Tube mono-block amp with SIX 6L6 outut tubes

Well, yeah, it just scales. More in parallel, more (total) bias and load current, lower load impedance.

By the time you get to 8 or 10 or more, you should probably be asking yourself if a one of those fancy DSP pedals plus a class D amp is cheaper (not to mention more efficient, as we start running into the limitations of a 120V 15A circuit around this level..), or at least another tube type with better economy of scale (i.e. a transmitter tube of some sort), give or take whether we're talking production here (6L6s are still around, if shitty -- the cheap ones that is) or one-offs from NOS.

Heh, funny, come to think of it, that matched tubes are very common, and independent grid bias is reasonably common, but independent grid drive level is not at all common. Really all that matters is balance at the OPT, for magnetic reasons. The more tubes you wire in parallel, the less critical their matching is (assuming independent variables, and assuming any individual does not exceed its plate dissipation rating).

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Design Website:

Tubes may age at different rates, so the characteristics might be quite different after a long time. It might not make sense to replace a single (failed) tube and you may have to replace the whole set of tubes with new tubes preferably from the same manufacturing batch. This can be quite expensive :-)

In the 1960's, I made a guitar amplifier with six EL500's. The tubes are not characterized for linear use, and they were not matched in any way. Each tube had an own cathode resistor, and there was no evidence of any overload from unbalanced operation.

The tubes had a tendency of parasitic oscillation, so I added a ferrite bead on the control grids and an inductor-resistor parallel combination on each plate (for constructions, see nearest ham handbook).

--

-TV

Using individual cathode resistors to make the whole grid bias helps a lot to equalize the current in each tube.

In general one tries to avoid cathode resistor bias in power stages, since quite a lot of the output voltage swing is lost in the cathode resistor(s) and use fixed low power negative grid bias supply instead, but this may require separate bias adjustment for each tube.

I am not familiar with 6L6, but the similar EL34 is used in pairs for

30 W for HiFi, using 4 or 6 for "100 W" guitar amplifier with 450 V anode voltage. One bass qui tar amplifier is rated at 100 W using only two EL34 tubes, but runs at 800 V, but only Telefunken "Special quality" EL34s seemed to survive more than a few gigs.

Unless the tubes are arranged truly symmetrically, e.g. in a circle with vertical air flow, they won't all run at the same envelope temperature, which means they won't have the same cathode temperature either.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

All amps of that scale should have active protection circuitry/monitoring of currents, biases, and temperatures via microprocessor to avoid potentially catastrophic faults.

Not putting that into an amp with the kind of non-inconsiderable expense one will sink into building it on the grounds of "purism" or whatever is mad, the kind of irresponsible behavior that the less well-heeled of the world get called out on. Work hard. Save money. Don't burn up 2 grand of parts/risk your life cuz one thinks silicon is too new-fangled.

If you want to see an interesting audio amp, look up the Editors and Engineers Radio Handbook from the mid 60's. There's a design that which uses 6 or 8 6080 double triodes to build a direct coupled amplifier, no output transformer and dual power rails, positive and negative. Never built built it here, but a very original design. Could probably update that to use power mosfets...

the two big issues with OTL tube amps is the intrinsic impedance mismatch of a tube cathode vs a tube plate in a totem pole arrangement looking into the load, and related lack of any intrinsic power supply noise rejection in the output stage the way a push-pull transformer-coupled output stage has when both halves have similar output impedances working into the same reflected load.

Global negative feedback can't do anything about the second and there isn't usually enough open-loop gain available to do a good job of correcting for distortion caused by the first across the audio band. often leading to a kinda poor-performing amp.

These issues can be addressed but it requires more work than I'm guessing they put in in the mid 60s (can't immediately find the article in question online)

I found it in my copy, 16th edition, 1963.

The amplifier is not direct coupled, there are several stages with coupling capacitors. The final is series-connected with

3 series pairs of 6082 triodes in parallel and direct feed to a 16 ohm speaker. The power supplies are + and - 140 V.

--

-TV

The largest production tube MI amp I am aware of is the Mesa Boogie 400+. It uses ( I believe ) 12 power tubes, essentially 6 push-pull pairs ).

SFAIK, it exists only to compete with the Ampeg SVT.

I wouldn't want to lift it.

--
Les Cargill

sounds like a gimmick. Tube plate/cathode currents are just way too small to properly drive a speaker without an impedence matching transformer.

