Thoughts on tubed preamps?

[Frank Dernie]

Well I think the fun part is often overlooked. At the same time it's not fair to imply that music listening and equipment appreciation are mutually exclusive.

For me there is a broad spectrum of interests in our hobby, from people who are mainly interested in the equipment and comparing equipment a lot to those who settle on something they like then use it to play music for the next 10 to 20 years.
I know people of both extremes and plenty in the middle.
Tube rolling places somebody very firmly at the equipment interest end of the spectrum IMO.

 

[watchnerd]

Tube rolling is like listening to different pressings of the same LP.

Sometimes it's like listening to different masterings, if the tubes are somewhat distant from each other in specs, condition, or provenance.

 

[egellings]

If the amp or preamp uses negative feedback, then tube rolling would make little or no difference, as the feedback would neutralize any differences. For equipment with no feedback, then tube rolling could well cause audible differences.

 

[mhardy6647]

If the amp or preamp uses negative feedback, then tube rolling would make little or no difference, as the feedback would neutralize any differences. For equipment with no feedback, then tube rolling could well cause audible differences.

This is actually at least somewhat (or, more to the point, sometimes) true* -- and is a little ironic given the jaw-dropping differences in vacuum tube circuits ascribed to tube rolling by many audiophiles.

_____________
* at least from amp designers' perspective -- the classic components were designed to be tolerant of aging or sample-to-sample variation of the active components (tubes) used in them.

 

[egellings]

Audiophiles, if they can barely hear a difference between two setups, will often rage about how yuge! the difference is, even if it's barely detectable.

 

[watchnerd]

This is actually at least somewhat (or, more to the point, sometimes) true* -- and is a little ironic given the jaw-dropping differences in vacuum tube circuits ascribed to tube rolling by many audiophiles.

_____________
* at least from amp designers' perspective -- the classic components were designed to be tolerant of aging or sample-to-sample variation of the active components (tubes) used in them.

I don't think I've ever had a tube rolling experience that was jaw dropping, unless it was in a bad way (noisy or microphonic tube).

Background noise floor, at max gain, with no music playing, is the most obvious difference to detect.

But with music on my little-to-no feedback amps, the difference can range from noticeable, but subtle, to barely detectable.

It's usually easier for me to tell apart very different families (e.g. NOS Russian vs NOS Dutch), and difficult to tell apart varying production years of the same tube from the same manufacturer (a 1960s Siemens vs a 1970s Siemens).

Highly dynamic or heavy bass content can sometimes also show differences, probably due to variable tolerances in gain.

But it's nothing even close to the changes caused by changing transducers (headphones, speakers, cartridges) or messing with EQ.

 

[MakeMineVinyl]

I don't think I've ever had a tube rolling experience that was jaw dropping, unless it was in a bad way (noisy or microphonic tube).

Background noise floor, at max gain, with no music playing, is the most obvious difference to detect.

But with music on my little to no feedback amps, the difference can range from noticeable, but subtle, to barely detectable.

It's usually easier for me to tell apart very different families (e.g. NOS Russian vs NOS Dutch), and difficult to tell apart varying production years of the same tube from the same manufacturer (a 1960s Siemens vs a 1970s Siemens).

It's nothing even close to the changes caused by changing transducers (headphones, speakers, cartridges) or messing with EQ.

Recently I ran distortion analyzer tests on a dozen JJ 12AX7s which I purchased for my tube preamp. The measured distortion was all over the place and I had a rough time picking the two lowest distortion tubes from the batch.

I have no doubt that, given the wide range of measured distortion, that audible differences between these tubes would exist.

 

[SIY]

Recently I ran distortion analyzer tests on a dozen JJ 12AX7s which I purchased for my tube preamp. The measured distortion was all over the place and I had a rough time picking the two lowest distortion tubes from the batch.

I have no doubt that, given the wide range of measured distortion, that audible differences between these tubes would exist.

Interesting- what test circuit did you use? I tested a bunch of their 12AT7/ECC81 with CCS plate load and LED-biased cathode (equivalent to bypassed resistor) and found them very consistent and better than any NOS 12AT7 on hand (which was quite a few types).

