Thanks, MrDerrick. The link has lots of good information but does not really answer my poorly-communicated questions. Let me pose my questions again in a different way: 1) which tubes tend to sound better and exhibit lower noise, those with the higher-than-spec gain or those with the right-on-spec gain or even slightly lower-than-spec gain? and 2) in the BHK preamp, is there any benefit to the individual triode plate sections being matched in gain, or is it only important that the tubes’ overall gains be matched? I understand that it is important to have each tube in the pair have similar total gain so the left and right channel will have similar volume.
I have read that tubes with higher-than-spec gains may be noisier, suffer limited bandwidth and/or distort easier, yet in tube ads sellers often boast about the high Mu (or Gm) test values of their tubes, as though exceeding manufacturer new tube spec is a good thing worthy of an extraordinarily higher price tag. In my system, as in most, the BHK preamp really is not amplifying the signal coming from the DAC; instead–I think–as part of its preamp function it is reducing (attenuating) the strength of the signal. The DAC volume set at 100 without the Preamp would probably destroy my system or eardrums in short order. Intuitively, a lower gain tube in the Preamp should theoretically perform as good or better than a high gain tube, as long as it is not far from factory spec.
On the subject of Mu (gain), 12AU7 tubes have a new tube spec of 20, whereas 6DJ8/6922 tubes have a new tube spec value of 33. Clearly the latter tubes have a higher amplification factor. Does this mean that 6DJ8/6922 tubes may typically as a general rule be noisier than 12AU7 tubes? I realize the Preamp is designed with the adjustable voltage and bias to make everything work out well, but in principle, does one tube family have audible benefits over the other? (I know…try them and see! But in exploring the tubes available in the new and used marketplace, it is nice to know if “premium” higher-than-spec Mu tubes with individually-matched triode sections have any advantage over more mundane near-factory-spec tubes.)
On the subject of tube parameters and other recent questions:
In the preamp it is desirable to have the two halves of the tube have the same gain. If the gains are too far off, the cathode to cathode voltage equality servos won’t be able to pull the cathodes to the same voltage and then volume control pops will be LOUD. Also, this will obviously affect channel balance to some extent. This seems to be happening with some preamps and with the 7DJ8/PCC88 tubes along with potential “thumping” problems. This happened in my preamp with some 12AU7 varients. In that case, the Rch really didn’t work with very LOUD pops. Surprisingly, when I reversed the tubes channel to channel. All was well. So, when something seems very amiss like this, swap the tubes channel to channel!
About questions about what test parameters are most relevant, for those being measured at 15 mA and above are not directly relevant to the preamp or the amplifier as their currents are in the 4-5 mA range. And the questions about if parameters are greater or less than at spec are related to noise or sound quality, I really don’t know having not studied that per se.
In both the preamp and the amplifiers, the tubes form an input gain stage and are followed by a solid state output stage composed of all MOSFET devices. The gain of the tube stage is fixed in the power amplifiers and is variable in the preamp as the partial means of volume control.
The tubes are not wired in series or parallel in either case (amp or preamp) as each tube half is needed for one of the two signal channels.
q8728 above asked about the tube change jumpers in the preamp and was correct about it BUT the power amps only use the 6.3V 6922 type and varients and the 7DJ8/PCC88.
I hope this answers some of the questions. Ask more and I will respond
Thanks, Bascom, for your as-always insightful response. So, in addition to matching total gain of the tube pair, it is also good to match the two triodes within each tube for the reasons you give. The tube pairs I’ve tried so far have perfectly matched triode sections, so I have not encountered any pops or thumps. I paid extra for the triode section matching without knowing why I should.
As for any noise advantage of selecting tubes that test with higher-than-spec gains, it may be that their higher gains are a lucky result of plates with lower-than-spec resistances. Lower plate resistance means less noise, according to a formula I saw in one article. On the other hand, for the same listening volume wouldn’t the resistance of the volume control have to be higher to control the higher output of a higher gain tube, introducing more noise and offsetting the noise advantage of lower resistance tube plates?
Your suggestion to “swap the tubes channel to channel” as a potential remedy for certain problems is interesting. I recently had a similar (mysterious) experience. In my 2-channel tube amp I had a nagging ongoing problem with a delay in the start-up of one of the four driver tubes. On a whim I swapped that tube with a corresponding tube in the other channel and the problem was no more! Now all tubes start at the same time.
