The Phase inverter! The most important tube in your amp?

[PeaceMaker]

Hi all, I found this pdf about the Phase Inverter on the net and thought someone might find it helpful: http://www.guitaramplifierblueprinting.com/files/Phaseinverter.pdf... I put a EH 12AY7 in my Pignose G40v and I can't believe the difference- less noise, a lot more clean and headroom.

I also found this FAQ about the Fender Hotrod and Blues Deluxe Amps which also has info on preamp tube placement and other interesting info: http://www.justinholton.com/hotrod/faq.html#preamp.

I hope someone can get some use out of them.

Take care, Love,
PeaceMaker

[phsyconoodler]

I actually use a 12BH7 in my phase inverter to get a sweet clean amp that takes pedals extremely well.
  I've tried all the tubes and a 12AX7 makes my little amp grind like a bastard,but it's got no clean headroom.Good for small clubs where I just do blues or classic rock but not so good at the medium sized clubs.That's when I pop in a 12AT7 or the 12BH7 and it shines with a pedal board.

I prefer changing the PI tube rather than V1 cause V1 alters the tone too much for my liking on this amp anyway. YMMV

[HotBluePlates]

Hi all, I found this pdf about the Phase Inverter on the net and thought someone might find it helpful ...

Thanks for posting this.

There are some half-truths or incomplete information though in it. I agree with some of his big-picture end results, but not the how/why of it occurring.

Like what?

1. "V1 ... has the largest impact on your tone and gain but has less impact on your output distortion touch dynamics and output stage distortion than the phase inverter."

- Okay, I can buy that. An amp's tone has the biggest change when you try various tube in V1, but those expensive NOS tubes seem to have less impact on the amp's tone when you plug them in the phase inverter or output tube sockets.

2. "Some basic tube facts
      • 12AX7
          o Has a published spec gain of 100
          o Has a publishes spec current output of 1.2 milliamps
      • 12AT7
          o Has a published spec gain of 60-70
          o Has a published spec current output of 10.0 milliamps (ten times that of the 12AX7 as a side note)"

- Sorta. They have a published spec Mu of 100 and 60, respectively. Actual gain in the circuit is often a bit above half of these numbers, and depends on use in the circuit.

The phase inverter can use various circuits, and the gain of this position in the circuit can vary from less than 1 (split-load/cathodyne inverter by itself) to ~1/4 of Mu (long-tail pair, as used in most amps) to ~1/2 Mu (paraphase inverter, but just the first triode; second should have "no gain" or roughly the same result as the long-tail when considering the effect of both triodes).

"Has a published spec current output of ..." is either wrong or a bad way to think of the situation, because:
a. There's no spec for output current on the data sheet (plate current does not equal "output current")
b. Tube plate current depends on how you use the tube in your particular circuit, and
c. Most data sheets have show-off conditions which are far removed from the reality of how we use the tube in a guitar amp

What I mean is there are few times a guitar preamp tube is idling above a milliamp or so because we also use large plate or cathode resistors, and to have high tube current implies higher supply voltages than are common. It would have been much more accurate for him to say the 12AT7 has lower internal resistance and therefore lower output impedance.

3. "When you push your amp hard it is not as much the output tubes distorting as it is the phase inverter breaking down and distorting.  ...   The phase inverter may be the hardest worked tube in most amps."

- The second sentence looks like repeating Aspen Pittman's claim that you should replace the phase inverter along with the output tubes. It simply isn't true, and I believe was really about selling more 12AT7's when it was first said.

- The first sentence makes me think, "Are you kidding me?" Output stage distortion is mostly about output tubes distorting.

The saving grace for him is that it isn't only about output stage distortion. We'll revisit in a moment.

4. "The output tubes are less important than many folks may think. ... In the Hi-Fi world ... The output tube type had very little to do with anything. In guitar amps we purposely push the output tube beyond their design limits to make them distort."

