Super Reverb Blackface Mod Question

[Rhodesplyr]

First, let me say that my expertise is in Hi-Fi, Hammond Organ, and Leslie tube amps.

A friend of mine asked me to take a look at his 1970 Fender Super Reverb, AA270 circuit; he's had problems with an output tube going into thermal runaway. He told me (several times) that he'd had the amp converted to Blackface specs by a previous tech, which, as far as I've been able to discover means making the relatively minor, but sonically significant circuit changes to the AB763 circuit.

The amp had been converted to adjustable bias and had the 100k 6L6GC grid resistors changed to 220k, but the PI plate resistors were still both 47k rather than 82k/100k. To my mind, changing the phase inverter resistor values seems one of the more significant parts of the mod--other than snipping the shunt capacitors on the 6L6GC grids.

Since I'm ignorant of this, did my friend get what he paid for when he paid for a "Blackface Mod" if the tech left the 12AT7 phase inverter plate resistors at their AA270 circuit values? To change one circuit to another makes sense to me. To change it halfway, but not all the way leaves me puzzled.

David

[stokes]

When "Blackfacing" a Silver Face amp,you arent entirely changing the amp into a Blackface,you are making it closer to Blackface specs and you will often find that it is done to a different level by different techs.If you read the procedure to do it by different authors you will likely see some differences in how far it is taken by different techs.Myself,I would change the PI plate resistors along with other mods,and I would tell the client exactly what changes I will make.I had one client who had "read someplace" that it only required changing the bias scheme,and insisted that was all he wanted.So whether or not your friend got what he paid for,depends on what conversation he had prior to the work being done.If you look at the schems,the AA763 has two 100k plate loads and the AB763 has the afforementioned 82k/100k plate loads,both Blackface amps.Granted they arent 47k's but only the tech that did the job can say why he left the 47k's.

[Rhodesplyr]

Stokes,

Thanks for the reply. To my mind, if you only changed the bias from balancing the output tubes to adjusting the range for both tubes, you now have a Silverface amp with adjustable bias. I mean, let's face it, you could adjust the bias to be pretty much the same as the Silverface bias, so what's the difference?

With Leslie amps, we occasionally change a 147 unbalanced input amp to a 122 balanced input configuration, but leave the 147 output circuit because they do sound different from the 122. In that case, we call it a hybrid amp.

I find "Blackface conversion" a confusing term if it can mean any mod that brings a Silverface amp closer to Blackface circuits.

Another friend had me help him replace output tubes in a Silverface Twin Reverb, and I tried to use the bias control to balance the tubes. Eventually, I discovered that it had been changed to adjustable bias by a previous tech with no notation on the amp. When I change a 147 Leslie to 122 specs, I change the label so no one can make that kind of mistake.

Best,
David

[hasserl]

IMO the 47k plate loads are the "correct" value for the 12AT7 PI, Fender got it right with the sliver face amp PI; I'd leave them alone.

[mark]

Why are 47K plate load resistors the correct value?

No flaming, I'm wondering what your opinion is based on.

Yours Sincerely

Mark Abbott

[hasserl]

Because of the lower plate resistance of the AT. Compare the data sheets between a 12AT7 tube and a 12AX7, the AX7 has 5x the Rp of the AT7, so the larger values work fine for a PI using an AX7. With an AT7 in use for the PI the lower 47k value gives increased headroom and will allow the PI to drive the grids of the power tubes better. The amp sounds more open when playing clean, will stay clean longer and when pushed into OD will drive the power tubes farther into non-linear operation before grid blocking occurs. It's a beautiful thing, leave it alone.

[stokes]

"Thanks for the reply. To my mind, if you only changed the bias from balancing the output tubes to adjusting the range for both tubes, you now have a Silverface amp with adjustable bias. I mean, let's face it, you could adjust the bias to be pretty much the same as the Silverface bias, so what's the difference?"
Dave,I wasnt trying to imply that merely changing the bias balance to a BF bias control would constitute a BF conversion,IMO,just pointing out the different opinions out there.I also wouldnt refer to anything as a "BF conversion" just a "BF mod".You are just changing some components to get somewhat closer to BF specs.It is done by people to "improve" the tone of the SF amps,so what needs to be changed to that end will vary from one tech to the next.As you can see from hasserl's reply,he's one who feels the PI plate loads are fine,and offers a very valid explanation for it,so I assume if he did a BF mod he would leave them as they are,and who's to say he's wrong."Improving" and amps tone is a very subjective matter,so if the amp sounds better to someone then they "got what they paid for".

[Rhodesplyr]

IMO the 47k plate loads are the "correct" value for the 12AT7 PI, Fender got it right with the sliver face amp PI; I'd leave them alone.

Thanks for expanding the discussion.

What's your position on 100k vs. 220k grid leak resistors in this amp? I ask because the reason my friend asked me to look at the amp was that the 7581As he was using were going into thermal runaway. It is a grid emission problem, not a bias failure problem. The amp runs fine for a few minutes until the tubes really start to heat up; then grid emission starts to snowball (bad metaphor, I know).

His previous tech changed the 100k AA270 value to the 220K AB763 value, ignoring, as Fender did, the published limit for 6L6s in fixed bias (100k). I can't help thinking that these two things may be related.

I tried several pairs of 6L6GCs in this amp, and only a used/good pair of GE 6L6GCs (grey plate, clear top) showed no sign of grid current. (I'm talking about small amounts--and stable--not going into runaway.)

Other clues: one output tube coupling capacitor was replaced by a previous tech, and the octal socket contacts were really filthy. I've meticulously cleaned the socket contacts and am replacing both coupling caps with a matched pair of Mojo's Dijons, tested for outer foil, oriented towards the 12AT7. I don't go for fancy marketing. All I ask for is film & foil construction and solid copper leads :-)

I'm in the same state as Mojo, and they got me the order by mail the day after it was placed. Good service. The Dijons have extremely low DF according to my impedance bridge. I'm almost suspecting that they're polypropylene, not polyester.

