28 August, 2009

Extended Audition: The BIG 300B Tube and Amp Report

Overview: The BIG 300B Tube and Amp Report

First, an overview, and then a 7-part listening chronicle. My listening preference has shifted slightly over time with regards to certain tubes, but these articles written in 2001 still ring mostly true. Later I shall post my current feelings about even more 300B tubes. Almost 10 years later, I'm still listening to 300B, but things have changed a little.

Comparison of 300B tubes: 2001 Overview
(Revised from thread posted here, on 05/07/01)

The current production 300B tubes illustrate one basic point: that competition by manufacturers fuels progress. Let's hope the competition does not proceed to cut-throat level which will surely provoke a decline in quality, reversing the trend of product improvement we have enjoyed.

When I started with the 300B tube several years ago, Chinese and Cetron were almost the exclusive game. Then we got Sovtek, Svetlana, Tesla, re-issue WE, VV (defunct), KR, AVVT, and now Tianjin. Not only that, manufacturers, particularly Chinese ones, have come up with
commendable variations and, more importantly, improvements. The current state is an embarassment of riches.

No 300B test can have ALL 300Bs. I believe our several articles have the LARGEST variety of tubes available in any test, and we have tested the tubes in more amplifiers and configurations. Most of us, especially me and Andy have lived with a variety of tubes for many years and have intimate knowledge of our own, as well as each others' systems. I have not tested the Sovtek this round as a friend recently destroyed 2 filaments of one of mine and the other is orphaned (had NEVER happened to me; just don't pull on those things when hot) but that is an excellent tube for the price. I also had run a pair of Cetron for a long time many years ago. It was smooth but
lacked some resolution and air and I preferred the Sovtek in the AN kit one, at that time my only 300B amp.

No test can seriously A/B all tubes in one setup. Indeed it is my personal belief such A/B test is non-optimal and ultimately not helpful. I do believe there are no best tubes, and those serious about sound know things are system dependent. Our tests offer an interesting variety of amps and operating conditions, and carefully readers will even note that I repudiated myself from time to time. Such is the nature of things.

We should be humble in scrutinizing our own opinions. We should be proud that we enjoy the music so much and that we lsiten to such a large variety of serious music. Indeed it was a joy to have a teenager bringing his so-so recording (sound-wise) of Shostakovich largo to test out tube gears! And yes it was sometimes a chore to have to wait for the violin to emerge from Beethoven's long lead-in. But all was ultimately joy. Yes, the joy was in the music and 300B is the great messenger, better than most over-rated tubes.

I enjoyed most of the tubes. There's plenty of room for "error" in the views, but I do believe NO ONE is immune. If I am forced to make a summary, here it is:

Original WE, Tianjin meshplate, Svetlana
These are my absolute favorites. The TJ mesh is surprisingly similar to the WE. I do not agree with those saying it has a different sound. As a matter of fact the TJ is closer sounding to NOS WE than the re-issue WE. The Svetlana has a different sound, with some of the characters of other tubes, but it is a good boogie tube that keeps the refinement. I would keep it in this august company and do not regard it as inferior to TJ mesh, at the same price point too.

Sovtek, Valve Art (both versions), non-mesh Tianjin (ST and globe)
These are basic tubes with balanced virtues but a somewhat over-excited top, but used correctly would give a lot of pleasure. The best starter tubes for the budget minded.

Tesla, KR, AVVT, re-issue WE
These share some interesting characteristics that have NOTHING to do with the original WE 300B. Our listening favors KR which may be the best sounding tube in this group, with Tesla last. I have heard good performance of re-issue WE300B in 1 system, so care is needed. And it
cannot be emphasized enough that run-in is perhaps most important for this group and evaluation likely to be most erroneous. Problem is, people will only run in very expensive tubes like re-issue WE...

Very warm and suitable to lean and over-revealing systems.

Evaluation is not meant to be inflammatory. Your mileage may vary, even greatly. With due respect, reading Thorsten's and the old TNT and Stereophile reviews, I agree with many points and disagree with as many. Your inputs are MOST welcome.

Lastly, we have no ties with ANY company, though some testing was done at Opera.

