A 6J5-807, single ended, class A1 amplifier

Dimitrios Tsifakis (dtsifakis AT gee-mail dot com)

The 807 has always been my favourite valve. I have used it on a number of RF transmitters and modulators and I thought it was only fair if my hi-fi audio amplifier uses it too. There are plenty of good designs out there showing different audio amplifiers using 807. The Williamson design is a famous instance. My desire was to design something simple from scratch, so I turned to the 807 datasheet looking for single ended designs. There are many datasheets out there but the one that is by far the most complete is the STC one.

Triode or Tetrode?

The majority of hi-fi valve ethusiasts will rave about the quality of the triode valve. The 807 is not a triode, but it could be connected in triode mode by connecting the screen grid directly to the anode. In this case, the maximum voltage that can be applied to the plate will limited by the maximum allowable screen voltage which is 300 V. That's a low value and that's the reason why the 807 wouldn't be an ideal valve to use in an ultralinear design. The datasheet shows that a triode strapped 807 is not expected to produce much more than a watt of audio. The tetrode connected 807 produces over five watts. I have found that given the sensitivity of my speakers (89 dB) and my listening room and preferences, five watts is more than enough.

While the valve enthusiasts will argue that one triode watt is plenty and they would prefer that to anything else, I was of the opposite opinion. In the STC datasheet, there are tables with suggested configurations for tetrode operation. In the tables above, we see that the 807 will produce more power if used as a tetrode. So what's the difference? The difference appears to be in the distortion of the final amplifier. The tetrode amplifier is expected to have more harmonic distortion, as seen below. In general, even harmonics (2nd, 4th, etc) are said to be more "pleasing" to the ear, while odd harmonics (3rd, 5th, etc) are believed to be an unpalatable form of distortion. One probably has to experience this and form their own opinion. In any case, I decided the described amplifier will produce a healthy dose of even harmonics!

Driving the 807

Given that this is a class A1 amplifiers, the driving requirements are simple. In Class A1, there is no current flowing on the 807 grid. A resistance-coupled voltage amplifier seems to be all that is needed. In the 6J5 datasheet, we get some instructions on how to do this. The schematic is as follows:

The proposed values for the components of the 6J5 amplifier are seen in the schematic below. They result in a voltage gain of about 18.

Final schematic, as built

Design comments

I would have preferred a valve rectifier but the first version was built with silicon diodes. This was dictated by the availability of the power transformer that did not have a centre tap or 5 V winding. In the future, the amplifier power supply will be re-designed for a 5AR4 full wave rectifier after a suitable power transformer is found.
I had a lot of hum initially. After some investigation, the filaments were found not to be the source of the hum, it turned out to be the screen. Tetrodes are very sensitive to the screen voltage variations, in fact you can produce an amplitude modulation transmitter with the modulation applied to the screen grid. I added R8 and C5 to further regulate the screen and that solved the problem. R6 and C4 do the same thing for the 6J5.
The output transformer, T1 (L1-L2), is made by Edcor and it is the single ended, 3750 ohm to 6 ohm model CXSE25-6-3.75k. It has an Ultralinear tap which is not used. It was purchased for a different design, so it's not optimal, but the mismatch is small and the result is acceptable. If I use 4 ohm speakers, then the ratio is closer to the one recommended by the datasheet. But then again, the only speaker that I know that has a constant impedance on the audio frequency range is a resistor! In reality, speaker impedance varies wildly with frequency, and the manufacturer specified impedance is probably just a hint that most of the time, the impedance is close to that value.
The full wave rectifier sees a 100 uF capacitor which presents a very low impedance on startup. The transformer is a modern, toroidal type one and doesn't seem to mind this. The resistance of the winding is such that the startup current is limited to a lower value to the peak current of the diodes. The diodes are 1N5408 which have a surge current of 200 A. R9 and C6 provide more AC ripple filtering and R9 is adjusted to bring the voltage close to the datasheet suggested value. This is a class-A amplifier which implies that the current requirement is almost constant and independent of the load. Most hifi monoblock designs would have a separate power supply for each channel, but in reality, that may be not required as the crosstalk introduced by the shared supply will be very very low.
All filter capacitors have 330k bleeder resistors across their terminals.
Cathode voltages for the 6J5 and the 807 are shown on the schematic. Small deviations are acceptable, as this is a single ended amplifier. If you are happy with what you're hearing, there is probably no need to find a closer matched valve pair. Speaking of valves, I am using two new old stock Westinghouse (Made in Canada) 807 valves which I purchased for $2 each from the local radio club trash and treasure. The 6J5s are made by RCA. I have a couple of 6J5G valves which look much better, having the "shoulder type" shape. They are reserved for special occasions, as they are hard to find. I have also tried 6C5 valves which are also a medium miu single triode with similar performance to teh 6J5. Another alternative would be a single 6SN7 or 12AU7, both double-triodes, which could have one triode per channel.
The amplifier is connected directly to an Akai CD-A7 compact disk player. This is a vintage CD player which has an output volume control knob. This is an unusual feature but it is handy in this case, as I don't have to implement the volume control function on the amplifier. In the future, this may change as the CD player may be replaced with a newer one.
C2 is a Sprague "Orange drop" 715P polypropylene capacitor. It's a cheap and cheerful capacitor which seems to work pretty well. Many valve audio enthusiasts may recommend fancier, audio-grade, expensive capacitors, but I can't quite say I have heard any difference in the various capacitors I have tried. Your miles may vary...
All resistors are 1 W rated except R5 and R9. R9 is 10 W rated and R5 is two 5W 330R resistors in parallel, as I didn't have the 170 ohm cathode resistor suggested by the datasheet. I think 165 ohm is close enough!

Listening

One could write all sorts of subjective comments on the performance of an audio amplifier, but I will resist that. In general, the amplifier sounds pleasing to my ears. Is it because of the even harmonics or because I like the looks of it? Or is it because I built it? In any case, the result is positive and that's what it's all about!

Photos