David

[hasserl]

What's your position on 100k vs. 220k grid leak resistors in this amp?

Good question, here is another area where I think Fender got it right w/ the SF amps. The 220k grid leak resistors (I prefer to call them the bias splitters or bias feed) increase the tendancy towards grid blocking, or farting. A well set up silver face Fender will drive the power tubes better and farther into non-linear range for more output tube distortion with less of the dreaded "Fender Flatulance".

Randall Aiken explains it like this:

In addition to an increase in crossover distortion, if the time constant of the AC coupling is large enough, and the transient waveform is of low enough frequency, a transient distortion known as blocking occurs. This happens when the transient signal quickly pushes the clamped grid waveform down far into the cutoff region, and there is a finite time that is required for the grid waveform to recover and float back up to the correct bias point once the transient signal is removed. Until the bias comes back to the correct point, the output stage is effectively cut off for a major portion of the signal. This results in a choppy, "farting" sounding distortion.

Fender amps are particularly susceptible to this because of the large values of coupling capacitors on the grids of the power tubes (0.1uF). Blackfacing your Super Reverb can actually make the problem worse, because you change the grid bias feed resistors from 100K to 220K, which increases the time constant of the AC coupling to the output tube grids.

I would swap out the PI input cap though to .001, that is an area where they got it right on the black face amps. The larger .01uf cap on the sf amps allow too much bottom end thru to the PI, increasing the tendancy toward the flatulance problem again. Cut that back to .001uf and you will help control the farting when the amp is pushed hard. Beano for amps!

[Enzo]

In the first post you mention "an" output tube running away, not all of them. If that is the case, it is not an issue of the bias arrangement. Latre you did use the plural. Which is it? And when saying "tubes," does that mean all four or just two?

If it is indeed the case of one tube running away, then swap places with it. Does the same tube run away in a different socket, and does a different tube in that socket also run away?

Monitor the control grid voltage when this happens. Is it stable or does it drift downward? A leaky coupling cap could easily be at fault there. Amd that would affect two tubes on a side.

One bad tube on a side can kill the bias to its mate on that side.

[Rhodesplyr]

Hasserl,

I tend to agree with you on all your points about what's good about the AA270 Silverface design. I actually build my own Hi-Fi tube amps, so I'm familiar with blocking distortion, etc.... On Leslie amps, you can tune how some of them begin to distort by varying the cathode bypass cap for the output tubes. If you make that capacitor too big, it takes the circuit too long to recover from an overload condition, and you get the "farting" distortion.

I just gave the Super Reverb back to my friend, and I told him that if he wants it tweaked, I'd look precisely at the .01uF capacitor feeding the phase inverter. It seems like you could tune the amp by choosing any value from .001uF to .01uF. I hadn't read your post then. You could also choose different dielectric materials.

I got him a NOS 1950s Mullard phase inverter tube with pretty closely balanced sections, and we decided on new production Russian Tung-Sols for the inputs to both channels. Outputs are a good/used pair of GE 6L6GCs. I replaced the severely mismatched coupling caps to the output tubes (one had been replaced) with a matched pair of Mojo's Dijon brand. I don't go in for boutique hype, by the way. They just have film & foil construction and solid copper leads. That's all I ask.

The one thing I'd give Fender an 'F' on is the location of the bias pot. After putting the amp back in the cabinet, I realized how hard it was going to be to get to. It has to be the case that Fender simply didn't want people messing with it. Do I burn my hands, or put a screwdriver through a speaker cone? If it were mine, that pot would get relocated to the back panel for easy access.

Thanks for all the benefit of your experience!

Best,
David Anderson

[Rhodesplyr]

My friend came over to play his Super Reverb, but....he didn't like the way it sounded, complaining that it sounded too distorted. What he didn't like was common to both channels, so that narrowed the possibilities.

So, back to the drawing board I went. Sweeping the amp with a sine wave, it's putting out 50-55W before it starts to clip, but there is an edginess to the sound that I can hear and see on the oscilloscope.

It occurred to me that I'd broken one of my own rules of replacing coupling capacitors: Thou shalt not upgrade the coupling capacitors without paying attention to the components in the feedback loop around them (if there is one). Realizing the 0.1uF shunt capacitor on the 12AT7 phase splitter grid was in the feedback loop, I tacked a good polypropylene film & foil type in place of it. Instantly, the edginess vanished both to my ears and on the o'scope screen. Ahhhh.....

He's going to come back, and we're going to play with values and types for the .001uF (BF) .01uF (SF) capacitor. I have a capacitor decade box, so I can dial in those values and any in-between while he plays. I've got polystyrene, polypropylene, and paper capacitors in those values, so I can make it as smooth as he wants. I wouldn't feel 100% comfortable using paper capacitors between the PI and the 6L6GCs with ~350V DC across them, but the ceramic disc sees much less. (When I say paper, I mean hermetically sealed paper in oil, not wax dipped!!) Since that capacitor has no feedback around it, it's a good candidate for voicing the amp.

One other thing caught my attention this time: at some point, before he bought the amp, someone had one of the original Fender CTS alnico speakers reconed, and whoever did it used a noticeably softer paper cone and softer suspension than the stock speakers. It's almost like someone was reconing it as 10" home audio woofer rather than a musical instrument speaker. I think it's contributing to distortion he complained about in the bass.

I'm beginning to understand the appeal of building from scratch and selling your own amplifiers. I'm sure I could have built a brand new Super Reverb in the time it's taken me to figure out what was wrong with this one.

David