Listening Log
1. Part 1.
(Revised from thread posted here, on 22/06/01)

I was lucky that I got hold of a pair of Tianjin (Fullmusic) meshplate 300B to play for a while. Preliminary listening is very encouraging. As personally I always thought 3-note shootouts to be not desirable, I decided, what the hell, I'm going to seriously evaluate these, and in the process, test again all my 300B amps and 300B tubes. This is going to take many parts.

As summer has fully pressed upon us, I have gone into my small room for daily listening. This 80 ft room has the quieter and smaller (energy-saving) air conditioner. Following is the stuff I used for testing, which may change as test progresses, as the aim is to bring out the best in each.

-Micromega Drive 1/VdH Optocoupler/Micromega DAC.

-None used for amps with volume knob (Advance and AN kit one)(CD direct in, passive attenuation).
-For pure amps (AES-1) I use either the Audio Synthesis Passion V passive preamp, or if necessary, the Fisher 400C.

-Taiwan military teflon silver cables into amp.
-When preamp is used, VdH Second from CD to preamp


-AES-1 (Cary) 300B kit, original version, but upgraded with a load of boutique parts like Hexfreds, Cardas and Tiffany posts, Kimber caps and a pair of AN copper-in-foil caps. ss rectification. Fixed bias, with one adjustment for total of both channels. 1x 6SL7 and 1x 300B
per channel. Ken Rad silver plate 6SL7 used. This amp is highly revealing but a bit lean and un-involving, so a challenge. More details here.

-Advance 300B kit. This Japanese kit is one of my favorites. 1x 6AN8 (don't tell that to Joe Rosen) and 1x 300B per channel. Sylvania 6AN8 used. Stock except for Black Cat couplings caps. Meter and 2 pots bias the 2 tubes individually. Fine, warm sound, no doubt partly attributable to some NFB (no-no in some people's book, but not mine). Not as powerful, but nice.

-Audio Note kit one. On loan to a friend but soon to be back. Stock. 5U4 rectification. 1x 300B per channel, together driven by 1x 6SN7 and 2x 5687 (in SRPP). Will use RCA smoke glass VT231, Sylvania JAN 5687WB. This IMHO is THE most powerful 300B amp.

-TAD 300 used mainly, placed close to wall.
-Spendor original SP1 for comparison.

-For TAD, original simple twisted stranded copper.
-Goertz cheapest copper for comparison.

I'm going to spend a lot of time listening to various CDs, but for each combination I'll listen to a few fixed CDs:

-Brahms piano concerto #2, Katchen/LSO/Ferenscik (Decca)(Superb sound in bargain 2fer, sounding like Kingsway/Wilkinson type of thing, but uncredited, a recent favorite)
-Burmester CD II, various tracks, Ricky Lee Jones, Inbal/Weiner Symphoniler/Shostakovich #9, Von Otter sings Handel.
-Bach partita #2, chaconne, Menuhin (EMI), mono (sound good but difficult to make appealing due to Meunuin's style; a challenge)
-Teresa Tang (Polydor)

First 2 nights I played with both the AES and the Advance. With 2 amps one has the advantage that one can insert tubes in the idle amp (it is NOT advisable to change tubes hot, sure to kill weak filaments). I started with the Advance amp. For 300B I started with the long familiar Valve Art (VA), then Tianjin (TJ).

I quickly determined that using the preamp was not beneficial to the Advance, which already has volume knobs. With CD direct in. Sound into TAD was a good match. The TAD has ultra-analytical power. You hear SO much it is almost annoying sometimes. Fingers striking on the board, pages turning. The Advance smooths this out a bit. Without the preamp the Advance had the more natural flow. With loud passages the congestion was still a little annoying with the VA. This disappeared almost completely with the TJ. One immediately noticed that things are
a bit rounder at the leading edges. But importantly the scale was preserved (no mean feat to improve smoothness of attack and preserve scale), and the energy and presence of the Valve Art was not lost either. At this point I listened to several hours of CDs and I had absolutely no fatigue and enjoyed all the discs. No doubt the TJ is great for long term listening.

