Power/Volume difference Tube vs Solid State

[J M Fahey]

Hi Steve, congratulations.
Beautiful job.
Now we need to be treated to some of your playing.
Would you kindly post some MP3s?
Thanks and again, congratulations.

[888guitars]

I gotta hear that amp! I wish mine looked that clean!

[Krwkka]

My own little theory is that there is a stronger 2nd harmonic produced by tubes than by SS and that harmonic impacts one's perception of the loudness of the instrument.

Playing mostly clean in a band with two guitars, sax, piano, vocals (and at not very high volumes I might add) I was able to hear myself far better with a tube amp. Of course, as mentioned above, there could be many factors - different voicing of the amp/speaker combination, etc. - but again and again, as I experimented with various amps, I always found that the sound produced by the tube amps always floated comfortably above the other instruments and into my ears. When playing through solid state amps I always had to crank up the volume to find my comfort zone, usually to the point that I was forcing the other guitar player to crank up his volume to hear himself (he was using a SS amp) and the sax player would complain that I was too loud!

Maybe it's just my ears...

[J M Fahey]

Hi krwwka.
If it's "your ears", fine !!
That's the point.
Please post which amps you are comparing, to have a better idea.
There are many other factors beyond Tube/SS , specially speakers.
Thanks.

[Krwkka]

Which amps, well let's see, there were quite a few. I was into experimenting during that period (1998 - 2003 or so). The solid state amps I tried were a Crate GX(or something)30, Fender Stage 100, a Peavey something - it was an older model, maybe from the early 90s - bigger than the PV Bandit, and a little Fender H.O.T. Out of these, the only one that really didn't sound good to me was the HOT (except when you hit the distortion button with the right guitar).

The first tube amp I tried was a '71 Fender Twin Reverb (in which I tried several different speakers). That got real heavy so I then scored a second-hand Fender Blues Jr. After that was a Fender Hot Rod Deluxe, and then a Reverend Hellhound. I think that's the whole list.

Moving to the present, I still hear the tube amps better, although I haven't been in a band for awhile. Lotsa amps now and a room full of speakers and cabs, but the only solid state jobs are a Stage 1000 (my son's amp, actually) and the Line 6 X3 Live, which I play through PA speakers. Even with the X3 live I notice a big difference in the - I don't know what to call it - the "fullness" or "presence" when powered by a tube amp (50 watt 6L6 push-pull - bypassing the tone controls on the amp) as compared with playing through a solid state Mackie FR-800 amp. And this is using the same cabs (Eminence Delta Pro 12 crossed over to either Eminence PSD-2002 or Selenium D220Ti.

I get the same result when I compare the tube amp with the Delta 12 Pro cabs to an EON Power15. The solid state amps sound good, but when I A-B compared with the tube amp there just was no contest - the tubes just make the X3 Live sing!

[loudthud]

I've been messing around on the "dark side" lately and I tend to agree with the above comments. I don't see any fundamental reason why a transistor or hybrid amp shouldn't sound as loud and as good as an all-tube one.

I thought I could count on Steve for a little backup on this subject.

Here's the setup: An A/B footswitch is connected to switch between an 8 ohm dummy load and two 16 ohm speaker cabinets connected in parallel. One cabinet is a 16 ohm JBL E140 and the other is four 16 ohm no-name 10 inch speakers in a ported cabinet. A digital scope (Tek 7854) monitors the input to the A/B switch and is set to single shot mode. Signal source is a bass guitar.

The first photo is a Peavey 80s bass amp with a 260 power module. The upper trace is with the speakers connected, the lower trace is the dummy load. The second photo is 6550 bass amp. Again, the upper trace is with the speaker load, bottom with the dummy load. Note that the speaker gives more voltage than the dummy load but with the tube amp there is more difference. The third photo is the same as the second except the waveforms are directly overlaid and the deflection factor is increased.

More voltage across the speaker means it sounds louder.

Steve: Build a little solid state power amp to duplicate the output of the EL84 into a dummy load and then compare the outputs with a speaker connected.

Another way to look at it: Connect a 5E3 to a dummy load and note the peak to peak voltage it can deliver. Then connect a speaker. You'll notice the difference immediately.

