sound practices reading club /30

Gone cold turkey? Here is your fix: the second ‘season’ of sprc kicks off, fittingly, with power amps (part 2). In the previous article Thomas mentioned Sakuma and here is the man himself. We are reading The Sakuma System by Susumu Sakuma, from sound practices issue 13. In it, Sakuma enters a flashback and takes us through the journey of creating his 845/845PP amplifier. Meanwhile, we get to know his system: driver tube = output tube(s), generous and unorthodox use of iron, amorphous cores, two-point earth method. Sakuma finishes the article with impressions, associations, prose and poetry.

teaser quote: ‘[…] I had to work through the barriers of convention myself. Now I assert that theory in electronics reference books took my favorite music from my life for a long time. I think an amplifier builder should consult with himself and his amplifier before asking the reference books and authorities.’

my take

This article is a manifesto for experimentation over theory, for using one’s ears and intuition over brain and calculation, for innovation over received wisdom. Reading the article I was struck by in what different terms Sakuma speaks about amp design and construction—for instance compared to the Bugle SE 45 article in the same issue. I could explain which parts I can relate to in what way, but I think everyone should just do that for himself.

It is not that Sakuma’s approach has left him completely out in the wilderness. Today, amorphous cores is a standard, though deluxe, option at several transformer winders. And when someone says that he has build a Sakuma-style 300B amplifier, then most of our cult knows that means a 300B driver. Being a pioneer means being seen as a fool in the beginning.

Yes, size matters where it comes to tubes. I am not talking about Freudian factors, but of generously sized and spaced structures inside the tube, that lead to the ability to drive, with high linearity. Although using beefy tubes as driver has a long, pre-WWII heritage, Sakuma’s ‘same tube for driver and output’ did get in an excellent way the points across that

• no tube is too big or powerful to be used as driver; and
• the driver better be similar sized or one tick smaller than the output tube; no small-bulb 9-pin driving the mighty 845.

No, it does not have to be exactly the same tube for driver and output (Sakuma was and is also not too dogmatic about it) but it is a good stating point for a design exploration. If you combine Sakuma-style with Shishido inspired exploration of A2 territory, then you are on the road that I am traveling at the moment.

The website mentioned at the end of the article is still the place to go to for Sakuma info. Wonderfully mid-90s. Although design updates stop there in 2004, Sakuma seems to be still at it: I have an MJ magazine from 2008 and one from 2009 and each has a Sakuma design article in it.

bonus tracks

Want to see what Sakuma’s restaurant looks like start here at youtube and be prepared to get lost for a while.

Now go and read the article, see you next week.

ps: Ah, I see I got a bit rusty and forgot one of my two cents.

Of course Sakuma puts himself beyond reproach where it comes to technical stuff, but still it must be said that his two-point earth method results in overly long current paths. For instance: the signal current of the diver tube goes from its filament to signal ground. But then has to travel through the signal and power ground busses to get to the 8µF cap that connects to the interstage transformer, to make it back to the driver tube. For me a good chunk of the beauty of using interstage transformers is one can keep the signal current really local, not flowing through earth reference lines at all. Here this would happen if P and S grounds would be the same point, for each stage.

Something else that I noticed that for all the experimentation/innovation of the Sakuma system, there are areas of design where Sakuma can’t be bothered/goes with the flow/doesn’t care: resistors, kathode and filament capacitors and wire. Sakuma uses what is safe and available. Fair enough for wire, I say, and all resistors are bypassed by caps in this design. But yeah, after going through these ‘extremely clear’ amorphous transformers, all the signal goes everywhere through standard electrolytic kathode capacitors. And by being on an input of every tube (the kathode) these caps get their mushiness amplified by these tubes. Every designer seems to go ‘practical’ at some point, using comfort solutions for areas that are perceived to be ‘outside the signal path.’ Heater supplies, filament supplies, bias supplies and power supplies are—in that order—the first candidates for receiving a comfort treatment. It is hard to go whole hog, down to (or beyond?) the power plug.