winged C vs Svetlana el34s?

[Steve Conner]

I agree that tubes don't "filter" in the frequency domain. However, due to their non-linearities, they can generate new frequencies that weren't originally in the signal. A tube that clips sharply, generating a lot of high-order harmonics, will sound brighter than one with a more gradual clipping characteristic.

And these characteristics depend on fine details of manufacturing (like the uniformity of spacing of the control grid wires) that aren't too well specified or controlled.

[km6xz]

If a tube clips more sharply than another, it is not the same type of tube regardless of the silkscreen printing on the glass and box.
If it does not share conduction curves with the defined designation, it might be a fine tube but just not the same tube. The shared licensing of tube types in the past between manufacturers to provide second sources required that they be interchangeable. That does not happen now so they all go off on their own path.
I specifically mentioned a signal placed within a linear portion of the curve and when done so, differences in sound disappear, unless a completely different style of tube is substituted such as a remote cutoff.
Take apart a variety of tubes currently being sold as EL34 or 6L6 and you will see that they are not of the same spacing or coverage as the definition of the tube type that became standardized by cross licensing. No tube would have been picked for new designs unless additional sources were available so every time one with unique characteristics was developed it was offered to competitors to assure that a second source was available. That standardization does not exist now so anything with a glass envelope can be called anything the seller wishes. I do curves on many tubes and it is obvious that plate resistance and shapes of curves vary greatly. But how they sound or perform will be circuit specific so no matter what its is, there is some amp out there that just happens to work well with those combinations of traits. That is why referrals for tubes in general are worse than useless, they are downright misleading. If they said, "in my Peavey XXXX serial number on xxx mains, at xxx bias and the stock transformer, I like the sound I get", I would be more accepting of referrals. But the "brand xyz are great because they are brighter sounding" is not only too general of a claim to have any value but will almost assuredly not work as well in something else.

[MWJB]

So different tubes do sound different (when installed in the same circuit/conditions)...isn't that kind of where we came in? ;-)

[Steve Conner]

Well, yes I agree completely. I offered an explanation of why different tubes with the same type number might sound different. I never said that the differences in sound would be consistent from one circuit to another.

Case in point. I have a funky old Selmer amp that runs its preamp tube with 5k cathode resistors, so it biases up colder than the normal bias point for a 12AX7. Then it chains the two triode stages directly together without the usual tonestack or volume control in between. I modified the amp a lot, but I kept this feature because I thought it defined the fat, dirty, scuzzy character of the amp, and set it apart from being yet another Champ clone.

Unfortunately it means that choosing a good sounding tube for it is a crap shoot. I tried several known good Brimar, Mullard and Mazda tubes (pulled from scrapped lab equipment) that sound great in my other amps, and they sounded crap. I finally found a Brimar that I liked and stuck with it.

[MWJB]

Hi Steve, sure, I concur with everything you have said (Aaah, the old Corvette?).

It just strikes me, with respect to what others have written, that there seems to be a consensus that "similar" tubes should sound the same. I see no reason why even tubes of the same designation & brand should sound the same given acceptable deviation in manufacturing tolerances and plainly measurable differences in characteristics when installed. If 2 tubes were literally "the same" then I daresay differences might well be imperceptible (a meaningful test, all factors being equal, would seem to me to be time consuming & largely impractical), but if there are in any way different (irrespective of the description on the box & any sales hyperbole) then why wouldn't they sound different. If a certain brand, when viewed as a representative sample compared to other representative samples of other brands, displays a typical trend (e.g. all JJECC81 I have ever put in a circuit draw less current than any Phillips 12AT7 in the same circuit), then it is reasonable for the end user (making comparisons in one amp/circuit) to attribute those trends & characteristics to that brand (even though they may also dovetail to a degree with non representative samples from other brands).

Any tube outside of a circuit doesn't make any sound (unless it rolls off the table on to a tiled floor, of course) - in the boxes, undisturbed, they all sound the same. In the real world all bets are off. Change the operating conditions/circuit for a certain tube and you're just second guessing as to what it will sound like...tubes I have hated in one amp have frequently sounded great in others.

Ultimately, I think many of us are saying the same thing in essence, just a difference of opinion on what is the critical, determining factor. My feeling is, a single determining factor is pretty well moot, because the playing field is rarely level.

[Steve Conner]

What he said. My own pet theory is that tiny mechanical differences in the control grid winding account for the differences in tone. The more unevenly spaced the wires are, the more of a remote-cutoff characteristic the tube will have.

