[K.C.]

RAM cherry-picking new production is good practice, but I have seen postings here and there by people who have done some rolling that they have preferred NOS in quite a few cases over RAM tubes.

You can find 'some people' who prefer anything old over new and who are willing to post their subjective opinions freely without too much trouble. Without the value of a blind test there's nothing more being reported than one man's opinion.

And if you want to talk about the shameless willingness to over charge then Kevin Deal is a good name to start with and Michael Elliot is a good one to finish with.

[6BQ5]

I'm sure that is happening (people having pre-disposition or a bias towards NOS).

But I also see people who have the pockets to try out many different kinds of tubes (including) RAM, and have no other objective than describing their experiences, positive and/or negative, with RAM tubes, new production and NOS.

One thing I've seen said before on various boards: just because a tube tests well does not mean necessarily that it sounds good/better. Fact or fiction, I don't know, but I've seen that comment often enough to think that perhaps RAM tubes may test very well,  but may not necessarily sound better than some of the NOS varieties fetching premium prices. RAM SLN tubes are not all that cheap either for that matter.

[pbrstreetgang]

Im not getting what you are saying. RAM tubes are just a test procedure and guarantee. RAM doesnt make tubes, the ones RAM offers are Chinese, EH, Sovteks, ect just like everyone else, except picked for tight tolerances. The rest not so great testing ones prolly go to the other New production tube vendors. You dont see any comments on a "better" sound of the RAM tubes. They dont sound better by themselves they sound like Chinese, EH, or other russian tubes because thats what they are. The tangible difference and the universal acclaim of RAM tubes are measured by the users love for the MR amps. Thats the whole point and its put into full effect by its utilization in the circuit Roger masterminds. Im sure RM doesnt use 1% resistors that measure 1.5% and Im sure he tries to get matched mirror channels in his amps- this is were the RAM tubes come in completing the integrity and the actualization of the circuit he is designing.

FWIW a clue that RM does recognize the sound difference in tubes is evident by the suggestion the Shugang KT88s on the tube page are an UPGRADE to the RM200 stockers.

Also you wont ever mistake a superb testing RAM EH6922 as sounding like a 50s or 60s Amperex. I cant imagine a person favoring the 6922- but I guess Ive seen crazier things

[K.C.]

...people who have the pockets to try out many different kinds of tubes (including) RAM, and have no other objective than describing their experiences, positive and/or negative, with RAM tubes, new production and NOS.

But their opinion is still totally subjective.

[6BQ5]

Im not getting what you are saying.

Sorry for the confusion. I think we are saying the same thing.

Also you wont ever mistake a superb testing RAM EH6922 as sounding like a 50s or 60s Amperex. I cant imagine a person favoring the 6922- but I guess Ive seen crazier things

But this statement confuses me. Are you saying someone would favour the Amperex or RAM?

[6BQ5]

But their opinion is still totally subjective.

I guess, such it is. Though I would argue, this is not necessarily bad or wrong if done without self-interest or a bias.

I remember Vacuum Tube Valley magazine once published a rating of various classic tubes such as 6L6GCs and 12AX7s (I think authored by Eric Barbour). Do you take issue with this too?

[pbrstreetgang]

Also you wont ever mistake a superb testing RAM EH6922 as sounding like a 50s or 60s Amperex. I cant imagine a person favoring the 6922- but I guess Ive seen crazier things

But this statement confuses me. Are you saying someone would favour the Amperex or RAM?

Yes 10 out of 10 times.

[K.C.]

I guess, such it is. Though I would argue, this is not necessarily bad or wrong if done without self-interest or a bias.

I remember Vacuum Tube Valley magazine once published a rating of various classic tubes such as 6L6GCs and 12AX7s (I think authored by Eric Barbour). Do you take issue with this too?

Yes, I would take issue because it's a limited perspective and highly subjective opinion.

I currently own a Threshold SS amp, Music Reference RM-9 and RM-10, Quicksilver monos, a Dynaco and Pathos and Counterpoint hybrids.

