Technically, what exactly is TUBEY sound?

[Pluto]

Read and liked

Thank you!

 

[milosz]

Maybe this was mentioned earlier in this thread, but I did not see it if it was:

The biggest difference I hear when listening to speakers driven by a tube output stage vs. solid state is related to Ohms-law interactions between the higher output source impedance of a typical tube amp and the load impedance of the speaker, which varies with frequency. Many tube amps introduce a frequency response artifact from the interaction of their high-ish output source impedance with the speaker's load impedance. Have a look at the following two graphs- the black trace shows amplifier output power output as affected by speaker load impedance interacting with the amplifier's source impedance:

PRIMALUNA DIALOGUE AMPLIFIER - TUBE from https://www.stereophile.com/content/primaluna-dialogue-premium-power-amplifier-measurements
± 8.5 dB response variation !

PARASOUND INTEGRATED AMP - SOLID STATE from https://www.stereophile.com/content/parasound-halo-integrated-integrated-amplifier-measurements
± 0.125 dB response variation nearly two orders of magnitude lower variation than the Primaluna!

One can argue about low-order harmonics that tubes can inject into the sound, etc etc but those effects are very small in magnitude compared to this source impedance- speaker load impedance "Ohm's law interaction." I mean, ±8.5 dB of frequency response variation is a pretty huge coloration.

Not all tube output stages show such high source impedance, but with the recent fashion of avoiding negative feedback we see more and more tube amps with high output source impedance. This is fine, but people who prefer this sound need to understand that they are using their amplifiers as a TONE CONTROL.

(Actually I rather like the way a pair of 2-way speakers I built using a 5.25" SEAS magnesium cone woofer and a SEAS Millenium tweeter sound on an Audioromy FU29 tube amp which shows EXACTLY this kind of coloration. But I KNOW I am using the amp as a tone control. The other thing I like about the amp is that when it is clipping in the low bass it still sounds OK, the amp doesn't produce all kinds of high-order trash. This is a system I use for casual listening. It sounds "nice" but not neutral or accurate. I wouldn't say it sounds especially "musical" either, it just sounds "nice.")

 

[Wombat]

Maybe this was mentioned earlier in this thread, but I did not see it if it was:

The biggest difference I hear when listening to speakers driven by a tube output stage vs. solid state is related to Ohms-law interactions between the higher output source impedance of a typical tube amp and the load impedance of the speaker, which varies with frequency. Many tube amps introduce a frequency response artifact from the interaction of their high-ish output source impedance with the speaker's load impedance. Have a look at the following two graphs- the black trace shows amplifier output power output as affected by speaker load impedance interacting with the amplifier's source impedance:

PRIMALUNA DIALOGUE AMPLIFIER - TUBE from https://www.stereophile.com/content/primaluna-dialogue-premium-power-amplifier-measurements
± 8.5 dB response variation !

PARASOUND INTEGRATED AMP - SOLID STATE from https://www.stereophile.com/content/parasound-halo-integrated-integrated-amplifier-measurements
± 0.125 dB response variation nearly two orders of magnitude lower variation than the Primaluna!

One can argue about low-order harmonics that tubes can inject into the sound, etc etc but those effects are very small in magnitude compared to this source impedance- speaker load impedance "Ohm's law interaction." I mean, ±8.5 dB of frequency response variation is a pretty huge coloration.

Not all tube output stages show such high source impedance, but with the recent fashion of avoiding negative feedback we see more and more tube amps with high output source impedance. This is fine, but people who prefer this sound need to understand that they are using their amplifiers as a TONE CONTROL.

(Actually I rather like the way a pair of 2-way speakers I built using a 5.25" SEAS magnesium cone woofer and a SEAS Millenium tweeter sound on an Audioromy FU29 tube amp which shows EXACTLY this kind of coloration. But I KNOW I am using the amp as a tone control. The other thing I like about the amp is that when it is clipping in the low bass it still sounds OK, the amp doesn't produce all kinds of high-order trash. This is a system I use for casual listening. It sounds "nice" but not neutral or accurate. I wouldn't say it sounds especially "musical" either, it just sounds "nice.")

A well respected and very experienced member of a one-brand Hi-eff speaker forum I frequent says that if you use vintage high-efficiency loudspeakers with SS amps, to get that tube sound, just insert a 1 to 3 ohm resistance in the line. It is all about damping factor with these old babies that were designed in the good old tube amp days.

 

[solderdude]

That is partly true. Partly because this only f-s up the damping factor.
As tube amps deliver power rather than voltage they react differently to higher impedance swings compared to only a lower damping factor.
Tube amps with the same nominal power rating as SS amps can provide a much higher voltage swing in higher impedances where SS amps would clip.
When both amps play at levels below clipping levels of th SS the added resitor trick will probably have the desired results emulating some tube amps.