- So this is one of those key places where I agree with real-world results that he heard, but disagree with why the results happened. Read on.

If you swap Svetlana 6L6GC's for RCA blackplate 6L6GC's, you might be disappointed by the lack of audible change in a given amp. I agree that the cost/benefit isn't as good as for preamp tube changes. The reference to hi-fi seems to have little relevance, given he then states that guitar amps don't operate tubes the way hi-fi operates them.

[HotBluePlates]

5. "In real life we rarely get to the point of pushing our output tubes to their limits.  Our front end is going into distortion."

- Folks who build Champs, tweed Deluxes, and other non-master volume amps might beg to differ. If there's no master volume, which stage distorts first depends on the design of the particular amp. Most (maybe all) of the classic amps were designed to have the output tubes distort first, so that you got every watt you paid for (especially since output tubes then, as now, were the most expensive tubes in the amp).

6. "Going from a 12AX7 to a 12AT7 in the PI (phase inverter) will yield a change in output tube distortion, touch, and output dynamics in most amps.  ...  The lower gain is a factor but the larger factors are:
      • We have almost 10 times the current available to drive the output tubes before the phase inverter starts to break down.
      • We have a transconductance of 5500 vs 1600 of the 12AX7.   Keeping this simple, it means it takes a lot less input signal for a given output signal."

- You're telling me that due to the transconductance, we get more output signal with less input from a 12AT7? Horsesh$t!!

So here's where he dropped the ball:

Most of the classic amps use feedback around the output stage, from speaker to the phase inverter (or the stage ahead of the inverter in a split-load circuit). The feedback acts to linearize the output stage and reduce audible variations with different tubes, as well as due to the tube aging. If your amp doesn't have output stage feedback, you will hear more variation with different tubes and will notice the effects of a tube aging and wearing out more readily.

Unless the output stage is run class AB2 or class B (or B2), the phase inverter wasn't designed to deliver current to the output tubes, so reference to "current output specs" is meaningless.

What I think he has observed/heard was that when we run the output tubes to distortion, the grids are driven positive. When the output tube grids approach 0v and are driven positive, they no longer look like an infinite input impedance and begin drawing grid current. Because the phase inverter was not designed to deliver this grid current, the phase inverter sees a much heavier load when the output tubes distort.

Because the 12AT7 has a lower internal impedance, it is better able to cope with the lowering impedance of the output tube grids when they distort. However, at some point, the apparent input impedance of the output tube grids will drop low enough to drag down the 12AT7 too.

Depending on the specific design of the inverter/amp, the heavy load either causes the phase inverter gain to drop, or distort or both.

We can't draw any general conclusions that apply to every amp, but it is fair to say that lower mu tubes with lower internal impedance are more likely to maintain their gain and output as the output tubes distort and draw grid current.

Whether that is "good" or "bad" depends on your perspective I suppose.

[DummyLoad]

we get more output signal with less input from a 12AT7? Horsesh$t!!

depends on the load. with 47K load AT7 AV shines over AX7, with 100K load AV with AT7 is slightly higher, with 220K load AV with AX7 is slightly higher. Vo max is greater with AT7 with same external component values, same B+, and under same load condition.

--DL

[HotBluePlates]

Vo max is greater with AT7 with same external component values, same B+, and under same load condition.

Yes.

To be clear, you're saying maximum output signal voltage is greater with the 12AT7. I'm on-board with that. But the original assertion was more output with less signal, in a situation where the input is a voltage signal. The 12AT7 can allow a bigger output voltage, but will also require a bigger voltage input. So what I'm really trying to say is the explanation is lacking.

Now if the load is a transformer, the 12AT7 will eat the 12AX7's lunch. You could actually make use of the bigger current available from the 12AT7. 12BH7 would swamp them both, even though it has lower gain and will need more input signal voltage.

