Finally the work on his pair of amplifiers is done. Let's have a look on what's different..
The spots that used to hold the navy grey UTC chokes were taken for other purposes hence I mount three power chokes inside the chasis. The following figure is a simplified version of the schematics I use. The gain of a single 6C45 is not enough to drive a 50 tube so I inserted a choke-loaded 6C4 gain stage before it. The spec for the interstage transformer was 5k:5k ordered from James Audio. The measured output swing was not big enough so I use a 10k resistor as the the load to make it effectively 10k:10k. It will impact the low frequency response but the influence is manageable. The negative voltage source comes from a circuit mentioned in my another article: High negative voltage power supply
I plan to replace the first 6C4 gain stage with something like a 600:10K step-up transformer from Sowter. I am interested to hear the difference.
The schematics for power supply is omitted for there is nothing new. Two globe 81 rectifier tubes supply the voltage for the 50. The power supply for the two small-signal tubes comes from a 80 rectifier tube. The filament of 50 are AC powered with a 100Ohms/3W Allen-Bradly potentiometer parallel connected for hum balance control. The resulting hum can be only heard if I put my ear close to the diaphragms of my pair of Tannoy Reveal 601 speakers which have an efficiency of 89dB/w/m. I will change the filament supply to DC powered in the near future.
For the wiring I use OCC (Ohno Continuous Casting) single-crystal solid core silver wire only with Teflon tubes for isolation. I use the following wire gauges: 1.2mm, 1.5mm and 2.0mm. For the routing I am a total liberal. I don't do right-angle routing and I don't align components in a row. My philosophy is routing should follow what the signal likes. Not how it looks. The routing philosophy in conjunction with a strict star-grounding topology makes the interior a little bit cluttered. But I don't think it can be worse the the work done by my hero- Mr. Sakuma.
Wiring of a Sakuma amplifier. Source: Internet
Now allow me to do a little advertisement. I found a source that manufacture OCC wires. I plan to order some 5N pure (99.999%) solid-core silver wires. Every batch comes with a certification of purity as well as microscopic photos to show the single-crystal structure issued by ITRI (Industrial Technology Research Institute). The wires are 100% made in Taiwan and are reasonably priced. Please let me know if you are interested.
As to the vacuum tubes I have 6C4s made by GE and RCA. I also purchased a couple ECC90/CV4058 made by Mullard from eBay. For the driver tubes I use the big 9-pin sockets for WE437s on the chassis. I made two adapters that convert 6C45's pin-out to WE437's. I also plan to build a pair of adapters for my ITT 3A/167M tube stocks. These tubes will keep me entertained for a while.
Following is a series measurements taken by Praxis from Liberty Instrument. Since I only have 7 Ohms power resistors the output power is based on 7 Ohms instead of 8 Ohms. James Audio improved a lot in the transformers they build over the years. I ordered a pair of nickle-core interstage transformers from James like 15 years ago and the measurement was so bad. It is a night-and-day difference compared to what they offer today. I purchased another pair of interstage transformers from a local supplier which is quite famous here in the local DIY circle. The test result was not good so allow me to skip it.
The measurements at low power is quite good with a THD of -40dB and a frequency response of within 0.5dB from 20Hz to 20kHz. Quite impressive for an amplifier that uses a plate-loaded-choke, an interstage transformer and an output transformer.
Frequency response at 1Watt output
The two figures above shows the measurements when the amplifier is outputting 1Watt. Having good performance here is essential because the amplifier output is under 1Watt 99% of the time.
The above two figures are measurements when the amplifier is close to it's full-power output. The frequency response is well kept within +/- 1dB 20Hz to 20kHz. The THD went up a lot. Not too surprising for a single-ended amplifier with zero feedback. How does it sound? Well, I am not going to talk about it since it really depends on personal taste. I posted the measurements on Facebook and one guy asked me why there is ripples in the frequency response curve especially in the high-frequency region? The answer is: it looks server in the high-frequency region because the X-axis is in LOG scale. If it is on a linear scale than the variation will be even spaced. This variation is a combination of stray inductance and stray capacitance inside the transformers. It can be mitigated by introducing some snubber-circuits (RC network) but the high-frequency response will be sacrificed.
In the end of this article I would like to talk about the spec for frequency response for an audio amplifier: Before CD era it is sufficient to reproduce 40Hz to 10kHz and considered high-fidelity. But today I think we need 20Hz to 20kHz. Personally I consider a frequency response of 20Hz~20kHz within +/-1 dB variation to be a good design while 20Hz~20kHz within +/- 3dB to be a minimum requirement. If the frequency response deviates by 3dB it can be easily identified by human ears. With much attenuated high frequency and low frequency for sure the mid-range (vocal) get richer(What else left?) . But the cost is sound staging and lots of details. This is of course my humble opinion. I know there are many people disagree with me.