The e-mail I have received lately both encourages and discourages. Many letters have been filled with enthusiasm and joyful experimentation— unfortunately, most of these have also held unadulterated solid-state circuits and come from solid-state partisans. It seems that I had unknowingly opened a thick vein of IC-power-amplifier enthusiasm by publishing my articles on simple amplifiers, which contained enough references to the LM12 IC power amplifier to be found by Google and by about one hundred audio hobbyist who are having a great time exploring the possibilities dormant in prepackaged silicon amplifiers. This exploration makes a lot of sense in that the IC power amplifiers are cheap, reliable, and easily understood—if it weren’t for the sound… (It’s interesting that solid-state loving folks came to me for advice. Are there no solid-state-audio design websites?)
I wish many more tube fanciers resembled these solid-state hobbyists who have written me; I wish that they reveled in a carefree and inquisitive enthusiasm. There have been a few glorious exceptions, but not enough to make me overly optimistic about our tube-audio future. Unfortunately, the contrast has been sharp. The solid-state hobbyists are never nervous or timid; they boldly experiment. In many ways, these solid-state hobbyists remind of the tube hobbyists I knew twenty-five years ago, when tubes were cheap, reliable, and easily understood. Well, they are still reliable and easily understood.
Perhaps because tubes are now so expensive today (some 12AU7s go for $40 USD and the once cheap 45 can cost as much an entire five-channel home theater system at a discount store), few can afford to experiment and explore. I used to believe that the lack of exposure was the biggest problem tube audio faced and that what was needed was a prominent TV star to promote tube audio on his show: for example, Jerry Seinfeld in the role of a tube audio enthusiast on his show. His pursuit of tube nirvana could have easily been worked into the scripts. Of course, Jerry's enthusiasm for tube audio would have been ridiculed, as there is much to ridicule in our hobby (imagine George preferring cheap solid-state amplifiers that didn't cost thousands and didn't catch fire), but at least tube audio would be presented as something that a hip, intelligent man might become enthusiastic about. Today, my new fear would be that such exposure would only further increase the price of tubes, as there wouldn't be a compensatory increase in new tube manufacturing. Tubes are closer to being an exhaustible rather than a renewable resource.
How do you bypass the high cost of tubes? Don't use the expensive ones. The 8SN7, 10SN7, and 12SN7 all sound as good as the 6SN7, but cost much less. The Russian 6B4 being made today is a fine tube. And there's still a good supply 12B4s, 5687s and 6N1Ps to keep their prices low.
How do you bypass the high cost of fancy enclosures and tweak parts, such as silver-foil coupling capacitors and tantalum or bulk-foil resistors? Don't use them. Instead, go to Target and buy some aluminum baking pans; they make great prototype enclosures, as the metal is soft and easily cut. Because the pan is only a test bed, you can alter the circuit and chassis without fear or remorse. Use average-quality parts; failing that, use the cheapest parts you can find: film resistors, Mylar or cheap polypropylene capacitors, cheesy connectors, and junk-box inductors.
Why? Why forgo the best possible? The perfect is the enemy of the actually built and the perfectly expensive is the ally of the mentally feeble. Let's be frank: most components are garbage; this includes $800 gold/oil caps and $30 Japanese resistors and NOS WE 300Bs and transistors, MOSFETs in general. The secret is that there is no perfect resistor or capacitor or tube or MOSFET. Yet, there is a cult of part worshiping. Many claim that it is better to listen to $20 speakers through $4,000 cables than to listen to $4,000 speakers through $20 cables. Others insist that only one capacitor brand is worth listening through, but in three years time, they will insist that brand is terrible. It is illuminating to read old reviews in light of new reviews.
Part fetishes flourish where understanding hasn't taken root and the intellectual topsoil thin and barren. An analogy might be found in a car-racing enthusiast explaining how the decals' color and placement on the car determine its speed and agility. Or a better analogy might be the found in cooking. All good cooks know that quality of the ingredients (the parts) are essential. But high-quality ingredients are not enough. for example, the poor cook can be given the highest-quality ingredients, but still produce a pedestrian dish. Why?