But that's apples-to-oranges. I think I covered pretty well that it's the lower internal resistance of the 12AT7 that allows it to stay cleaner longer than 12AX7 in the same phase inverter, and as the output tubes start drawing grid current.

[John]

Just reading the parts that HBP quoted, I'm thinking maybe the author was aiming at guitarists who aren't builders really, just giving them some very basic info about swapping tubes?

[HotBluePlates]

True, but the amplifier blueprinting notion is geared towards individuals who are often novice technicians.

What I mean is it originally put forth the notion that an amp could be fixed if all resistors, caps and voltages were within a gnat's-ass of the schematic values. So unless he's posting this material to convince non-technical users to pay him to perform the blueprinting service, it appears to speak to those getting their feet wet in amplifier mods/repairs.

I realized that my earlier posts sound like I'm trying to put him down, and honestly I'm not. In fact, as I mentioned, I agree with his overall observations. I just wanted to clarify what really caused the results he describes, and try to correct some misleading information.

I learned a lot of my earliest amp knowledge from a series of books that, well, had as much marketing as true info. I was somewhat frustrated to find out I had a fair amount of false or incomplete knowledge to un-learn as a result. I'd prefer budding amp techs begin on a solid foundation of facts.

[John]

I'd prefer budding amp techs begin on a solid foundation of facts.

And why I'm glad I found this forum!

[Katie 77]

I use 12BZ7s in all my PIs....hits the output tubes harder.....kinda like an S&M situation...I should get a studded black leather tube shield for it
But DL is right on...you can't just swap tube willy nilly w/o looking at the entire circuit topology....but whatever works is the correct method.

[Tone Junkie]

I use 12BZ7s in all my PIs....hits the output tubes harder.....kinda like an S&M situation...I should get a studded black leather tube shield for it
But DL is right on...you can't just swap tube willy nilly w/o looking at the entire circuit topology....but whatever works is the correct method.

So when you use the 12bz7 do change values or just plug and play.
Thanks Bill

[HotBluePlates]

So when you use the 12bz7 do change values or just plug and play.

I'm gonna guess KT77's response. Plug n play.

The 12BZ7 is like a 12AX7, but with half the internal resistance and twice the transconductance. But tubes are listed with a mu of 100.

So if you pop a 12BZ7 in to a socket set up for a 12AX7, you'll get somewhat closer to the maximum gain of 100 as well as a lower output impedance for the stage. That's because mu (maximum possible gain) stayed the same, but internal resistance went down. The internal resistance of the tube works against the load resistance to determine how much of the possible gain we actually get in a voltage amplifier.

The lower internal resistance also means the 12BZ7 stage has a lower output impedance and is able to drive somewhat heavier loads while maintaining the size of its output signal.

You probably could optimize the circuit a little more for the 12BZ7, but the plug-n-play convenience is nice.

[plexi50]

I love this forum! Such a well explained breakdown of any thing you could ever want or need to know about amplifiers. 

[DummyLoad]

keeping all other components at the same values, except Rload, sim shows following:

with a 47K load gain is about 25 with BZ7, 20 with AT7, and 17 with AX7.

with a 100K load gain is about 30 with the BZ7, 23 with AT7, and 22 with the AX7.

with a 220K load gain is about  35 with the BZ7, 25 with AT7, and 26 with AX7.

just for S&Gs let's see what a 470K load does to all three:

gain is 36 with the BZ7, 26 for the AT7, and 28 for the AX7.

somewhere around 150K load the AT7 and AX7 will deliver similar gain. however, as i've already stated, overall the AT7 will swing the largest peaks because of the low internal resistance and relatively high mu. IMO, the AT7 is the best all-around choice for higher power amps where you must maintain output stage stability with smaller output valve grid leak values (that add more loading to the PI). for smaller amps where the grid leak loading of the output stage and gain is more desirable, then the BZ7 or AX7 are the go-to. 

KT, you could hang a cat-o-nine-tails in your amps to get the BZ7s to behave...right?