Had they thought of it "back then" they could've used toroidal transformers with PWM push-pull amplifiers modulated up to 70 or 80 percent as a class 'G' amplifier. By the time that was invented, it was all transistors, though. [and no benfit to using tubes].

I suspect that the push-pull configuration's advantages [and disadvantages] disappear when you use a transformerless configuration, and new ones take their place.

If you want high power, you should consider using KT88's and a matching transformer from Hammond. A single pair should get you at least 100W RMS. I saw a stereo amplifier built using these at a state fair once, in the 70's, basically a clone of the GE tube manual's reference design. It was on an oversized chassis, though. I would've preferred it as 2 "monoblock" amplifiers that could fit together on a shelf...

even the high heater current requirements of a KT88 would be smaller than 6 or 8 6L6's [and is physically smaller]. But I suppose it looks 'cooler' to have all of those bottles grouped together.

--
(aka 'Bombastic Bob' in case you wondered) 

'Feeling with my fingers, and thinking with my brain' - me 

'your story is so touching, but it sounds just like a lie' 
"Straighten up and fly right"

I think it works ok without "all that". fuses where needed of course, to avoid things catching on fire. But tubes are amazingly strong at absorbing conditions that transistors would melt under. Sure the plates turn pink [like when half of a push-pull transformer melts away] but if you don't push them to the edge of physics they generally take it ok and survive getting a new output transformer. Or if a single tube fails in multi-pair configuration, the others will take up most of the slack without too much bad behavior. You'll probably hear the bad quality sound at high volumes, but without extended operation "that way" the other tubes should survive.

the more you have paralleled, the less impact a single tube failure will have on the other tubes.

Even briefly overvolt or overcurrent on a typical transistor, and you'll be replacing it VERY soon. I made the mistake of designing a circuit that operated close to the maximum Vceo (these 60V transistors should be able to handle 45-50V right?), and the transistors never lasted long under load (replaced 3 times, and 3 blown fuses that were supposed to protect them). Replaced with transistors that had twice the Vceo and no problem.

--
(aka 'Bombastic Bob' in case you wondered) 

'Feeling with my fingers, and thinking with my brain' - me 

'your story is so touching, but it sounds just like a lie' 
"Straighten up and fly right"

No-one tell him about electrostatics, piezo arrays or high impedance moving coils.

NT

A lot of sweep tube datasheets contain the rating, "short term overload duration: 220W for 60s" or something like that. Transistors can only dream of such abuse, gone in milliseconds!

But the purpose of that rating must be understood. These were TV tubes, and the sweep tube in particular often took a beating as other tubes heated up and other signals stabilized: horizontal oscillator and sync, and the damper diode especially -- its high cathode voltage isolation takes a long time to warm up.

There was definitely no money in adding a protection circuit! Burn a tube, pop it out, take it down to the corner drugstore and buy a new one for a buck or a few.

Nowadays, with both tubes and transformers being rather pricey, let alone the repairman -- the balance changes, and especially with how little hardware is involved in adding a protection circuit (if one does not mind that it contains silicon), it's well worth it.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
Website: https://www.seventransistorlabs.com/

So what, nothing comes out?...

They were definitely below the maximum power point (at clipping), if that's more accurately what you meant.

6AS7/6080, 6S33S and the other regulator tubes have peak cathode current in the ballpark of an ampere, so a modest number of tubes offers a modest power output, say 20 or 40W, preferably into a higher load like 16 ohms. The efficiency is poor, with more heat dissipated in the heaters alone, than delivered to the output.

If you put dozens in parallel, the efficiency keeps going up as you get closer to matching, but now your whole system consumes multiple kilowatts...

There was also the Philips "SEPP" with a pair of EL86? driving a 100s-ohms voice coil at good efficiency, but those speakers are so rare that this is practically a unique case.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
Website: https://www.seventransistorlabs.com/

Yep, the OP is talking about monoblock amp with 6, 8, ten power tubes. The power supply iron and output transformer will be large, possibly custom, and not cheap.

$5-10 worth of microcontroller or ICs and relays to at the least monitor tube cathode currents and grid voltages and cut the HT if things start going t*ts-up compared to a melted half of an output transformer that might cost $3-500. Big BUMMER!

Search for 6082 "50 ohm" and you find quite a lot hits with single pair driving 50 ohms.

With tree pairs, the load could be 16 ohms.

Usually closer to 1000 ohms.

These speakers were sold for a few years, but later on, they went on sale. Adding an 8:1000 ohm transformer and you could cheaply build a good quality speaker box for an ordinary 8 ohm tube or transistor amplifier :-)