If you are looking for a simple first build amplifier that is easy to put together, low in cost and forgiving in operation this may be just for you. Alternately if you are like me and looking for a fun break from complex projects with the need for a nice sounding low power amplifier for background music in your office perhaps this will fill the bill.
Power output is limited to a few watts but as it runs as a SE class A amplifier that power is very easy to listen to for many hours. For background music in your office or den I find it a gem.
I based this design on a Russian 6P1P-EV pentode for the output stage and a 6AK5 for the input gain stage. The 6AK5 nice pentode for audio that is quite linear and can produces a impressive 48dB of gain at low distortion. Developed by Bell Labs / Western Electric it is a bit of a sleeper for audio use. As a bonus it is still available as NOS at low cost. Used ones abound and in a large lot I purchased I found almost all were perfect. The Russian 6Z1P is also a good option. For the power stage I used the Russian 6P1P-EV. It is close to the 7 pin 6AQ5 in a larger 9 pin base. The 6P1P-EV is available at very low cost as it is a bit of a odd duck with a strange pin-out. As well the peak plate current the smallish cathode will support is a bit limited compared to a 6BQ5 or the Russian equivalents but this is no limiting factor in this design. The 6P1P-EV with a 12 watt plate dissipation it can provide pretty well the same power output in a simple SE class A circuit as the much more expensive 6BQ5.
Below is the spice circuit that the design was developed with. The spice models were derived from sample tubes using a Utracer 3+ and the model creation software ExtractModel. These tools are serious good at creating simulation models. See DOS4EVER.com. In the case of this design the spice circuit performed exactly the same as the spaghetti monster that I created on my test bench to verify before a PCB was created with the same stage gains, same DC operational points, same frequency response and distortion. No component values needed adjusting from the spice circuit to the real world.
Circuit description
Let start at the input stage using a 6AK5. DC operation point is set by the 220K plate resistor, the 800K screen resistor and the 950 + 47.7 cathode resistor. With the values used the plate sits and about 90V and the screen at about 40V. Note that the screen is bypassed to the cathode by a 4.7uF capacitor to maximum gain. Getting the most gain is why there is a 470uF capacitor across the 950 ohm cathode resistor. With the values show and light loading by the 1Meg 6P1P grid resistor gain is about 48dB. The large 470uF bypass capacitor as does the 4.7uF screen capacitor helps maintain excellent phase margin down to very low frequencies. Many tube amplifiers suffer from marginal phase margin at subsonic frequencies and that can lead to such effects as "loose" bass, "burps" with some applied signals, motor boating with some loads or in a few cases the destruction of the amplifier if the speaker becomes disconnected. This amplifier is stable into any load even with no speaker attached. Spice shows there are no "bumps" or peaks in the frequency response suggesting instability from 10e-4 to 10e7 HZ. There is a nice smooth first order roll off below 20Hz and above 20Khz. The output stage is a simple self biased pentode stage operation as single ended class A amplifier. The cathode self bias resistor is bypassed by a large 1000uF capacitor again to maintain the excellent subsonic phase margin. The linearity of the 6P1P is reasonable in this mode of operation but not stellar. If you are willing to trade damping factor of the speakers for linearity leave out the self bias capacitor. DF will drop by 1/2 but the linearity of the output stage increase substantially. Total THD of the amplifier will be unchanged as the open loop gain drops with no cathode bypass in the output stage as there is now much less negative feedback. To some this may seem a worthwhile trade off but for my speakers the higher DF was more audibly accurate than concerns about the amount of negative feedback used. The grid of the 6P1P is dc grounded by a 1Meg resistor the maximum value allowed on the data sheet and coupled to the input stage by a 470nF metal film capacitor again protecting that subsonic phase margin.
Below pictures of the PCB, the proto type amplifier and the amplifier schematic.
Comments