[J M Fahey]

Hi Loudthud, thanks, *excellent* work.
That's the way to analyze what's going on, coupled to a *good* listening test.
A lot of people have scopes, many don't use them, or don't believe/care what they show, preferring to repeat what's floating around on the Net.
That's the way to check; changing ONE variable at the time.
In this case, only the load, but keeping amplifier and signal constant.
Fully agree with pictures 2 and 3, they match my findings.
I have a problem with picture 1:
that kinky waveshape (in the real meaning of the word) can be found on tube amps, but not on SS power amps. Are you sure some earlier stage is not distorting?
Please repest the test with a clean sinewave (same frequency and amplitude) to have something more repeatable than a plucked bass string.
By the same token, I find it strange that the lower (resistive load) waveshape is kinkier than the upper one, and asimmetrical to boot.
Very unlikely on an SS amp, looks all the way like a *Tube* power amp, with not much negative feedback (or none at all) and not very well matched power pentodes .
Maybe you uploaded another set of pictures by error.
Anyway, thanks again for an a very useful test.
PS: anyway, I can easily build an SS power amp which matches the behavior shown in Photos 2 and 3 []

[teemuk]

I have a problem with picture 1:
that kinky waveshape (in the real meaning of the word) can be found on tube amps, but not on SS power amps. Are you sure some earlier stage is not distorting?
Please repest the test with a clean sinewave (same frequency and amplitude) to have something more repeatable than a plucked bass string.

There you have it (I guess), the important line: "signal source is bass guitar".
So as far as I understand, the input signal was not sine wave to begin with and that might easily explain the skewed wave shape. Our mind plays a trick on us and tries to compare that wave to a sine wave, when actually one important scope capture is missing: What was the signal actually driving those amps?

The output signal of tube amps overdriven with sine waves also often skews a bit like that due to bias shifts causing crossover distortion, which then skews the wave shape at crossover region. Solid-states do not generally do this - unless you deliberately make them do it like e.g. Hughes&Kettner's Dynavalve power amps or Peavey's Tube-Dynamics power amps. Also, doesn't the Peavey 260 module come with the DDT limiter? I imagine that at such levels of overdrive it would have some effects that would further skew the wave shape.



Also, another important point is that the 260 module doesn't employ any current feedback so the SS amp and the tube amp will have a widely different response to a reactive load. The test about developing more voltage across such load makes more sense when you compare a SS amp with "mixed-mode" feedabcak to a generic tube amp, because with such circuits the differences in frequency response are much more levelled out. Naturally, in such comparison the differences between tube amp and solid-state amp behaviour tend to decrease even more. It's a rather sensible comparison too because most modern solid-state power amps (for musical instruments) tend to employ mixed-mode feedback rather than plain voltage feedback.

[Enzo]

But as we shift the focus over to whether a SS circuit can duplicate a particular form of tube circuit distortion, and indeed the whole spectrum of them, and whether such an emulation circuit could transition from clean into that state smoothly, we then are no longer addressing the main question: will a "50 watt" tube amp make higher SPL from a speaker than a "50 watt" SS amp under anything reasonably considered normal use.

[J M Fahey]

Hi Teemu, thanks for posting.
As always, offering an accurate insight on new technology.
I hope this will get incorporated in some new edition of your book.
What you posted shows exactly the waveshapes I was expecting, even the (usually) non symmetrical tube clipping.
Although they centered the output waveform in the scope, fact is that the real center point is not the center horizontal screen line but the crossover kink points; referred to which the lower half has a larger peak value than the top one.
H&K faithfully reproduces even that.
I would love to hear one of them live; haven't come across one yet.
I asked for a sinewave simply to have a more repeatable test signal, of course it is also very important to test with "real world" signals.
Loudthud, thanks again.
You two guys are an example (among others) of what makes this board awesome.
PS: If I remember right, the 260 used a limiter based on the CA3080. That one, as well as the photocell based ones, is quite symmetrical; but Fet ones are not (by definition) and might furnish different amplitude on different halves of the compressed signal.
Anyway I think it was not the case here.

[J M Fahey]

Hi Enzo.
Yes , you are right, if we compare an "average" Tube amp to an "average" SS amp of the same measured "just clipping" power , under the same conditions (same preamp driving them, same speakers), yes, I agree the Tubed one has an audible advantage, for the same reasons posted above and shown by Louthud's scope pictures ( which match "audible" results).
What Teemu and I drifted into, was into how to make an SS amp match (or even beat, why not?) a similarly rated average tube amp .
My personal point is that it can be done, but it will take a lot of work and we have only advanced a short walk along that path, there's still too much to be done.
Anyway, what's life without a challenge?

[loudthud]

Ok, now we have something to talk about. It was my intent to study clipping behavior, not crossover distortion. Crossover distortion certainly plays a role in how an amp sounds and tube amps have a dynamic crossover characteristic that solid state amps don't. The issues of compression and the harmonics the amp generates also play a role but it is my contention that the primary reason a tube amp seems to be louder is because the tube amp can deliver more voltage to a speaker than a comparably rated solid state amp.

