Tube Rolling: Does it Make a Difference?

[ThoFi]

General question. Why not measure FR (SPL) of the system at listening position? Tube integrated amp + speaker?

 

[SIY]

Then again, maybe the plate voltage is really low so that it is not totally unsafe, and hence there is almost no linear region.

Low B+ is super common for so-called "tube buffers." Not because of safety but because it's cheap and no-one on the buyer end cares if the performance sucks as long as they can bask in the knowledge that they have tubes in their systems. There ARE some tubes which perform OK (i.e., mediocre but not horrible) at low voltages, generally targeted at the 1950s auto radio market, but they are not fashionable.

 

[JBNY]

Why did you measure the distortion of the different tubes? No one rolls tube to get lower distortion, the only purpose of tube rolling it to change the sound. Different tubes can produce a range of dramatic changes from frequency response to imaging to dynamics.

 

[solderdude]

General question. Why not measure FR (SPL) of the system at listening position? Tube integrated amp + speaker?

because the measured differences at the listening position are much bigger than the actual differences.

See it as trying to measure millimeters with a measure tape that only has cm scale on it.

 

[solderdude]

Why did you measure the distortion of the different tubes? No one rolls tube to get lower distortion, the only purpose of tube rolling it to change the sound. Different tubes can produce a range of dramatic changes from frequency response to imaging to dynamics.

Where is the evidence of this ?
I mean have the perceived differences between tubes been objectively verified ?
Also do not confuse rolling power tubes or not really suited tubes or tubes with substantially different gain and or transformers / loads (in non- or low-feedback designs) with a simple gain stage.

 

[Gorgonzola]

Why did you measure the distortion of the different tubes? No one rolls tube to get lower distortion, the only purpose of tube rolling it to change the sound. Different tubes can produce a range of dramatic changes from frequency response to imaging to dynamics.

Haha! @JBNY, you're hollering down a well. I totally agree with your statement of the purpose of tube rolling and the results in subjective listening. But around here your assertions will tend to dismissed and likely ridiculed.

However "Evidence of the Ear", no matter how consistently experienced by the individual, nor corroborated by the shared experience of other listeners, will be dismissed out of hand by much of the ASR constituency. Fair, I suppose, because ASR is about the science of audio.

 

[Snafu]

"Does it Make a Difference?"

didn't read full thread but yes it does. a lot.

i have mostly used 6SN7 variants like CV-181, and self noise or static noise varies a lot. for example some NOS tubes have horrible hiss and new CV-181's don't. also longevity isn't same with all tubes, some last years and some are gone in weeks.

sound wise, yes hiss effects sound quality quite often, tube sounds "round" and unnatural.

tube types i have used in my amps 12AU7/12AX7, E88CC/ECC88/6922, 6SN7/CV181, EL34/KT66/6CA7, EF86/EF806/E80F

cheers

 

[mhardy6647]

Low B+ is super common for so-called "tube buffers." Not because of safety but because it's cheap and no-one on the buyer end cares if the performance sucks as long as they can bask in the knowledge that they have tubes in their systems. There ARE some tubes which perform OK (i.e., mediocre but not horrible) at low voltages, generally targeted at the 1950s auto radio market, but they are not fashionable.

Plus, there's little doubt in my mind that plate-starved triodes will have a... heaven help me... sonic signature.

PS The above-mentioned, low voltage, purpose-designed "Space Charge" tubes are fashionable enough in some circles to have become rare, expensive, and/or hard to get (e.g., the 12N7 12K5 "power tetrode")! Weird but true. It's a funny old world whenever the Internet's involved.

https://frank.pocnet.net/sheets/137/1/12K5.pdf

EDIT: In fairness, these were meant to be used as drivers for AF output transistors in automobile radios and not as outputs per se.

