I'm Freaked Out by this Tube comparison


I was comparing two amplifiers - ARC VT100 Mk II & SF Power 2. They both have Sovtek 6922's in the input stage that were manufacturer selected and matched of about the same vintage. Both have 8 new EH 6550's, matched & measured.

On cymbals, brushes etc, the Power 2 had a more distant and smaller presentation that lacked detail. Chesky guitar recordings sounded thinner, had less ring, less body, less resonance. Overall, less involving sound.

I attributed the sonics to the design of the amplifiers. I had a third set of Sovtek 6922's that I put in the Power 2. No noticeable difference.

THEN ------ I decided to take the 6922's out of the VT100 & put them in the Power 2. I couldn't believe my ears -- now the Power 2 took on much, though not all, of the character of the VT100. The cymbals, etc. became more detailed, more present, more resonant. So did the Chesky guitars, and now with more body and resonance. More decay. In some respects, the Power 2 sounds better than the VT100.

Am I crazy??? I can't believe that one set of the Sovteks could be that much better than the other. Or is Audio Research's claimed intensive selection process responsible?

There is no biasing for the 6922's in the Power 2. SF told me to replace tubes, you just plug and play. The output tubes (6550s) were accurately biased in both amps.

What gives here?
kevziek
I don't use 6922's, but I have heard differences like are described here - just not between tubes of the same new manufacture and vintage (some can be noisier, more microphonic, etc., but the basic character is similar). Where I have heard these differences is between new tubes of unlike manufacture or vintage, so I would have to assume that the tubes described above are not really identical, despite outward appearances. I do know this: When I have bought "tested" tubes direct from conrad-johnson (who also claims a high rejection ratio and carefully controlled burn-in), I have found the results less than satisfying, and have had to go back and exchange certain tubes multiple times before giving up on them (and their higher prices) and just doing all the selecting myself by ear. Also, I would think that Sonic Frontiers, who sells tubes through their Parts Connection catalog, would have to have some sort of testing and QA program in place as well, and I for one would be surprised if any of these manufacturers' testing programs were that much better or worse than their peers'. Rather than placing faith in one company's pre-selected tubes other another's, I think it just boils down to the particular tubes in question in any given instance - one just has to listen for one's self, and if you're lucky enough to find a batch that sound better, then that's what you go with that time out. Next time, you may find you have to start the proccess all over again. Now, explain to me again, Why don't transistor users have ever-growing stocks of rejected and stand-by input and output devices cluttering up their closets, or have to pay return shipping charges on the losers, or run all over town searching for fresh ammunition for comparisions, or constantly wonder whether the sound has subtley degraded over time? Oh yeah, I almost forgot - They don't understand what *real music* should sound like. At least they have the time to listen!
I did make a slight error. The Sovteks in the VT100 are not of the same year as the other sets. The other sets are stamped 98-12 & 99-01. The VT100 tubes have no date, but the lettering and internal construction are identical. They are AR sourced, in since amp was manufactured (1997). So, my feeling is that either they are so critically measured & chosen, that they sound better. Or, the manufacture of the tubes varies that much.

I have some older (1996) Sovtek's where the construction looks different. Plus some unmarked ones that I think originated from Kevin Deal. I'll try these and let you know my results.

The amp was made in 1997
Kev- I have been told that Sovtek is just a generic term chosen for export to the West and that there were (are?) many different manufacturers. If that is true, take that and add that to the other variables, and what you report is not at all surprising.
Swampwalker, Sovtek is a brand made for New Sensor Corporation of USA. The owner was in business twice. He came into bad times in the mid 1980's, closed shop, then he started New Sensor several years later. The tubes are made in Russia by, I believe, Reflektor Factory, a large tube manufacturer. I don't believe they were always made there, especially back in the 80's. Someone correct me if I'm wrong.

Back to my tubes, all the 6922's I compared were made by the same factory and are the same design. If you look at them all, they are identical. That's the part that boggles my mind.
Welcome to the world of tubes. A manufacturer makes tubes, to pass electrical tests. A vender, such as Ram, New Sensor, Audio Research, etc., tests tubes, to meet certain electrical parameters. Certain venders, burn in tubes, so that the ones, that will fail early, will fail in burnin and will not wind up, in someones gear. Then they are tested again, and graded. 99% of the time, the customer, is the first to actually listen to music, through the tube. No manufacturer today, comes anywhere near, the consistency of manufacture, that companies had, during the tube making heyday. Different runs, off the same line, on the same day, will sound different. Every tube maker today, makes a great sounding tube. Its just that they have problems, consistently making a great sounding tube. No manufacturer, is going to destroy tubes that test great, just because they don't sound great. A tube that tests perfect, does not neccessarily, have to sound perfect.