D Class amps - anyone got their minds around these yet?

[Steve Conner]

Hey! So turns out we were all working on hybrids all along.

I've said before that I don't think there is anything magical about tubes from a pure engineering point of view. It's just a case of "mojo". Musicians think that tube amps sound better, so they hear what they expect. When nobody makes good tubes any more, maybe they'll change their minds.

My feeling is that a small single-ended tube amp with maybe 2 tubes in it is more acceptable. The cost of tube replacement is low, and the user doesn't have to pay to get it biased, or learn how to bias it himself.

Then I bolted a solid-state output stage onto that to make it loud enough for gigging. The problem I ran into is that the single-ended tube stage clips very asymmetrically, and once the transistors start clipping too, the result is a DC offset on the output. (The positive and negative clipping levels are equal, but the duty cycle is not 50%.)

I've built hybrids before, but they were stinkers. This is my first one that sounds any good.

[km6xz]

Yes, the key to hybrids is making the SS current amp having enough headroom as to never clip. There is nothing pretty about the distortion that results from both types of clipping(the Class A tube and AB hard clipping).
If there was some way to get around the OT limitation I would love to build a large tube power amp, because the tubes are readily available and are really really tough in over 1kw range. I built 833a Class A hi-fi amps, which were manufactured by a high-end tweaky brand. This oil cooled unit about was also built for customers by the same company. Being wildly esoteric and believing in fairy dust, they knew it would sell best priced at $40,000 per pair. They sold a few dozen of the science fiction looking 833 amps for $30,000 a pair which to me was robbery. I did it just for fun and did not make anything from it. They did pay for parts.
They have sold thousands of Class AB amps with 6L6's, 50 watts for $5400 a pair. They came to me and asked me to design an amp. OK, but just for fun. A week later they showed up with chassis with transformers already mounted and hole for tubes. They laid out the amp and ordered chassis and transformers without every having a design! I worked around their growing list of design restraints and ended up with a working amp. It was not optimized due major parts being ordered and installed that would never have been anyone's choice. But it did look good. I gave them full spec as measured in my lab. It was an OK, nothing special amp. They sent a pair to a major tweaky audio magazine to review. And to another.....The reviews that were published were unbelievable, full of flowery praise that the amps did not deserve and were cited as the best amp in their class and a steal at "only" $5400 a pair. Somewhere I have a pair in storage but no matter what, in a blind test, would anyone pick them over just about any solid state $200 receiver out there. The whole sell any myth mentality seems to be alive and well with well heeled buyers.
I see a lot of myth involving tube M.I. gear also, a lot more than 10 years ago, mostly revolving around tubes. I have been at this long enough to know that most truisms are busted by double blind tests and a lot of highly touted gear, and parts would not pass and properly done comparison test. Most tubes sold today are of completely different type and spec than as sold as. Just look at the curves created by any 12 or so different "12AX7", there is no way they all have the basic specs to fit the 12AX7 nomenclature. What are they? Who knows but we do know that they are not much related to old 12AX7s whose published curves are being used in designing units. Whatever they are, some are fine, just different and good specs and curves would be helpful in getting what as expected from designs. It is like building a house using 2X4s as spec's in the architect's drawings. You buy lumber specified as "2x4" but when you built the house you discover nothing works as planned. Measuring them they turn out to be 2.345X3.9. That would be fine if the architect knew that and designed with those specs in mind.
Some of the tubes available seem to be good and were no doubt developed by the Russian or Chinese industrial or military and if the tube companies would just tell us what they are, and the manufacturers of equipment what they really are, all would be great. There would be a body of factual information and design experience with them to make'um sing. A company in Russia recently sent me 6L6's to try, the spec sheet in Cyrillic had their real designation. They were tough looking thick glass tubes, looked pretty rugged with lots of support and all welded construction. But looking up the real specs, they were not much like a 6L6, with a max anode potential of 250v! But they look good and using that information I think I will build something with them. Cost? $1.75 each I KNOW one or more of the re-branders is selling this tube to the M.I. market as a $18 6L6. It appears to be the most rugged 6V6 on earth....but the size and plate structure of popular higher power tubes.
I need a little bench test amp, these would probably make a great 25 watter that could not be blown up.