Also, a quick comment about double-blind testing: just as in any social science research, experminental design is absolutely critical. E.g., how does one meaningly control for, e.g., the potential effect of removing and reinserting a tube in a socket, which might (?!?!) improve electrical connection between pin(s) and socket (due to 'cleaning by friction' of the connections)? Randomly including A to A "swaps" (sham surgery, as the procedure is sometimes referred to in my old line of work)?
I think (?!?!????) sometimes the difference in cable swaps may be attributable to such "metaphenomena"

(Note well my very deliberate weasel-wording above!)

 

[watchnerd]

Haha! @JBNY, you're hollering down a well. I totally agree with your statement of the purpose of tube rolling and the results in subjective listening. But around here your assertions will tend to dismissed and likely ridiculed.

However "Evidence of the Ear", no matter how consistently experienced by the individual, nor corroborated by the shared experience of other listeners, will be dismissed out of hand by much of the ASR constituency. Fair, I suppose, because ASR is about the science of audio.

I don't agree -- controlled listening tests are welcome at ASR.

I've posted cartridge rips on ASR with positive feedback on audible differences from forum members. The experience was single blind, i.e. the forum members didn't know which file was which cartridge.

I think forum members would be very interested in a tube ABX listening test, especially if they could try themselves.

If I get some time over Christmas vacation, I'll make some tube rips.

 

[watchnerd]

Plus, there's little doubt in my mind that plate-starved triodes will have a... heaven help me... sonic signature.

PS The above-mentioned, low voltage, purpose-designed "Space Charge" tubes are fashionable enough in some circles to have become rare, expensive, and/or hard to get (e.g., the 12N7)! Weird but true. It's a funny old world whenever the Internet's involved.

Also, a quick comment about double-blind testing: just as in any social science research, experminental design is absolutely critical. E.g., how does one meaningly control for, e.g., the potential effect of removing and reinserting a tube in a socket, which might (?!?!) improve electrical connection between pin(s) and socket (due to 'cleaning by friction' of the connections)? Randomly including A to A "swaps" (sham surgery, as the procedure is sometimes referred to in my old line of work)?
I think (?!?!????) sometimes the difference in cable swaps may be attributable to such "metaphenomena"

(Note well my very deliberate weasel-wording above!)

Tube swapping runs afoul of time issues and auditory memory. It takes many minutes to swap tubes, let them heat back up again, etc.

I think ripping the tubes to digital files and using a software ABX tester would solve this.

 

[PierreV]

Where is the evidence of this ?
I mean have the perceived differences between tubes been objectively verified ?
Also do not confuse rolling power tubes or not really suited tubes or tubes with substantially different gain and or transformers / loads (in non- or low-feedback designs) with a simple gain stage.

Yeah, that's the main problem. Assuming a convincing blind test, the reactions would be "must be defective", "tubes out of spec", "amp out of spec" and we'd go back in a circle to "one should use appropriately verified tubes on properly biased amplifiers and the differences will vanish".

 

[mhardy6647]

I think ripping the tubes to digital files and using a software ABX tester would solve this.

Oh, them'd be fightin' words in so many circles!

I mean, heck, once you raped and pillaged that pure analog signal that much -- you might as well just take the difference between the two datasets and see if it is zero at every point. Who needs ears!

 

[DanielT]

They should be silicone and not regular rubber if they're to be used on power tubes.
So... I mean... the idea is to damp internal vibration (root cause of "microphonics") of the tube's -- shall we say -- infrastructure, which certainly can be audible. Do they really 'add value'? Beats me. I don't use any and never particularly felt motivated to.

BUT, tubes have been made with some coveted properties, including this one with vibrations. Vaccum tubes have been used in a lot of different applications, so it's probably not so strange:

The 6AQ5 [1] (Mullard – Philips tube designation EL90) is a miniature 7-pin (B7G) audio power output beam tetrode vacuum tube with ratings virtually identical to the 6V6 at 250 V. It was commonly used as an output audio amplifier in tube TVs and radios. There are versions of this tube with extended ratings for industrial application which are designated as 6AQ5A [2] (with controlled heater warm-up characteristic), and 6AQ5W / 6005 [3] or 6005W (shock and vibration resistant).

