There is no test that correlates with sound quality except you want transconductance to be well within the good range and the two sections of a 6SN7 or any dual triode to be about equal in transconductance, but for two tubes that test well, no measurement is going to tell you that one sounds better than the other. Furthermore, with octal triodes you need to be aware of microphonics. Some are very sensitive and THAT can color the sound.
Do vacuum tube test values reflect sound quality?
I recently ordered a pair of output tubes to try in my LTA MicroZOTL preamp. Two Sylvania 6SN7WGT tubes, Gold Brand. I’ve had good luck with Sylvania 12sn7gt bad boy tubes, so I thought I’d give them a try.
After I bought them, I saw a similar pair for more money that had slightly higher test values. Does that mean they might sound better or that they just have more life left? What properties do you look for in tubes?
Below are the values of the tubes that I bought.
Tubes are matched for dynamic gain and structure, testing at 3000-2750 and 3000-2750 where pass/fail is 1625 micromho. They have the same structure inside and out, having matching date codes of "922", the 22nd week of 1959
37 responses Add your response
@lewm Thank you. So, would it be better if the 3000 and 2750 values were more closely matched? The two tubes match each other closely which I assume is good. |
Have you heard differences between tube brands thru THAT unit? I would listen to the ones I bought, AFTER I break them in for 60 hours by simply leaving the unit on, no signal required, and if they sound great, expect many years of life. IF they don't sound great, try a different tube not the same one, no matter what it measures. Sound good, but not better than prior good ones, there's your 'never without music spares', find some that you prefer some day. Keep in mind, each time we push pins in a socket; rotate a dial; move a switch on a tube tester, it's similar but can be slightly/perceptibly different; and published numbers from two separate testers are not scientifically exact. If you had your own tester, then you could test them when received, and every 6 months, keep a chart, have confidence in your tubes when problems occur. |
@elliottbnewcombjr I've tried several different input tubes as well as output tubes. I can hear noticeable differences. I'm pretty set on Telefunken ECC801S for input tubes. Siemens ECC801S being a close second. I hadn't tried very many different output tubes until recently. I now have a couple of pairs that work well. I just thought I'd try something different for grins. |
These days, for the exotic types and brands you mention (Siemens and TFK), I’d hazard a guess that there are more fakes than real ones on the market. I’d stay away from eBay as a source for such tubes. Most reputable dealers don’t pretend to have such tubes to sell. I assume you know that “ECC801”indicates a selected ECC81, even more rare than the run of the mill version of Siemens or TFK. What few exotic small signal triodes that I own I bought in the 1970s, so I know they are for real at least. However, maybe you are an expert in examining these tubes to prove to yourself that they are what they say they are. I certainly am not, even after 45 years in this hobby. To the question you did ask, yes I think 3000 and 2750 are sufficiently well matched sections. |
As lewn stated, one cannot tell by numbers what a tube will sound like. The sound will be determined by the particular amp and the particular brand and construction of the tube for any given type. Depending on the application, a tube can test quite poor, but can sound perfectly good in an amp that does not push the tube very hard. It may also sound like a sibling tube even if the two test very differently. I notice you are using very expensive Telefunken ECC801S input tubes. I hope you got it from reliable sources because the Chinese have become quite good at making fakes, including fakes with the raised Telefunken diamond in the bottom of the glass envelope. I got my pair for my phono stage a long time ago, and even then I worried about fakes. The boxes they came in looked very clean and bright, which worried me until I tried to open them and they started to fall apart from age. That was a good sign to me. |
@lewm @larryi Thank you for the replies. The Telefunken came with my LTA which I bought used. The seller seems to be knowledgeable about tubes. He paid $375 for the Telefunken tubes. He's a straight shooter, so I believe that to be correct. But, as far as my knowledge, I'm no expert with tubes. I'm learning. I buy from sellers that have lots of feedback and specialize in tubes. I guess I rely on their expertise and honesty.
|
There are actually a number of test values that can be measured with a good quality tester. Although I am not sure any have sound quality correlations.
