If you’ve been following this thread you seen some of the back and forth concerning the amp design. This is not unusual as design and development always involves trade offs between customer requirements, available components, different design solutions, etc. And, after a couple of embarrassing missteps by yours truly
, we finally have a main power supply design. Actually, I have two. One is for 240v, 50Hz mains and one is for 120v, 60Hz mains. Please allow me to explain.
First, I fell into a common trap when doing the power supply design. Remember way back at the second post in this thread I listed requirements for this amp? Well, I seem to have missed a pretty important one. On that list should have been a requirement for 240v, 50Hz mains. You see Mark, for whom I am designing this amp, lives in South Australia where mains are typically 240v, 50Hz. I, on the other have, live in the Pacific Northwest region of the United States where mains power is 120v, 60Hz. Without the requirement in the list I just assumed that the mains power was just like mine. OOPS! The result of this little error was that I ended up designing the power supply three times! But that’s good for anyone who wants to try building this design because I’m going to post both final PS designs and a builder can simply pick the appropriate design.
My second point is about the order in which things get done. You see, I though I had finished my design to the point of having a set PS topology and told Mark he could go ahead and order the filter capacitors. Subsequently, I discovered the whole mains voltage and frequency issue mentioned above. Having a mains frequency that is 16.7% below the initial design point means that the ripple frequencies are 16.7% lower as well. The consequence is that the filter sections which were acceptable to meet the target ripple voltage at 60Hz, were insufficient to meet the ripple requirement at 50Hz. I had to add an R-C filter section to meet the requirement at 50Hz. And now Mark has to order an additional filter cap.
The lesson here is that all of this could have been avoided if I had thought to ask the question when setting the design requirements. Well this is just another lesson learned. I’ll know better next time (I hope).
Now on the the actual design. I have set the requirement for ripple at 0.003% or ~ -90dBv. This is a level suggested by Herbert Reich in the book “Theory and Application of Electron Tubes” for microphone circuits. Normally one might think this excessive for a simple audio amplifier however, as this is a high fidelity amplifier, I want to err on the side of caution. In addition, the step down for the 6SN7 drive will include and additional stage of filtering to ensure that the noise in the gain stage is as low as possible.
The one thing to note is that these designs assume some things about the characteristics of the power transformers themselves and the output transformers. On the design sheets you’ll see where I have estimated the Rs parameter based on some assumed primary and secondary resistances and the windings ratios. Also, in setting the requirement for 454v B+ I have assumed a voltage drop in the output transformer of 19v based on an effective resistance of 200Ω (364v+71v+19v=454v). I have asked Mark to measure the transformer DC resistances after he receives them so that I can makes sure everything is ok. If not, it will just mean that the dropping resistors in the power supplies may need to be altered some what. This is not really a big deal.
Here is the design sheet for the original 120v mains supply.
Final 120v PS Design.jpg
Here is the design sheet for the 240v, 50Hz mains supply.
Final 240v PS Design.jpg
A couple of quick points. First, these designs are based on the 5U4-GB rectifier tube. Contrary to what some people believe, the 5U4-GB and the old 5U4 are not the same tube. The 5U4-GB has much lower internal voltage drop then the old 5U4. If one were to use a 5U4 in this design, it would probably still work, but the B+ voltages would be as much as 30V lower. Second, the output voltages and ripples are not the same between the two designs and are not exactly what was specified. This is ok. The differences are minor and once built you’ll be very hard pressed to tell the difference even with two otherwise identical amps side by side.
And now for what everyone has been waiting. The final main PS schematics.
Now we are almost done with the design. The only thing missing is the heater design for the two 300B filamentary cathodes. I have a preliminary design that I have posed a little earlier in this thread. However, I am prototyping it right now to make sure that the circuit responds correctly at startup. I should be ably to post the final schematics for this last bit in a few days. At that time I’ll also post a complete schematic of the design including all components.