6AQ5 - Wikipedia

en.wikipedia.org

Attach pictures, a small tube amp with 6AQ5 as power tube that I had, tested for a while in the summer cottage

 

[watchnerd]

Yeah, that's the main problem. Assuming a convincing blind test, the reactions would be "must be defective", "tubes out of spec", "amp out of spec" and we'd go back in a circle to "one should use appropriately verified tubes on properly biased amplifiers and the differences will vanish".

That doesn't invalidate the exercise, though.

It would show that, with a given topology and set of tubes, yes, differences were audible.

Granted, it still leaves it a matter of "it depends", but I think everyone already admits that.

 

[MattHooper]

You are calling yourself a objectivist and do blind tests.
Why don’t you do measurements?
Maybe you did and can share them.
cheers

I wasn't labeling myself but rather referring to how others seem to perceive me in those forums.

As for measurements you may seek, I'm sorry but I haven't the equipment nor the aptitude. Fortunately there are folks like Amirm who have both

 

[MattHooper]

Yeah, that's the main problem. Assuming a convincing blind test, the reactions would be "must be defective", "tubes out of spec", "amp out of spec" and we'd go back in a circle to "one should use appropriately verified tubes on properly biased amplifiers and the differences will vanish".

That does seem sort of waiting in the wings. The comments I've seen from most engineers and those with EE knowledge through the years is that, while of course listener bias can contaminate any particular test, nonetheless tube amps can sound different than solid state. It's mostly in this forum where I have encountered some casting doubt on the very phenomenon of tube amps sounding different: that it's almost inevitably listener bias and a myth. That's an interesting proposition. But it does seem perhaps hard to pin down because as you say if someone reliably detects a difference between tubes, or tube amps, or tube and solid state, the reply can always be as you list. And then the claim becomes "properly built and functioning tube amps will sound the same." Which would take on the feel of a "no true scotsman" situation as it relates to people's claims of hearing differences between tube amps.

I don't know how accurate that is and I'd be happy to see clarification.

 

[Doodski]

no true scotsman

I've never heard of this, "No true Scotsman" before but I have experienced the practice. It's funny that it is used this way.

 

[pma]

It's mostly in this forum where I have encountered some casting doubt on the very phenomenon of tube amps sounding different: that it's almost inevitably listener bias and a myth.

It is probably due to not enough experience of those who post such claims. Tube power amplifier, depending on topology and design, may severly affect frequency response especially with speakers with violent impedance characteristics. This is of course audible in a DBT test. They may, sometimes, also exceed audible threshold of distortion and may exhibit compression of dynamics especially on classical music with big differences between ppp and fff, if they are pushed to their "soft limitation". Tube preamplifiers are a different case, it is not difficult to make them transparent.

 

[MattHooper]

I've never heard of this, "No true Scotsman" before but I have experienced the practice. It's funny that it is used this way.

Person A: "No Scotsman puts sugar on his porridge."
Person B: "But my uncle Angus is a Scotsman and he puts sugar on his porridge."
Person A: "But no true Scotsman puts sugar on his porridge."

Re tube amps:

Person A: "No tube amp produces audible differences."
Person B: "I reliably detected audible differences between tube amps in blind testing."
Person A: "But no true tube amp produces audible differences."

(Where "true" entails those tube amps built to sound the same - in other words "properly built" tube amps sound indistinguishable).

 

[Spkrdctr]

You guys are too far down the rabbit hole. If in a blind test you can reliably tell a difference between the tubes then it is an audible difference. No one can then say well, if it was built better/different to different specs it would sound the same. You are testing the different tubes in a fairly regular use scenario. If there is a difference that is audible and can be confirmed, then you have a difference. Other people trying to change the test is not relevant to the test that was done. Only if an item is used outside of its normal usage would that be valid. Just having anything be audible is a very big deal except for speakers.