The TJ also needed a slightly longer warm-up to reach steady state. Within one minute it will clip. One tube has slightly dimmer outer filaments, but sound was unaffected.

Moving on to the AES proved a lot more troublesome. To be honest when I got this amp, with original Sovtek 6SL7 and Sovtek 300B I played it with the TAD. I was astonished by the brutally revealing nature of it, but I was not happy as it did not have the boogie factor. Perhaps it was "upgraded" too much that musicality vaporized. Recently I returned from the USA and brought back some Ken Rad silver plate 6SL7, the best in the business (yes, those UK "equivalents" are very bad). OK, with the KR 6SL7 on board and VA 300B the sound was quite a bit more refined, indeed almost musical. I started using the Audio Synthesis passive, and had to advance the volume to 3 o'clock to have good sound pressure, so the input sensitivity is lowish. But there still was considerably more grittiness than using the Advance amp. The lean nature had some advantage. Piano was better rendered, with the resonances clear, like the piano top had been opened. Clearly there were some good things but the AES must have particular attention paid
to matching. The Goertz fleshed out things a bit but lost the improvement in the piano. Switching to the SP1 mellowed out the sound quite a bit. Really more musical but the slight veiling and loss of details were palpable and difficult to accept after TAD. Switching to the Fisher preamp, volume around 12 o'clock, the effect was similar to the change to SP1. I decided to stick to the AD, original cables, and Audio Synthesis for the AES.

And then I changed to TJ in the AES. Predictably the leanness was ameliorated, not completely but by half. Things were still sharper than the Advance but the pro's and con's now became more even, and listenable too. Things were really quite fine except massed strings which still were a bit too thin, but this may be the TAD's Archilles' heel anyway. The Burmester CD, Teresa, all beautifully rendered. The only area in which the VA was not bettered was piano tone, with the TJ a bit fat and muddy there. Playing Menuhin again just confirmed that the playing was still not appealing. TJ could not do that, so was not over-decorating the music (IMHO Menuhin needed that).

One thing I need to mention. This used pair of TJ 300B biased quite a bit differently than my other 300Bs. I had to turn the bias pot quite a bit more clockwise (a few hours) on BOTH Advance and AES. The difference is greater than any of the 300Bs, usually within 1-2h. That
probably indicated the conductance characteristic is possibly quite different, or possibly aging effect. The 2 tubes also did not bias quite the same in my Advance, indicating not quite a matched pair. This has implication. As the AES can only bias 2 tubes together, that means for unmatched tubes it was wise to bias low for safety. I adjusted the AES to be 100-105 mA. Ideally it should be 130 mA for a matched pair, meaning 65 mA each tube. In this
situation who knows I may still be getting 60 mA in 1 tube and 40 mA in the other, with the total 100mA (though I doubt the difference is that much). So I played safe. The bias held rock steady in both amps though, with no nasty fluctuation, and that was nice.

2. Part 2.
(Revised from thread posted here, on 26/06/01)

Over this weekend I finally had a bit of time, and the weather was not THAT hot. So I took the AES amp out to the living room and hooked it up to the La Scala. Totally different setup:

-either the Quad 67 cdp or PS Audio Lambda/Illuminati/Monarchy 18B. The former is very musical but the latter is more hifi, with excellent resolution.

-Wai Lee. 1/2 6SN7 transformer output. Using generic GE 6SN7. I used an old RCA 5Y3 rectifier.

-AES-1, with KR silver plate 6SL7 and TJ meshplate 300B.

-MIT Terminator 3 interconnect throughout. Cloth WE (thin red) for speaker cables.

-Beethoven Symphony #9. Krips/LSO (Vanguard)
-ditto. Zinman/Zurich Tonhalle (Arte Nova)
-ditto. Wand/NDR orchestra (RCA)
-Chevalier de Saint Georges, violin concerti, Nishizaki/Cologne chamber orchestra/Muller-Bruhl (Naxos)
-Bruckner Symphony #3. Bohm/VPO (Decca)

This marked the first time I listened to the La Scala in months. Ever since I got the lowther I concentrated on the lowther to make it sound better and better. That was NOT because I preferred the Lowther. On the contrary, I prefer the La Scala, but it was interesting working on
the Lowther.