In the studio, actual power has little relevance to sound, with gigantic PA systems everywhere, a great sounding practice amp at 20 watts can generate all the tone in the world and fed into the reinforcement system. Stage volume drops and everything sounds better, and easier to haul....I see no downside lowering stage volume and improving sound. Vocals improve dramatically. Guitar strings ring truer not being modulated by high sound pressure levels, and the list goes on. I am all for small great sounding amps even though there is something about a large stack...well, those can be inflatables for image alone...
In fact the house PA system is too big and too powerful for enjoyment of the music now, and audiences are thinning out. Only the half deaf house mixer benefits from the mega-power and SPL wars. After acustomed to the level, a listener can't tell the difference between a 136 db average level and a 100db average level in systems that both allow the same 3-8 db headroom. But one system costs $1,000,000 and the second costs $85,000. No one every left the old Filmore saying "boy, I wish it was louder" yet they had less power than typical car stereos now. They has tube 807 Altec amps and high efficiency speakers that had a low freq cut off of 60 hz. 20 Hz costs a million more.

[Steve Conner]

I think you're preaching to the choir here.

In high-end audio, the review you get in the magazine correlates with the size of the ad you took out for it in the same issue. And to the end user, the perceived sound quality correlates with the price. In double blind tests, nobody can tell any of that sh!t apart. I personally failed to tell my own super MOSFET hi-fi amp from a $150 Cambridge Audio unit in a blind test.

Those tubes you were talking about, the 6P6S or 6P3SE by any chance? The 6P6S is halfway between a 6V6 and a 6L6GC, really more like the original 6L6, and I've seen it rebranded as both. The 6P3SE is commonly rebranded as a 6L6GC or 5881. The datasheet says 250V, but in practice they can take a hell of a lot more.

I hear you on stage volume. I spent a couple of years gigging with what was probably the quietest band in town. That's quiet as in, I had a 60 watt bass rig and I couldn't turn it up. It was a folky band with a singer who was incapable of singing loud, and an accordion, so we just had to build the mix around that.

It was a learning experience, but I eventually quit because I couldn't stand the squeals of feedback any more, caused by sound guys who were used to mixing rock bands. On the other side, I've been to metal gigs where the sound guy was wearing ear defenders, which pretty much sums up the state of live sound for me.

I love the idea of the inflatable Marshall stack. If I ever get round to offering kits of my new amp, I should include a free one.

[km6xz]

Yes, the ones I got are marked "6п3с" which is "6P3S" in Latin characters. It is larger than a 6V6 but shorter than the 6L6. The 6п3се was also available but "really" expensive at $5 each. If I buy a hundred that comes down a lot. The base is wider, like the wide base Sovteks and Ruby's that do not fit most retainer clips.
I tested some similar tubes for Ruby when they were located in my town in California(they moved to 20 miles away) but they were marked 6L6xx if I remember right back in 1999 or so. I tested some of their new tubes because I had a full set up of gear to test them. My tube test setup is limited to 1kv but I wouldn't be surprised if these in hand can't handle 475v without stressing anything. I remember there was a 6V6 amp, but I can't remember its name, that ran 6v6s at 500++volts and none of the current tubes of the time could handle it. I should remember it....think of it in a minute...These tubes handed basic tests but I would not have turned over to a player for gigging. In March I am going back to the States and bring a few pieces of test gear that I miss, a HP3580a spectrum analyzer and either a Fairchild 6200B or Tek curve tracer. For home brewing gear I have an old tube regulated power supply that was built for providing meters outputs for anode supply current and voltage, variac controlled and metered screen supply, metered bias supply 0-100, 5 volt heater, ct, 6.3 heater Ct and 24vdc non-metered. It was made by Lambda in the early 50s. Testing a new tube circuit idea without dealing with kludging a supply was so handy. The plate supply was regulated to 0-600volts and 550ma. I never saw another one like it. I would love to get it over here but it weighs 100lbs. It also has 110 volt mains. I could ship by sea but it means building a solid crate for it.

[gbono]

Great discussion - some of my favorite criticisms about hi-end, MI and pro sound all validated in a single topic.

I'm looking at an FPGA/CPLD solution for a tube "emulator" in audio design. There is unlimited possibility - seems the front end development (more access to IP) and chip cost are getting lower.

[ThermionicScott]

Hey Steve, a couple thoughts:

1) many/most of those old single-ended amps don't seem to be carefully set up -- just pick a reasonable load, bias the tube at max or so, and ship the thing. Have you tried tweaking your Selmer's bias for more symmetrical clipping?

2) that 250V rating on Russian 6L6's baffles me too. At first, I thought people were misreading the "typical operation" data, but I rechecked, and it does give that as the max rating for 5000 hour service. Maybe they just had to be conservative with their tubes. They can take a hell of a lot more than 20.5 watts, too!

Sorry for the tangent. I dunno how they expect to sell "Class D" amps when "Class A" amps are obviously superior.