https://www.flickr.com/photos/jdflood/54409642846/in/dateposted-public/ |
The only thing a tube tester can tell you is whether a tube is worn out or nearly so. That you can determine by comparing the manufacturer data to the tube tester result. But the red flag is transconductance. As tubes wear, transconductance goes down, for any given bias and plate voltage. Consult an RCA tube manual or any of several websites to download the pdf file for the tube under investigation. These are very old data from manufacturers that no longer exist or no longer make tubes, but the numbers are still valid ways to evaluate tube condition. If the measured Gm (transconductance) is below the range of values given for a given bias and plate voltage, then the tube is worn or wearing out. Yet, as Larry said, that tube may still sound "good". Better test is to replace such a tube with a new one and then listen. If you hear improvement in terms of a livelier sound or better treble and bass response or a sense of better gain character, dump the worn tube or save it for a spare. The Chinese-made fake ECC81/12AT7 often are labeled as "801S", just to make them look even more desirable. On the other hand, these fakes often sound fantastic. I compared about 6 different NOS 12AT7s in the input stage of my Atmasphere amps, many years ago. Other than the Chinese made 12AT7s, all the other tubes were true Euro or US NOS brands such as Amperex, Sylvania, Mullard, GE, etc. The Chinese tubes sounded by far the best. I bought them from Billington in the UK, and they were openly sold as Chinese tubes under the label "Billington Gold" (they had gold pins); there was no attempt to mislabel them as TFK or anything like that. The downside of those Chinese 12AT7s was only that they had a short lifespan compared to true NOS tubes. I used them anyway for a while.
|
You really cannot tell anything about sound quality from test numbers. In fact, test numbers can be misleading. For example. if you test two tubes of the same type and one has a higher transconductance than the other, it is assumed that tube will have more gain. But only if the plate resistances are the same, otherwise the tube with the higher gain will be the one with the highest combined Gm and plate resistance. To further complicate things, if two different tube tester are used... do they both have the same voltage parameters for plate, screen and grid to cathode? Your best bet is to purchase tubes from reliable sellers such as Vintage Tube Services or Upscale audio where they test their stock against each other with the same testing equipment. This way you can compare tube readings among the tubes they sell. Getting different readings from different sellers -- especially eBay -- is futile IMO. |
My experience is nope…testing or matching seems to have minimal impact on SQ especially with preamp tubes.Triodes that are unbalanced sound just as good as long as one side tests strong. Matching and gain balance on output tubes does matter more. I have a tube tester that is very helpful for comparing tubes and validating whether a tube is good. Just my humble experience on my system… |
I do hope you scored genuine Telefunken 803S tubes. These days, they sell for up to $2,000 a pair. The fakers have become increasingly sophisticated. They use to offer their fakes at a price so low as to tempt even the wary. These days, many of the fakes are being sold at the going rate of genuine articles which makes it harder to distinguish real and fake offers. With things like rare Western Electric transformers, the fakes use to be in too good condition visually to be realistic. These days, the outer cases are rusted a bit to look old, but often the Western Electric decal is left in nice condition to tempt collectors. A local builder who buys such parts on line prefers to buy such parts from sellers from Japan. |
"Triodes that are unbalanced sound just as good as long as one side tests strong." That’s true, IF the gain circuit is single-ended as opposed to balanced, and if the section with high transconductance is the one you are using. This is often how tubes become unbalanced; they have been used in a single-ended circuit which only requires one of the two sections where the other section (of a dual section triode) is not doing anything. Some canny manufacturers of SE preamps build their units so that one channel uses section 1 of a dual triode, and the other channel uses only section 2 of the same type dual triode. (Two tubes, one for each channel.) That way, when the tubes become worn, you can swap them from one side to the other of the stereo circuit and in essence you have new tubes. Balanced sections is much more important for a balanced circuit. Also, transconductance is not necessarily a correlate of "gain". Depends on the circuit and the tube. For example, a 12AX7 has highest gain among typical dual triodes, It is commonly used at the input of MM phono stages, because it has highest mu. Mu is equal to Gm*RP (transconductance X plate resistance), and the 12AX7 has very high plate resistance but relatively low Gm. Mu is a better correlate of gain. Gm is more a correlate of how the tube handles current. |
@larryi They are ECC 801S. Clearly labeled with the Telefunken logo. I assume they are legit, but they aren't the $2000 pair that you mentioned. |
There are instances where someone gets tubes for well below market price, so it is entirely possible for you to have real Telefunkens. If they are not new in box, the price could be substantially lower than the price I mentioned. The good news is that these tubes have a very long life, so even old tubes can be good for a loong time. if you see the Telefunken diamond embossed in the glass on the bottom of the tube, that is a good sign; only the most sophisticated fakes have that embossed pattern. |