After several weeks of vintage tube rolling I’m basically in agreement with JM’s excellent post #239. (Thank you, JM). The best tube I’ve tried in the BHK Pre is the early Siemens Halske Cca. The sound is the clearest, most balanced, least sibilant, and most natural of all the tubes I tried. A matched NOS pair of this tube is almost impossible to find. I picked up a lightly-used matched pair on Ebay. Fortunately, the early Siemens Halske E88CC—if tested and selected for closely matched triode sections, low-noise and low-microphony—sounds identical to the Cca (because it is the very same tube, manufactured with the same parts, at the same time at the same factory). I was able to acquire two pairs of these as back-up for the Cca. Second place in my system goes to early Amperex 7316 (Beckman label) with long plates, D-getter and matched triode sections. This tube is well-balanced, smooth and clear, with similar frequency extensions but a tad less revealing than the Cca. Third place goes to the Amperex 7308 US-made USN (JAN) white label. This tube has an excellent full bodied sound with a strong bass and presence, but not quite as natural and detailed as the Cca and Amperex 7316 in my system. [Ironically, some people like the 7308 US JAN better than the Cca, claiming that the Cca sounds sterile and bright in their systems. This is good news for them, since the 7308 US USN (JAN) tube is much easier to find and not very expensive.] The Tungsram 7DJ8, the most economical tube of the bunch, sounded slightly more forward and less refined than the Siemens and Amperex in my system, but that may be the “texture” and “liveliness” that many people like. There is no right and wrong answer here. It’s a matter of taste and system synergy.
These are just my impressions of these vintage tubes in my system. For now, I’m just glad to be done with tube rolling but very glad I did it. Thanks, BHK and PS Audio, for giving us such a tube-friendly preamp that allows us to optimize the sound to suit our individual systems and tastes.
I used to have 4 matched pairs of Siemens Halske CCa tubes from the early 60’s that I used in a headphone amplifier. These were my favorite tube. I bought several pairs of perfectly matched Siemens Halske E88CC tubes thinking that they should sound just as nice as the CCa labeled tubes. After all, they have the same construction, right?. Well, no pair ever came close to the CCa tubes in sound quality. Others in the tube headphone amp community had the same experience I had.
The only conclusion we could draw is that Siemens Halske had a special line for these tubes. This is borne out by the fact that the CCa tubes, even never opened NOS tubes, often have mismatched triodes. Not as mismatched as we would see with the E88CC tubes , but still mismatched. Something is different in their construction.
I had similar experiences with Telefunken CCa tubes. From the same time period, the Telefunken CCa and E88CC tubes looked to be the same. But the CCa tubes sound different.
Anyway, the high prices of these CCa irritated me. So I went in a whole different direction and tried some 5670 family tubes in my amp using a 5670 to 6922 adapter. Why? I heard from an amp designer that he liked the 5670 family tubes more that 6922 family and they were a lot cheaper. I didn’t have anything to lose and I figured it would be fun to experiment!
What I found is that I liked the Western Electric “JW” 2C51 and 396A tubes (square getters) better than the Siemens Halske CCa tubes and they cost from 1/4 to 1/3 the price. They are detail monsters like the CCa tubes but are much more musical…similar to the best Heerlen tubes in that regard. The best of both worlds. I sold all of my 6922 tubes and have never been happier.
Someone should ask BHK if the 5670 tubes with adapters are compatible. They should be as I have never heard of an amp or preamp in which they weren’t.
JLG - thanks for the hat tip and additional research.
In the interim since the last testing round, I tested out the Telefunken ECC802S (12au7). In my opinion, these bested the Amperex 7316s/Telefunken 7308s that had previously held 2d place in the previous testing cycle. The Tele 802s added a bit more soundstage and musicality. The Siemens Cca grey shield early 60s maintain the throne, however.
Consistent with previous findings (and BHK’s earlier comments), the 12au7 family of tubes were quieter than the 6dj8/6922 family variants.
Up next for testing will be the Amperex 7308 (USA, 60s, white PQ label) and Siemens 12au7 (early 60s) nickel plates. Given competing priorities, probably won’t happen for a while. Will add when completed though.
My amps are Pass Labs xa100.5, not BHK, so I do not know how the tubes I tested would sound in a BHK amp or an all-BHK amplified system.
I have not tried Brimar and Telefunken. Their typical house tonal balance is probably not as well suited to my system as the Siemens and Amperex. My amps and speakers are neutral in coloration and have sufficient bass.
I mistyped JM’s post number. It is 230 not 239.
