DIY Amps: A Roadmap for Beginners (2017)

There's people that have done some independent testing of audio outputs. The quality varies widely so paying a lot is one way to make sure you have good outputs. A lot of phones have notoriously weak/bad audio jacks. Some built in motherboard audio outputs suck as well.

However, in general, any dedicated audio card, even the cheapest soundblaster will have audio outputs just as good as extremely expensive audiophile stuff. Maybe not as much power output, but that only matters with high ohm headphones. Ditto with any Apple products, they've been known to have extremely high quality Cirrus Logic audio chips for a long time

I have a Fireface and never had only the slightest issue with the headphone outputs.

As somebody who understands quite a bit about analog circuits and has seen Firefaces from the inside I would definitly trust RME on that one, this thing is beautifully engineered

Np clip is one of the greatest headphone amplifiers ever produced lol

Another good place for tubes, especially NOS weird ones, is ESRC: http://www.esrcvacuumtubes.com/

Originally the Bassman was a 50 watt amplifier with a pair of 6L6 tubes, this is enough power to allow a solid-body bass guitar to play cleanly only about as loud as a regular acoustic string bass.

For bass guitar, there was no need for reverb or tremolo so those tube effects were not included on Bassmans. 6-string guitarists sometimes found these Bassmans more gainful at maximum volume than the non-bass Fenders having equivalent power and tube complement. A single pair of 6L6 (about 50W) for guitar will be loud enough to keep up with an acoustic drum kit for lots of blues and soft rock, while bass guitar actually needs about 300W.

With a loud drum kit and a heavy-hitting hard rock acoustic drummer, then I naturally recommend a 100W fullstack for guitar and 400W for bass.

The circuitry of the Bassman is not designed for big differences in frequency coverage compared to the non-bass vintage Fenders. The Bassman was primarily a regular no-frills 50W head, and for actual bass use within its power level (like in studios) it achieved lower lows because of use with a closed-back cabinet.

As I understand it, the frequency response of a tube power amp is influenced by the output transformer, and to a lesser extent by minor design parameters such as the value of coupling capacitors. To a very rough approximation, lower frequencies need more iron.

On top of this, the preamps in some historic amplifiers tend to have a decidedly non flat response curve when the "tone" knobs are centered.

Personally I don't think they (the old tube Bassman) are that much better than any other guitar tube amp. Old Marshall amps (JTMs) are based on that same design, btw.

The slight touch of harmonic distortion from a tube amp can sound nice, even though it might not be totally transparent. Tube amps are also a lot of fun to build and tinker with because of the huge diversity in tube models, the operating point at which the tubes are run, the gain stage topology, power supply designs, etc. Personally I find that different tube designs almost always sound a little different from each other while well-designed transistor amps sound indistinguishable or almost indistinguishable from one another.

Pete Millett, who runs a DIY Tube site, made a good presentation on this: http://www.pmillett.com/etf_sod.htm

Those "defects" is just non linearity in the circuit. At low gain factors (much lower than a guitar amp) it creates prominent 2nd, 3rd, and 4th harmonic distortion (depending on the circuit) which corresponds to the musical intervals of an octave, octave + fifth, and 2nd octave. That sound is typically described as "warmth" and can be quite pleasing. It also increases the RMS of the signal without (greatly) affecting dynamics, which makes it perceptually louder.

Some people like the ultra hifi systems with crazy THD+N metrics, with super flat speakers in the deadest room they can find to get the most "pristine" sound they can. Other people like listening through systems that subtly color their favorite recordings to match their taste. Different strokes for different folks.

Tube amps (as do transistor amps) can vary quite a bit in how much they change the sound, e.g. part of the appeal of tube guitar amps is that you can drive them out of the area they work "most precise" into distortion, and that distortion sounds interesting. And some people like the sound changes tube amps cause for listening.

I am no EE, but transistors also have problems giving precise accurate gains as well.

Technically speaking there is zero advantage in using tubes instead of solid state when the amp is driven below distortion, that is for listening to music at normal level, and anyone claiming they can hear the difference between a decent 100 bucks class D amp and the most expensive tube amp in the galaxy is lying (probably also to themselves if they bought the latter). However, when tubes are driven to distortion, the waveform is clipped in a much gentler way compared to what is produced by solid state parts, so in a Hi Fi amp those high power peaks would indeed sound a lot cleaner, and in a guitar amp where distortion is looked for by design, using tubes would would produce a much nicer clipped sound, which is even more audible when playing chords. In HiFi amps the problem can be easily solved by designing (or buying) for more power than what is needed, so that the headroom reserve will be used to cope with those peaks without getting into clipping, while using tubes for guitar amps apparently still makes sense. Technology however has progressed a lot since the old days, most parts have become a lot cheaper, and even without resorting to digital manipulation todays solid state pure analog guitar amps can sound great without using tubes.