I was literally floored listening to the La Scala again. Unparalleled presence and rhythm and timing, and just a big wallop of sound. How delightful.

With the admittedly milder cabling in the living room, sound was eminently musical. As a matter of fact, The AES with TJ mesh sounded great! I now have the additional query of whether I had under-estimated the AES, or if the TJ made things much more musical, or both. But that is for later.

There is no doubt that the qualities observed of this TJ meshplate before (vs VA) shone even more in the large room. The smooth leading edge is very much in evidence. I put on the Vanguard Beethoven (have never listened before, excellent performance) and was very pleased by the excellent scale and smoothness. What REALLY astounded me was the very defined vocal image. This was much sharper than I remembered with the La Scala. To confirm I played the other 2 versions of the 9th that I have played many times before. No question, the sense of scale and definition had better subtle gradation than before. A delight in the last movement of the 9th because of its large and varied forces. Similarly, the Vienna brass in Bruckner (LOTS of them) howled with fabulous full tone, surely a reason why we use horns.

After the Beethoven and Bruckner, I changed the source from Quad to my just-returned PS Audio/Monarchy combo, and played the violin concerti. This front end can sound a trifle hard with the violins but not this time. Beautifully clear and involving violin.

I would think all this due to the TJ mesh, but I shall change tubes later in the AES to confirm. But so far so good.

3. Part 3.
(Revised from thread posted here, on 27/06/01)

Thanks to the seemingly non-ending rainstorms I got to spend a surprising amount of time at home. So last night, after a pint or 2 (none for the road, :-)), and after cooking myself a vegetarian dinner, I continued the 300B test. Setup exactly the same as Part 2. I refreshed myself with the sound of TJ meshplate. Then I tried to change to Tesla.

What a bummer! The base of the Tesla was a little too big for the depressed socket, so this tube CANNOT be used in the AES. Will have to do this with another amp.

What then? OK, my favorite Svetlana then! So I installed them. This is a matched pair but I decided to bias them at the same point used for the TJ mesh (105 mA for 2 channels). Fair?

What do you know! Sound is really quite different from the TJ meshplate. There is no doubt that, despite the smooth leading edge and excellent scale of the TJ mesh, these are surpassed by the Svetlana. The Svet has the denser texture and better control. And though the leading edge is a tad wirier than the TJ mesh it does not detract from the overall dynamic superiority. The boogie factor is in full force. But most of all, the Svet plays louder passages more composed. With the TJ mesh I haven't previously mentioned that with large volume the sound still whites out and flattens out a bit (though already superior to VA). This does not happen with the Svet. The Svet/AES kept its composure with loud passages, with excellent instrumental separation
and bass texture. There is something else difficult to describe. The Svet simply has superior 3D feel, imparting the feeling of greater texture.

The performance of the AES is more than a little unexpected. As a matter of fact, it is on full par with my other amps. I should retract some of my less kind words on AES before. As used here, it has excellent timbre rendition and instrumental outline. Imaging is superior.

I suspect the results will vary with the bias. At 105 mA this is lowish for 300B, but I think I like it very much, and shall continue testing the other tubes in AES at this bias point. Perhaps a little
lowish but there may be something to gain too. If I do get another pair of matched TJ mesh I shall try a higher bias on both of these tubes again.

4. Part 4.
(Revised from thread posted here, on 28/06/01)

Thanks again to the cool weather I was able to continue testing in my big room.

I substituted the Audio Note Kit One for the AES, and inserted the Sevtlana ("autobias"). Contrary to my usual practice of direct in, I kept the preamp to make a comparison with the AES. Sylvania 5U4GB, NU black glass 6SN7, Tungsol 5687.

Additional music played: Buena Vista Social club's Ibrahim Ferrer and Ruben Gonzalez CDs (World), and Sibelius symphonies #4, 2, and 3, Babirolli/Halle (EMI) (yes, 3 in one evening).

