I'm putting this out there as my next project. During my covid layoff I built Triode Dicks Topcap 807PP amp from parts off ebay. They turned out pretty darn good for my first ever build. Fun project. I found out though that, for my DACs output, the 6SL7 had too high mu and not enough swing at the OP. 6SN7s fixed that but the OP is really low. The 807 THD was also high. So, looking at Triode Dicks later projects, I decided that I could adopt what he learned over time and make some big improvements to the early 807 design. The attachment is a pdf of what I'm about to do. I decided late to make a PCB for the CCS and Neg Bias circuits so while I'm waiting I thought I'd share this and have others poke holes in it so I end up with the best mods possible. What do you think of Mono Bills Topcap?? Please let me know!
1. It seems to me that when you use an LTP (with a real CCS), IF you want to use global negative feedback:
Use a 2 resistor attenuator:
The first one connects from the secondary output tap to the un-driven grid of the phase inverter; and the second resistor from that feedback resistor/grid node, and the other end of the second resistor to ground.
Given a fixed value second resistor, then . . .
I suggest that whatever secondary tap the speaker connects to, is the tap to connect the first feedback resistor to.
The resistance of that feedback resistor is proportional to root(secondary tap Z):
4 Ohm tap: 1.4 times the resistor value for an 8 Ohm tap.
16 Ohm tap: 0.707 times the resistor for an 8 Ohm tap.
Then use a square wave signal and scope, to see if you need a high frequency lead capacitor across the first feedback resistor.
2. It seems to me that when you use an LTP (with a real CCS), and Triode Wire the output tubes, IF you do Not want to use global negative feedback:
Although the output power will be less than in pentode mode, and less than in Ultra Linear mode;
A triode Wired 807 has a lower plate impedance, rp.
So, without global negative feedback, Triode mode has much higher damping factor than pentode mode.
And without global negative feedback, Triode mode has about 2x to 4x higher damping factor than Ultra Linear mode.
Does that sound reasonable?
Use a 2 resistor attenuator:
The first one connects from the secondary output tap to the un-driven grid of the phase inverter; and the second resistor from that feedback resistor/grid node, and the other end of the second resistor to ground.
Given a fixed value second resistor, then . . .
I suggest that whatever secondary tap the speaker connects to, is the tap to connect the first feedback resistor to.
The resistance of that feedback resistor is proportional to root(secondary tap Z):
4 Ohm tap: 1.4 times the resistor value for an 8 Ohm tap.
16 Ohm tap: 0.707 times the resistor for an 8 Ohm tap.
Then use a square wave signal and scope, to see if you need a high frequency lead capacitor across the first feedback resistor.
2. It seems to me that when you use an LTP (with a real CCS), and Triode Wire the output tubes, IF you do Not want to use global negative feedback:
Although the output power will be less than in pentode mode, and less than in Ultra Linear mode;
A triode Wired 807 has a lower plate impedance, rp.
So, without global negative feedback, Triode mode has much higher damping factor than pentode mode.
And without global negative feedback, Triode mode has about 2x to 4x higher damping factor than Ultra Linear mode.
Does that sound reasonable?
Last edited:
Heck, that's good information for me. So, I thought a lot about global negative feedback in this design. Here's where I landed. Neither of the two Triode Dick amps that I'm sort of mashing together used any negative feedback. I built and listened to the 807 PP pentode monoblock amplifiers for a couple of years. I never thought that they absolutely needed it either. Triode Dick had eventual negative feedback on his schematic but he never updated it. His Mono Bill amplifiers did not use negative feedback either. My answer then, is build these with no negative feedback and see how they sound. As far as power, I'm not in the least worried about enough power. My newest amp is my design (standing on the sholders of those before me) has a mu follower featuring a 201a gain stage driving an E80L triode strapped with no negative feedback. It sounds unbelievably good. It only has around 3 watts but is plenty loud on my cornwall speakers. The 807 amp will end up around 9 or 10 watts. Way more than enough power for me. So I am holding off on it for now.
It would have been interesting if you had tried the original schematic with 6SL7 (too high gain) and GNFB (would have reduced the gain). I think it is likely that negative feedback is a good thing with 807's in PP. Some people have to fit small inductors on the top cap to prevent 'snivets'. Did you need that?
Oh, I did try the 6SL7 tubes. The volume couldn't be turned past 2 without earplugs! Very loud. You are probably right about the negative feedback lowering the gain. But enough?
If I understand my tube design basics (which I might not) I don't think the 6SL7 is well matched to the requirements of the 807 in that it provides far too much grid voltage to the 807. If I'm wrong about this tell me, that's why I'm here.
But, if you look at the 6SL7 load line, the operating point grid voltage in the TD design was -2.1 volts. My DAC signal swings 2.83 volts. This guarantees problems in the preamp section. Then the amplified signal has something like 9 times the voltage swing of the 807 at its grid voltage. This creates more problems. Even the 6SN7 with its CCS will provide a ±55 volt swing to the 807 grid. I know it isn't actually that big but losses aside, it is plenty. Am I wrong about this?
I guess before I cut out the old design I could add GNFB and give it a listen.
And thank you for your input!
If I understand my tube design basics (which I might not) I don't think the 6SL7 is well matched to the requirements of the 807 in that it provides far too much grid voltage to the 807. If I'm wrong about this tell me, that's why I'm here.
But, if you look at the 6SL7 load line, the operating point grid voltage in the TD design was -2.1 volts. My DAC signal swings 2.83 volts. This guarantees problems in the preamp section. Then the amplified signal has something like 9 times the voltage swing of the 807 at its grid voltage. This creates more problems. Even the 6SN7 with its CCS will provide a ±55 volt swing to the 807 grid. I know it isn't actually that big but losses aside, it is plenty. Am I wrong about this?
I guess before I cut out the old design I could add GNFB and give it a listen.
And thank you for your input!
Miles Prower has (had) a page on here (somewhere?) where he discusses the 807 and hints for it. That confusing first sentence is about having seen it, but after a search I can't find it. You can look up his 'Vixen' amp, which uses PP 807s, where he mentions using a very small choke, wound about a resistor, at the 807's plate to remove the RF noise (it was an RF tube). He also recommended a 1.5 K resistor on the screen.
I've actually made an update this week that I will post tomorrow. I made some changes to the schematic and am having issues that I can not figure out. I am getting no output whatsoever from my speakers. I've checked all of my connections and the design is basic as all getout. I thought that maybe it was the output transformers but they check out as fine. I am measuring the correct voltage negative bias. The correct mA through 10R cathode resistor. The correct mA through the 6SL7 common cathode phase inverter. Bartola valves CCS load on 6SL7 plates. Yet nothing out of the OT. I'll put up the schematic and pictures tomorrow. Would love input, observations, criticism whatever.
jayrodge, should there be a resistor returning to ground on the grounded-grid side of the splitter (like R2)?
see https://wtfamps.com/phase-splitters/
see https://wtfamps.com/phase-splitters/
James, I use the Universal Load Line Calculator from Vacuum Tube Amplifiers DIY for the load lines. I also clip the curves from TDSL tube data PDFs and paste them into a PowerPoint slide deck and "draw" lines on them in PowerPoint as well as doing all the math with calculator. I have had great results on past projects doing it this way.
https://www.vtadiy.com/loadline-calculators/loadline-calculator/