@lewm +1 |
@larryi You made me curious, so I pulled one. There is a diamond on the glass bottom. |
@lewm one section of a dual triode could be used for the gain stage and the other section could be used as a cathode follower in a single ended preamp. So both sections still being used. |
How a tube sounds in any given amplifier, especially a preamp level tube, highly depends on how much local feedback is used on the tube in that specific circuit. Generally, high gain stages, like the first stage of a phono preamp, use little if any feedback on the tube. In a cathode follower circuit, there is nearly 100% feedback giving little if any gain, but a much lower output impedance which, of course, is the reason for its existence. In other words, high gain stages are particularly fussy about the quality of a tube whereas low or no gain stages with lots of feedback like cathode followers are pretty much agnostic about the providence of the tube. |
You can notice (bad measurements == bad sound) when something’s really wrong:
If you want to compare NOS or extremely strong testing tubes versus much less strong testing tubes - the latter will (probably) not last as long as the former in service, but will generally sound just as good (if not better), and may still last a long time. If your matchings (triodes, tubes) are within say 10% or better, don’t worry about it beyond that. In the past I’ve bought super premium matched / tested / noise-selected / cryo / curve-matched / brawndo enhanced / etc tubes before, and found NO difference to just decently-testing tubes. Yes it can be dependent on the component & slot you’re plugging them into - but usually not that much. I’ve never needed better than Upscal’es "Platinum Grade" screening and testings service. I also like to buy new production power tubes that have been burned in a bit before matching (to screen out early failures). |
Tubes have a huge effect on SQ, in every level of the circuit. Tubes sound wonderful and buttery. |
For a cathode follower you WANT high transconductance. Tubes like the 12AX7, used by many companies as a cathode follower, are not a good choice. They do it because there are billions of 12AX7s in the world, and more are being made in Eastern Europe. The CF changes voltage to current. 12AX7s are puny for current. |
@donavabdear LOL. I'm probably close to 30. I have ten 6922 tubes that I bought for a c-j preamp that ended up having issues and I returned it. Found a set that liked (four + 1 extra), so I bought five more. I don't have anything that uses 6922 tubes. |
@sls883
|
In a well designed circuit, tubes will last quite a long time, certainly many years, sometimes more than a decade. One of the few good uses for a tube tester (and I specify a tester that measures transconductance in units rather than as “good” or “bad”) is to record the values for Gm when new and then retest the tubes periodically if you want to ameliorate your anxieties about tube aging. Hickok made the best testers and they all can measure Gm, for example. Should one prefer SS because of tubes aging? With SS you get sudden total failures where the discrete transistor or integrated circuit that failed is no longer available. Nor is there anyone left with expertise to do the repair. With tubes you just plug in a new one. |
Many did it because the high µ was needed and the other half of the tube was right there without a good use so it was turned into a CF to buffer the high Rp that comes part and parcel with the high µ. The marantz 7C uses three AX7s per channel. 4 sections are used for gain and two are cathode followers. It would be a mess if a more suitable tube were used for the CFs dave |
"A mess" meaning there would have to be at least one or two more tubes, preferably of high transconductance to use as CFs. But could you then delete one of the 12AX7s and end up with the same number of tubes? While I recognize the importance of the 7C in audio history and its high sonic quality for its day (and its collector value), I don't think it measures up well against any. of the best phonolinepreamps of the past 25 years purely in terms of sound quality. Not that I can prove that the 12AX7 it uses as a CF is an audible weak link. More likely the power supply and parts quality hold it back, in the sense that we have much better capacitors and resistors of all kinds available these days. Also, the one and only time I ever saw a schematic, it seemed overly complex, but I cannot recall specifics. |
Most tube testers don't test tubes at normal in-use voltages, and that's especially the case with output tubes. Other than noise associated with short circuits or excessive gas, how a tube tests won't really indicate sonic characteristics. That said, I have 4 tube testers because I enjoy the process of testing vintage and/or new production tubes before I deploy them in a vintage amp. Most recently bought a Maxi Matcher which is a little more helpful with matching output tubes, as it will apply 400 volts to the plate. The final test of any tube, whether in a gain stage or output, is the gear in which it's used. |
I get my tubes these days from a guy who drops them off, I listen for a while and then decide if they are the real deal. Hearing is believing. How tubes measure matters, but only as a baseline. You could have tubes that are identical in brand, and measure identically, and still sound different, though not likely better or worse if they are from a reputable provenance. |
