Tube Vs. Solid State

So is the impedance curve of the 558, like, really different than the 600?

 

Yeah, a transformer coupled amp would certainly be different. I wonder what it would be like through the Laconic NBM.

 

Actually, I wouldn't say so. From my measurements even headphones like the HD600 and HD800 (and most certainly the stats) have less than 0.2% distortion at 1kHz at sane levels - even at OJ levels. The key here is at 1kHz, of course distortion will rise past 0.2% in the bass for most dynamics headphones. (I measure about 0.025% at 1kHz at 90db for the HD800, but that's probably limited by my microphone.)

Is 0.2% 2nd order distortion audible? Probably not.

 

Some find -60 dB THD (0.1 %) objectionable. See Belcher (1978) so to some the 0.2 % of the DG300B will be audible.

Interesting with the headphone measurements. Note how H2 creeps up with the tube amp but the other distortion products stay put.

Tom

 

Technically, no. While it is possible to create a solid state amp that performs worse than a tube amp on a range of parameters, it is not that hard to make a solid state amp that performs better than a tube amp. Tubes are not that linear and don't have that much gain. Thus, there isn't a lot of loop gain available to drive the parameters towards perfection. Solid state circuits, in particular those in modern opamps, are quite linear and make it possible to make solid state amps with gobs of loop gain (= available error correction), thereby, achieving stellar performance.

It's mostly a matter of personal preference. Some, me included, enjoy the little something-something (low-order distortion) that tubes add. On some recordings that works really well. On others, not so much. Other people, me also included, like the lean and clean sound of a precise solid state amp. I find the precision, while unforgiving, works better across a wide range of musical genres. Each to his own. We don't all have to like the same stuff.

Some will argue that good measurements and good performance are not the same thing. That's not my cup of tea. To me, that's like arguing that you can't calculate your tax return using a calculator. You should estimate your taxes based on what you feel like in the moment. There's so much ambiguity in audiophile language that I've given up on tracking it all. In the 1980ies we were lifting veils. "It is as if a veil has been lifted". Now, I guess, we're talking about micro-dynamism and such.

A tube amp will always win on looks, though. There is nothing as cool as a mesh plate 300B glowing dimly in a dark room.

Tom

 

Any thoughts on Korg's new NuTubes?

 

Are those the "solid state tubes"? Tube emulators?

There's quite a bit of tube emulation out there in the guitar amp space. To my knowledge, some of the emulation circuits are pretty decent. Decidedly not bad. Many musicians still swear by the old tube designs, which could indicate that the emulators don't quite get it right.

In general I don't get the "solid state tubes" or "discrete opamps". If you want to use a tube, use a tube. If you want an opamp, buy one! I've yet to see a discrete opamp that outperformed a modern opamp.

Tom

 

Those are actual vacuum tubes, just in a very different form.

 

I guess the discrete designs for audio achieve higher open loop gain and still respectably low noise density. Though it looks like the OPA1611 does win in the noise density dept. The price delta is also downright silly.

 

Ah! Those. Meh. Show me a data sheet...

The discrete designs (both for "discrete opamps" and discrete amplifiers) I've seen have all had considerably lower loop gain (50-60 dB) than modern opamps (120-140+ dB). On silicon you can get small devices and high-impedance nodes. You also get stellar matching between individual components if you pay attention. This allows you to get high loop gain. I spent 10+ years as a semiconductor design engineer (the LMP2021 is my (and Rod Hughes') design and I have circuits in a bunch of other National/TI chips). I can't even in the simulator get loop gain in a discrete circuit that competes with what you can get on silicon. Then you add that the "discrete opamps" are really poorly spec'ed. They might list a few parameters, but it's nothing that comes even close to the full data sheet you get from TI, which is based on measurements of thousands of units.

The discrete opamps cost 25x as much as a good audio opamp, such as the OPA1611 or LME49710.

Tom

 

Ok, I checked one randomly... The Sparkos Labs has 140dB, 3nV/rt(Hz)... So not bad performance comparatively, but horrible price to performance ratio unless you're one of those insatiable "opamp rollers". And even then.

 

Weirdly sparse.

http://korgnutube.com/pdf/Nutube_DatasheetV1E.pdf

 

That's a very linear tube. It even looks like it could be driven into Class A2 (grid current). Interesting with the uA anode currents. It could work well with a CCS load. Resistive loads would be so large that they'd dominate the noise floor.

Tom

 

imo, for headphones, tubes are better, but when it comes to speakers, it gets much more complicated. Tubes with many speakers are at a disadvantage.

while I understand people saying ultimately they prefer tubes, imo when you get into real high end SS, both are very good and I can live with both. Reviewer of 6 moon, Srajan, went from the best of SET amps to a pass lab SIT SS for example.

Problem with SS is that many go to SS to save money, but fact is, to build a good no compromised SS will cost almost as much as a good tube amp. therefore, many people opinion of SS is based on compromised SS design. when it comes to Pass labs or Jeff rowland SS, you will struggle to find a better (tube) amp.

 

I'd like to see some references for your claims, in particular that to build a good solid state amp you need to spend as much as you do to build a good tube amp. You can build my Modulus-86 for $400, chassis included. The only commercially available amp I'm able to find that meets the performance of the Modulus-86 is the Benchmark ABH-2 ($4k). You won't find a tube amp with performance coming even close to that of the Modulus-86. The distortion in the OPT is enough to blow it out of the game.

Now, that doesn't mean that you can't have a personal preference for a tube amp. That's your choice. But in terms of technical performance, semiconductors win every time (assuming the designer has a clue, but then that's the case for tube amps too).

I'm not impressed by the designs by Rowland and Pass. Rowland's enclosures milled from solid aluminum are pure marketing and incredibly wasteful. His LM3886-based amp adds no fancy tricks. It's just many LM3886es in parallel. Pass has some really interesting ideas and I like his minimalistic approach, but the performance isn't there. Each have their followers. I'm not one of them.

Tom