Regarding the degree of equivalency of early 60s Siemens Cca to premium (low noise, matched plates, low microphony) early 60s Siemens E88CC, my understanding is that the tubes that got the Cca label were E88CC tubes that were pulled from production runs, tested and further sorted according to certain strict criteria for use in German telephone equipment. That meant selection for uniformity of gain, matching triodes, lowest noise and lowest microphony. It stands to reason that a pair of E88CC-labeled tubes made in the same production runs from which tubes were pulled as Cca candidates can and should sound as good as the Cca-labeled tubes if they are tested and meet the same test criteria. I recall at least one reviewer preferring the sound of the E88CC to the Cca. Maybe tubes are like violins, pianos and camera lenses—even when made from the same parts in the same production run their excellence varies somewhat from copy to copy.
Other than the labelling, the only difference I could see in the above tube construction was a difference in the shape of a tiny little hole on one of the insulating spacers, but that could be because the above tubes are not from the same run.
My idea of a superb tube is one that makes me want to keep the BHK Pre in my system as opposed to no-Pre. Fortunately, there are many tubes that accomplish that. Having now experienced at least a few tubes that excel in my system, my mission is accomplished. My enthusiasm for further tube rolling has waned.
JosephLG: I don’t have the CCa tubes in my stash to try, but I stopped my tube rolling last June when I installed a pair of Amperex 7308 US-made USN (JAN) white label tubes - - they just are a match to marvel at for my system. My speakers do better for me when I am able to pull them a little to the warmer side (but not lots, like Mullard style!), and maybe that’s why the Amperex 7308’s are so good IME. Your hearing a little less sparkle up top with them is just right in my system! You’ve gotta love tube rolling for its flexibility (and you’ve gotta know when to stop!).
That said, I am acquiring some 5670 to 6922 adapters and will try some of that family of tubes, just because I’m tube curious.
If you like those Amperex tubes, I think you will really like the Western Electric tubes I mentioned. Also, older Tung-Sol 2C51 tubes with a rectangular getter and black plates are great tubes to try. They have spectacular mid-range and staging.
I used to have 4 matched pairs of Siemens Halske CCa tubes from the early 60’s that I used in a headphone amplifier. These were my favorite tube. I bought several pairs of perfectly matched Siemens Halske E88CC tubes thinking that they should sound just as nice as the CCa labeled tubes. After all, they have the same construction, right?. Well, no pair ever came close to the CCa tubes in sound quality. Others in the tube headphone amp community had the same experience I had.
The only conclusion we could draw is that Siemens Halske had a special line for these tubes. This is borne out by the fact that the CCa tubes, even never opened NOS tubes, often have mismatched triodes. Not as mismatched as we would see with the E88CC tubes , but still mismatched. Something is different in their construction.
I had similar experiences with Telefunken CCa tubes. From the same time period, the Telefunken CCa and E88CC tubes looked to be the same. But the CCa tubes sound different.
Anyway, the high prices of these CCa irritated me. So I went in a whole different direction and tried some 5670 family tubes in my amp using a 5670 to 6922 adapter. Why? I heard from an amp designer that he liked the 5670 family tubes more that 6922 family and they were a lot cheaper. I didn’t have anything to lose and I figured it would be fun to experiment!
What I found is that I liked the Western Electric “JW” 2C51 and 396A tubes (square getters) better than the Siemens Halske CCa tubes and they cost from 1/4 to 1/3 the price. They are detail monsters like the CCa tubes but are much more musical…similar to the best Heerlen tubes in that regard. The best of both worlds. I sold all of my 6922 tubes and have never been happier.
Someone should ask BHK if the 5670 tubes with adapters are compatible. They should be as I have never heard of an amp or preamp in which they weren’t.
