid as well as their counterparts in other

fields of engineering expertise- good designs take hard work,
experimentation and refinement. Otherwise, we should all take up somethng
*relaxing* like golf .

Here’s the info you seek courtesy of RCA and Amperex. I’ve tried to compare
apples and apples by selecting comparable sets of operating conditions-
triode connected vs. pentode.

Amperex 6CA7/EL34

Triode connected, push-pull, ‘class AB’, values for two tubes:
Plate and screen voltage, 400
Idle plate and screen current, 130 ma, max signal, 142 ma.
Maximum signal output, 16.5 watts. THD, 3%
Plate-to-plate load, 5000 ohms.

Ultralinear connection: (screen taps located at 43% of plate
winding turns on OPT) . A 1 K resistor is required in series with each
screen grid to prevent its overload.
Plate and screen voltage, 500
Idle plate & screen current, 114 ma, max signal, 224 ma
Maximum signal output, 60 watts–(at)–THD of 2.5%
Plate-to-plate load, 7000 ohms.

Pentode connected: (using a common 750 ohm resistor
between the paralleled bypassed screens and the HV supply)
Plate and screen voltage, 500
Idle plate current, 60 ma, max signal, 204 ma.
Maximum signal output, 58 watts–(at)–THD of 6%
Plate-to-plate load, 5000 ohms.

Note in the case of the EL34, with more or less constant operating
conditions, the oprimum load Z appears, if anything, to go *down* as the
operating conditions are shifted from pure triode, to semi-triode to
pentode.
(although the only pure triode operating specs they published are for 400,
not 500 plate volts, the output Z for 500 volt pure triode operation would
be higher than that specified at 400 volts)

Since the screen current of a pentode such as the EL34 is not small
(compared to beam-power tubes such as the 6550 and 6L6GC) I would have
intuitively expected the load Z for triode operation to be significantly
lower than that of pentode operation due to the additional load of the
screens added to that of the plates. Instead, the opposite appears true.

RCA 6550

Push-pull triode connection

Plate and screen voltage: 450
Idle plate current, 150 ma, max signal, 220 ma
Maximum signal output, 28 watts–(at)–THD of 2.5%
Plate-to-plate load, 4000 ohms

Push-pull pentode connection

Fixed Bias:

Plate voltage, 400, screen voltage, 275
Idle plate current, 180 ma, max signal, 270 ma
Maximum signal output, 55 watts–(at)–THD of 3%
Plate-to-plate load, 3500 ohms

Cathode Bias:

Plate voltage, 400, screen voltage, 300
Idle plate current, 166 ma, max signal, 190 ma
Maximum signal output, 41 watts–(at)–THD of 4%
Plate-to-plate load, 4500 ohms

So, in the case of the 6550, RCA doesn’t show much significant
difference in optimum load Z attributable to triode or pentode operation;
the two pentode specs bracket the triode operating specs
as far as plate current goes, and the triode load Z is right in the middle
as well. So in the case of the 6550, the load Z doesn’t appear to be
altered very much solely due to to a triode of pentode connection.

Finally, let’s take a look at the 6L6GC. I can only find comprable triode
vs. pentode operation specs for single-ended class A operation, but the
same principles should apply if they were in a push-pull configuration.

6L6GC

Fixed bias, single-ended class A, triode connection.

Plate and screen voltage, 250 volts
Idle plate current, 40 ma, max signal, 44 ma
Maximum signal output, 1.4 watts–(at)–5% THD
Plate resistance, 1700 ohms
Load Z, 5000 ohms

Fixed bias, single-ended class A, pentode connection.

Plate and screen voltages, 250
Idle plate current, 52 ma, max signal, 57 ma
Maximum signal output, 6.5 watts–(at)–10% THD
Plate resistance, 22,500 ohms
Load Z, 2500 ohms

Check it out! Here, the optimum load resistance is inversely proportional
to the plate resistance, and the pentode connection requires a lower load Z
than a triode connection.

So- there’s the practical operating stuff. There is a means of calculating
the optimum load resistance for a tube given a particular set of operating
conditions, but there is still going to be some actual testing required to
nail it down exactly. But in any event, I hope the above examples show that
the correct load for a tube is *not* necessarily proportional to triode vs.
pentode plate resistance as you were guessing. The best way to get in the
ballpark when trying to match the tube to the load is always going to be
the published manufacturer’s data.

Have fun.

-BK
Boulder, CO

 

Buy the Book!

I cleaned up my tab for Sonny Boy's Help Me and made it into a short book. There's a Kindle version for 99 cents, and if you buy the paperback you get the Kindle free.

Playing "Help-Me" In the Style of Sonny Boy Williamson II: A step by step, note for note analysis of some of Sonny Boy's Signature Riffs