EL84 Amp Design...Why Such High B+ Voltages?

[kg]

KG: Yes, I remember doing that. The screen current reached a minimum at Vg3=0, and increased both ways.

steve: that's interesting that vg3=0 was the minimum.

did you recall power output changing as well?

i wonder if the results could be different if the load Z were varied at the same time...

[Merlinb]

Alittle power tube history.
The Beam Power tubes (Tetrodes by the way) came out on top.
The tube is considered easier to drive, less prone to distortion & more efficient than either the true Tetrode or the Pentode.

Not true. Pentodes have higher gm and are easier to drive. Beam tetrodes can often achieve lower distortion into fixed loads than pentodes, but pentodes give lower distortion into variable loads. These are, of course, massive generalisations and not universally true for all bottles.

There is no winner, beam pentodes were available in America and so favoured there, while pentodes ruled Europe.

[kg]

how many BT voltage/signal tubes has anyone seen?

[Wilder Amplification]

I'm not sure who's bright idea it was to tie the suppressor grids to the raw bias supply but it was definitely not a good one.

I did some testing with that a year ago. Didn't do squat to the tone. However, one particular set of =C= EL34s decided to have its suppressor grid short to the heater. Well guess what? Also knocked out the bias supply, which sent the power valves into red plate city once that happened.

Yet there are guys out there recommending it with the defense tactic that "well Traynor was doing it", as if to state that just because a manufacturer is doing something that that automatically makes it right.

Bottom line...it's a dumb idea and doesn't do squat for the tone.

[Steve Conner]

Hehe, I'd like to see u post this at diyaudio.com

Well, I chose my words carefully!

A 300B is intrinsically a very low distortion tube with a reasonable power output, provided that the grid is driven from a zero impedance source with a huge voltage swing. But that is impossible, and attempts to approximate it add a lot of cost and complexity to the amp. (Unless you cheat and use transistors. )

So in the real world, say you have $1000 to spend on building a tube hi-fi. If you chose push-pull triode-strapped KT88s, for example, it would eat the equivalent $1000 300B amp for breakfast in all performance metrics. Firstly, because the KT88s are better amplifiers than the 300Bs, and secondly because they're easier to drive, so you can take the money from the 300B driver stage and spend it on some better transformers, or a Tice Clock or whatever.

I am aware however, as a long-term lurking member, that DIYAudio isn't a rational world. I intend to post my new Blameless project there when it's done.

KG: I did the test with DC lab supplies, at conditions that simulated the peak signal. Vg1=0, low plate voltage, high plate current. I guess I can do it again in a real amp under signal conditions. From a design and optimization point of view, the overdriven power output of EL34s is limited by screen dissipation, so I doubt the trend for (safely usable) power output will be any different to the trend for screen current.

[imaradiostar]

good thread, keep it rollin'.

jamie

[ThermionicScott]

The RM-10 Stereo amplifier has 4 EL84's running on 700 Vp into a 13K primary.

Seriously?

I bet the screens didn't run that high though. You can get away with running the anodes high, if you liek that sort of thing, but it's the screens that suffer.

I'm a little late to the thread, but for a brief period, I did a lot of reading about the RM-9 and RM-10 amps in order to build something similar for guitar (which i never did). At any rate, the screens were run at half the plate voltage, and so the load was about double what you'd expect for a 2xEL84 amp. Output was 25-35W/pair, and tube life was reportedly great!

- Scott

[Ronsonic]

I know I'm coming late to this thread, but it's a good one and there are a few points that really could have been covered, but weren't.

Most mfr tube specs ended up being conservative, in some cases they were extremely conservative. For many tube numbers the specs were defined very early in the evolution of tube design and even with suffixes added to the number they didn't try to keep up with the changes. 6V6 is a good example, it was introduced early in beam tube development and as production methods and materials improved the changes weren't reflected in the specs. The 2A3s I use for hifi are an extreme example they originated in the 1930s and the specs never got updated. The makers would have had to go to some trouble to build a tube that was limited to the older specs. Even with a max diss of 15W and 250V they run at 25W and 320V long and happy.

As for the ubiquitous tube tester that was more a function of heavily Muntzed design by the product manufacturers than the quality of tubes. Design is a lot easier and cheaper if you don't account for drifted tubes. Even where adjustment was possible, a fresh tube or two fetched and tested by a teenager is cheaper than having a tech come in and tweak a trimpot.

[archaos]

As ever, you can always find lucky exceptions to the rule. Still, more dead tubes means the tube factories won't go out of business!

You're unfortunately so right Merlin. I have 2 DIY AC30s whose EL84s don't run over 12W max., & I always use to play the amps in stereo. However, one amp eats the tubes every 5/6 years, the other every 5/6 months ! I believe I'll replace the 100R screen-grids with 1ks...



P.S. : I've just received "Designing Power Supplies for Tube Amps", & I'm really enjoying it : as good as "Designing Tube Preamps"

[Chuck H]

You're unfortunately so right Merlin. I have 2 DIY AC30s whose EL84s don't run over 12W max., & I always use to play the amps in stereo. However, one amp eats the tubes every 5/6 years, the other every 5/6 months ! I believe I'll replace the 100R screen-grids with 1ks...

If this is factual and both amps are wired the same now there is definitely something else going on other than screen grid resistors.

[diagrammatiks]

the power supply book is great. except the chapters on voltage and current stabilizers and regulators.

habberdash merlin. How do I do that with tooobs?

i got all these toobs.

[tedmich]

supposedly the aforementioned amp company, Traynor had the Guitarmate running el84s at 420v and could make 25w rms, for a while at least...

remember when amp companies would sell wattage? they still do in PA amps, witness Mackie's hilarious "ns brown noise burst" ad..

and many a guitar amp company has made a ton of money selling bedroom players stadium amps... although now some oddballs want...tone without pedals

[Steve Conner]

Important point. The more power you try to squeeze out of a tube, the closer to Class-B you have to bias it, to prevent it burning up at idle. And the worse it will sound, generally speaking, although some musicians like cold bias with lots of crossover distortion. It gives a harsh, "expanded" sound, as the gain is greater for large signals than small ones, which can improve a high-gain amp by tightening up the mush of too many overdriven 12AX7s. In fact a cold biased output stage does the same as Peavey's diode "coring" circuit.

But for a hi-fi amp like the previously mentioned one, it seems like a really bad place to start.

[archaos]

If this is factual and both amps are wired the same now there is definitely something else going on other than screen grid resistors.

What for instance ? Perhaps you have an idea ?

A buddy of mine whose AC30 is from 93 has known something similar -see pic- .
Everything was OK until last year, when the started eating up EL84s, & burning out screen-grids.
Therefore, my mate swapped the 100Rs with new ones + new tubes, nothing changed : resistors + tubes killed.
Then he decided to install 470Rs metal oxyde instead -see pic- , & to this day it's been O.K.

[Steve Conner]

Maybe one of the speakers died and the load impedance is doubled.