Today, another premiere: an online edition of the reading club. When it comes to issue 1–16 of sound practices, my position is simple: either you have the printed mags (cool!) or you have the CD. With Joe’s offer, there is no reason not to own the CD.
But then there is the ghost-issue 17 a.k.a. Sound Practices 1999 Audio Yearbook. It was never printed, but Joe released the articles in batches on the net and they all ended up on the CD. But yeah, if you have all the printed mags but not the pdf of today’s article handy, then buying the CD just for ‘issue 17’ is a bit rich. So I have decided to link to online pdfs—that happen to float around—of the issue 17 articles that we discuss here.
Our system du jour is described in the Homebrewer of the Month article by Joseph Esmilla of Suburban Maryland in sound practices issue 17. It is basically an extensive needle–to–conus tour of the system, complete with several pre-and power-amp schematics and even plans for an open-baffle speaker.
teaser quote: ‘My circuits are pretty straightforward and conventional, nothing radical or innovative. I am a violinist, not an electronic engineer… I just paid close attention to using tubes that I am familiar with sonically, preferring Art Deco era preamp, input and driver triodes for voltage amp applications in a given circuit topology based on years of listening and experimentation.’
my take
No, one does not have to be an electronic engineer to successfully home-brew ultra-fi gear. Willingness to learn from the masters, combined with plenty of experimentation, musical taste and trust in one’s own ears is also a valid route to DIY bliss.
You can see that at that time Joseph had five SE DHT amps at his disposal. I think this makes him what I call a vertical builder. That is, the approach to build an piece of gear—and sure, it takes a while to develop it, also through experimentation—and then it is done. Then the next piece is built. Like Joseph, vertical builders then compare these complete pieces of gear, like five power amps, and draw their conclusions from that.
The opposite is the horizontal builder. Here a piece of gear is never done. Evolution and revolution of its design and construction is always ongoing. Comparison is between the last and current incarnation of the same piece of gear. I tend to fall into this category.
The weakness of the purely vertical builder is trying to draw conclusions from comparing machines that are themselves made out of dozens of design decisions and dozens of other factors. And if you say: ‘but in my vertical builds I keep everything the same except one thing,’ then I say: then you are either forgetting to optimise these builds for the what is different between them, or you are a part swapper.
The weakness of the purely horizontal builder is that then every experience is relative. There is only one instance of each piece of gear (one preamp, one amp, one record/cd player) and they are constantly under construction. It could have been that the power amp was at its best sounding three years ago, but after a series of bad and good moves that is impossible to tell today.
Sometimes one can get the wrong impression. In the late 90s I was mightily impressed with the steady stream of amps that Jim de Kort was building at the time. Talking to him at the first ETF in Århus, I found out that a lot of these amps were iterations of the same amp. He was more of a horizontal builder than a vertical one. So I did not have to feel too bad about my five-year-plan of getting something built. Since then I have had stages too where in a matter of months some completely different ‘amps’ were playing in my home, horizontal style.
Last week we saw Ed Billeci complain about DC heating of triode filaments, and this week it is Joseph’s turn. I am convinced that the ‘leaner sound’ comes from voltage-source DC heating (the type where a capacitor is at the end of the supply, connected to the filament). In this case the low impedance supply is an attractive route for the audio signal, which then runs through the so-so parts (4700µF caps, etc.) of the filament supply.
The solution is current-source DC supply. The high impedance of these is no option for audio signals, which then refuse to flow through the filament supply, taking a cleaner route. Building a current source is simple, you can see two in action at the bottom of this page of Jim de Kort. Check the LM350 datasheet (page 10) how to calculate these.
bonus tracks
Check out more writing and schematics by Joseph over at (the wayback archive of) angela.com: PP 2A3, SE 300B, SE 45/2A3, SE EL34 Hi-Fi (!). And just as I was preparing this article, Joseph turns up at the Joenet mailing list and announces the Tribute to Sound Practices Flickr group. Pure coincidence, I swear.
Now go and read the article, see you next week.
No comments:
Post a Comment