One important point though: when making guitar amps with newer technologies such as class D, we must redesign everything so that the amp -I mean the power module- will never clip; no matter what we throw at the amplifier main input, the power module must never be driven into distortion because class D distortion is plain awful. That requires a lot of redesign: no more clipping tubes, coils driven into saturation, purposely weak designed power supplies etc, but simply speaking we distort before and keep the level put in order not to distort after. This makes things even more interesting because one builds the guitar sound much before it gets to the final stages, the only further contribution to the sound being the speaker and the cabinet, not the amp module.

In audio most objectivists believe that currently known measurements (frequency response, power handling abilities, THD, PSSR, other non linear distortions, noise floor, etc.) can 100% conclusively define subjective sound. But then on the flip side there are very reputable people (eg. Paul McGowan from PS audio) agree that we may not have figured out all measurements necessary to nail down subjective audio quality. It is upto the ears of the listeners and their exposure to super high end audio to decide whether that is true.

I fall in the camp that thinks we haven't fully figured out what conclusively decides subjective sound quality. It makes things much more interesting and open to more research to settle debates on solid state vs tube, sigma-delta vs r2r dacs, full range driver vs other speaker configurations, cables/interconnects/power-cords matter or not and things like that).

Some interesting links (I don't endorse what these guys are saying but find them VERY interesting):

About DACs: http://www.streamplayer.io/v1/sharing?v=8Mn5PrnZV-k

About whether audio power cables make a difference (Ask Paul from PS audio): http://www.streamplayer.io/v1/sharing?v=8QuToO9JUfw

Roughly speaking, tube amplifier goes in distortion adding pair harmonics when the solid state add odds ones. Odds harmonics aren't nice to the ear (see by yourself by listening to a Sawtooth wave) while even ones are much more flattering (see by yourself by listening to a triangle wave). This distortion relates to the entire theory that's behind the creation of chords and also to psychoacoustics.

Objectively speaking, if the subjective enjoyment of your music is generally increased, then that's a net positive. So it at least depends on the target audience.

My rule is: If it's real, then it can be measured, and if it can be measured, then it can be reproduced in solid state circuitry, perhaps using digital techniques if necessary. With that said, whether anybody is actually doing so satisfactorily is an open question. Actual test data for things like musical instrument amplifiers is remarkably sparse. And I won't begrudge anybody the pleasure of owning or building historic circuits.

Not really. Tube amps use quite a lot of power and their sound can be captured very easily by some of the more expensive modelling amplifiers.

It's a little bit like the tonewood debate in guitars, no one has been able to objectively prove it in electric guitars but some people swear by it.

Did you want to make hifi amps, or guitar amps? The electronics knowledge and construction techniques are nearly 100% transferrable, but the design lore diverges, and the online DIY communities can be very diverse. I've built a dozen or so guitar amps, preamps, etc.

I don't personally build hifi stuff, but I love John Brodskie's tubecad.com for the analysis of classic circuit topologies, and tons of original ideas. He offers nice looking PCB's for preamps and power supplies.

For me it was useful to determine what I wanted to build first, and find the related online sites and communities.

Pete Millet's site is a pretty good resource, he's posted a lot of different designs ranging from preamps to power amps to headphone amps. His "Starving Student" design was my first foray into building a tube amp about a decade ago, though that one is really a tube/transistor hybrid and it's designed for headphones but I used it as both a speaker and headphone amp.

http://www.pmillett.com/

Yes, I should have said "building tube based headphone amps" makes no sense. Opamp based ones make sense for low impedance headphones if the source is not designed for low impedance loads.

Guitar/bass player here. I own a few tube guitar amps (Fender Blues Jr, Peavey Ultra 410) and also have a vintage 70s Ampeg V4b. For nearly 100% of my daily playing and recording needs I use Helix Native; it's simply easier, more versatile, and sounds better than anything I can do with my real amps.

I use Axe FX myself. It's good, but for me it does not really replace _all_ tube amps. I still have a few that do what Axe FX can't do.