technical questions - tube vs. solid state, etc

[baald]

(x-posted on webervst)

calling randal, john, OTM, and others:

hi guys, giving a presentation as part of a job interview process (i was
able to choose any topic as long as it's "technical") and I'm hoping
someone has succinct answers or links to answer some of the
foolwing.

1). it is said that one reason why tubes sound "better" than transistors
is because they clip softly. why would a tube amp clip more softly than
a SS amp?

2) it is said that tube amps sound "better" because they generate even-
order rather than odd-order harmonics (when distorting). I could see
how this would be related to the above, but it seems it could also be a
mechansim of most tube circuits using minimal feedback (as feedback
lowers the amplitude of distortion spectrum, but drqaws it out in terms
of more and more harmonics...)

3) what accounts for the greater effeciency off PP topologies besides
the fact taht you can run them in class a/ab, ab, or b, and that the
power supply requirements are easier as there is complementarity
between the two sides?

any help appreciated. need to finish writing this tomorrow (Tues) for a
wednesday presentation. thanks

baald - baald {AT} auriculum {DOT} com

 

[John Phillips]

IMO the single largest difference that makes each sound the way it does is the fact that tube amps almost always have an output transformer and solid-state amps almost always don't. It's not all of the difference - if you fit a solid-state amp with an OT it doesn't sound exactly like a tube amp - but it makes a very large difference, especially on overdriven sounds.

It's not really true that tubes clip more softly, and definitely not true that they generate even-order harmonics whereas SS amps generate odd. In fact, push-pull amps (tube or solid-state) by their very design cancel even-order harmonics.

BUT if you pass a clipped signal through an output transformer, interesting things happen. A clipped signal is basically a square wave, and if you look at the waveform, a square wave is effectively short sections of DC, separated by sharp vertical transients. But these are the two things that transformers can't pass - they only work with changing currents, so DC is blocked, and their large inductance stops very fast changes, so really sharp transients are blocked.

So instead of rising instantly and then holding a DC voltage, the waveform rises less fast (ie sloped) to a peak, then falls back (also sloped), in each half cycle, which turns the square wave into something much more like a sawtooth wave. In a SS amp with no OT, the plain square wave is supplied directly to the speaker - this is also why distorting solid-state amps are very hard on the speakers.

Sawtooth waves are composed mostly of even harmonics, whereas square waves are mostly odd harmonics. So the final output to the speaker from the two types is very different harmonically. FWIW, there are no 'natural' acoustic sounds that are square waves - it's something that can't occur in the physical world outside of electrical amplification, which is why we hear solid-state distortion as harsh and unnatural, and tube as sweet and natural. In fact, a distorted tube amp is quite like a brass instrument in terms of output waveform.


Tube amps are less efficient - even discounting the filament power consumption - also because of losses in the transformer, as well as the higher internal impedance of the tubes; even though the currents are smaller, they dissipate proportionally more power.

 

[baald]

Originally posted by Old Tele man
...well, how'd it go?

why thank you for asking, OTM

first let me tell you a little about the interview process....
1. phone interview
2. email quiz w/ technical adn brainteaser questions
3. 3 hour-long interviews (CTO, hiring manager, dept. peer)
4. 6 hour-long interviews (3 peer-level engineers form other depts, two managing engr's, one product manager)
5. the presentation yesterday

my presentation was 3 parts: how tubes work, why they disappeared (from the mainstream - hifi anyway), and why they have made a resurgence.

20 minutes into part 1 (it was the main part - about 40 minutes worth or so) the CTO interupts me and says "ok, can you just skip to the meat? i wanna hear about why they disappeared and why they are coming back". I'm convinced this was just part of his plan -- see how the candidate deals with a curveball. The rest of the presentation (30 minutes) was nothing but questions. one of which i couldn't answer (why does harmonic distortion happen - have read the reason but not retained it).

15 minutes after the presentation, they made me an offer. Best offer i've ever had in my life. I'm pretty stoked. btw - it's a software QA/test engineer position - sr/lead level.