Playing continuos sine waves into a speaker in not something I like to do. It's really hard on the ears, even with ear plugs. A short burst is ok but I don't have a function generator that will do it properly. I have a device from Groove Tubes that can act as speaker load which is what got me started down this whole train of thought. Aspen Pittman said that there is something that happens when a tube amp is connected to a speaker that makes it sound different that a solid state amp. It's the effect the lumpy speaker impedance has on how a tube amp clips. Since the impedance changes with frequency, single frequency analysis will play tricks on you.

I must confess that I distorted the data in a way nobody picked up on. The two speaker cabinets in parallel have staggered resonances that combine in parallel to present a higher (than 8 ohms) impedance across most bass frequencies. This is the key difference between the resistive load and the speaker load. When a tube amp clips driving a resistive load, there is a significant voltage across the tube that is trying to pull one side of the OT to ground. (Especially 6550s.) If the instantaneous current that is demanded by the load is reduced, either because of a phase shift between current and voltage of the load or just a higher impedance, the voltage across the tube will be reduced and more voltage delivered to the load. A resistive load always demands peak current and peak voltage at the same time so the voltage that can be delivered to a resistive load is limited.

One way to emulate the tube amp's behavior at clipping is to simply add a resistor in series with the speaker. My experience says this resistor needs to be somewhere between 1/4 and 1/2 the speaker's nominal impedance. That resistor is going to dissipate significant power and it will reduce the power that a solid state amp can deliver to a dummy load and thus the solid state amp's power rating. For lack of a better term I call this the "Damping Resistor".

The damping resistor of course reduces the damping factor seen by the load so it would seem natural to apply a small amount of negative feedback around the damping resistor to tighten up the output a little. A simple resistor from the speaker side of the damping resistor back to the inverting input of the power amp could do it. This kind of feedback confuses me unless modeling software is used.

[J M Fahey]

Ok, now we have something to talk about.

Good !

it is my contention that the primary reason a tube amp seems to be louder is because the tube amp can deliver more voltage to a speaker than a comparably rated solid state amp.

Agree, it's one of the mechanisms, the most obvious one.
I have been experimenting on SS replacing tubes since 1972, since power tubes disappeared from Argentina for over a year. And I mean disappeared, you had to buy what was left at an auction, paying 6x the price.
Most others switched to locally produced 6DQ6 (ugh), but I thought that even SS might be better than *that*.
One of the first things I noticed (obviously I was not deaf to the tone differences) was that my Twin clones
a) had around 4 ohm internal impedance and
b) could deliver almost the same power on 2x the regular impedance.
Beware that both statements are *not* the same !!!
Or to be more precise, they *can* be the same if one uses a 4r resistor in series with the 4r nominal load and if you do not include it within the loop (there's the answer to Loudthud's last doubt) because this would defeat its purpose.
I found two separate ways to achieve both goals; the first one is what was re-invented later and called "valvestate"; the second one is not used today, except, maybe, by accident.

Aspen Pittman said that there is something that happens when a tube amp is connected to a speaker that makes it sound different that a solid state amp.

Yes, it's true, but it goes even beyond the load variation effect on instantaneous voltage along time.
On purpose I'm avoiding the use of the word "peak" because , for me, it's "static", it only shows one specific point on the waveform, and I mean many points which, to boot, change dynamically.
It will cause a variation on the instantaneous peak voltage across speakers terminals, that's what we see on the scope pictures, and as I said, can be somewhat easily imitated by a proper SS circuit.
Anyway, there's more than that happening, there's an interactive effect which modulates what the tube does when overdriven which changes what the tube does along time, which can't be shown by any regular screen capture, which is bi-dimensional, rather needing a tri-dimensional graph.
Sorry, can't find the words to explain it better than that, although the effect is quite clear to me.
It's like trying to explain a colour over the phone. Oh well.

[Steve Conner]

Also, another important point is that the 260 module doesn't employ any current feedback so the SS amp and the tube amp will have a widely different response to a reactive load.

This is exactly what's happening IMO. The speakers have an impedance higher than their nominal value in the bass register, because of resonance. The tube amp responds to this with more output voltage because of its high output impedance, but the solid-state amp keeps delivering just the same voltage.

The solid-state part of my amp is set up as a current source, and doesn't damp the speaker at all. I had to add some voltage feedback to tame it, because it made too much presence boost for my taste.

I'm willing to bet (but not a large amount ) that this is the "something" Aspen Pittman was talking about.

[kg]

look, tube watts are bigger than solid state watts.

that's all you need to know.