The difference of the 2 amps are clear enough. The kit one remains the more powerful amp, handling any climaxes with more headroom. It brings even further smoothness and refinement to the Svetlana, removing the (very) little bit of sizzle in the top (when in the AES). It is
noticeably slower than the AES though. The separation of instruments is a tad (but just) not as crisp as the AES, but dynamic gradations and turns are even more graceful. Only when the quavers get really complicated, as in symphony #3, did I miss a little more separation and a faster speed.

So another splendid outing. I would like to try next the Tesla, Sovek and KR. And if I get another pair of TJ meshplate I would love to test it in the 450V operation of kit one.

5. Part 5;
(Revised from thread posted here, on 29/06/01)

Last night I took the big plunge. And what a pleasant surprise!

Eager to hear more of the TJ mesh, I threw caution out of the window and inserted them in my autobias AN Kit One. This is having them work at considerably higher disspation than the AES SE-1. And HOW beautiful they sounded!

I had to listen really hard to compare the performance with the Svetlana. Execept ONE thing, which is soon noticeable. The TJ mesh sounded faster than the Svetlana in the Kit One. This is without sacrificing ANY of the smooth leading edge.

And unlike in the AES, the TJ mesh yielded little to the Svetlana. If anything it was a wee bit smoother in the upper midrange. It was fully dynamic, with none of the whitening-out noted in the AES. The Svetlana still have the upper hand in density of texture. The TJ meshplate was a shade lightweight in comparison, but that, together with the faster speed, had the occasional advantage of sounding more buoyant and airy, especially with jazz cymbals and percussion, though this was by no means always evident with other music. As I said you have to listen real hard to tell. With bopping music it was also apparent that the boogie factor is a bit below Svetlana.

Additional software: City of Sadness, soundtrack (Flying Saucer). Shostakovich Preludes and Fugues, Scherbakov (Naxos).

With this superior performance, it appeared that TJ mesh preferred to be run at higher dissipation. It would be interesting to run it in the AES at 130-140 mA if I have a real matched pair. We shall find out more about it when I take it to Andy's rig this weekend. His DIY 300B
amp runs at only 300V or so!!!!

6. Part 6.
(Revised from thread posted here, on 03/07/01)

Thanks to the bad weather and the successive long weekends, I managed to get a lot of work done this past weekend (due to rain, no laundry though, :-()! System stayed the same.

Additional music played: Mozart symphony #40. Wagner, Celibidache/Munich (EMI).

I need to clarify one fact. I just confirmed the AN kit one actually works at ~380V and not 450V. So nothing unusual.

I inserted the Valve Art into the kit one. One immediately noticed coarsening of the treble. With warm up this improved but was never on the same level of TJ. Indeed the whole texture was slightly coarser than both the Svet and TJ. The ability to play loud as well as the scale were both slightly restricted. It was faster than the Svetlana but did not have quite the same boogie factor. But it was an even performer that had no overly weak areas. Unlike...

Then I tried out the Tesla, which was my least run-in pair, with probably <80 style="font-weight: bold; color: rgb(255, 0, 0);">7. Part 7.
(Revised from thread posted here, on 05/07/01)

OK, write this part before the impending typhoon reaches me...Has anyone read "Man in the holocene"?

Sunday night and Monday I took the TJ mesh to 2 persons' homes and tested it out. This is possibly the most important part of this mammoth series.

First system (Andy's):

Marantz CD7/Michell Orbe/SMEV/Benz Ruby/Benz phonoamp/DIY 6SN7 transformer oputput preamp (using RCA 5692)/DIY amp using 5U4 rectification, 1x 12AX7 (Forgot which expensive one) and 1x 300B per channel [Note: these have unusual operating conditions as it was really an 2A3 amp with changed filament; Opearting at 240V and 70 mA]. Into Tannoy Canterbury HE (so this can approximate Thorsten's setup a bit more than mine). Tara labs Decade and Air cabling.

CDs used: Burmester CDII; Mozart piano concerto #23, Bilson/Gardner (DG). LP used: Paganini violin concerto #1, Rabin (Classics re-issue).