High-quality ingredients are not enough. Understanding and art are needed. For example, a simple stew in the pot of a great chef dazzles, where the same ingredients in pots of the less capable in comparison prove insipid. The great chef might first caramelize the onions, mince, rather than crush the garlic, lightly brown the garlic in olive oil, reduce the wine, and grill the meat before filling the pot, things that a poor cook wouldn’t dream of doing. Additionally, high-quality ingredients are not as readily discernable as you might imagine; for example, perfect looking tomatoes and apples can, and often do, taste bland. And cheap ingredients sometimes taste better in a dish than the expensive alternatives. In audio circuits, high-quality parts are just as difficult to discern, as price and exteriors reveal little.
I listen to everything my audiophile friends say about parts and believe not a word of it. Someone tells me that BrandX caps sound best in the highs. Do they? How could he know? How could anyone know? By listening? How do you listen to only a capacitor except through a long chain of equipment? Let's say that his phono cartridge's VTA is way off and all his records sound screechy; would a capacitor that filters away high frequency information be best in absolute terms or only in his system? Or let's say that his CD player is broken, producing too much high-frequency energy, but he insists it isn't because it has received rave reviews (how's that for logic?). No one can listen to his setup for more than ten minutes before blood pours from the ears. So what does our part-swapping audiophile do? Of course, he replaces every part he can with duller sounding parts, as far as he is concerned, these dull sounding parts are the best sounding and he is right—but only as far as his system is concerned.
Selecting the best-sounding component in a particular circuit is work. As a consequence, I am a big believer in blind tests, when they are conducted over days, not minutes. I often build an amplifier that holds a toggle switch that switches between plate resistor types, say bulk foil vs. wire-wound or carbon composition. I then have someone rotate or not rotate the toggle switch, so that I don't know which resistor type is being switched in. After a few days of careful listening, and after several preference reversals, when I am finally sure which position is best, I will look inside to see which resistor won. My audiophile friends hate this procedure passionately, as they are terrified of preferring the wrong part, the one not approved of by some audio authority or the one that costs less. (I do the same thing with wine tasting by insisting that all bottles be bagged and the same friends who hate the blind audio testing hate this as well.) Without having in hand a pre-filled-out scorecard, i.e. a magazine review of the wines—or amplifiers—being evaluated, most are lost.
The only-one-tube-type fetish is common. You know what I am talking about: “Only the 45 or 12SX7 or WE417 is worth listening to; all other tubes are unsuitable for audio use.” Really? I get e-mail that says the circuits I have published look interesting, but as I didn’t specify the one of the many “only” tubes that can be listened to, the design must be profoundly flawed. “Well, use your favorite tube instead,” is my standard answer. The reply is always something along the lines that the circuit must be entirely remade to suit that tube. The fact is that just changing a cathode resistor value is usually all that is needed.
A triode is a triode; is a triode; is a triode. As long as cathode-to-grid voltages are adjusted and plate voltage and current limits are not exceeded, one triode can readily take the place of another within reason, an 845 shouldn’t be used in a phono preamp, for example. Yes, there at times when additional subtleties that must be acknowledged, such as using a remote-cutoff triode in an automatic gain circuit (or compressor circuit) or using a low-microphonic tube in a phono stage.
I know this is not prevailing view with the armchair tube circuit designers, as they see it: every time a generic, textbook-boring, almost-one-hundred-years-old grounded-cathode amplifier is outfitted with a different tube type or a different capacitor or resistor brand, a bold new topology is born.
Here's an analogy: say that you read on the net that solid-oaks doors, not solid-walnut nor oak-veneered doors, are the best, so you write an architect asking if he has any good house plans for a solid-oak door. He replies that all his plans, save for the designs for an outdoor tool shed and dog house, can be used with solid-oak doors; well, he just lost this commission, as he obviously doesn't know the first thing about architecture and solid-oak doors…or does he?