As an example rolling EL-34s in the Quicksilvers versus the RM-9 would lead to vastly different opinions about the same tubes. Both are respected, great sounding amps, but very different in the degree to which a tube has influence on the sound. The same 6922s rolled in both the Pathos and Counterpoint reveals a much greater influence of the tube on one of the amps as compared to the other. Which amp would you like me to base my opinion about either tube on ?

[pbrstreetgang]

Great points, and I agree.

[6BQ5]

... reveals a much greater influence of the tube on one of the amps as compared to the other. Which amp would you like me to base my opinion about either tube on ?

So let's see if I understand ... Are you saying that a 6922 that sounds good in one amp will not necessarily sound better in another amp relative to perhaps another 6922? And if so, then opinions of tubes will vary amp by amp?

[pbrstreetgang]

I have come across some pres especially that all but negate the tubes in their circuit. Meaning they hardly sound like a tube pre in any sense, I think this is what he is meaning.

[K.C.]

I'm saying that the level of influence a tube has in the sound varies significantly by the amps design and an overall opinion about the tubes sound is highly subjective as a result.

[6BQ5]

OK. I see. Thanks, I think I now understand.

[Roger A. Modjeski]

I am flattered to hear that Kevin has taken up my nomenclature for noise levels. Does he give numbers and limits or just his perception?

Price of NOS tubes appears to be a function not only rarity but of what is "the kind to have". Five years ago it was all about Telefunken and I couldn't find any enthusiasm for Mullard and Bugle Boys. Now it seems Telefunken has taken the back seat and the others have moved up front. Right now a fellow who specializes in supplying tubes to guitar players is cleaning me out of my Tungsram 12AX7s because they are the current favorite though a few years ago no one cared.

I scan eBay now and then and recently I was amused to see that Amperex PQ were bringing big bucks. Too bad I threw so many away when I was developing the Beveridge RM-1, they were unreliable, prone to developing noise in short order and very discouraging all in all.

You are correct, no sane manufacturer is going to fool with NOS. Though I have steady supply of current production tubes there are distinct differences batch to batch from the same maker. Factories are forced to substitute materials from time to time and hope nobody notices. The tubes may even be in spec but I pick it up right away because the testing I do is different and in addition to theirs.

The theoretical noise for the 12AX7 design is less than 1 uV making their average value about 5 dB above theoretical and their max almost 20 db above theoretical. My limits for the SLN are within 2 dB of theoretical (that is you can't find a 12AX7 that is better than 2db quieter. My LN is within 3 dB of the SLN and SG within 6 db of the LN.

Yes, I do the cherry picking for you. In a flat of 100 tubes there might be 5 that are Super Low Noise, there might be none as is the case with the Sylvania/JAN 6DJ8s that go for very little money. Other than being noisy they are great tubes, but you can't use them in any input stage. I went through enough to convince me that any low noise examples would be a miracle. Sylvania made good tubes but made them to perform well for their intended purpose which for the 6DJ8, was the tuner in a color TV. They are fine at high frequencies but have a lot of low (audio band) noise which did not effect the TV tuner. In a way, we are lucky that there are any low noise 6DJ8s.

This is why I developed a computer driven rack to gather and print the data for each section of each tube.   Of course everyone has some "computer driven tube tester" but not for noise. I wouldn't even have attempted it had I not the experience of chasing down noise in the RM-1 preamp and RM-4 headamp where the noise referred to the grid is less than 0.3 uV. To give you some perspective on that I like to use money because we all understand and appreciate money in our everyday lives. Since we are going down to the small change we have to start big. So let's have a volt be 1 million dollars. Your CD player has a typical output of 2 volts so that's $ 2 million. That means a microvolt is $1 and thus my grid noise is 30 cents. Which is a very small amount compared to the $2 million output of the CD player. In the world of vinyl a moving magnet cartridge is typically 5 mV so its voltage output is $ 5,000 while a moving coil will be 0.5 mV or $500. The RM-4 employs 2 sections of a 6DJ8 which reduces the noise by 3 dB thus reducing the grid noise to 0.21 uV or 21 cents, 9 cents better than a single section and very small compared to the $500 output voltage of the MC cartridge. If you like sums more than fractions just ask yourself how much difference is there between $5,000 and $5,000.21? Hardly any, but I have to work pretty hard to get the noise down to 21 cents out of $5,000.