Responding to the last paragraph here about the 5870 tube. Made me think of another favorite for amps in the past, the 5687. I started to research the 5670 and found the most amazing tube on vacuum tube theory: Conductance Curve Design Manual by Keats A Pullen. It is free and a great repository of tube characteristics. On page 78 we find the plate characteristics of the 5860 tube. Here is what one does in evaluating whether or not a tube would fit into the operating conditions of the tubes in either the amp or preamp. Plate supply voltage is 120V, plate resistors are 10K. If one puts a straight line overlaying the plate characteristics so that it intersects 120 V on the horizontal axis and 12 mA on the current axis, the operating point will be where the overlaid straight line intersects 4 or 5 mA for the preamp and 5 mA for the amp. See the attached curve for the 6922 tube in which I used in the design. I had actually plotted the load line on these curves and one can see the operating point shown circled. in this case the corresponding grid bias voltage is a little more negative than -1 V.Now, if one does the same plot over onto the 5670, the operating point is about right on 0 V grid bias which is no good as the tube will draw some grid current and lower the input impedance. At 4 mA in the preamp the grid bias is still too close to 0 V and I would say it is too marginal to use. On the other hand, the 5687 would work with a lower circuit gain. This one is left to the student so to speak as one of my professors used to say ;-)
The Amperex 7308 US-made USN (JAN) white label tubes that you (and so many others) praise are unquestionably nice tubes. JM says he is also going to try them (the 60s USA PQ white label, which I believe are identical to the USN/JAN label) when he gets a chance. I’ve read that the Amperex 7308 Holland-made tubes have better midrange, but the US-made military ones have better bass and are preferred by many over the Holland-made tubes. If you want to enrich your speakers’ midrange, the Holland-made tubes might sound better to you than the US-made ones, but I doubt the difference between the two is dramatic.
@BHK what is the max heater current that the BHK Pre can supply? I have found a wonderful sounding pair of Russian 6n6p (6h6n) tubes that are essentially 12bh7’s,… plot wise they work as the bias point falls around -2v (@5ma),… Heater current is ~700ma (per tube) though,… pretty much double a 12au7,… I don’t want to over tax/stress the heater supply if it’s not up to it…
One other question,… when using a cc source/drain on a tube, doesn’t that flatten out the load line from an operation standpoint? realizing that the bias point will still be the same and you arrive at that point in the same way (Pr\V),… the slope flattens out some because of the constant current being delivered, correct?.. thus operating the tube in a more linear area between the grid lines?
Anywho,… if you confirm that the Pre heater supply is up to snuff for the higher ma draw, then I would highly recommend the 6n6p as a tube to try,… you can get a nice matched pair for ~$40 so not crazy expensive either. Please note that you do not want the 6n6p-I as it is for impulse use and not suitable,…
@BHK what is the max heater current that the BHK Pre can supply? I have found a wonderful sounding pair of Russian 6n6p (6h6n) tubes that are essentially 12bh7’s,… plot wise they work as the bias point falls around -2v (@5ma),… Heater current is ~700ma (per tube) though,… pretty much double a 12au7,… I don’t want to over tax/stress the heater supply if it’s not up to it…
One other question,… when using a cc source/drain on a tube, doesn’t that flatten out the load line from an operation standpoint? realizing that the bias point will still be the same and you arrive at that point in the same way (Pr\V),… the slope flattens out some because of the constant current being delivered, correct?.. thus operating the tube in a more linear area between the grid lines?
Anywho,… if you confirm that the Pre heater supply is up to snuff for the higher ma draw, then I would highly recommend the 6n6p as a tube to try,… you can get a nice matched pair for ~$40 so not crazy expensive either. Please note that you do not want the 6n6p-I as it is for impulse use and not suitable,…
Best!
I will check out the tube heater regulator to see what kind of current it can supply before the regulator drops out or, in my opinion, the regulator heat sink gets too hot. This opens up the possibility that I had been thinking about of a considerable number of tubes that could be used with the BHK circuitry like 12BH7, 6N30, 6CG7 5687, and more. Of course circuit gains will vary but in general, gains would most likely be lower than with the 6922 but that in itself is OK. One thing to note though, high gain triodes like 12AX7 won’t work because they can’t support 4 or 5 mA with out drawing grid current with a positive grid bias voltage. Construct the load line for the BHK tube circuit conditions (120 V plate voltage supply and 10 K plate load resistors) on a set of 12AX7 plate characteristics and you will see what I mean.
To your other question, yes a constant current load in the plate circuit of a tube will make the load line horizontal and more linear. However in the case of a differential amplifier configuration like the BHK circuits, a constant current source in the cathode circuit doesn’t have this effect but does cause the plate currents to be constant at the desired value i.e. 4 or 5 mA IF the tube can support it, see above.
Thanks B! Appreciate the help and the insight on CCS/cathode in regards to the differential circuit… Really hope the regulator can handle the increased load because the 6N6P sounds really good, just don’t want to leave it in without knowing long term effect… when looking around I did have on a short list the 12bh7, 6cg7/6fq7, 6gu7, etc… to try,… all of which can be bought cheap… comparatively speaking of course, and have suitable load lines…
