12au7 as reverb driver

[John Clifford]

Looking to tame the reverb in my BFDR clone, I tried a 12au7 tube in place of the 12at7 in v3, as many people have suggested. I liked the sound, but the tube got super hot - much hotter than the 12at7. Could I fix this by changing the 2.2k cathode resistor to a higher value? Any suggestions?

[tubeswell]

Yeah you should bias the 12AU7 to be within Pmax. (5W for parallel 12AU7 triodes in same bottle)


This link should help https://www.vtadiy.com/loadline-calculators/loadline-calculator/

[John Clifford]

Thank you for the link! I’m not sure I know enough to use this tool properly, but here are my current assumptions:

1. The anode voltage on v3 is approximately 410v.
2. The reverb transformer is a reactive load of approximately 22k ohms, with essentially 0 DC resistance.
3. Therefore, the quiescent operating point for a single triode (1/2 the 12au7) would correspond to the intersection of 410v with the chosen grid bias voltage on the average plate characteristics chart for the 12au7.
4. This would result in approximately 3.2mA at a grid voltage of approximately -24v, which is well under the maximum plate dissipation.
5. So the cathode resistor for one triode at that operating point would be 24/.0032 = 7500 ohms. So far, so good (I think).

However, my circuit would use both 12au7 triodes in parallel. From what I’ve read, the graph stays the same, except the plate current values on the y axis are doubled. So that would mean dividing 24 volts by 6.4mA, for a combined cathode resistor value of 3.75K ohms, for which I would use a standard 3.9K resistor.

Have I analyzed this correctly?

I appreciate your help!

[tubeswell]

Yep FWIW, pushing the load line to the Pmax, at 410V and 22k I got a tube current of about 4mA for one triode at about Vg=23V. So 8mA for the pair. So  23/0.008= 2k8. (So in a standard series resistor, that’d be 2k7 if you’re not fussy or 3k3 if you want to be conservative) YMMV

[sluckey]

Looking to tame the reverb in my BFDR clone, I tried a 12au7 tube in place of the 12at7 in v3, as many people have suggested. I liked the sound, but the tube got super hot - much hotter than the 12at7. Could I fix this by changing the 2.2k cathode resistor to a higher value? Any suggestions?

I would just a dwell control.

[John Clifford]

Thanks again guys. Sluckey, I did put in a fixed "dwell" control of sorts, substituting a 330k before the grid and 680k after, which made the reverb pot more useable, but with pretty much the same sound. I liked the sound better when I changed tubes in addition to the above. I'm not really a huge fan of Fender-style spring reverb, even though I've had it in my amps for decades. So the more subdued sound of the 12au7 was more to my liking. At least this week. Next week I may have a different opinion. ;-)

[SoundmasterG]

Was just thinking about this....does the voltage gain of the tube have much effect in this circuit or is the current drive the tube can provide more important? 12AT7 is around 60 for voltage gain and 12AU7 is around 17-20. I suppose a 12AV7 would split the difference at around 41, but the current these tubes are capable of is probably more important I am thinking. In that case the 12AV7 can provide the most current drive, but I'd expect the voltage to drop. Since the circuit is designed to work with the 12AT7, I wonder how well the 12AU7 and 12AT7 match the specs of the reverb tank? I've always intended to try some other reverb driver tubes like the ones I mentioned, but I don't have a Fender amp and the amps I made for myself don't have reverb, so I guess it still remains a one of these days things. 


Greg

[John Clifford]

I finally got around to actually doing this. I ended up replacing the 2.2k cathode resistor with a 3.3k for the 12au7, and left the bypass cap in place. I'm very happy with the result. The reverb is less "springy" and more "roomy" if that means anything to y'all. Some of the highs are cut, so it doesn't get in the way of the clean signal, and just adds a nice, uh, "ambience." Exactly what I wanted. As previously mentioned, I also replaced the 1m grid resistor with a 330k/680k voltage divider, so the reverb pot has more usable range. The 12au7 is a little noisier than the the 12at7, but it's not noticeable unless you turn the reverb up high. So anyway, I would recommend this mod to anyone who finds the AB763 reverb too harsh and overpowering. Thanks for the help. This board is a terrific resource, and I'm in awe of the collective knowledge available here!

[HotBluePlates]

Looking to tame the reverb in my BFDR clone ...

A "better" way is to simply use a 100kΩ Audio pot for the Reverb control rather than a 100kΩ Linear pot.

Where is "plenty of reverb" on the Reverb control in the stock circuit?  Like at "2" or "3"?  When you change to an audio taper, the signal level of "3" now happens when the knob is turned to about "7.5".  You now have much more fine-control at lesser amounts of reverb.

[PRR]

...a 330k/680k voltage divider, ....The 12au7 is a little noisier than the the 12at7...

I suspect the tubes are within a hair of each other's hiss.

Looking into the 222k divider impedance is significant thermal hiss. Perhaps 6dB more than a more-sane impedance level. Are you certain you can't get your desired gain-range just by omitting the cathode cap?

===== OR= what HBP interjected while I typed.

[HotBluePlates]

...does the voltage gain of the tube have much effect in this circuit or is the current drive the tube can provide more important? ...