At first I tested these tubes one at a time. The RM-4 was selling well and I had to have tubes for them.
Listening for noise and judging their quietness is not something you can easily teach someone nor will your senses be uniform from day to day. How about the ambient noise? So here is a case where measurements really do make sense and my philosophy is if you are going to charge someone $60 for a SLN tube you ought to measure that tube's noise and if you can measure it let the user have the measurements.

Who would want to go through 100 tubes to find the 5 best and what to do with the rest, but that's my problem, not yours. There are maybe 20 that are LN and there is very little demand for the Standard Grade although I am happy to advise where these tubes will perform flawlessly and contribute less than 0.1 0dB additional noise to the overall circuit. In any two stage amplifier or single stage with a cathode follower only the first tube need be low noise. The second will have to be quite noisy to make any noticeable contribution to the total.

There are only a few tubes that actually have ever had a noise spec. Noise testing is very difficult and not part of a factory's test procedure, I know because I visited the factories and asked them. At EI, the most syphosticated equipment was a Tektronix curve tracer that hadn't worked for several years. I asked them how they produced the curves for the KT-90. The chief engineer said, "we did it steady state with meters, and we did it fast" They had to do it very fast or the tube would go up in flames. This is not the way to generate curves.

What makes noise testing difficult is separating the tube noise from the noise of the tester. White noise from a tube doesn't look any different from white noise from the input of the tester. These are very small voltages going into very high gain amplifiers. The gain of my tester is about 100,000 times which is 100 dB. The gain when I listen via a speaker to check things out is 120 dB. At that gain I can hear everything and I can verify that the numbers printed on the label indeed represent the performance of the tube. Besides the microphonics I can hear the cathode expanding and pushing itself through the mica in little bursts. I can also hear the noise come up as the tube comes on and go down as the tube's transconductance goes up and steadies out in about 15 seconds. Even though there is not much change from there on, some tubes take a good half hour to show some particular problems so they get to soak an hour before the computer runs the test. I get very very few returns, less than 1%.

[Roger A. Modjeski]

Historically there have been fewer than a dozen tube types that have a noise spec at all. A good example is the 12AX7/7025. In many cases the manufacturer simply got the 7025s from the 12AX7 run picking out enough to get what they expected to use for the year. These got branded 7025 and sometimes 12AX7/7025 which is called a "double brand". For that tube the spec is rather lose at 1.8 UV average and 7 UV (microvolts referred to the grid) maximum. In the same manual, the average noise spec is the same 1.8 uV with no stated maximum. Noise is always "referred to the grid" as that takes the gain of the tube into account. Thus all the manufacturers be it RCA, GE, Sylvania or anybody had the same published spec and the same retail price to the penny. Consulting a 1983 price list the 12AX7/7025/ECC83 is a triple brand, it's one tube, no difference at all. By this time either the noise spec was being ignored or all the tubes met the max limit. Of course they weren't testing all the tubes, usually they would test a few out of every thousand. Most tubes were made only once a year in one batch that was determined by past annual sales. If RCA ran out of 12AX7's before the the next run they would simply buy them from their friendly USA competitor or from Europe. The run for a popular tube like the 12AX7 would typically be several hundred thousand. Fender alone was buying 100,000 a year.

They all used the same specs, gave the same applications and were proud of things they had learned. At one point Sylvania was buying plate material for rectifiers from RCA because it had lower back current than their plate material. Those were the days of standardization by the Radio Electronics Television Manufacturers' Association (RETMA). See http://en.wiped.org/wiki/RETMA_tube_designation for more information.