First up was REAL NOS WE, 1951 matched pair. These have been used maybe 50 hrs? For those with long memories, initial sounding was good (better than new re-issue WE when new) but not quite beautiful nor relaxing as my very old pair. Sound now was excellent, indeed best
sound Andy had obtained with this amp (yes, better than this amp with RCA 2A3 which we found ultimately inadequate for Tannoy). It sounded rich, with no artificial forcing of any passages, and indeed remarkable with LP (yes, better than CD of course). Looked like TJ would be in trouble.

So we swapped the TJ in. Surprise!!! Excellent and on a par! Much of the smoothness was comparable. Indeed the TJ was smoother in some passages though violin had a wee bit of edge that yielded slightly to the WE. The TJ sounded a bit faster and more involving and you could
listen at a lower volume. The TJ was a wee bit muted at top compared to the WE. But the astonishing thing was that they sounded mostly ALIKE! How about that! For Andy, as TJ was not really better, he would continue using the WE (lucky him), and for me I see no reason not to
go with TJ. Everyone was happy.

Then we tried out the KR 300B. We expected maybe some disappointment? But surprise, it performed rather well! It had the denser texture mentioned about Tesla and re-issue WE (NOT present in the NOS WE300B) but it was more balanced, with reasonable rendering of harmonics and less shut-in quality. Bass was tight but good. Scale and rhythm and pace were good. A very clean sound. Indeed of this type of more uptight sound we both have found KR better. With some more playing (these were 30 hrs?) maybe a bright future ahead? Not quite as
beautiful as WE and TJ, but a dignified performance.

And thanks Andy and his wife for an excellent dinner. Much appreciated.

Second system:
Sony ?B940 SACD/CD player (this is the cheap but top-rated one by UK mags)/Audion Silver Night (Amperex Holland 6DJ8, Tungsol 5687) into Lowther Acousta (Opera Audio's version). Mostly DIY cabling.

SACD software: Vivaldi, Four seasons, Stern/Israel chamber (Sony); Kunzel/Cincinati pops DSD. Supplemented by CDs: Beethoven, violin concerto, Perlman (EMI); Shostakovich symphony #5 and Tchaikovsky symphony #6. Mravinsky (Victor); miscellaneous vocal recordings.

Paradoxically we (friend of host and his classical music fanaftic teenage son [who brought the Mravinsky's], and I) found the SACD player to be quite shrill with SACD, and the on deck DAC was not as good the digital out connected to a Musical Fidelity X24k DAC, which we ended up using except for the SACDs.

First up was the host's newly bought WE re-issue 300B. These are brand new and <15 hrs run-in, so take these with a grain of salt. The sound was a bit like the Tesla sound described last time. At small volume, good and dense mid-range texture which is however somewhat shut-in,
with shortchanging of harmonics and short decay (not a good thing for Lowther), thus accentuating the upper treble. Scale was somewhat compressed and bass was one-note and so-so. With louder volume there was shouting. As I know the Audion quite well, I estimate the rhythm and timing to be a bit better than Tesla though. These WE need much run in but some of the sonic characteristics is just as I remembered.

The TJ mesh were substituted and, BINGO, all problems were gone. Music smoothed out, regained scale and dynamics, with better and elastic bass. Air and harmonics returned. What more could we say! Unanimous decision.

Then the crowd drove to Andy's place and enjoyed his system (with 1951 WE 300B) and he played a few LPs. Viva analogue: Beethoven, piano concerto #5 Bishop/Davis (Philips); Rimsky-Korsakov, Scheherazade, Reiner/Chicago (Classics); Wagner, ride of the Valkyries, Boult/LPO (EMI).

With this part I think I shall draw a line and finish the 300B series. Indeed I had tried my best to try out various tubes in various amps and places. Not all combinations were possible (nor deserving). Together with the 300B tube and amps summaries, which I shall write soon, these will form the inaugural 300B issue for the magazine, and hopefully people will find it useful. No meaningless 3-note shootouts here.

I shall continue experimenting with other combo's, especially the tubes that did not do well (this is what I really like rather than proclaiming something as the best). E.g., I had put in the Tesla in my Advance and the results were promising, and I will run the Tesla in some more (and the re-issue WE 300B owner will too for sure), so more fun are coming.

No comments:

Post a Comment