Sometimes preamp tubes are run at 50% dissipation, if the location in the circuit has modest demands for voltage swing or current delivery.
However, even preamps benefit from output sections that have good current linearity, particularly if part of the load is the RIAA network. In the bad old days, 12AX7's operating at a feeble 1 mA were forced to drive both the RIAA network (which is a capacitive load) along with the output cable. This results in slewing, considerably increased HF distortion, and poor performance with moving-coil cartridges. For that matter, there are popular active crossovers that use 12AX7's (running at 1 mA or less) that drive the CR networks.
12AX7's have one purpose only: to give reasonable first-stage gain in a phono or mike preamp ... but in turn can only drive a load of 470k or higher. No RIAA networks, no cables, just a short wire and the grid (and grid-resistor) of a following stage.
Most of the Golden Age preamps used 12AX7's in every stage of phono and line-level preamps. That results in slewing, excessive HF distortion, and extreme sensitivity to cable capacitance. Raise the operating current ten times, the problems go away. But then you're using a 12AU7, 6CG7, or 6SN7, which are higher-current types.
As mentioner earlier, there are effects boxes used in studios that impart a "warm" (distorted) sound to an incoming signal. These use tubes biased at operating points that are intentionally nonlinear.
P.S. The plate dissipation spec is one of the most important parameters of a tube. Fortunately, unlike the Safe Operating Area (SOA) curves of bipolar transistors, this spec can be violated for a few seconds. (This would be an eternity for a transistor operating outside the SOA). In practice, this means the quiescent (steady-state) operating point is set in the 70% range, and an appropriate ratio of current and voltage is selected (this can be done with reference to the plate curves).
Transistor lifetime is decreased with increasing die temperature. A cooler transistor is one that lasts longer. Room temperature would be ideal, but active refrigeration isn't usually practical. Tubes are designed to operate at high temperatures, and will accept gross overheating up to the point where the plate starts to glow red and mechanically deform.
This is pretty similar to looking at a tweeter and working out what the prohibited operating region is. There's a minimum highpass filter that protects the tweeter from destruction, but if distortion and sound quality matter, then a higher-slope (or higher frequency) highpass filter is called for.
Last edited:
For frequencies above the modal range, I agree with your overview and when considering common placement options attempted by various development labs for upright or wall mounted implementations the results were less than optimal. But if considering 120hz and below and even gravitational forces working on the panel, the problem becomes much easier. Bass enclosures/subwoofers have always been an intrusion........so much so that many a declared audiophile write off nearly an entire octave of content as the logistics are prohibitive. I'd like two or even four bass horns in my space too.............but I don't like lawyers and matrimonial court either so...............
Lynn,
thanks for the answer, you actually answered my question early on in your design thesis here. That 3000-5000 hours rating was what I was expecting and to me that would be many years of use before a set of tubes would need to be replaced.That was why I was curious about Gary's comment about needing to change the tubes as I thought that those amps were relatively new. I didn't know he was using older used tubes in the amp, so it all makes perfect sense now. I see what you are saying about how some of the older designs are using feedback to cover up basic design flaws in their thinking. The question would be how well your alternative design would work with a little feedback after your optimization without it?
thanks for the answer, you actually answered my question early on in your design thesis here. That 3000-5000 hours rating was what I was expecting and to me that would be many years of use before a set of tubes would need to be replaced.That was why I was curious about Gary's comment about needing to change the tubes as I thought that those amps were relatively new. I didn't know he was using older used tubes in the amp, so it all makes perfect sense now. I see what you are saying about how some of the older designs are using feedback to cover up basic design flaws in their thinking. The question would be how well your alternative design would work with a little feedback after your optimization without it?
Mayhem,
okay, I didn't realize you were going to suggest the very limited bandwidth in your panel concept. What I had to work with was a full bandwidth system and I can tell you then the entire concept falls apart, more problems than the thin cross section has any advantage of. I would still have to think about the idea though, if you have an unterminated sheet you would have many of the same issues of a free air drivers cancellations, I would think you would still need some form of baffle and some type of edge termination. I am no fan of the OB type of enclosures, reminds me of an open back guitar cabinet which are really completely miss used if you know what the original intent was.
okay, I didn't realize you were going to suggest the very limited bandwidth in your panel concept. What I had to work with was a full bandwidth system and I can tell you then the entire concept falls apart, more problems than the thin cross section has any advantage of. I would still have to think about the idea though, if you have an unterminated sheet you would have many of the same issues of a free air drivers cancellations, I would think you would still need some form of baffle and some type of edge termination. I am no fan of the OB type of enclosures, reminds me of an open back guitar cabinet which are really completely miss used if you know what the original intent was.
Reminds me that SY tested the 12ax7 under optimal conditions and it performed exceedingly well. Before revealing the DUT, he asked for guesses and all the usual suspects for high-speed,low-drag tubes were bandied about.
Guess That Tube
You have to be careful with this though - Rms is a "bulk" figure - attempting to represent the total mechanical resistance.