Many current production tubes are not within those specs and there is no official organization to police them. Some of those variations are purposely designed to be different, principally for the guitar amp industry. I have been told many times "For every one tube we sell for a Hi Fi amplifier we sell 100 to the guitar players". We are indeed a small segment of now small market. Although I am not the "Tube Policeman" I do work closely with several importers, letting them know when things have gotten out of hand at the plant. I also get pre-production samples of new designs for evaluation. There are at least 8 distinctly different 12AX7s in current production and they will sound different among their applications. The sonic range of these variants is greater than the range of NOS tubes.

None of this is meant to discourage those willing to take the time to experiment with different tubes, both NOS and current production. There are no "standard" operating parameters (plate, grid or cathode voltages and currents) and many preamps have some parameter which can be very sensitive to either gain, plate voltage or plate current. In a feedback RIAA phono preamp I have measured 2 dB frequency response changes due to gain (mu) alone. That will be quite audible by making the bottom end rise, fall or stay flat in the lower octaves. How one hears this and expresses what he hears with words is an individual thing. We do our best to describe what we hear with language that is drawn from other arts. If a Telefunken tube has more bass than some other brand it may be that their average gain is higher or lower than the standard. But this is a tube/amplifier interaction and not attributable to the tube alone or expected sound in some other amplifier.

When I was developing the RM-9 I realized that the popular output tubes could all be used in the same amplifier if I made the bias range large enough, the filament supply strong enough and the amp stable enough to handle the range of transconductance in popular output tubes. In addition I added a 3 position switch to allow an increse or decrease of about 6dB from the nominal value. For those who are not familiar with the Manley-Modjeski letters in Stereophile around 1985, David took issue with my granting freedom to compare these different tubes in the same amplifier. He contended that things like plate load impedance, operating voltages etc. were written in stone in the tube manuals and must be adhered to the letter. One might think that a 6550 has only the few applications listed in the data book when, in fact, there are an infinite number of possible applications. The purpose of these applications is to give the engineer some guide lines of what sort of voltages, currents and load impedances will give reasonable results. The applications usually cover a range that includes single ended, push pull class A, AB1, AB2, and class B configurations which for some tubes range from a few watts to 100 watts per pair. New tubes, like the 8417 and the 7591 were designed to be particularly attractive to new designs. In the 8417 the transconductance is twice the average of most power tubes thus allowing lower bias voltages and lower drive requirements. It allowed the amp design to have one less driver tube. Rather penny wise and pound foolish. In my opinion they went to far with this idea the result being a rather unstable, squirrely tube. It found its way into only a handful of designs, most notibly PA amps where power and economy were paramount and fidelity was not important. None of the European tube companies bothered with making it and I believe Sylvania and GE were rather sad they ever released it. 

So I made an amp that would properly run the major output tubes types. At the time I could find no other amplifier that would run EL-34s, the whole KT series (KT stands for Kinkless Tetrode), the 6L6/5881 series and the 6550. I didn't think is was reasonable to say a EL-34 was more musical and a KT-88 was more authoritative when the comparison had to be done in amplifiers of very different design. All the Marantz amps used EL-34's while Audio Research and CJ employed the 6550.

One final comment about tubes, numbers and sound. A 6922 is a 6DJ8 equivalent. They have the same specs. The 4 digit number system designates "industrial" tubes which were sometimes constructed with extra micas, some wire straps holding the elements in compression between the micas all of which was up to the particular maker of the tube. It was even permissible to put 6922 on all or part of a 6DJ8 production run with no internal differences and no extra tests. As tube production became smaller and smaller the tubes might carry both numbers on the tube and on the box. Therefore, a 6922 is not necessarily better than a 6DJ8, nor does it have any inherent sonic advantages.

For the 28 years I have been providing RAM tubes to tube enthusiasts around the world I have been rewarded for my efforts by many a customer for two things in particular. The quotes go something like this:

"I've had this preamp for 10 years, tried lots of tubes, noticed 'tube rush' and accepted as fact that tubes cannot be as quiet as solid state. Now with your tubes the noise is gone"

"Your tubes let me hear more detail and the silence between the notes"

When asked to comment on the subject of NOS verses current production I demur. On cryo treatments I chuckle and on the white gloves I say. "Skip the silly gloves, these are not quartz lamps, just wash the peanut butter off your hands first, that gets kinda messy."  BTW, Who started this silly notion that you can't touch the glass?

