[zenpunk]

Many people continue to argue about whether cables/ op-amps/ DACs really  do sound different but discussions about tubes just seem to be ignored or brushed aside as tubes sounding different seems to be accepted as a fact by everybody. We all heard it, didn't we? but I failed to come across any measurements. 
I am also curious about why NOS tubes are supposed to be much better that new issue as today's technology would suggest new one should be better designed and built.
Is there any research explaining the differences? plate construction? materials used. Can you design a tube to have a particular sound?

 

[Tacoboy]

Let me guess (my two cents)
Before the transistor came about, everything was tubes, lots and lots of tubes. So it was worth it to design, engineer and build quality tubes, because you sell a lot of tubes, which would cover your high R & D  and quality control costs.
And there was a fair amount of competitors which would take your market share if your tubes were subpar.
Now as transistors replaced tubes in everything, what ever market was left for tubes had little demand and used the same 20 (or 30 or 40) year old designs.
One good thing is that Russian and Chinese electronic equipment kept using lots of tubes even after the western world went transistor.
The Russian and Chinese military kept using older equipment designs that used tubes, usually the military stuff need higher quality control and reliablity.
Russia and China switched over to more transistor equipment, so warehouses just sat full of unused high quality tubes for the past 20 or 30 years.
I would also guess everyone had a slightly different way of making the same tube, which accounts for the slightly different sound.
 

 

[Head Injury]

Quote:

Let me guess (my two cents)
Before the transistor came about, everything was tubes, lots and lots of tubes. So it was worth it to design, engineer and build quality tubes, because you sell a lot of tubes, which would cover your high R & D  and quality control costs.
And there was a fair amount of competitors which would take your market share if your tubes were subpar.
Now as transistors replaced tubes in everything, what ever market was left for tubes had little demand and used the same 20 (or 30 or 40) year old designs.
One good thing is that Russian and Chinese electronic equipment kept using lots of tubes even after the western world went transistor.
The Russian and Chinese military kept using older equipment designs that used tubes, usually the military stuff need higher quality control and reliablity.
Russia and China switched over to more transistor equipment, so warehouses just sat full of unused high quality tubes for the past 20 or 30 years.
I would also guess everyone had a slightly different way of making the same tube, which accounts for the slightly different sound.
 

That explains why there are varying levels of quality. But how does tube quality relate to sound quality? Most tube technology had nothing to do with sound.

 

[scuttle]

Quote:

 But how does tube quality relate to sound quality? Most tube technology had nothing to do with sound.

A valve doesn't know whether it is amplifying a radar signal or the Sex Pistols. A signal is a signal.

 

[Head Injury]

Quote:

A valve doesn't know whether it is amplifying a radar signal or the Sex Pistols. A signal is a signal.

So it stands to reason that better tubes distort less. Is there an explanation for things like increased bass, or brighter/warmer tubes?

 

[Uncle Erik]

There are three reasons:

1. Build quality and materials used play a role. Also, some tubes are physically different inside than other types. Look at mesh plates for an obvious example. All these play a role. Also, it's not easy to get a good vacuum inside a tube. Some repros are a little gassy - manufacturing isn't as good as it used to be. I do think that if tubes continue to be popular, good ones will be manufactured again.

2. The circuit is really important. Without getting technical, there are a number of way to use a tube. Changing the circuit to a different type will change how a tube behaves.

3. This is related to point 2, but how you bias the tube changes how it performs. Every tube has to be run with a bias voltage and every tube has a range of where that can be set. For example, one hypothetical tube can be biased anywhere between 100V and 200V. If you were designing an amp with that tube, you'd pull up it's curves (available on the Internet for all popular audio tubes) and look to where it is most linear. Say for this tube, it is 150V. So you can set the tube up for a linear response with a 150V bias. You could also set it anywhere between 100V-200V for a particular sound you want or to meet a design requirement of other tubes used. So, depending on the bias voltage, the same tube will sound different. There's a lot more to this, but you get the idea.

All of this is real-world stuff. Measurable, testable and audible. We've known this for almost 100 years. That's why there's no controversy. Just like there's no controversy over headphones and speakers sounding different, record players, or the quality of recordings.

Cables are something else entirely. There are no measurements or tests because you cannot demonstrate a difference. No one ever has and no one ever will. They're made by companies who are out there strictly to make money. False arguments are cooked up to make money. DACs have science and engineering, but the problems were solved years ago. Most of the arguments are over stats that humans can't even hear. Humans can hear differences in output voltage, which is why people do hear a difference. However, you can compensate by turning the volume control. You don't have to buy a $2,000 DAC with a 1" CNC machined faceplate and $200 RCA jacks for good sound.