John - i used some of your ideas, though i'm not sure i agree with all of them - mainly because i was tech mgr for a company who made a tubed preamp that passed 10kHz square waves like it wasn't even there (again, talkin hifi stuff). Great point about the OPT's role in tone shaping though.

my overall gist was that my theory is that it has more to do with simple circuits and especially low NFB in tube gear vis a vis SS gear (which is based on distortion spectra and behavior of NFB-laden circuits). but to be honest, i think there's genie in the bottle

thanks for your input John, and your interest, OTM.

 

[John Phillips]

Originally posted by baald
John - i used some of your ideas, though i'm not sure i agree with all of them - mainly because i was tech mgr for a company who made a tubed preamp that passed 10kHz square waves like it wasn't even there (again, talkin hifi stuff). Great point about the OPT's role in tone shaping though.

Preamp, yes. A tube preamp can have pretty much the same frequency response as a solid-state one, and handle square waves very well (although they're more usually decoupled between stages compared to solid-state, which limits it to some extent) - a DI'd tube preamp can sound pretty harsh and solid-state-like too.

But feed that same signal through a tube power amp and look at what happens .

 

[baald]

Originally posted by John Phillips
But feed that same signal through a tube power amp and look at what happens .

ok. using an output transformerless amp w/ a dozen or so 6as7g's driving magnepans (purely resistive load).

just kiddin' with ya. getting a job after a year or so of being unemployed can make one crazy giddy .

 

[baald]

thanks for the congrats, OTM.

unfortunately, TRS-80's were a bit before my time. well, actually, we had some in the high school computer lab, but i wasn't interested at the time. Strangely, for a while I majored in computer science in college, but switched to philosophy (which I rcvd my BA in) because i didn't think i wanted to code 8 hours a day. HAHA!!! now i enjoy coding, especially when it's MORE than 8 hours a day (as it means i'm very focused and absorbed).

i'm an analog guy at heart too (started on my first turntable when i was 16. it's still being improved upon. as it stands it's mainly made of about 50 lbs of delrin, though it may be rethought at sometime) but play mainly CDs (have a wall of records though). And will be investing in a 192/24 front end for my computer for recording. I have a friend in my hometown who makes gorgeous recordings on his 2" 16 track MCI, but I have neither the $$$, space, or inclination for tuning that he has.

and even though i'm a tube guy at heart, my stereo is based around this sweet little italian hybrid job (pathos classic one). i had a gorgeous 2a3 amp built by Don Garber (Fi electronics, brooklyn) that i was playing through tweaked altec valencias, but that was BEFORE marriage, when i had a 20 x 30 ft room for them in the barn cum loft i was living in at the time. now stuff has to have a high WAF (wife acceptance factor). she's pretty tolerant, but even the fancy-cab'd A-7's were a bit overbearing for our current place.

anyway, thanks for the thoughtful replies. I'll have to post pics of the PPP 2A3 amp for guitar i'm working on (well, not ACTIVELY, but i have drawings on paper and the iron on a chassis in my storage place and the glass in my tube boxes) if it ever gets finished. It's based on an old hammond organ amp, but has absolutely no front end - just iron, PS, and a pair of 26's driving the 2A3's. Still deciding if i want to keep it period correct and do 26 or 76 gain stages, or get "modern" and go 6SL7 hehe...

baald

 

[QuickDraw]

http://www.renditionaudio.com/evolution.html

this might be of some interest to this discussion

 

[drbob1]

Transformerless means "direct coupled output" right? Isn't there a schematic for that in later versions of Aspen Pitman's book? What would it sound like? 4 8" drivers in a stack like that suggests that they're using line array type linking (ala the Bose PAS) to reduce attenuation of the signal with distance. Has anyone ever done that with a guitar amp?

 

[QuickDraw]

well i'm not going to pretend too much to know how it works, cuz it still seems like voodoo to me but the way i understand it, the key to what he does is the power tubes, the way he wires up the poweramp i believe it puts out an impedence of 32 ohms which can go straight to the speakers. they sell these in a store by me. (they make them right by my house as well, i plan on going to the shop to check out everything, sounds very interesting.

http://www.atma-sphere.com/products/ma3.html

they also make these hi-fi amps that look crazy!!!