[6BQ5]

re: No fingers on the glass tubes?

I think I know that one. It also ages me.

Remember when .... you had the good old flash bulbs for use with cameras? Or, 1st generation halogen lights? Circa 1960s and even early '70s.

I remember studying photography back then in school, and that was the first thing you were taught: NEVER ever put your fingers on the glass bulbs. The reasons why ... well, supposedly the oils from fingers would remain on the bulbs, resulting in a hot spot when the flash would go off and the bulf could explode. Or with halogens, this would limit the life of the bulb or even cause them to explode too. They were much more delicate back then I guess.

Somehow, someway, the same logic wound up extending itself to vacuum tubes.

Roger, there are markings on my RM10 RAM EL84 tubes. Can you explain what those mean?

[Roger A. Modjeski]

Let me clairfy: Oil from fingers is not and never has been a problem for tubes. They are nothing like halogen bulbs new or old. As to flash bulbs, I have some of the old ones that are as large as a 100 watter but I have no desire to fire them off. The magnesium wool inside is quite attractive.

Tube mythology is a backwards thing when compared to Greek Mythology as it runs in the other direction getting more and more fantastic as time moves forward. Just chat with an old timer radio repairman or radio amateur and he will laugh at the things that are currently taken as fact today.

Due to the large numbers of tubes that have to be tested for a run of 100 amplifiers I wrote a program specifically to test RM-10 tubes at the actual operating voltages and currents. The number refers to the group to which that tube belongs. The groups are very tight with no more than 1/4 volt of grid bias difference in any group. We often ship replacement tubes by their groups if the amp is an RM-10.

[jules]

Roger,

you've touched on something here that I've long been wondering about:

There are only a few tubes that actually have ever had a noise spec. Noise testing is very difficult and not part of a factory's test procedure, I know because I visited the factories and asked them. At EI, the most syphosticated equipment was a Tektronix curve tracer that hadn't worked for several years. I asked them how they produced the curves for the KT-90. The chief engineer said, "we did it steady state with meters, and we did it fast" They had to do it very fast or the tube would go up in flames.

Using Telefunken as an example, there are common specs for the E88CC, 6922 and the CCa. On paper there is absolutely no differences in figures or tolerances between the E88CC and the CCa. Current mythology suggests that the CCa tube was selected on the basis of better measurements and low noise. The former belief appears to be wrong. The latter belief appears to be founded on a "low noise" standard that CCa tubes represented the best 10 or 15% [different figures from different sources]. The standard was apparently a requirement since the tubes were used for telephone services [a fine example of hi fi audio in the fifties  ] but my question is this:

a) am I right in thinking that "low noise" was the sole difference between a CCa and a 6922

and

b) can you shed any light on the actual testing procedure used for selecting low noise 6922/CCa tubes?

Thanks for the thoughts you've been sharing above. Knowledge of the thinking behind early tube design is something it would be sad to lose.

Why, for example, did manufacturers move away from D getters? Was it for economy or was it regarded as a design improvement.

Sorry, I'll stop now. This is supposed to be an amp design thread 

What are the markings on your tubes 6BQ8?

jules 

[Roger A. Modjeski]

[quote author=K.C.

You can find 'some people' who prefer anything old over new and who are willing to post their subjective opinions freely without too much trouble. Without the value of a blind test there's nothing more being reported than one man's opinion.

And if you want to talk about the shameless willingness to over charge then Kevin Deal is a good name to start with and Michael Elliot is a good one to finish with.

[/quote]

Yes I finished with M J Elliot many, many years ago.

[pubul57]

Roger, what do you think of this "myth" or "truth"?