 

[BlackbeardBen]

UE is right - the topic of tubes (or anything else) really should be approached like this:
 
 
Q: Why do tubes sound different?
 
A: They're measurably different on a large scale, well within previously established audible levels - blind testing has proven this.
 
Q: Why do they measure different?
 
A: Reasons X, Y, and Z...

 

[Head Injury]

Quote:

There are three reasons:

1. Build quality and materials used play a role. Also, some tubes are physically different inside than other types. Look at mesh plates for an obvious example. All these play a role. Also, it's not easy to get a good vacuum inside a tube. Some repros are a little gassy - manufacturing isn't as good as it used to be. I do think that if tubes continue to be popular, good ones will be manufactured again.

2. The circuit is really important. Without getting technical, there are a number of way to use a tube. Changing the circuit to a different type will change how a tube behaves.

3. This is related to point 2, but how you bias the tube changes how it performs. Every tube has to be run with a bias voltage and every tube has a range of where that can be set. For example, one hypothetical tube can be biased anywhere between 100V and 200V. If you were designing an amp with that tube, you'd pull up it's curves (available on the Internet for all popular audio tubes) and look to where it is most linear. Say for this tube, it is 150V. So you can set the tube up for a linear response with a 150V bias. You could also set it anywhere between 100V-200V for a particular sound you want or to meet a design requirement of other tubes used. So, depending on the bias voltage, the same tube will sound different. There's a lot more to this, but you get the idea.

All of this is real-world stuff. Measurable, testable and audible. We've known this for almost 100 years. That's why there's no controversy. Just like there's no controversy over headphones and speakers sounding different, record players, or the quality of recordings.

Awesome information, Erik.
 
So does the voltage bias alter an amp's frequency response in big ways? Or are the effects subtler, like distortion and roll-off? Or any of the above depending on the tube, the bias, and the circuit?

 

[Uncle Erik]

Pull up your favorite tube over at www.nj7p.org. You'll find a graph that shows the linearity of the tube at certain voltages. You want a curve that's as close to 90°, flat, as possible. None of them are perfectly flat, but some are pretty close.

So, yeah, +/- 10V can significantly alter the frequency response. Tube impedance and other things alter the sound, too. There's a fine balancing act going on - you have to design for several factors. I don't have my head around everything, but I keep reading and trying to understand more. It's really interesting.

 

[zenpunk]

Thanks, very interesting points. But somehow I find difficult to believe today's manufacturing isn't as good as it used to be. I know that there is a lot less manufacturers those days but I wonder where such beliefs originated. It there some resources showing today's tubes don't measure as well as NOS ones. 
Also it seems obvious that because tubes of of the same types have different characteristics and curves, amp manufacturers have to design their amp around specific tubes so tube rolling is only likely to mess with the designer's intended sound and measurements.
@Blackbeardben,
Have you got links  to those blind test of tubes? 

 

[jcx]

computer modeling, mechanical precision, automated process control are all better, cheaper today
 
but tubes also need obscure materials, coatings with processing dependencies beyond just the alloy composition/chemical formula that may have been written down - so higher purity chemicals, alloys may not be "better" without the lost or never made explicit process knowledge
 
the special materials may no longer be made in enough quantity for the suppliers or the tube manufacturers to have retained the "cultural" knowledge among their production personnel, small batches that have to be specially set up will have higher variability than continuous production of the same material
 
many tube factories were shut down, personnel dispersed, the working knowledge lost before the equipment/factory was sold off to someone wanting to make tubes for audiophiles
 
 
even good tubes, same manufacturer, date code may not match well enough in all electrical characteristics to sound exactly the same in common tube amp circuits
 
higher loop gain for effective negative feedback requires more relatively expensive tubes, is limited by output step down transformers, so tube amps seldom use as much feedback as SS amps - each tube costs you Watts of heater power even for mW signal levels
 
cheaper, less power hungry SS parts allow much higher gain from more gain devices in series, higher global feedback allows serial production SS amps to be more similar to each other,

 

[Uncle Erik]

zenpunk, that's why I don't bother to tuberoll. If a tube works, I use it until it dies. They're moving targets, too, because their electrical properties change as they're used.

And there are plenty of "lost" technologies. We have great computers and a better understandin of science.

But a lot of the manufacturing ability and know-how disappered when factories closed and when engineers died. My hope is that some of this will be rediscovered as tubes become more popular.

If you want to read up on some interesting reverse-engineering and lost arts, look into the reproductions of fine violins, like Strads, Guarneris, et al. The repros are getting better and some of them sell well at auction, too.