 

[watchnerd]

Recently I ran distortion analyzer tests on a dozen JJ 12AX7s which I purchased for my tube preamp. The measured distortion was all over the place and I had a rough time picking the two lowest distortion tubes from the batch.

I have no doubt that, given the wide range of measured distortion, that audible differences between these tubes would exist.

We know all tubes, even of a given type from a given manufacturer, don't measure the same.

So if every tube has the potential to have a lot of variability due to slop in tolerances, and then you add in variances in condition due to age, it doesn't seem like voodoo to have subtle audible effects.

Especially in a no- or low-feedback circuit.

 

[SIY]

We know all tubes, even of a given type from a given manufacturer, don't measure the same.

So if every tube has the potential to have a lot of variability due to slop in tolerances, and then you add in variances in condition due to age, it doesn't seem like voodoo to have subtle audible effects.

Especially in a no- or low-feedback circuit.

And yet... where's the data?

 

[MakeMineVinyl]

Interesting- what test circuit did you use? I tested a bunch of their 12AT7/ECC81 with CCS plate load and LED-biased cathode (equivalent to bypassed resistor) and found them very consistent and better than any NOS 12AT7 on hand (which was quite a few types).

I tested them in the actual preamp (a custom design, not a commercial product) and measured using the QuantAsylum analyzer. These were purchased from Amazon, and were JJ Electronics tubes. I returned all but the two I needed. The distortion ranged from about 0.2% up to 0.9% from what I remember.

This exercise wasn't intended to be a study of tube variability - I simply was trying to get a good matched pair for my preamp.

 

[watchnerd]

And yet... where's the data?

I have none. I don't have the equipment or knowledge.

Are you saying it's theoretically impossible?

 

[watchnerd]

I tested them in the actual preamp (a custom design, not a commercial product) and measured using the QuantAsylum analyzer. These were purchased from Amazon, and were JJ Electronics tubes. I returned all but the two I needed. The distortion ranged from about 0.2% up to 0.9% from what I remember.

This exercise wasn't intended to be a study of tube variability - I simply was trying to get a good matched pair for my preamp.

Amazon let you return them after testing?

 

[SIY]

I have none. I don't have the equipment or knowledge.

Are you saying it's theoretically impossible?

I'm saying highly unlikely. The only equipment you need is your ears and a willingness to implement controls- and to accept the results, like 'em or not. IME, the only times I ever reliably heard a difference with subs of the same type of tube was when at least one of them was defective (e.g., gassy, noisy, microphonic...).

 

[MakeMineVinyl]

Amazon let you return them after testing?

I'm sure they assumed I was just a typical Joe and returned them because they weren't what I wanted. Except for the two I wanted. I didn't attempt to do any 'listening tests' because I was going on measurement only.

 

[watchnerd]

I'm saying highly unlikely. The only equipment you need is your ears and a willingness to implement controls- and to accept the results, like 'em or not. IME, the only times I ever reliably heard a difference with subs of the same type of tube was when at least one of them was defective (e.g., gassy, noisy, microphonic...).

I would be glad to accept the results.

The controls is the hard part. Swapping tubes just takes too long and audio memory is short.

I've thought of running the output into an ADC and using software ABX testers to provide the control.

 

[pma]

When I measured 12AU7 x ECC82 x ECC802 in a SRPP based preamp, both distortion spectrum and gain were ALWAYS slightly different. Did not bother with listening test, just saying that @watchnerd is right about the differences between tubes of even same type.

 

[SIY]

I've thought of running the output into an ADC and using software ABX testers to provide the control.

Excellent idea.

 

[watchnerd]

Excellent idea.

I think I can just run the TRS output of a tube headphone amp into my RME ADI-2 Pro ADC.

I might need a Y-cable, though.

RCA pre-out would be easier, but a different animal.

 

[Gorgonzola]

For me there is a broad spectrum of interests in our hobby, from people who are mainly interested in the equipment and comparing equipment a lot to those who settle on something they like then use it to play music for the next 10 to 20 years.
I know people of both extremes and plenty in the middle.
Tube rolling places somebody very firmly at the equipment interest end of the spectrum IMO.

Humm ... well I listen to music 1-2 hours a day to the exclusion of other activities, (except once in a while I doze off ).

Of the couple of hours a week I spend concerned with equipment, most of that is spend here at ASR amongst the engineers and scientists .