Tubes: Modern vs Vintage

[dfuller]

So this is kind of an interesting topic.

Modern tubes are wildly inconsistent. The most consistent preamp tube I know of is the JJ ECC83S and it's still easily outside of +/-20% tolerances on mu, Gm, Rp, Ip. So there is actually something to rolling modern tubes, assuming there isn't much or any NFB, because the tube characteristics vary so much.

Vintage tubes, on the other hand? They're remarkably consistent. Good small signal tubes from the likes of RCA, Raytheon, GE, Mullard, and so on are often within 10% tolerance of electrical spec without any sort of post-factory testing, which is remarkably good for something with as much hand assembly involved as a vacuum tube. Rolling them in something intended for linear repro is utterly pointless, IMO.

Guitar amps are another story because they're running well into nonlinear regions, and the distortion characteristics do vary somewhat between different designs.

 

[mhardy6647]

Guitar amps are another story because they're running well into nonlinear regions, and the distortion characteristics do vary somewhat between different designs.

A very good point -- and I think that lots of tube (hifi audio) tyros conflate the mystique of tubes in musical instrument amplifiers (which create sounds) with the role of the same kinds of active circuit components (triodes, pentodes, even rectifiers) in a hifi amplifier which is (should be) designed and purpose built to reproduce sounds.

 

[watchnerd]

someone say tubes?!?
Sort of on topic. I just kinda smile & shake my head when folks tout tube rolling in, e.g., vacuum tube buffers (cathode followers). A properly designed and constructed buffer won't do much more than buffer impedance (mismatches) -- draw one's own conclusions.

Oh, but some designs don't have buffers.

My Fi Yph passive RIAA doesn't have an followers.

 

[BlueTunes]

So this is kind of an interesting topic.

Modern tubes are wildly inconsistent. The most consistent preamp tube I know of is the JJ ECC83S and it's still easily outside of +/-20% tolerances on mu, Gm, Rp, Ip. So there is actually something to rolling modern tubes, assuming there isn't much or any NFB, because the tube characteristics vary so much.

Vintage tubes, on the other hand? They're remarkably consistent. Good small signal tubes from the likes of RCA, Raytheon, GE, Mullard, and so on are often within 10% tolerance of electrical spec without any sort of post-factory testing, which is remarkably good for something with as much hand assembly involved as a vacuum tube. Rolling them in something intended for linear repro is utterly pointless, IMO.

Guitar amps are another story because they're running well into nonlinear regions, and the distortion characteristics do vary somewhat between different designs.

I watched this video from Mullard (gotta love the old BIG corporations )

It's quite incredible to see the work, effort and detail that goes into the manufacture of a tube. The economies of scale here are mind blowing and it's no surprise to me that new production struggles to be comparable. However, entropy is ever present and this video is already 5 years ahead of the transistor so could be argued to be "peak" tube era but is itself, now 70 years old.
That is a lot of years. I understand that sat in a box, on a shelf, nothing much should happen over those 70 years but chaos will always win. Dud valves, microphonics, seem to become more common. Getting good NOS is getting difficult.
I like your comments about consistency. Its not something I've heard before but again, my amp only has one power and one pre amp tube so consistency between tubes wasn't my concern. If you are talking about staying in spec (I think you are?) That's something else. I wonder how tolerant a valve amp is of valve spec performance itself. Its not something I know about. (Time for some Google fu)

The Mullard video has a couple of interesting highlights.

# Assembly is done by hand and without gloves. I thought this surprising as I have "rusty hands" myself and I couldn't imagine manhandling of these elements not leaving some contamination behind. Maybe it isn't an issue

# Material purity is paramount. The nickel anode tube is folded, not rolled to reduce contamination.

# Chemical treatments seem crucial (spraying of the elements, etc). I suspect above all of the difference between manufacturer and designs, this chemical treatment is the biggest and most important variance and may be key differentiator.

# research, improvement, development was a big thing. These valves seem to be constantly evolving. I would be surprised to see any old valves measuring better than later versions. This was my keep factor in deciding on new production valves. I figured the clamour for old NOS valves was technically misguided and was simply users going for a certain flavour rather than actual technical competence.

I'm still wary of tube rolling tbh. I'm struggling to see how the process is nothing more than ensuring (as you say) that your tube is working properly and to spec. And perceivable gains above that are subjective?

Hopefully my new valves will be "in spec". And if so, good enough for me. I'm trying more and more to listen to my music rather than the equipment itself so tube rolling isn't going to be my thing.

If You've some link to further reading regarding tube consistency, I'd appreciate it!

 

[BDWoody]

Oh, but some designs don't have buffers.

My Fi Yph passive RIAA doesn't have an followers.

That's what you get with that kind of budget gear...

 

[SIY]

Oh, but some designs don't have buffers.

Poor designs. Or I should say, units that are designed rather than engineered.

 

[mhardy6647]

Oh, but some designs don't have buffers.

My Fi Yph passive RIAA doesn't have an followers.