 

[Wombat]

That is partly true. Partly because this only f-s up the damping factor.
As tube amps deliver power rather than voltage they react differently to higher impedance swings compared to only a lower damping factor.

Power delivery is due to voltage and current capability into load(impedance). The current flow in the resistive component of the load, determined by the impedance, will produce power. I don't understand that tube amps deliver power rather than voltage. Without voltage there is no power delivery in amps. Ohms law. I'm sure you meant something different to what you said.

What the guy was essentially saying is that the vintage HiEff speakers were designed to work at their best with higher output impedance (tube) amps.

 

[DonH56]

https://www.whatsbestforum.com/thre...esponses-into-speaker-loads.8103/#post-139289

 

[solderdude]

Power delivery is due to voltage and current capability into load(impedance). The current flow in the resistive component of the load, determined by the impedance, will produce power. I don't understand that tube amps deliver power rather than voltage. Without voltage there is no power delivery in amps. Ohms law. I'm sure you meant something different to what you said.

What the guy was essentially saying is that the vintage HiEff speakers were designed to work at their best with higher output impedance (tube) amps.

Nope what I said is correct. Tube amps supply power.
An example. A SS amp provides 30 W into 8 ohm and 15W in 16 ohm.
A 30W rated tube amp rated 30W in 8 ohm can supply 25W in 16 Ohm so substantially more (granted depends on the design) .
So around resonance frequencies of speakers more power (and thus voltage) is available. That's simply how tube amps work.

 

[Julf]

Nope what I said is correct. Tube amps supply power.

So do solid state amps. The difference is that tube amps have a significant output impedance, so the output current is limited.

An example. A SS amp provides 30 W into 8 ohm and 15W in 16 ohm.
A 30W rated tube amp rated 30W in 8 ohm can supply 25W in 16 Ohm so substantially more (granted depends on the design) .

Yes, that SS amp would be rated a 15 W amp. Let's use a 25 W rated SS amp, so both the SS and tube amps give 25W into 16 Ω. With a 8 Ω load, the SS probably provides close to 50W, while the tube amp is down to 30W. At 4 Ω, the SS amp will probably still give 30W, while the tube amp would really be struggling.

To see why this is relevant, have a look at the impedance curve of a nominally 8 Ω speaker.

 

[MRC01]

I would like to know what is the scientific reason behind the "attractiveness" of tubey sound.
My impression is as below: ...
...

I've listened to several tube amps over the years but the only one I actually owned was the Wheatfield HA-2, a SET OTL headphone tube amp that I used to drive Sennheiser HD580s, whose 300+ ohm impedance makes them a good load for an OTL amp.

That tube amp had frequency response that was essentially as ruler flat as most good SS amps. To my surprise, in level matched blind tests it was difficult to differentiate from a good SS headphone amp (at the time, a Headroom Maxed Out Home). I was able to tell them apart in blind tests but only with careful critical listening in a quiet environment to certain high quality recordings having full spectrum sound. Casual listening, even attentively casual listening, they were indistinguishable. This suggests that oft-described "tubulicious" sound is a caricature and well engineered and built tube amps don't necessarily fit that mold. Subjectively, the differences I noted were the deepest bass (like 25-30 Hz drum hits) had faster cleaner attack on the SS amp, and the extreme HF (like castanets or jangling keys) was better articulated (not louder, but cleaner, lighter, more natural) on the SS amp. Also the tube amp had a low level (about -66 dB at moderate listening levels) 60 Hz hum that could just barely be detected in quiet passages.

Overall I went with SS because it had essentially equivalent sound quality, at a lower price, with greater reliability and lower maintenance.

I think the scientific or psychoacoustic reason that some tube amps are described as attractive is is twofold. First, whatever distortion they have, most of it is 2nd harmonic which is less dissonant and can actually be euphonic. Second, driven near their limits they clip soft rather than hard similar to recording on analog tape using levels just a bit on the "hot" side. Not enough for obvious distortion, but just enough to add some presence on dynamic peaks.

 

[DonH56]

Most tube amps have tapped output transformers so deliver the same power into a resistive load be it 4, 8, or 16 ohms (assuming you are using the tap matched to the load). The transformer allows them to boost the voltage for the higher loads. Most SS amplifiers look more like voltage sources so power capability varies with load (McIntosh's output balun a notable exception). SS amps' low output impedance generally means they have less frequency variance driving typical speakers that have fairly wide impedance deviations over frequency.

IME/IMO/just another unsubstantiated misinformed unscientific viewpoint - Don

 

[MRC01]

Most tube amps have tapped output transformers so deliver the same power into a resistive load be it 4, 8, or 16 ohms....

Yep, and the term "OTL" refers to the ones that don't. "Output Transformer-Less". These may be more transparent (no transformer in the signal path) but generally, they can only drive high impedance loads.