--DL

[jjasilli]

I think a lot depends on amp topology, and the desire for overdrive tone.  Simple Fender topology:  input stage > tonestack > tone recovery stage > PI > power tubes.  In this circuit, the only likely place to get overdrive in the small bottles is for the tone recovery stage to overdrive the PI; so the choice of PI tube is significant.  A post PI MV can keep the power tubes clean for overdrive tone at lower volumes.  OTOH, modern cascading gain stages can shape overdrive in the preamp while keeping the power amp, including the PI, totally clean.  This seems to be the approach favored by KOC. 

I use a 12BZ7 for the PI in my Ampeg ReverboRocket RI.  The amp has way to much gain with the 12AX7 preamp tubes -- the vol & gain controls are way too fast.  I tamed this with lower mu preamp tubes; but this left the amp anemic especially in the clean channel.  The 12BZ7 PI cured this.  Just some ideas re the choice of PI tube to complement the rest of the amp. 

[Katie 77]

So when you use the 12bz7 do change values or just plug and play.

I'm gonna guess KT77's response. Plug n play.

The 12BZ7 is like a 12AX7, but with half the internal resistance and twice the transconductance. But tubes are listed with a mu of 100.

So if you pop a 12BZ7 in to a socket set up for a 12AX7, you'll get somewhat closer to the maximum gain of 100 as well as a lower output impedance for the stage. That's because mu (maximum possible gain) stayed the same, but internal resistance went down. The internal resistance of the tube works against the load resistance to determine how much of the possible gain we actually get in a voltage amplifier.

The lower internal resistance also means the 12BZ7 stage has a lower output impedance and is able to drive somewhat heavier loads while maintaining the size of its output signal.

You probably could optimize the circuit a little more for the 12BZ7, but the plug-n-play convenience is nice.

yes

[Katie 77]

keeping all other components at the same values, except Rload, sim shows following:

with a 47K load gain is about 25 with BZ7, 20 with AT7, and 17 with AX7.

with a 100K load gain is about 30 with the BZ7, 23 with AT7, and 22 with the AX7.

with a 220K load gain is about  35 with the BZ7, 25 with AT7, and 26 with AX7.

just for S&Gs let's see what a 470K load does to all three:

gain is 36 with the BZ7, 26 for the AT7, and 28 for the AX7.

somewhere around 150K load the AT7 and AX7 will deliver similar gain. however, as i've already stated, overall the AT7 will swing the largest peaks because of the low internal resistance and relatively high mu. IMO, the AT7 is the best all-around choice for higher power amps where you must maintain output stage stability with smaller output valve grid leak values (that add more loading to the PI). for smaller amps where the grid leak loading of the output stage and gain is more desirable, then the BZ7 or AX7 are the go-to. 

KT, you could hang a cat-o-nine-tails in your amps to get the BZ7s to behave...right?

--DL

hmmmm, along w/ the black leather tube shields....that would be very nice 
I've tried the BZ7 in preamp, follower, gain recovery...and it sounds like shit...but it shines in PI position. A friend of mine out here uses them in his DR103s and SC100s to get a little more 'roar'

[HotBluePlates]

... read the article, but what flavor of PI is being discussed?

Yes.

[Katie 77]

I think Myles is referring to LTP Fender/Marshall type.......he is correct, in that in most amps, the PI clips first...and is generally hammered much more than the rest of the preamp tubes...but the power tubes aren't far behind...so there's not much merit in that 'white paper', to me.

My sentence structures suck, don't they 

[Tone Junkie]

 Once again I walk away knowing a little more, thanks to you guys.
Bill
                                                                                               

[HotBluePlates]

HBP; yes? 

...  The article does not discuss what type of PI is being used. ...

The amp builders on this forum use several varieties of PI, so anybody know which PI is being discussed.  ...

You didn't miss anything. What you're asking is my point; I think the article is a little too much sweeping generalization. So we can't say he's wrong exactly, but I'll bet money we can design an amp circuit where everything in the amp will distort prior to the phase inverter.

But yes, I think we might infer he's talking about the long-tail principally.

[Willabe]

I just noticed that the Fender Rumble Bass amp uses a 12BH7 for the CF driver after the PI to drive 6- 6550's.

http://www.el34world.com/charts/Schematics/files/fender/rumble_bass.pdf


                       Brad      

[HotBluePlates]

I just noticed that the Fender Rumble Bass amp uses a 12BH7 for the CF driver after the PI to drive 6- 6550's.

Yes, but it looks like a 12AX7 is used as a split-load inverter, with a 12AT7 boosting each drive signal (though not in a diff-amp configuration, which seems odd), and a 12BH7 as a cathode follower to drive the relatively heavy load imposed by 3x 6550's max grid resistance.

[12AX7]

Heres a question from someone who understands a LOT less than most of you here, so please bear with me. If i put a AT7 in the PI spot of a typical JCM 800 marshall PI, would it work better with the AT7 if i change the 10k to 47k?

[12AX7]

Well, it's said he T7 gives better dynamics. I just figured if the circuit is designed around a X7 the T7 might not give it's optimal performance in that regard. So dynamics, and clean is good too. This amp uses preamp for distortion so PV doesn't need to distort.  hate PI distortion too.

[HotBluePlates]

If i put a AT7 in the PI spot of a typical JCM 800 marshall PI, would it work better with the AT7 if i change the 10k to 47k?

Try the 12AT7 first with no changes.

Raising the 10kΩ to 47kΩ will generally make for a more balanced long-tail inverter, but the original value probably suits your amp and its power supply voltages best. Making the "tail resistor" bigger requires higher supply voltages if all else with the inverter truly stays the same.

[marshallguy]

So when you use the 12bz7 do change values or just plug and play.

I'm gonna guess KT77's response. Plug n play.

The 12BZ7 is like a 12AX7, but with half the internal resistance and twice the transconductance. But tubes are listed with a mu of 100.

So if you pop a 12BZ7 in to a socket set up for a 12AX7, you'll get somewhat closer to the maximum gain of 100 as well as a lower output impedance for the stage. That's because mu (maximum possible gain) stayed the same, but internal resistance went down. The internal resistance of the tube works against the load resistance to determine how much of the possible gain we actually get in a voltage amplifier.

The lower internal resistance also means the 12BZ7 stage has a lower output impedance and is able to drive somewhat heavier loads while maintaining the size of its output signal.

You probably could optimize the circuit a little more for the 12BZ7, but the plug-n-play convenience is nice.

That’s what I had in one of my amps which at first swapped from 12AX7 to a 12AT7 driving 2 KT90’s with 120k grid resistors.  Bought 2 at ham radio exhibit a few years back. Decided to put the 12BZ7 back in the P.I. and audio balanced it. Swapped it for the spare one and reading was also audio balanced. Good balanced tubes for sure at least in my case.

[kagliostro]

I've an old Davoli Krundaal Show II that has EL500 as power tubes and ELL80 (double pentodes connected as triodes) used as PI

As the ELL80 tubes are scarce and expensives a friend adapted a 6CG7 tube as PI and he told the power section was pushed up and reached an higher output power



Reading here I'm near to understand the reason for this good result but still not arrived to the 100% of understanding

Freanco

[tubeswell]

Clickbait warning- this thread is over 10 years old. The original poster gave up checking replies already.

[kagliostro]

Ciao Tubeswell

I readed from the telephone and didn't realized what was happening

thanks for the warning

---

BTW: Today I posted this 

https://el34world.com/Forum/index.php?topic=16963.msg338577#new

Franco

[marshallguy]

Clickbait warning- this thread is over 10 years old. The original poster gave up checking replies already.

My bad, HotPlateBlues wrote something I was reading and I replied.