That's how remote-cutoff tubes are made, by intentionally varying the grid winding pitch. And an old worn-out Cold-War era grid making machine could well vary it by accident. With a high-mu tube like the 12AX7, the grid wires are so thin and finely spaced that it would only take a few thousandths of an inch of slop to make a difference.

I expect to be threadjacked by the pickup winding guys any time now.

[g1]

I see no reason why even tubes of the same designation & brand should sound the same given acceptable deviation in manufacturing tolerances and plainly measurable differences in characteristics when installed.

Not sure if I have misunderstood you here. If 2 tubes of the same brand and type clearly sound different, how can that possibly be an acceptable deviation in tolerance? Or are you speaking of subtle, barely perceptible differences in sound? Back when everything ran on tubes, if someone went to the drugstore to get a new output tube for their radio, would they expect a new one of the same brand and type to sound different than the original? I would think the manufacturers would have demanded a tolerance that would ensure this was not the case.

[MWJB]

Not sure if I have misunderstood you here. If 2 tubes of the same brand and type clearly sound different, how can that possibly be an acceptable deviation in tolerance? Or are you speaking of subtle, barely perceptible differences in sound? Back when everything ran on tubes, if someone went to the drugstore to get a new output tube for their radio, would they expect a new one of the same brand and type to sound different than the original? I would think the manufacturers would have demanded a tolerance that would ensure this was not the case.

No, G-One, I'm not talking about "barely perceptible difference in sound". Different tubes of the same brand & designation bias up differently in the same amp. This has always been the case (I guess if you were a manufacturer with enough clout you could specify certain requirements, like tube vendors today who sell pre-selected, drop-in, "retube" kits for certain amps, but I think it more likely that the manufacturers aimed for a "safe" tolerance). Wall voltage is also "nominal", so even if they did set by the milliamp at the factory & ordered tubes to that exact spec, as soon as the amp left the factory they had no control over the operating conditions in different regions and at different times of day (wall AC fluctuates over the day).

Any mass produced items are built to reasonable tolerances, not to exact tolerances (this would push up the price to unsustainable levels for everyday use). I'd expect the typical radio, or entry level amp, to be cathode biased, which is less critical in terms of outright reliability, but loose enough tolerance wise that you may well hear a difference. Why would the radio/amp manufacturers have any control over what stock a particular drug store, possibly in an overseas market, kept? They probably didn't give 2 hoots about what tubes people used after the warranty expired...especially with new/revised products coming off the line every few months.

Fixed bias amps in the late 50's & 60's were often biased to a safe margin (some manufacturers learned their lessons there in the early & mid 50's on that front)...this can be evidenced by the fact that the amps still run today, despite a hike in wall AC. There are some exceptions of course & people still get it slightly wrong today (e.g. I live in a 230VAC country apparently, but my wall socket can put out 244VAC - many amps I see are well above schematic voltage), as long as the amp & tubes survive warranty period there is no need to do anything about it.

And here, I am merely talking about the one aspect of bias vs. plate voltage (because this is what I have the most experience in identifying & measuring)...bear in mind tubes have multiple components and multiple ways in which tolerances can vary, within reasonable levels.

[Steve Conner]

If 2 tubes of the same brand and type clearly sound different, how can that possibly be an acceptable deviation in tolerance?

I invoke Murphy's Law of Datasheets: "Any unspecified parameter will behave such as to ruin your day." This happens because when parts are made to a specification, the manufacturer has to deliver on the specified parameters, but he can get creative with the unspecified ones. Typically he'll cut every unspecified corner to save money. This includes letting the unspecified parameters fluctuate wildly in production, if controlling them would cost money. (The most famous example of this is transistor beta.)

Vacuum tubes were designed for linear amplification, and the datasheet specifies them as such. So for a 12AX7 it might say something like, "50Vp-p output with 2% distortion". It doesn't specify the harmonic makeup of that 2% distortion and it doesn't specify what should happen when you drive the crap out of it for 200Vp-p output. Therefore by the above law, I expect the results to vary randomly quite a lot from tube to tube.

In a 1950s phonograph design, the designer wouldn't take a preamp tube even as far as 50Vp-p and 2% distortion. He would maybe only want 15 or 20Vp-p out of it. Under these conditions, all the tube brands would sound the same. He would probably settle for 2% or 5% distortion in his output stage so he could get decent volume from a cheap pentode.