Yes, of course, sorry, I was a little too elliptical.
I was referring (ahem, I meant to be referring ) to the cheap one or two tube add-on 'buffers' that were all the rage in the early 21st Century. "Add tube warmth to sterile digital for $79.95!" That sort of thing. As best I can tell, those boxes were poorly implemented buffers -- no more, no less. But many of the folks who bought 'em seemingly got off on rolling (e.g.) $150 vintage European twin triodes into 'em to perfect (ahem) their performance.
Oh, and most of those products, to the best of my knowledge, ran the tubes at low plate voltages (i.e., "space charge" mode) -- so way, way away from their linear operating conditions.

That's the kind of stuff that makes me shake my head.

EDIT: this sort of thing
https://geardiary.com/2017/12/01/audiophile-budget-fx-audio-tube-01-best-tube-buffer-money/

This one -- unlike the (ahem) "better" ones, looks like it uses the hyper-cheap, generic Chinese small signal pentodes that have become all of the rage in hyper-cheap vacuum tube add-on products (many sold under that "FX-AUDIO" brand name).

Poor designs. Or I should say, units that are designed rather than engineered.

Buffers can certainly do a lot of good when matching source and load (especially if long cable runs are needed) -- but of course, so can mo(u)lding do a lot of good for poor carpentry (oh, and I speak with great authority on poor carpentry -- trust me on this!)
I think the add-on buffers are the audio equivalent of... I dunno... really large, garish, moulding.

 

[SIY]

Any phono stage or preamp needs to have some sort of buffer at the output. Often, if one is doing passive EQ, one should also use a buffer internally to maintain RIAA conformance with tube aging.

 

[dfuller]

That is a lot of years. I understand that sat in a box, on a shelf, nothing much should happen over those 70 years but chaos will always win. Dud valves, microphonics, seem to become more common. Getting good NOS is getting difficult.

This is more because NOS has been picked over for the last 30 years pretty heavily than anything else.

The Mullard video has a couple of interesting highlights.

# Assembly is done by hand and without gloves. I thought this surprising as I have "rusty hands" myself and I couldn't imagine manhandling of these elements not leaving some contamination behind. Maybe it isn't an issue

# Material purity is paramount. The nickel anode tube is folded, not rolled to reduce contamination.

# Chemical treatments seem crucial (spraying of the elements, etc). I suspect above all of the difference between manufacturer and designs, this chemical treatment is the biggest and most important variance and may be key differentiator.

# research, improvement, development was a big thing. These valves seem to be constantly evolving. I would be surprised to see any old valves measuring better than later versions. This was my keep factor in deciding on new production valves. I figured the clamour for old NOS valves was technically misguided and was simply users going for a certain flavour rather than actual technical competence.

Yup, lots of R&D went into these when they were in every bit of consumer (and more importantly military) electronics.

I'm still wary of tube rolling tbh. I'm struggling to see how the process is nothing more than ensuring (as you say) that your tube is working properly and to spec. And perceivable gains above that are subjective?

Pretty much, yeah. Circuit matters a lot more than any particular brand of tube.

Hopefully my new valves will be "in spec". And if so, good enough for me. I'm trying more and more to listen to my music rather than the equipment itself so tube rolling isn't going to be my thing.

Good man. That's what matters really.

If You've some link to further reading regarding tube consistency, I'd appreciate it!

from Myles Rose (ex-Groove Tubes SAG)'s blog: He bought lots of I believe 20 tubes from a well-regarded dealer who does test their tubes and these were the results he got. 12AX7s because he's fairly guitar amp focused, but this applies to lower gain small signal tubes too.

 

[BlueTunes]

This is more because NOS has been picked over for the last 30 years pretty heavily than anything else.
Yup, lots of R&D went into these when they were in every bit of consumer (and more importantly military) electronics.
Pretty much, yeah. Circuit matters a lot more than any particular brand of tube.
Good man. That's what matters really.
from Myles Rose (ex-Groove Tubes SAG)'s blog: He bought lots of I believe 20 tubes from a well-regarded dealer who does test their tubes and these were the results he got. 12AX7s because he's fairly guitar amp focused, but this applies to lower gain small signal tubes too.

Well thanks for the link. Interesting reading.

So, as I understand, all tubes vary from the design optimum, but tubes manufactured in the hayday would be much more consistent and have a tighter tolerance around those optimums. However, it seems that it is possible to get new production valves as good as or better than NOS, but one has to cherry pick them from the current supply.
Whilst New production tubes are inconsistent due to manufacture, NOS tubes are now becoming inconsistent due to age, condition and that as the "good stuff" is expended, the dregs (as good as they might be) are all that is left.

That's a particular battle huh.

I'd still er on the side of new production if only for the fact that production variance is likely less risky than variance from age, vintage, etc.

I suppose the upshot here is that in reality, if you are putting expensive valves into expensive equipment, you absolutely must get them as tested and from a good dealer. I would see no point in spending big bucks on eBay.