Expanding on this analogy: for every architect there are hundreds of interior decorators (both professional and amateur) that believe that they are at least equal to architects, if not better than architects. "Can you imagine: Frank Lloyd Wright liked to put beige curtains in bay windows? What an idiot! If only he had my skill in making interesting houses." So it is with the one million armchair tube circuit designers (audio decorators): facile opinions, very little understanding or experience.
Back in 1980, I expressed interest in building a 300B push-pull amplifier; my audio-decorator friends had either never heard of a 300B and thus dismissed 300Bs out of hand; or they had heard that 300Bs were used in old movie theater amps and thus dismissed 300Bs out of hand or had read that the 300B was designed for telephone use and thus dismissed 300Bs out of hand. Eventually, a new fashion came in and these very same friends told me that only 300B single-ended amplifiers are worth listening to, and now, 25 years later, that only 300B push-pull amplifiers are worth listening to. Fortunately, I never heed anything they say, as I could probably find two or three who would argue violently against ever using the 45 or 12SX7 or WE417 in any serious audio role. Why? Because it is either a 9-pin tube or because it uses a 12-volt heater or because it was designed for TV use or because they never heard of it and thus it must be dismissed out of hand.
Yes, the expense can be daunting and part fetishes needlessly constraining, but more must be at fault. Instead of adventurous and free-spirited e-mail from tube enthusiasts, I all too often receive nervous and fretful e-mail that betrays either a profound insecurity (“My amplifier doesn't work without grid-stopper resistors in place, but my friends tell me not to use them, so I removed them, but then I fried my power amplifier...”).
Audio suffers from two types of puritan: those who follow the "absolute" sound and those who follow the distortion meter. A shared characteristic between these two warring groups is not the pursuit of purity, but timidity. The supplicants of the Absolute must receive guidance from their High Priest; the specification slaves, from their meters. In both cases, the audiophile is exempt from criticism and blame, i.e. responsibility.
I was once at a high-end audio store with two audiophile clubs, a total of 60 people (all men of course), half of whom I did not know. We all sat attentively waiting to hear something great. The first few notes had just sounded from the obscenely-expensive speakers, when I stood and complained that the speakers were wired out of phase with each other. A gasp rippled across the crowd. The store's owner glanced at the back of each speaker and protested that they were correctly phased and the crowd seemed reassured. I asked him to check it anyway; after shaking his head, he did; indeed, one speaker's wires was out of phase at the amplifier's connections. The music resumed for a few seconds and once again I stood and this time I complained that the left and right channels were swapped. If stones had been handy, I would been stoned to death right then; I had violated the sanctity of the shrine and called into question the theological underpinnings of the faith. Well, after much cable unwinding, it turned out that the left and right channels had been swapped. Do not be confused: this is not a story about how good my ears are (as no doubt many heard the same thing that I heard), rather it is an indictment of how timorous too many audiophiles have become. Later that day, while driving home with three audio buddies, one turned to me and said, “That was the bravest thing I have ever witnessed.” Imagine that something as basic to being an audiophile as standing behind what you hear becomes a brave act; yet, for most audiophiles nothing, absolutely nothing, could be scarier.
Thus, my recommendation is to let go of fear, and to build with average quality parts; then listen for a few days, cataloging the amplifier's failings; then fiddle with part values and topology; and, finally, swap parts as needed until you are satisfied with the sound. After two hours, you will have graduated to audio decorator status; after two years, with some study and many solder joints, audio circuit designer. You will be surprised at how little is required to produce great sound.
The Church of the Cheap
First off, I'm glad to see your mailer once again
populating my inbox, and the pleasure that follows
when I read one of your articles or "blogs"(ugly
That said, I feel the need to reassure you that the
hobby of non-boutique DIY tube audio is alive and
doing well. Sure, the so-called tube renaissance of
the past decade has, indeed, generated its share of
fetishists, iconoclasts etc., but so be it. In fact, I actually belong to a cult of sorts: a cult that grows larger every day - the cult of the cheap.