But in guitar amps, we drive the tubes right off the edge of the map, and that brings out the differences between them in ways that the original application wouldn't have.

[bob p]

But in guitar amps, we drive the tubes right off the edge of the map, and that brings out the differences between them in ways that the original application wouldn't have.

In that respect, it makes sense for guitarists to engage in the HiFi Audio practice of "tube rolling." It still doesn't make sense for HiFi Audio, though.

[pdf64]

'Vacuum tubes were designed for linear amplification, and the datasheet specifies them as such'
At a bit of a tangent, but, on the principle that a spec without a tolerance ain't worth a hoot, what is the accepted tolerance that should be applied to the 'design center' values noted in manufacturer's tube info sheets?
I've heard mention of +100% to -50% but not seen a source cited; is it in RDHx somewhere?

EDIT. For instance, at what level do tube testers indicate a tube to be 'weak' for the various parameters? Is there a standard (set in RDH?) or does each tester manufacturer set their own acceptable performance level?
Pete

[Steve Conner]

In that respect, it makes sense for guitarists to engage in the HiFi Audio practice of "tube rolling." It still doesn't make sense for HiFi Audio, though.

Well, I also have a tube hi-fi amp (for research purposes of course! ) and I think swapping out tubes does make a difference.

In the input stages, some tubes hum more, and if they aren't matched, the stereo image will be off. See Pete's comment on design centre tolerances.

As for the power tubes, I started out with EL34s but they sounded a bit harsh at high volume. I ended up with GEC KT88s biased as hot as the transformers could stand.

With EL34s the distortion measured:
0.038% at 1kHz, 1W
0.22% at 1kHz, 10W
1.00% at 1kHz, 30W

It was mostly 2nd and 3rd harmonic, and was considerably worse when tested at 100Hz and 10kHz. Pretty much what I would expect from a tube amp. I never got round to measuring again with the KT88s.

When they eventually wear out, I plan to get the New Sensor "Tung-Sol" 6550s. They look nice, so they must sound good!

[Leo_Gnardo]

As for the power tubes, I started out with EL34s but they sounded a bit harsh at high volume. I ended up with GEC KT88s biased as hot as the transformers could stand. When they eventually wear out, I plan to get the New Sensor "Tung-Sol" 6550s. They look nice, so they must sound good!

Original made in the UK GEC KT88's I take it. Yes there's something mighty satifying to the ear with those. In my rebuilt Dyna MkII's they seem to extend bass response, almost sounds like there's a subwoofer added. And you might be a bit long in the tooth by the time they finally do wear out. I have a couple in my collection, so long used that no silvery getter flash is left on the glass. They still work fine & sound great.

What the New Sensor "Sovtek" versions sound like, so-called Tung Sol and "Gold Lion" posing as reissues, I have yet to find out. They have their fans, might be all right. There's only one way to find out, but I'm on a budget.

[Steve Conner]

Yes, the originals. I got lucky with the scrapping out of two high voltage power supplies, they yielded 7 tubes and I was able to make a reasonably matched quad.

They all have plenty of getter left. It would give me great pleasure if New Sensor had stopped making tubes by the time they finally wore out.

[daz]

Well now that you all have had your fun discussing everything BUT the thread subject, i can now chime in with the answer to my own question having received and played a set of svets. Maybe someone will be interested in the difference. I had a feeling these tubes would be similar because i had a set maybe 10-12 years ago and i recalled they had the same brown base and looked similar to the winged C. And they do sound a lot alike, yet at the same time theres a considerable difference. If that doesn't make sense, think of it this way....If you changed the tone settings on your amp considerably different than u normally do, it can change the sound pretty radically, yet the amp will still retain it's signature character. Thats how i hear the difference between the winged C and svet. If you took a winged C and made the lows sound a bit smaller and slightly duller, then made the hi mids and trebles a bit harder and lacking a bit in fullness and smoothness, thats what the svet sounds like to me. Yet it retains the same kind of basic tonal character as the C.

Not sure if i will ever use the svets, as i just prefer the C's by a good margin. I have sets of several brands of el34, all of which sound considerably different and while i prefer the winged C over all of them, i still get the itch to pop the others in from time to time for a change of pace. I usually end up putting the C's back in after a bit of fiddling with the other tubes. The svets will likely become another of those occasional 15 minute forays into a change of pace. But if the C's eventually do become scarce before i sink in to old manhood and stop playing forever, the svets may be my first choice above the rest being closest to the C's. Still wondering bout those russian Mullard RI's tho.....