diyparadise

It's a Mod Mod Mod Mod World

Hello folks. This time, your fearless diehard tube fan has been approached by a customer friend to "salvage" a horrendous sounding amp, of which the commercial make will be left unknown to you, in order to avoid blushes and embarassment (and flames and a lawsuit and many unhappy faces..). The challenge is simple: do what it takes to get good sound.

Interesting challenge? Lesser souls would have wilthered under the pressure and expectation of "good sound". But not your fearless intrepid "solder slinger" from the Not So Wild Boleh Land of Malaysia. How to go about this? We could go about "upgrading" components to boutique parts but I need to keep an eye on the cost side (customer request). Furthermore, how much mileage can you get from a leaky car? Let's check out the design first.

Taking a peek
At first look, the circuit is terribly complicated. 3 stages! A SRPP-ed 12AX7 followed by 6922 gain stage, then only the 5881 output tube. Of course, you could imagine the 12AX7 gain of 100x and 6922 gain of 30+ makes for some terrifying gain ! All this gain and we only need a gain of 20x to drive the 5881 to full power? Doesn't make sense right? And then of course, snobbish folks at diyparadise look down on pretenders like 12AX7 and 6922! Sheesh!

Making sense in a senseless world
Now all this gain is necessary as the amp uses some judicious amount of negative feedback to correct its errors. As we know "negative" feedback means reducing gain, thus we need more gain. More feedback, more gain. More gain, more feedback. But all this gain is, in the end, burned off in the feedback circuitry. But why use feedback? Well, to correct for errors/distortion generated by the driver stages. All this is well but I believe we are starting off on the wrong foot! I believe in what is fundamentally right. Start off right and you have a higher chance of success. Start of wrong and you need to do lots of work to get it right. So why start off wrong?

If you think I'm bullshitting "all that gain" stuff, well, try disconnecting the feedback circuitry. With gain in the thousands, this thing is more alive than you think! It's humming very loudly as every little noise/ripple on your power supply is amplified thousand times! It's also damn microphonic. Just touching the chassis and you can hear it! Didn't I tell you it's alive? Damn cranky as well. Touch the chassis and it goes "bo-ee-iingggggg". I don't want to insult your intelligence but what do you think the chances of feedback reducing ALL this power supply dirt, microphony noise down to the minimum? Very good? Think again. How often does stuffs that work so well in theory work well in reality?

So out goes the feedback circuitry!
Now we don't need so many gain stages anymore. I'll say out with the 6922 stage. Bypassed all of it and now we have a SRPP 12AX7 driving 5881. How's the sound? Nah. Still not up to my usual "high" (don't laugh!) standards. OK. Took out the 12AX7 and put in my favourite 5965. Ah! Much better! More life, more immediacy. But highs are terribly rolled off. Wait a minute. BIG resistors as grid stoppers? I always feel that grid stopper should be inserted AFTER listening rather than BEFORE, but who gives a shit to what I think? Took them out and the highs are back! Soundstage, imaging all improved! Grrrr!

But still there's a veil. Checked the circuitry. No, no, no. It's running too low a current. I have had quite an experience playing with 5965 and to me, it sounds better with a bit more current. Say 5mA. Changed the cathode resistors and... It does sound better! But there's still a veil. It's warm and sweet but ultimate clarity is still not there. Hmm... I have played with 5965 before so I'm pretty sure it could do better. Could the Solen coupling cap be the culprit? Replaced it with Auricap and even non-audiophile friends could hear the difference! Yeah! But but but. Still the veil is there. I'm getting cranky now. Very very cranky.

Talking 'bout a revolution
I think I want to forego the SRPP 5965 and go for something totally different. Got to do something drastic. This just isn't the way. I know 5965 sounds better so could it be the SRPP is not letting it "sing" Or is that the 5881 is not a nice tube? Consulted my customer friend and his words... "change the driver tube if you want to". Cool! That's the word I need! Why not use a medium mu linear tube like our favourite 5687? Since the B+ is 500V, we could use a sized up plate load resistor for even better loading on it. Note that the Simple 5687 uses only 10kohm. This, at 5x the plate resistance of the 5687, is already considered very good. But we are going to go macho here with 20kohm! That's 10x plate resistance!

Wired in 5687 and attendant circuitry. It's a breeze actually as it involves taking out more components than putting them in. Took this opportunity to install better parts as well. Using Welwyn wirewound for plate load (yes they are inductive but what is the problem when our resistor is 20kohm?), metal film for cathode and Elna Cerafine for cathode bypass.

Played music... jaw dropped, tongue rolled out, mouth agape, speechless, motionless, listening like an idiot, staring into the distance. I'm stunned. Everyone is stunned. Makes me think. WHAT THE HECK DO I EVEN WANT TO LOOK AT 12A*7 SHIT ANYMORE? WHEN MUCH MUCH SUPERIOR 5687s ARE READILY AVAILABLE?

Here's ~10W of SE Class A power. Cheap to build too. Who needs 300Bs?

Right now, the amp is dead quiet. No hum, no microphony problems anymore. Leave these problems to 12AX7 fans. : ) With further mods changing from self bias to fixed bias, we could make provisions to cater for other octal based power tubes like EL34, KT88, 6L6 etc.

Life is good when you have many choices.

diyparadise.com