What contributes most to a change in how an amplifier sounds?

About twenty years ago I bought a pair of old, used Altec-Lansing 1568A mono amplifiers in nice shape, they had spenent their lives in a rack.

Those 1568A’s are simple, lightly stressed designs, made to run 24 hours a day, using two EL34 tubes in a push-pull configuration to deliver a clean 40 watts.
They also have some of the finest transformers ever put in an amplifier, second only to the legendary output trannys found in the Harman-Kardon Citation II.
Any tube amplifier made today, regardless of price, except for some insanely expensive Ongakus, has transformers that cannot hold a candle to those on those Altecs and the H-K Citation.

But the Altecs also had anemic and primitive power supplies. Too simple.

Upgrading a tube amplifier power supply is really not too difficult, and these had plenty of room to work in.
So I went to town on these, replacing the PS caps with new oversize electrolytics with poyptopylene bypass caps.
I also also added a filter choke. A slight hum disappeared completely, All the signal caps were updated with Polypropes. All the resistors were replaced with new carbon films
Finally two NOS sets of Philips EL-34’s finished the job.

Those old Altecs put my McIntosh MC-60’s to shame. Greater slam in the bass, smoother mids, and the highs were just plan gorgeous, silky and detaled, into my JBL C50/S7 speakers.
I soon sold the Macs for $2000, and the total cost of those Altecs including the mods was about $700.00.

Downside: These Altec 1568A amplifiers are just plain ugly to look at.
Just put them somewhere you don’t have to see them.

Stock

Modded

Marc Stager
https://silversolids.com/

The listener

The room

The speakers

In that order.  

Of course it’s also nice to have an electricity supply, a switch to control that, and, I kid you not, a volume control is absolutely necessary.  
 

42

The cathodes are one of very important issues in SET tube amplifiers.

Some people use a fixed bias solution for the SET output stage. But there is the same issue in the input and driver stage and most designers do nothing to solve this problem. The problem is there are 2 signal paths (input and output) that go through each cathode. There is a Lynn Olson article about it.

There are 3 solutions: using fixed bisa, using batteries and using huge value capacitors (DIY enthusiast Yuri Makarov is another of this 3rd solution).

I tried the huge value capacitor solution (100,000uF with AN Kiasei NP 50uf bypass) in my SET amplifier first stage and driver stage.

The result exceeded my expectations. Much bigger 3D soundstage, better instrument separation, deeper and more controlled bass.

The most difficult was the solution for the output tube cathode. I afraid to use fixed bias. I think it is not safe. I also can't use 100,000uF because it is a strong output tube current for a long period of time after switching on while big cathode capacitors are charging. So I put 7000uf + 50uf AN Kaisei + 50uF Kemet DC-Link. It worked but I could hear the drawbacks of Kaisei capacitors. So I added Siemensk MKV 5uf + Duelund Cu-Sn bypass + 6000ps SGM3 silver mica.

In addition to my cathode experience,

I assembled a new 300B amplifier with external power supply. Most parts were reused from my previous amplifier but the basic break-in took 2 month. After that I started to recognize whatever I tried to do and tweek the sound of my SET amplifier was always sterile. And it was very strange for a 300B set amplifier.

In the end I decided to change the driver tube from 6f6 to 6v6. And I have to change the cathode resistor to increase 6v6 idle current. I bought a new AN tantalum resistor for it. At the same time I decided to change the input stage  (6sn7) cathode resistor from Vishay Z-foil to the Shinkoh tantalum resistor that I used previously in the driver (the value of the resistor is matched). The difference between driving tubes 6f6, 6v6, 6L6 is another story, BUT since I changed Vishay Z-foil to Shinkoh, my amplifier no longer sounds sterile anymore.

Application