Phase Inverter Choice

I have been doing a lot of book reading and studying schematics the last couple of months.


I am in the planning phase of building my own amplifier. While I am going to copy a popular topology, I don't want to just verbatim solder in component "x" because the chosen schematic says so. I understand how to draw load lines, and how to calculate the resistors and capacitors needed for my design. The only thing I am unsure about is the phase inverter. I know it takes the input signal and sends an original signal to 1 output tube and sends an equal but opposite signal to the other output tube. I know the usual choice is a 12A*7, but what I don't understand is exactly how that tube is chosen.


My desire is to build an entire octal tube amplifier. I intend to use 3-6SL7 preamp tubes, probably in a standard setup. 1 triode as initial gain going into a gain pot then the second triode going into the 3rd triode followed by a tone stack using the 4th triode as a "recovery" then into the remaining bottle. After the 3 6SL7s I'm thinking about a 6SN7 for the phase inverter with a master volume following. I intend on using 4 KT88 power tubes for almost 200 watts of output power. Can I use a 6SN7 for the PI? Also I would like to add a DI. Maybe using the 2nd 6SN7 triode? I am totally unsure of how to incorporate that yet.


So my proposed tube compliment would be 3-6SL7s, 1-6SN7, 4-KT88s and 2-5U4s.




Also, I've got a SS practice amp that has low, low mid, high mid and high controls. With the proper components, can such a tone stack be incorporated into a tube amp?
 
Last edited:
200W with 4 KT88, Is this going to be UL ?

What Voltage is on the plates?
Do you want to try to overdrive the output? I guess not...so only clean...Is this for bass?

Draw out the loadlines then to suit voltage and load, and then you can read-off how much swing you want to be driving the tubes. Once this is known, the working parameters of a PI can be worked out.

With 4 KT88 UL and clean, one would expect to see a cathodyne PI followed by two drivers, one for each pair of tubes. This is going to take you up to 12 tubes...
 
IMHO the better choice for a phase spliter is an unbalanced differential amplifier (Cathode coupled pair), and lets you a second grid to input a kind of NFB. Also a kind of anode follower will do the task. The cathodine or also called concertina is clearly the worse, it uses only one tube section but hum from the cathode and different output impedance and unbalanced capacitances in it make it the worse choice. Again, my own opinion only.
 
There's a dozen "phase inverter" schemes. Read.

How to design valve guitar amplifiers "Phase Inverters" (4 links)
https://dalmura.com.au/static/Phase Inversion Circuits Radio Craft 1948.pdf (old article)
http://www.tubebooks.org/Books/RDH3.pdf (page 9 onward)

There's no need to scratch load-lines. About any suitable tube will have a "Resistance Coupled Amplifier" table (G.E. most often includes that in the tube data) which you can/should just use. The main trick is estimating the load of the next stage. While "1Meg" looks good, in fact most audio nodes run close to 200k, at least at the top of the audio band. And specifically: KT88 has max grid resistor of 50k, so a 4-pony team is 25k per side, embarrassing to most small drivers.
 
Mhmm, the Hiwatt DR201 design demonstrates that KT88s can safely use 100k grid resistance (plus internal resistance of bias circuit) as a safe and reliable grid leak.

And these 4 tubes can safely be driven by a LTP phase inverter using an ECC81. Although not designed for that amp, even the ECC83 works fine as a phase splitter (to my ears the sound is better with the original ECC81).
Even a cathodyne based on ECC82 or ECC88 and maybe on ECC81 should be able to handle those 50kOhm input impedance. I would expect the best results from the ECC88; it should have the smallest output impedance of the three.


If You want to have more power You might go AB2. Which means grid current userland. And that requires biasing through ECC81 or ECC82 based cathode followers. followin a normal phase splitter. Which my be anything according to personal preference - a well designed cathodyne will easily provide enough output swing as long as it is ensured that to impedance of the following stages is high.

Case studies are the Hiwatt DR401, the large Ampegs and Fenders, all using basically the same design.

My suggestion: study these stages and try to understand what is going on and why. And then maybe adapt a few detals to Your needs. For example provide for the option seperate bias adjustment pots per output tube and a possibility to bias the amp without the need to open it:

original.jpg


And then simply build a copy of these stages - in case of 4 KT88 either the 200W stage as is or a 400W stage with only 4 power valves.

But honestly: with relatively little experience i would go the easiest way and use an LTP in front of a DR201-like output stage and avoid the more complex AB2 design.
 
I'm with Printer2, a novel 200W guitar amp is a huge undertaking, achieving a stable layout is not trivial. If you want to do this I would adjust your goals to just copying a commercial design, changing little. Even just changing a commercial design to only using octals at 200W will be quite an achievement.
 
I'm with Printer2, a novel 200W guitar amp is a huge undertaking, achieving a stable layout is not trivial. If you want to do this I would adjust your goals to just copying a commercial design, changing little. Even just changing a commercial design to only using octals at 200W will be quite an achievement.

I agree with people that agree with me. 😉
The best bet to build a working high power amplifier is to copy a design. All the people here that can roll their own amp had at least once made a high power oscillator.

Mind you looking over the post I see something worrying.

Also, I've got a SS practice amp that has low, low mid, high mid and high controls. With the proper components, can such a tone stack be incorporated into a tube amp?
Maybe more reading is necessary. But we will worry about if the OP chimes in again.
 
This is indeed more than worrying.

If the thread starter still reads:

My 1st project was a kit, EL84 PP powerstage. I would have had to make an additional input stage + tone stack.
Well, a small error in soldering parts together - despite double check - and the EL84 blew away. Both - it had a common cathode resistor.
And i did have just a multimeter - far too little equipment to actually test such a kit in case of an error.

I decided not to build it (well i am sure that i finally found the mistake, did even a test without tubes some years later). Bought an Epiphone valve junior instead and started to mod that. And learned a lot from that experience. restored a few larger amps and even made a preamp stage for the biggest (a Dynacord G2000 with 4 EL34 at 790V).

Message: start small. Learn the theory and apply it to some kind of prototype - a SE guitar amp, then a smaller PP amp.

And then maybe copy at least the output/PI stages of existing and really good designs, notably the Ampegs or the Hiwatts (and understand how these stages work, especially the 300+Watts amps). These have the HUGE advantage that it is easy to buy suitable transformers.
 
Last edited: