Wednesday, June 11, 2008

Testing the 12DW7 Phase Inverter Tube

A few years ago there was a lot of buzz on Harp-L and other Internet harp boards about the 12DW7 tube as a phase inverter in harp amps. I think Gerald Weber started it all when he discovered that an unbalanced phase inverter fattened up the tone and allowed the amp to be turned up to ridiculous volume levels before acoustic feedback occurs. Weber did this by inserting a potentiometer in the phase inverter circuit so he could dial in the exact imbalance level by ear, but the theory persisted that your could achieve the same thing with the unbalanced 12DW7 tube.

A little background…

In most tube amps with multiple power tubes, another tube (usually a 12AT7) takes on the job of splitting the signal to the pairs of output tubes and inverting (or reversing the polarity of) one of the streams. That way, one pair of power tubes pushes the speaker out and the other pair pulls the speaker back. This makes your amp more efficient. Weber discovered that an unbalanced signal – a signal where the push is stronger than the pull – improved harp tone and resisted feedback.

Tubes in the 12AX7 family, including the 12DW7, are dual triodes. That means they are actually two tubes in one. All the other tubes in the family are balanced, with both internal triodes presenting the same amount of gain or amplification. The 12DW7 is an odd duck, with one side having the same gain as a 12AX7 (a gain factor of 100) and the other side having the same gain as a 12AU7 (a gain factor of 20). As a phase inverter, this thing presents a very unbalanced signal to the power tubes.

On to the test…

Well, I was keenly interested in the feedback resisting part of this theory, but I never found anything online that persuaded me this really worked. I was not interested in using the 12DW7 to fatten the tone; there are other ways to do that. But if the 12DW7 was a feedback killer it would be gold.

So I ordered a JJ Tesla 12DW7 (ECC 832) from
Tube Depot in Memphis. I planned to try it in an amp where it could do the most good: A 1972 Fender Twin Reverb. This amp is a 100-watt high-gain beast, restored to original condition and unmodified for harp. It has a fresh quad of 6L6GC EH power tubes and an EH 12AT7 phase inverter. When turned up, it feeds back through my harp mic like nobody’s business.

For this test I placed a Zoom H4 digital recorder 4 or 5 feet in front of the Twin and let it capture the entire test. First I played for some time with the normal 12AT7 phase inverter in the amp, using a C Special 20 harp and my
Peavey Cherry Bomb mic modified by Greg Heumann. Here are the amp settings on the Twin:

Reverb Channel, Bright switch off.
Channel volume – 9
Treble – 1
Middle - 5
Bass – 10
Reverb and Vibrato are off

The 1972 Fender Twin Reverb is a master volume amp, but without the cheesy pull-boost. I rolled the master volume up to 3 ½ before feedback. The amp sounds good like this. The Altec speakers in my Twin are much less bright than the standard CTS speakers or the optional JBL 120s. To be sure, it ain’t Big Walter, but it ain’t bad, either. It sounds as you might imagine… kind of like an overdriven guitar. I like it.

I swapped the 12DW7 tube into the phase inverter socket and resumed the test. I sat in exactly the same spot with the same harp and mic, playing the same riffs. The results were not encouraging. The amp started feeding back at exactly the same place on the master volume. I could hear very little difference in the tone. I tried reversing the polarity of the speaker leads, but could hear no change.

When I first started thinking of doing this test I was talking about it with my friend
Joe Lempkowski, a talented harp player from the East SF Bay. He said he tried it a few years ago but could not hear any real difference. Neither could I.

When I loaded the recording into Adobe Audition I noticed the 12DW7 part of the test was actually slightly lower amplitude and slightly more compressed. When I normalized both tests and listened back to back I could hear a small difference in the tone, but I preferred the stock Twin configuration, not the 12DW7 phase inverter. If others have tried this and had a different outcome I’d love to hear about it

Here is a
link to a brief sound clip from this test. This was stock Twin Reverb, and I’m just tooting along testing different tones. I like the power and presence of the Twin. It worked for Magic Dick, but I don’t need to remind myself that he got that nickname for a reason, and I didn’t. It takes some harp Magic to tame the demon in this amp. The 12DW7 tube was no help at all.


Joe said...

I tried it in a Bassman LTD. I didn't notice a difference in tone or feedback resistance.

I tried the same tube in a Pro Jr. The results were the same as in the Bassman. No tonal difference or change in fedback resistance.

It is entirely possible that I have a hearing problem.

Rick Davis said...

Joe, thanks for checking in.

If you have a hearing problem, then I am mysteriously afflicted with the exact same problem. I could not hear a dime's worth of difference with the 12DW7 tube. And it doesn't take keen ears to hear that it doesn't kill feedback at all.

I'd like to hear from readers who tried this and got better results. It would be great if it worked...

Rusty said...

The tastiest sound I've ever achieved was through a Victoria 35210 at medium-low volume.

It was just me playing alone at a music store, and I am not sure how well the tone would have come across in a live gig situation, but it sure was nice solo - it nailed that Gary Primich sound from that tune "what's it gonna be" from the Company Man album, nice defined attack with a clean sound loosely cupped and a creamy distortion tightly cupped. That being said, the Victoria did not have the raw power and "cut" of the Meteor I am currently playing.

For small harp amps, I've played a couple of tweed champs, and both were magical. Would like to try out a Kalamazoo 1 or 2 but have not found one yet.

Anonymous said...

If you want to hear a 12DW7 inverter in a reissue 59 Bassman with no circuit mods, Kendrick blackframe speakers, 12AY7 input tube and CM mic, go to and listen to "Cadillac Alternate." That's one studio-recorded example I know of offhand. Mark Burness doesn't like the kind of distortion a 12DW7 inverter induces, but he put one in Pete's amp during that session and they both liked the resulting tone, though obviously a whole lot of the tone is pure Pete.

I'm afraid a master volume amp is the wrong kind to try a 12DW7 test on. Master volumes in push-pull amps are generally placed right before the phase inverter, with the pot determining the amount of signal entering the inverter. When a master volume is run in usual mode, turned down like Rick did, the influence of the phase inverter on the tone or feedback is minimized. It's not getting enough signal for the gain potential differences between its halves to matter. Like I said, wrong type of amp to test a 12DW7.

I don't think a 12DW7 inverter magically kills feedback in a RI Bassman. What it does there, compared to a 12AY7 or 12AU7 inverter, is keep some liveliness to the response and some 12AX7 gnarliness to the tone. Tends to distort monotonously rather than responsively, but a 12DW7 inverter, 12AU7 input tube and either RI Tung-Sol 5881s or TAD 6L6WGB shortbottle blackplates are a quick way to get a RI Bassman going. Squashier main tubes like these emphasize the uneven drive of the 12DW7 inverter. Kind of a band-aid fix, though, think I've done it to three friends' LTDs but it's no substitute for playing with biasing and reworking the preamp circuit even a little bit in those.

Joe, if you don't tell us what else was done to those amps, your observations on the 12DW7 don't mean much. No tonal difference compared to what? I would need to look at a Pro Jr. schematic, but if that amp's got a cathodyne phase inverter, then it's also the wrong kind of amp to expect much result from a 12DW7. Cathodyne inverters split the signal out of one side of the final preamp tube to drive both main tubes, rather than using one side per main tube, so the 12DW7 mu difference wouldn't matter there. My apologies if the Pro Jr.'s PI's not cathodyne.

I don't think it's necessarily a hearing problem at all, more of a testing problem. All the testing method in the world won't produce anything useful if the amp in question has a circuit that won't manifest the difference.

The thing about trying a 12DW7 as inverter is that it ~might~ do something useful there, in which case you can leave it there; but very likely it can do something cool in a typical single-ended Champ-style amp with a single 12AX7 preamp tube, more likely so if said amp has not been modified for less gain. If a silverface Champ is biting back, a 12DW7 can tame it without killing the tone; the 100 mu side drives the main tube hard enough to make things interesting, while the 20 mu side both reduces input stage gain and brings the input plate voltage down for easier breakup on attack there. Doesn't always work, but I try to always have one JJ ECC832/12DW7 around to try in such baby amps. It can also calm things around the homestead by getting the baby amp driving decently at a lower overall volume, reducing domestic strife.