Neither, exactly.

It's a single-ended power amp, but we're only looking to throw a part-watt of power into the input of the reverb tank.

[John Clifford]

Looking into the 222k divider impedance is significant thermal hiss. Perhaps 6dB more than a more-sane impedance level. Are you certain you can't get your desired gain-range just by omitting the cathode cap?

I'm not following where you see 222k impedance. I though I was implementing a fixed "dwell" control with the voltage divider - or does that always add noise? I could try the audio pot instead, but really this works OK for my purposes. The sound of the reverb is a bigger issue for me than the gain control. The tube change got me the sound I wanted. I didn't try just removing the cathode cap - maybe that would have worked as well. I might try it next time I have the amp on the operating table. Thanks.

[John Clifford]

On the theory that a soundclip is worth 1k words, here is what my modified BFDR reverb circuit actually sounds like through a Shure SM57 into Logic Pro with no other eq or effects. Might not be your cup of tea, but I'm very happy with it. Please let me know if the link doesn't work. Thanks folks!

https://soundcloud.com/user-878907836/bfdr-reverb-mod?si=3b394a30f8f349c8a9c8ee51cd537184&utm_source=clipboard&utm_medium=text&utm_campaign=social_sharing

[glass54]

Nice Clean Recording John 
Regards
Mirek

[HotBluePlates]

... I also replaced the 1m grid resistor with a 330k/680k voltage divider ...

I'm not following where you see 222k impedance. ...

330kΩ in parallel with 680kΩ is ~222kΩ total resistance.

Resistance has thermal noise, in proportion to the resistance.  It's why using very-high resistances at low signal level spots is best avoided.

But really, PRR suggested simply removing the bypass cap because you added parts to knock the signal down, but the bypass cap is an added part to keep the cathode resistor from knocking the signal down.  You could save the cost (and noise!) of a resistor by omitting a cap.

... or you could have used my approach, which is to change the taper of the Reverb pot to one that accomplishes exactly what you want.  I suppose Fender must have picked the "bad" linear taper pot so the Reverb effect was intense & impressive.

[John Clifford]

330kΩ in parallel with 680kΩ is ~222kΩ total resistance.

Resistance has thermal noise, in proportion to the resistance.  It's why using very-high resistances at low signal level spots is best avoided.

But really, PRR suggested simply removing the bypass cap because you added parts to knock the signal down, but the bypass cap is an added part to keep the cathode resistor from knocking the signal down.  You could save the cost (and noise!) of a resistor by omitting a cap.

... or you could have used my approach, which is to change the taper of the Reverb pot to one that accomplishes exactly what you want.  I suppose Fender must have picked the "bad" linear taper pot so the Reverb effect was intense & impressive.

I'm certainly not arguing with you, and I hope you'll forgive my ignorance, but I don't understand how the 330K and 680K resistors are "in parallel" (see attached diagram of my changes). It looks to me like they form a series resistance to ground of 1.1M. I do see that I have now placed a 330K resistor in front of the V3 grid, but isn't this the same thing as adding a 1M potentiometer as a "dwell" control, but without the adjustability? I know a lot of folks do that. Is it a bad idea?

Again, my goal was not just to have a reverb control that I could turn "to 11." I wanted to change the sound of it. For reasons I don't pretend to understand, the mods I've made do significantly change the sound of the reverb, even when it's turned up ridiculously high (which it still can be). It sounds less like springs and more like a hall or room reverb. I like it a lot. I haven't tried removing the cathode bypass cap, so I can't say whether that would sound the same or different than what I've done. Do you know?

I do appreciate all the suggestions and patience with explaining things to a noob. Thanks.

[John Clifford]

“1.1M” should read 1.01M.

[glass54]

Hi John,
In ref to AB763 circuit (1964-1967), your 500pF/1M is a high pass filter, (the 1M also serves as grid leak for Reverb Driver). The 500pF is driven by V2b ("signal generator") and 680k and 330k forms an attenuator (just as you mentioned "Dwell" control). The Grid on your Rev Driver sees 680k parallel with 330k and thus 222k "resistor" as your new grid leak  (bear in mind we are being simplistic and ignoring Generator Output Impedance, stray capacitance, etc).
As HBT suggested, you may raise your thermal noise due to the 680k resistor (being a high value). However the gain of your chosen tube (12AU7) is lower than a 12AX7 so you might get out of jail. Anyway, its good to experiment and find the "Rev FX" you most enjoy.
Thanks for sharing and Merry Christmas from "Down Under"
Kind regards
Mirek

[PRR]

> I don't understand

Draw the whole circuit.

The 500pF goes to a lowish impedance point. Probably a 12AX7 plate node? Roughly 39k to ground?  Then 369K one way and 680K the other way is 239k effective. Confirm this by tacking 33k or 47k from the cap to ground and read the DC resistance at the tube grid.

Yes, the Fender drive pot has the same effect at 2/3rd electrical rotation. But since it goes quieter at 0 or 10 you don't perceive it as "fault".

[John Clifford]

Ahhh, I see it now. I wasn’t looking “backward” from the grid. Lesson learned. Thanks guys!