 

[Wes]

My various components use 6DJ8, 12AX7, and 6SN7 tubes. So far, I've yet to find a vintage tube that is objectively better than the modern production. Different, yes. But better...?

How did you evaluate this?

 

[watchnerd]

How did you evaluate this?

Looking at the tube tester measurements that came with the tubes.

 

[egellings]

Nuvistors work but tend to be microphonic. They're best left to RF front ends. As for tubes and semiconductors, the semis win in terms of capability, small size and reliability. Tubes are used when the amplifier is considered more of a pet than an amplifier.

 

[dfuller]

So, as I understand, all tubes vary from the design optimum, but tubes manufactured in the heyday would be much more consistent and have a tighter tolerance around those optimums. However, it seems that it is possible to get new production valves as good as or better than NOS, but one has to cherry pick them from the current supply.

Yup. Though, it's not so much "optimum" as it is "design spec". Like, a 12AX7 has well-defined specs. At 250V B+ and -2V Ug1, it should have a mu of 100, transconductance of 1600 uMhos (1.6mA plate current per 1 volt swing at the grid), a plate resistance of 62.5k ohms, and a static plate current draw of 1.2mA, and it should follow the plate curves present on the datasheet. But, like all electrical components, tubes have some tolerance around their actual rating because no manufacturing process is perfect. The issue now is that tubes are generally quite niche. Only audio uses them still, and there just isn't enough cash in it to bother with throwing out tubes that will work just fine in a guitar amp (let's be honest, that's where probably 80% of current production tubes go).

Whilst New production tubes are inconsistent due to manufacture, NOS tubes are now becoming inconsistent due to age, condition and that as the "good stuff" is expended, the dregs (as good as they might be) are all that is left.

It's really not due to age, honestly - receiving tubes, as long as the vacuum doesn't fail, stay good basically forever if unused. It's more that all the good tubes have already been snatched up. You see what a Telefunken ECC803S goes for? It's nuts. People pay 1200 dollars for a 12AX7 that isn't really substantially different from a modern JJ E83CC.

 

[Sal1950]

People pay 1200 dollars for a 12AX7 that isn't really substantially different from a modern JJ E83CC.

Holy Smokes, really? And to think of stuff I've tossed or gave away. About 30 years ago I was heavy into tubes and old gear. During that period about once a week I'd be driving down the ally and see an old discarded tube radio or TV. I'd stop and pull all the tubes and bring them home. When time allowed I run them thru my tube tester and keep the ones that measured the best. I ended up with a pretty large stock of good tubes with which to repair the old gear I'd pick up on occasion. I had half a dozen functioning hifi's in just about every room in the house plus the garage. For me it was a lot of fun tweaking old 50-60s stuff in my free time. Later life changes came along and I gave it all away to a friend before my retirement move to FL.
Wish I would have had the room down here to keep some.

 

[dfuller]

Dead serious.
https://www.ebay.com/itm/Telefunken...446467?hash=item26227e18c3:g:n0gAAOSwZapdNllw

 

[Sal1950]

Dead serious.
https://www.ebay.com/itm/Telefunken...446467?hash=item26227e18c3:g:n0gAAOSwZapdNllw

And it sold for the 1204.00 and another 40.65 for shipping.
NUTS!

 

[watchnerd]

Holy Smokes, really?

I've got a few pairs of Telefunken and Mullard NOS 6DJ8 tubes that are going for $1200+/pair now.

I've got even more Tungsram, Siemens, Brimar, and Phillips SQ, in original boxes, that are in the $300-500/pair range.

I didn't pay that for them, but they've gone up a lot over the years.

Over the 14-17 pairs of tubes I have, it's decent chunk of change and I barely ever use tubes.

 

[paulraphael]

Years ago I had bass preamp made by James Demeter, who was known as the creator of the first modern outboard tube mic preamp. I loved the bass pre, but something went wrong and it got noisy. I called the company, and James himself answered the phone. As soon as I said "noisy," he said "send it in—somethings wrong." I asked if maybe it could be a bad tube, and if it would be worth getting some new ones to swap in. And then the conversation got interesting. He said, basically, no way—and that in his opinion, in a properly designed tube circuit, tubes don't wear out. He considered caps to be wear parts, but tubes to be permanent. And even more interestingly, he said this went for power amps as well (and he made a couple of tube power amps). He considered a dead tube to be either a fluke or a symptom of bad circuit design.

I've never heard this opinion expressed before. It seems to fly in the face of everything else I've read. Does anyone have a science / engineering response to this?

 

[Sal1950]

Over the 14-17 pairs of tubes I have, it's decent chunk of change and I barely ever use tubes.

I'd seriously think of bailing out of them as I really think the market has peaked.
I could be wrong but I'm feeling waves of sanity starting to run thru audiophool land.
The objective community has finally started to make inroads to the craziness that had taken over the high end for the last couple decades.
Their will always be a small hobbyist clan that will retain a desire for old tube gear and the tubes but I think the "more money than sense" cult will soon be moving on to something else.
Good Luck Bro