What's critical is how little the surround dampens the outer edge of the cone. Low dampening ultra compliant low mass surrounds are the best in terms of low-level detail, but of course are also the worst with respect to suppressing modal problems in the cone and are often poor with respect to keeping the VC centered under greater excursion. The problem is that driver manufacturers use it as a "band-aid" for poor cone geometry and material choice. Note that this is a reason why some a really attracted to "whizer" full-range drivers - the "whizer" is basically a surround-less cone. Additionally, its not uncommon for fullrange drivers to have very low dampening surrounds.
Ex. (..though I absolutely hate the vocals in this track):
https://www.youtube.com/watch?v=zXpRwRi1f2E
The loudspeaker uses a Fostex 206e - which has an almost non-existent surround, oddly it also has a whizer that is quite a bit better structured and dampened than a Lowther - which does give "better" linearity than a Lowther, but at the expense of some that low-level detail. (..my slightly smaller 166 ESR's are similar to the 206e.)
Also be careful with increasing sd - larger surface areas present as larger sources and often have inherent diffraction at the top of their pass-band. The effect then is as you move closer to the source (the loudspeaker), that imaging becomes less distinct/pin-point and increasingly "phase-y". This is however largely a lateral/horizontal effect - so you can still increase sd with smaller drivers in a vertical line and basically avoid this problem.
Last edited:
To me its all about the "order" of the LF source. Is this flat panel a dipole or a monopole. If a dipole then its efficiency will suffer dramatically and if its a monopole then the larger surface area requires a huge box. And I simply don't see how one panel is any different than several classic drivers put on a single baffle (not a good idea in either case IMO.) At these LFs damping is just not a big issue for the transducer because the damping is dominated by the room mode damping not the driver.
You said that it was "not easy in a small room" which clearly implies that it IS always possible.
Everything here is anecdotal and virtually everything that you say is likewise. There is no "evidence" for any of this "you are there" stuff, its ALL anecdotal.
Nope.(..but sometimes I'm "clear as mud".)
The reflections just don't have the timing/delay to be relevant for most small room conditions (..though not all), nor do you usually have the sort of "discreet" reflection (and intensity near the average or above it) that would compete with direct sound - unless you have unusual hearing.
Yes, this comment is "clear as mud". Are you saying that small room reflections do not compete with the direct field?
The very early reflections in a small room are most certainly competing with the direct field. They do NOT in a large room which is why a small room can never emulate a large room unless all of the early reflections are negated.
If we want to be clear, first specify the "numbers". Simple example, when we say "small room" or "large room", how small is "small" or how large is "large".
Don't forget that when the room is smaller, the speaker is smaller [bookshelf types, along with shorter LF extension], the listening distance is also smaller, the "couch/sofa" that you use for listening could be large enough to fight the early reflection, i.e. zero reflection from the back wall, and then because the sofa is close to speaker, zero also from the floor. With relatively high ceiling and the common "soft" type, nothing also from the above [especially if a bed is used for listening which will put ears closer to the floor]. How about the side walls?
Just do the Math, the best way to fight reflection could be to position the treatment "device" close to our ears, just like how ears [humans', dogs'] are created.
But whatever the result, it is still just an illusion, which is subjective in nature. But I can say that it is possible that the listener will hear as if the room is much bigger, the drum comes from behind the wall etc, such that when they have to guess the size of the room [in blind condition] they will guess it bigger.
I'm sorry, but the science here just doesn't add up. If there is a difference in the speakers working at low levels versus high levels then its nonlinear. There is no evidence that I know of that supports the Qms as being nonlinear. Further, the Qms is swamped by the Qes resulting in Qms being a virtually insignificant factor. The surround does not really enter into the design problem until it begins to resonate, which is usually the lowest frequency cone breakup problem. Here damping of the surround is clearly beneficial. I just don't think that there is any scientific support for your claims.
"inherent diffraction at the top of their passband" ... I do not understand this in the least - where is this "diffraction" coming from (that all drivers don't have.)
I think he meant that when large woofer produces HF, the cone becomes the inherent waveguide.
Psychoacoustic says that something is not heard as an echo until its delay is on the order of something like 20-30 ms (see for example Kutruff - Room Acoustics.) In a small room there will be dozens of reflections in this time (highly dependent on the speakers and room of course). It would be virtually impossible to have a singular reflection in small room that was perceived as an echo.
Now when you put the numbers, everything is clear and agreeable. Honestly, a factor of 10 is a great achievement that I don't think I have ever experienced.
He should speak for himself, but I hope that's not what he meant, because if its not a coaxial driver then this doesn't happen either. The cone radiates over its whole surface, not just the apex. It is not necessarily all coherent, but it is all in motion. Thus it is not a waveguide. At any rate most people would not use a woofer high enough for this to even be a consideration.
Although you can get slap "echo" with really impulsive sounds and cr@ppy room geometry and hard surfaces.
I guess however that I shouldn't have just asked what an echo is, but if it might be a "competing" sound with direct sound..