Squeal on cathode biased Deluxe Reverb, works fine in fixed bias

[afinitemind]

Hi friends,


Just finished up a Deluxe Reverb build today, and the good news is it's working perfectly in stock form and the voltages all match up within 5% to those on the vintage DR layout. Sounds great!


The trouble: my only mod was to install a fixed/cathode bias switch using Sluckey's schematic (on page 3 in that document). The amp still works perfectly in fixed bias mode, but for some reason I've got a painfully piercing squeal coming out when I power up in cathode bias. It's still passing signal fine and I can feel that squish, so it would appear that things are wired up correctly...but hey, I've been wrong before.


Had to step away for the evening, but I figured I'd ask the experts: Anything specific I should check for before I assume it's lead dress and start rearranging? The cathode resistor/cap are pretty close to one of the power tube sockets...

[afinitemind]

I guess I should mention that the OT is pulled from a dead PA chassis. It does have unused 70V taps that are all sealed off with heat shrink & secured.

[sluckey]

Temporarily disconnect the NFB wire. Any better?

[afinitemind]

Temporarily disconnect the NFB wire. Any better?

Good call. No squeal with that disconnected. Is this a ‘swap the OT wires’ situation?

[shooter]

that'd be a good start, leave NFB off, check, no squeal, add NFB

[afinitemind]

Works perfectly with primary swapped. Thanks for the direction!


I’m surprised, though...didn’t realize that the OT polarity/NFB issue would show up dependent on bias scheme.


Obviously I was wrong, but I assumed that I had the OT wiring correct because it worked fine in fixed bias. Can anyone help me understand  the ‘why’ in this situation?

[sluckey]

I'm thinking that you probably had the fixed bias set cold enough so the amp couldn't howl with wrong phase on the NFB. Then when you switched to cathode bias the tubes self biased hot enough to howl. Set the fixed bias to run at 70% and swap the plate leads back. I bet it will howl too.

A stock AB763 amp will definitely howl/squeal if biased at 70%.

[afinitemind]

Hmm, that idea makes sense. But I had the 6v6 fixed bias set right around 70%...20ma/tube with ~418 on the plates. Regardless, it’s working great now, so I’m  happy.


In cathode bias, it’s currently idling at about 11W per 6V6. How much would I gain by getting it to 100% (12W/tube)?

[sluckey]

My data sheets show 14 watts for max plate dissipation for a 6V6. That would put you idling at 60%.

Probably gain nothing your ears could hear.

[afinitemind]

I stand corrected...I was thinking max was 12W. It all makes sense now. Thanks!

[pdf64]

I’d expect a DR AB763 to oscillate if the feedback polarity was positive. That it required a lot of plate current to reach sufficient gain for that to happen may indicate something is a bit wonky.
Maybe the repurposed OT is putting a lower impedance on to power tubes than expected (thereby reducing the stage gain); have its winding ratios been checked?
Another possibility is a bad resistor value around the LTP or screen grids, though the voltage survey would hopefully have indicated something was amiss.

Note that cathode bias generally sounds a bit naff (overdriven tone thin and reedy compared to fixed bias) unless the operating point is close to class A (but don’t confuse ‘idling at 100%’ with class A; with a HT up there, I guess that class A would require an idle point way over that). Hence I suspect that cathode/fixed bias switches generally stay in the fixed mode after the initial installation and tinkering session , I know mine did

[shooter]

I guess that class A would require an idle point way over that

my self-biased, class A understanding, as long as the drive signal doesn't distort the output, class A.  typically ~~ = Vk minus a smidge 

[afinitemind]

Maybe the repurposed OT is putting a lower impedance on to power tubes than expected (thereby reducing the stage gain); have its winding ratios been checked?

You're probably on to something here. I tested it but please check my math! It's my first time working without a mfr's spec sheet for the OT, so mistakes are very possible 


Winding ratios for each tap: 42.5, 30.2, 22
Seems like 6800 ohm primary?
6800/winding²: 3.76 ohm, 7.45 ohm, 14.04 ohm speaker loads


I'm currently running 2x 16ohm 10"s in parallel, so I'm running off the middle (8 ohm?) tap.

[sluckey]

Winding ratios for each tap: 42.5, 30.2, 22

So, since impedance ratio is turns ratio squared, you have 1806, 912, and 484. Put a 4Ω speaker on the 1806 tap to reflect 7224Ω to the primary. Or put a 8Ω speaker on the 912 tap to reflect 7296Ω to the primary. Or put a 16Ω speaker on the 484 tap to reflect 7744Ω to the primary. I'd say all of that is a good match for those 6V6s.

A Deluxe Reverb uses the 8Ω tap for NFB. And the feedback resistors are 820Ω and 47Ω. If you want to run a 4Ω speaker you would move the NFB wire to the 4Ω tap and also change the 47Ω resistor to 100Ω. So, which OT tap are you using for NFB and what resistor value (47Ω or 100Ω) are you using?

[afinitemind]

Winding ratios for each tap: 42.5, 30.2, 22

So, since impedance ratio is turns ratio squared, you have 1806, 912, and 484. Put a 4Ω speaker on the 1806 tap to reflect 7224Ω to the primary. Or put a 8Ω speaker on the 912 tap to reflect 7296Ω to the primary. Or put a 16Ω speaker on the 484 tap to reflect 7744Ω to the primary. I'd say all of that is a good match for those 6V6s.

Thanks for writing it out for me! That math makes more sense - next time I need to let the 4/8/16 be constant and solve for primary impedance rather than the other way around.

A Deluxe Reverb uses the 8Ω tap for NFB. And the feedback resistors are 820Ω and 47Ω. If you want to run a 4Ω speaker you would move the NFB wire to the 4Ω tap and also change the 47Ω resistor to 100Ω. So, which OT tap are you using for NFB and what resistor value (47Ω or 100Ω) are you using?

That's how I'm currently wired up - 8Ω (912 impedance ratio) tap for the speaker out, 820Ω/47Ω NFB resistors off that same 8Ω tap.

[pdf64]

I guess that class A would require an idle point way over that

my self-biased, class A understanding, as long as the drive signal doesn't distort the output, class A.  typically ~~ = Vk minus a smidge 

That's the overdrive point of any type ₁ amp, ie that can't supply grid current to the power tube/s.
Class A requires that plate current doesn't cut off at any point, ie plate current is present for the whole 360 degrees of the waveform cycle.
See http://www.aikenamps.com/images/Documents/AmplifierClasses_from_RCA_RC12_1934.pdf

...A Deluxe Reverb uses the 8Ω tap for NFB. And the feedback resistors are 820Ω and 47Ω. If you want to run a 4Ω speaker you would move the NFB wire to the 4Ω tap and also change the 47Ω resistor to 100Ω...

Just to note that there no need to change the feedback tap used for the NFB when changing the tap used to drive the speaker, though I acknowledge that fender used slightly different feedback ratios for the 4 ohm and 8 ohm 2 channel amps, so to be an accurate copy, that would need to be accommodated.

[sluckey]

Just to note that there no need to change the feedback tap used for the NFB when changing the tap used to drive the speaker,

That is correct.