HIFI-Tuning Fuses

By Robert Deutsch   •   September, 2007 Tweaks come and go. When a new one creates a buzz in audiophile circles, I generally prefer to wait and see if it's still around after the initial excitement has subsided. I'd heard about "audiophile" fuses some time ago, and although the likelihood of them making a significant difference didn't seem as farfetched as such tweaks as the "intelligent chip" or the "clever little clock," I didn't feel inclined to try them. I was persuaded otherwise by the confluence of two separate influences: a report by Michael Fremer, in the February 2007 Stereophile, that the HiFi-Tuning fuses produced a "subtle but noticeable" improvement in the sound of his Musical Fidelity kWP preamplifier; and an encounter at the 2007 Consumer Electronics Show with Robert Stein of importer Ultra Systems (the HiFi-Tuning fuses are made in Germany), who said that they produced a big improvement and offered to send me some samples.
When you think about it, the notion that a fuse can make a difference is not that implausible. If, as most designers of high-end audio electronics will admit, the kind of wire used for a component's internal wiring can make a difference, then why not a fuse? After all, the job of the fuse is to melt when a certain level of current is reached, so might there not be differences in how various fuses react to current at lower levels, possibly constricting or otherwise influencing the power used by the component? The HiFi-Tuning fuses use pure silver wiring, gold-over-silver endcaps, and ceramic rather than glass casings for better resonance characteristics. But a plausible mechanism of action and the use of high-quality materials are one thing; audible improvements are something else.

My first test involved the PS Audio GCC-100 integrated amp, a product I hold in very high regard; if its performance could be improved simply by replacing a fuse, that would be great. I talked to the PS Audio folks to find out what the GCC-100 uses in the way of a fuse, and was told that it uses two, both located on a circuit board, and that accessing them was a bit tricky. Before tackling the job, I unplugged the GCC-100's AC cord for a couple of hours to allow its capacitors to discharge, then removed the screws securing the chassis cover and gently lifted it a bit, looking for the wiring harness that would have to be unplugged before I could proceed. Once I'd located and unplugged the harness, replacing the fuses was pretty easy.

Before doing any of this, I spent some time listening to the GCC-100 to refresh my memory of what it sounds like. Yes, the tonal neutrality, resolution, transparency, and rhythmic thrust that had originally impressed me were still there. If the HiFi-Tuning fuses were to improve on this, they had their work cut out for them.

But improve the sound they did, and not just marginally. The PS Audio GCC-100 now sounded clearer, more dynamic, with improved transients—simply better all around. The difference was big enough that I didn't feel I had to go back and forth between fuses to convince myself that I was hearing it. But I did so anyway, if only to satisfy myself that what I was hearing was not the result of the mere act of replacing fuses, which to some degree can't help but serve to clean the contacts of fuse and fuse holder.

It wasn't. The improvement in sound was far out of proportion with the $60/pair cost of the HiFi-Tuning fuses. (Under its Critical Link label, PS Audio markets special fuses for its own products that are priced about the same as those from HiFi-Tuning. I briefly tried a pair of Critical Links in the GCC-100, and they, too, produced an improvement over the stock fuses.)

As I note in the main body of my Onkyo A-9555 review, the A-9555 benefited similarly from the use of the HiFi-Tuning fuses. As with every tweak, your mileage may vary, but the HiFi-Tuning fuses, available from www.ultrasystem.com, represent what I found to be a very worthwhile sonic improvement for a relatively small investment.—Robert Deutsch

Go Blow a Fuse, Michael Fremer, February 2007 (Vol.30 No.2):

When I was a poor audiophile, I loved tiny tweaks. Without switching out components, I could improve my system's sound without spending too much money. Bypassing or replacing cheap capacitors, cleaning connections, applying Star typewriter-cleaning gum to vibrating headshells, and other tiny tweaks made memorable sonic improvements.

I recently replaced the cheap fuses in my Musical Fidelity kWP preamplifier with fuses made by HiFi Tuning in Germany, sent to me by The Cable Company. These have silver filaments, ceramic bodies, and gold-over-silver terminations, and damn if they didn't seem to produce a subtle but noticeable improvement in smoothness and coherence. For less than $30 each (available in various sizes they're worth trying, if only for the diversionary entertainment.—Michael Fremer