@SpreadSpectrum talked about those topics on this forum.
Yes, I just couldn't remember his name when I posted. Looks like a very clever circuit to me, assuming the goal is a neutral sounding amp.
@SpreadSpectrum talked about those topics on this forum.
In Spice increasing the blocking cap from 0.33 uF to 330 uF makes almost no difference to the interstage or output response from 0.2 Hz to 20 Hz. The dominant RC constant appears to be between the feedback blocking capacitor and the feedback resistor, below 2 Hz in my example. While Spice uses an idealized OPT bench measurements didn't show any issues in the sub-sonics.Or use a capacitor in series with the feedback resistor (use a large capacitance....
Yep, the low frequency response goes WAY down with a big blocking cap (rumble filtering anyone?) plus LF distortion can be kept low. The blocking cap can be dispensed with (it works) but I restored it to make the set-up simpler when swapping different output valves for comparative testing. There is also the issue of managing the standing current in the OPT, so that dispensing with the blocking cap is less of an issue in the push-pull Telefunken designs referred to in earlier posts. As per usual, RDH733 is not a cure-all and involves trade-offs:Or use a capacitor in series with the feedback resistor (use a large capacitance, the gain goes up at the RC pole frequency, can either be a wanted bass lift, or undesired lift. Watch out for 3Hz record warp frequency, for example.
Pro | Con |
Simple to implement | The design maths are not so simple |
Works best with high-gain devices | Most high gain devices are less linear to start with |
Lots of voltage gain allows lots of feedback | You’ll need all of it if you use cheap pentodes! |
Easy to adjust feedback factor over wide range | That big blocking cap has gotta be audible! |
OPT not in loop so doesn’t have to be “the best” | Can’t treat OPT distortion |
Drives OPT from low impedance for less LF distortion | Duh - it's simpler to just use a triode |
Very stable, even interposing a MOSFET/cathode follower | So what? So are other configurations |
That would be an interesting discussion!... Rowe amp rebuild ...
Aha! That was going to be the interesting part, using that centre tapped winding. When I proposed that idea here before, the people more experienced than me suggested it was a bit to much CFB for practical purposes, making it too hard to drive. I had been considering a tube with a low voltage screen requirement instead.convert the 70 volt winding to cathode feedback
Just remember that the UL84 is NOT a EL84/6BQ5 with a 45 volt 100 mA heater. It is a 45 volt 6CW5/EL86 which needs a much lower screen voltage to avoid becoming a flash - bang grenade.
My experiments are with cheap and cheerful output pentodes with high screen grid voltage ratings (I think I'll try UNSET when I come to play with big sweep tubes) and "not the best" OPTs. By way of interest, the EL84/6GK6 performs much like the 6CZ5 in my test rig ... perhaps I'll try a UL84 next ... or maybe some of my stash of 6V6 and 6L6 guitar amp "pulls".
convert the 70 volt winding to cathode feedback.
At this point not even sure if the quality of the 70 volt secondary is suitable for feedback. Does it tank at 8 kHz? Needs measuring. Fixed bias is preferable but if that doesn't work cathode biasing will provide a return resistor to load, pad and potentially shape the 70 volt winding signal. Still in shower thinking part of the process.it was a bit to much CFB
Just remember that the UL84 is NOT a EL84/6BQ5 with a 45 volt 100 mA heater. It is a 45 volt 6CW5/EL86 which needs a much lower screen voltage to avoid becoming a flash - bang grenade.
Fixed Bias is never a great idea for an SE stage running Class A.Another option could be to used fixed bias in the output stage
I got a request for the schematic of the amp, so it is included here. There are some caveats though. When I laid out this board, I used the "kitchen sink" approach. I put everything that I wanted to test out on the board except the kitchen sink.....that might put the project over budget.Yep, thanks Tubelab, on to that. And VERY interested in the UCC85/UL84 guitar amp - will PM you.