Ebay's our God, John. My power transformer? I
purchased it new for twenty cents on the dollar. PSU
caps? Motor runs at 25 cents on the dollar. Tubes? My
recently finished phono stage/line buffer doesn't
possess a tube above ~$8.00, most were $5... $4 constant current sources, $13 output transformers (actually toroidal power transformers), I could go on.
Funny thing though; because the parts were so cheap, I became more ambitious, and before I knew it, I was way over budget. Oh well.
Now, if you'll excuse me, it's time for my afternoon worship service.
ps - that 12B4A is a heckavu tube. Please don't mention it in public again.
Welcome back. I used to worry about every part being perfect but no more. I have built many projects over the last 15 years and each time I found a better—usually more expensive—part I would use it in my current project. Well, after 15 years of progress, my second most recent project must sound much better than my first one, right? Wrong: I dug out my first line stage out of the garage and put it up against my newer effort that cost 5 to 10 times more. Which sounded better? Well, I think I'll save my money from now on. Parts are important, but they are not everything.
TubeCad blog comments
My 2c worth is that there are plenty of mags / books etc covering SS and that the Tube CAD Journal should continue to focus on tubes. This differentiates it and provides a focus for contributions and discussions.
That having being said, the other main area of interest is AUDIO!!
SS is used extensively and highly successfully in audio. I believe most tube audio enthusiasts are open to using SS components and even ss stages within designs where these offer specific advantages otherwise not so well achievable by using tubes.
e.g. for power supply diodes, regulation, very low noise moving magnet phono stages, current sources, even perhaps output buffers for multiple pre-amp outputs, complex filters, complex multi-way crossovers.
Very well designed LOW LEVEL SS stages can usefully avoid NFB and well designed complex filters can use passive filters with active gain / buffer stages.
I think the key is to ask the question - why is this SS being used and what specific engineering advantage does it provide: e.g. noise, simplicity, cost etc.
Another recent trend I have noticed is new designs from those with some tube savvy which use SS to mimic older tube circuit ideas and keep out the complexity and NFB that came with the SS design era. e.g. I saw a rework of the old 80's Elektor 3 stage SRPP tube phono / line pre-amp which was almost identical but re-implemented using JFets in SRPP. Very cute and extremely cheap and simple. I think many tubeheads would be interested in such a design and it would have a place in the Tube CAD Journal.
Moderation and avoidance of extreme positions is probably the key. T
Perhaps another key point is that we can learn from SS design mistakes and find that extra complexity offering apparently even lower vanishing levels of distortion on our simulations and meters does NOT necessarily SOUND BETTER AND is most often avoided like the plague IF distortion is already acceptable hearingwise. This ALSO applies to tube designs!!!!!!!!! We also need to hunt down and find out why SS can sound less sweet e.g. would the SS power amp rectification sound "better" if a resistance was added to INCREASE its Zout?????????? Just how far can one hear back into a power supply?? e.g. If I do it right can anyone ever hear if I am running my trick phono MC stage from a tube rectifier, a battery string or a battery charger powered invertor???????
Hybrid SS project idea
A lot of us are interested in HYBRID designs where we can save HEAT??, COST, weight. e.g. tube driven, transformerless, single ended, Class A, SS output stages. The problem here can be that e.g. MOSFET capacitance can become so hard to drive that an SS driver appears also preferable (so use a BJT output?) next problem is we end up needing two power supplies (so use low volt tubes - 6C45pi? 6992? differential??) Add single supply, single SS stage only to the above requirements. SS diodes and zeners allowed. Suggest under-run fan tunnel heatsinking (saves cost, weight, size and heat could be piped underfloor or into wall / ceiling cavity.)
So here above is a nice project challenge for you to sneak in some SS and justify it. ;-)
Keep up the great work,