Ask Ian Collard of Collard Greens and Gravy about 12DW7s in Champs. I dunno if he's put that amp on a record (latest CG&G record very highly recommended even at ruinous exchange rate with Australia), but it made his Champ much more usable for him.

Rick Davis said...

Thanks for your interesting comments on the 12DW7. As I said in the article, my only interest in this tube was its much-ballyhooed feedback fighting properties, not its contributions to tone. I prefer the more traditional methods to voice a harp amp.

I repeated the test with the master volume dimed and the channel volume reduced, and once again the tube had no better feedback performance than the standard 12AT7.

I realize the SF Fender Twin Reverb is not an amp suited for good Chicago harp tone, but it surely is an amp that could stand some discipline with respect to feedback, to which it is hyper-sensitive. The 12DW7 tube did not appear to have any benefits in this regard, no matter what setting I used on the amp.

Anonymous said...

Well, that's another drawback of master volume amps: Put the master volume in clean mode and play harp and the amp is . . . clean.

I checked the Pro Jr. schem and that amp is not cathodyne, conventional long-tail Fender inverter there, so I'm wrong about that. I tried different preamp tubes in a Pro Jr. in a store around 2001 and nothing really helped with that amp. EL-84 amps aren't much for harp unless you make some circuit changes to get the most out of the main tube for harp (Kazoo Model 1/2, Harmony H410, many vintage Lectrolabs). Those main tubes overdrive really easily and unequal phase inverter sections aren't going to have the impact there that they would with 6L6s. Write it off, move on: it's Fender's fault, not ours.

"Much-ballyhooed" claims about reducing feedback when using the 12DW7 as a phase inverter? The ORIGINAL post introducing the 12DW7 to the online harp community is available at in its entirety, though its original division into three paragraphs is lost in transition: Google "Cool pre-amp tube swap for harp amp geeks" if the above URL does not work. I know, because I printed out the two original posts about using the 12DW7 with big amps and with small SE amps when they first appeared in Dec. 2001, and can refer to the hard copies as I write. The audience is invited to use the above URL, read the text, and decide for themselves whether the 12DW7 was "ballyhooed" as a feedback reducer ~as the phase inverter in a big amp~. It's right there in plain English, folks: somebody in Italy cut & pasted the original post. My hat's off to anyone who can point out a clear claim to feedback reduction with the 12DW7 as a PI in that text. Fattening the tone, yes, but not feedback reduction per se.

In rhetorical analysis, it's called the "straw man" logical fallacy when you distort someone else's statements, or ~attribute statements to them that they never, ever said~. If anyone can find any mention of feedback reduction with a 12DW7 AS A PHASE INVERTER TUBE apart from a passing allusion to G. Weber's claims about out-of-phase speakers, in the above text, I will publicly apologize to them in this forum. The part about 12DW7s in other Fender preamp positions DOES NOT COUNT. Neither does 12DW7s AS A SE AMP PREAMP TUBE.

If you're going to claim "much-ballyhooed feedback reduction as push-pull amp PI", Rick, you've got to give chapter & verse on the ballyhooing. Pull up the URLs & dump them on us. Otherwise it's just you talking smack.

BTW, Rick, who was the author of the original Harp-L post I gave a URL for above? I'd like to see his name in print from you, please.

If you are really serious about making that early SF Twin Reverb something to play harp thru effectively, rather than just look at, I'll post some suggestions at that SFTR post. There is stuff you can do, ain't gonna hurt the amp's value at this point. Tough call when an amp is already something some guitarists would crave, but once you did the cap job, the door was open.

Rick Davis said...

Dear "Anonymous",

Geez... Have you tried switching to decaf? Calm down!

My article achieved what it set out to do: test the theory that the 12DW7 tube had feedback-reducing properties for blues harp.

Here is a URL for you: How To Keep Hostile Jerks From Taking Over Your Online Community

johnny said...

Rick that guy isa troll. He keeps shouting about tone but you said up front your test was about feed back. He just wants to argue and he keeps changing his story.

Thanks for this test of the 12DW7 in the phase inverter. I wondered about that.