- Scott

[Steve Conner]

Hi all,

Great, we all agree that everyone is senseless and the whole industry is broken, that means it must be true!

Chuck H: Yes, I like them bottom ends big and grungy, as Lil Wayne or some rapper said. My amps could probably do with smaller coupling caps. Maybe I should rewire the tone control on the Selmer to act as a bass cut, instead of a treble slug. [Edit: oops, this reply belonged to another thread]

Gbono: I'm writing this from my development machine at work, which is hooked up to an instrument we designed, running a Spartan-3 FPGA, a 300MHz floating-point DSP, and a 500MHz ARM, all connected together in mind-boggling ways. DSP for business, tubes for pleasure.

Scott: Good question. The bias is a little cold, and the VVR upsets it too. But I think even if the clipping started off symmetrical, it soon gets asymmetrical as you overdrive it more, by the nature of a single-ended stage. I believe that has a positive effect on the sound, so I don't want to engineer it out.

Re the Russian tubes: I think those are 250 Soviet military grade volts. I've run the 6P6Ss at 475V and got 50W out of them. I've seen the 6P3SE run at 600V.
http://scopeboy.com/AB013.JPG

[ThermionicScott]

Re the Russian tubes: I think those are 250 Soviet military grade volts. I've run the 6P6Ss at 475V and got 50W out of them. I've seen the 6P3SE run at 600V.
http://scopeboy.com/AB013.JPG

Ooh, pretty! I've got a pair of late-90's issue 6P3S that I've run in pretty much everything I own. Now they're running at 450-460V in my 5E7 chassis that I've tweaked for bass use (reworked the output section like a 5F6A with a 5AR4 rectifier). One of them got a little gassy, so I zapped it in the microwave and now it matches the other one again. I don't think these buggers will ever die!

- Scott

[km6xz]

I finally remembered the amps that used 6V6's really hot, Jim Kelly. They had 525volts on the plate, and the Ruby( I did not know they were actually a common Russian industrial tube with different characteristics at the time) 6V6 survived and sounded good. Really good. We used to see a lot of those 4x6V6 amps with the pretty wood cabinets.
There really is no such thing as a Class A SE amp with symmetrical clipping, nor would guitarists who spent more per watt for them want it any other way. The plot of the curves show that, plot the swing of a large signal on the load line. Both extreme ends have different type of non-linearity.

5000 hours is the standard power tube rating for CCS rating, that is verified full pull for 5,000 straight hours....and still meeting that spec. RF power tubes would be really poor quality if they their, say, 20,000 watt CCS rating not perform within spec for each of those hours. By guitar amp standards a CCS rating is 3-4 times as stringent. IAS, Intermittent Amateur Service is even higher than what guitar amp companies and home builders expect out of a tube. By the way, that 3CX300 amp I posted images of had 600watts of CCS rating. So a guitar amp would probably using those two small tubes smaller than a doorknob at 1200-1500 watts of tube sound bigness.

[ThermionicScott]

Well, 485V per the schematic, but the point is still valid. He recommended NOS Sylvanias to handle the stress. The shared 4K screen resistor for each side probably helped, but I'd have qualms about designing an amp that required a very specific out-of-production tube...

Yeah, you'll not get a nice "rounded squarewave" out of a single-ended amp -- what I meant was an output stage where the grid is center-biased, so that it reaches cutoff and saturation at about the same time.

- Scott

[km6xz]

Sorry about the mistaken B+, it has been a long time since I saw one, or any amp except home brew units since I moved here close to 8 years ago. Now suddenly I am having amps stuck in my face to fix and about ready to open a part time shop.
The fact that we all marveled at the high voltage and how no contemporary tube would work...not only because of high anode levels but they would generate parasitic oscillations if the tubes were not hand selected. It did sound good but with every note I envisioned the tubes melting so I did not listen too long.
With a easy supply of cheap tubes such as the 6P3S and the extended life version 6P3SE it will be fun to play with the extremes. I would like to built a new version of a guitar or base rig using the original version of the 3CX300, called a GM-5 here. The only difference is that GM-5 cathode is tied to one side of the heater. Svetlana was trying to get manufacturers to use in designs but just before the distribution changes and law suits by New Sensor, they dropped production of the exterior anode metal ceramic AF 300 watt AF tube. The GM-5 will take its place. There is a GM-5EB version that is extended life plus high-reliability military version that should be even more conservatively rated. Combine that compact tube with a high power switch mode power supply and it could give new meaning to high-power guitar amps. My only interest is doing it, not having to listen to anyone actually use one....too little power is not a complaint you will hear from me.