 

[DonH56]

I deliberately avoided mentioning OTL in general and Futterman specifically due to bad experiences in a previous life... It wasn't all bad, but when it was bad, it was really bad.

I'd like to hear the Atma-Sphere amps.

 

[Julf]

Most tube amps have tapped output transformers so deliver the same power into a resistive load be it 4, 8, or 16 ohms.

A small qualification to that - they deliver the same power into the load that has been selected. They don't do that if a nominally 16 Ω speaker (with the transformer set to that load) drops to 8 Ω at some frequency.

 

[DonH56]

A small qualification to that - they deliver the same power into the load that has been selected. They don't do that if a nominally 16 Ω speaker (with the transformer set to that load) drops to 8 Ω at some frequency.

Thanks, corrected my post. Assumptions and all that jazz...

One of the other things I learned the hard way is that different manufacturers apply the "-" label differently. I had an amp that reversed polarity using the 4-ohm taps and the 8-ohm outputs actually floated... When I plumbed my 'scope to the amp's output it burned the probe's ground lead in two. At least it didn't fry the amplifier... Folk using subs like REL that add a ground wire need to be careful to follow their manual and not attach it in that case.

 

[Julf]

When I plumbed my 'scope to the amp's output it burned the probe's ground lead in two. At least it didn't fry the amplifier... Folk using subs like REL that add a ground wire need to be careful to follow their manual and not attach it in that case.

I too learned to use isolation transformers - the hard way.

 

[MRC01]

A small qualification to that - they deliver the same power into the load that has been selected. They don't do that if a nominally 16 Ω speaker (with the transformer set to that load) drops to 8 Ω at some frequency.

One consequence of this important point is you can't trust the frequency response measurements of a high output impedance amplifier. It might be flat as measured by the manufacturer. But conventional headphones or speakers have varying impedance with frequency, which will change the impedance ratios and thus the frequency response.

So when it comes to tube amps just because it "measures flat" doesn't mean it actually has flat frequency response when it's driving an actual headphone or speaker. I believe this is a contributing factor to tube amps sounding more different when driving different headphones, and if the impedance vs. frequency curves for headphones shares some common characteristics (such as impedance peaking in a certain frequency range) this could also contribute to what people think of as the characteristic "tubey" sound.

 

[MRC01]

One consequence of this important point is you can't trust the frequency response measurements of a high output impedance amplifier. It might be flat as measured by the manufacturer. But conventional headphones or speakers have varying impedance with frequency, which will change the impedance ratios and thus the frequency response.

So when it comes to tube amps just because it "measures flat" doesn't mean it actually has flat frequency response when it's driving an actual headphone or speaker. I believe this is a contributing factor to tube amps sounding more different when driving different headphones, and if the impedance vs. frequency curves for headphones shares some common characteristics (such as impedance peaking in a certain frequency range) this could also contribute to what people think of as the characteristic "tubey" sound.

PS: I've read that long ago there was some kind of standard assumption that headphone amps had 120 Ohm output impedance, and some headphones were designed for this. So they may sound too bright when played with a good SS amp having low output impedance. Have any of the old-timers or audio historians here heard of this?

PPS: some references on this:
http://nwavguy.blogspot.com/2011/02/headphone-amp-impedance.html
https://www.innerfidelity.com/content/musings-headphone-amplifier-output-impedance

 

[watchnerd]

I deliberately avoided mentioning OTL in general and Futterman specifically due to bad experiences in a previous life... It wasn't all bad, but when it was bad, it was really bad.

Any flames?

 

[rwortman]

Nope what I said is correct. Tube amps supply power.
An example. A SS amp provides 30 W into 8 ohm and 15W in 16 ohm.
A 30W rated tube amp rated 30W in 8 ohm can supply 25W in 16 Ohm so substantially more (granted depends on the design) .
So around resonance frequencies of speakers more power (and thus voltage) is available. That's simply how tube amps work.

I’ve seen current amplifiers and voltage amplifiers. You seem to be saying that the output response of a tube amplifier is power. Voltage signal in. Power signal out. I know a simple triode sure doesn’t do that. Explain.

 

[solderdude]

A simple triode, like a FET is voltage in and current out. It is the anode resistor that determines the output voltage.
A higher resitance is a higher gain.
Replace a resitor with a varying one, say a tansformer with a spreaker and when the impedance is higher the voltage is higher.
Of course it also works the other way around. A lower resistor has a lower output voltage. Depending on the circuit, amount of feedback and type of feedback as well as the used transformer and supply voltage this effect is more or less present.
Well designed tibe amps with lots of overall feedback and hybrids with voltage followers are not bothered by it.
It is not a general rule. It is a possible cause that explains why some tube amps have a higher than expected output voltage when looking at power ratings given at a certain load resistance while speakers have an impedance. Most SS amps have voltage followers as an output stage and limited voltage rails and plenty of overall feedback.