Mesa Boogie Studio 22 Phase Inverter

[christarak]

Greetings from the recently burning, now flooding Australia. At least the fires have stopped.

22_cal.pdf

I have a Boogie Studio 22 (non EQ). This may eventually escalate to the Troubleshooting forum, but I haven't given up on the repair yet. This is more in the "What the hell was Randall Smith possibly thinking" category. My initial thought is that it was some cheap-skate LTP, but no. I have been staring at the PI, drawing equivalent circuits, calculating currents and I am pretty sure its a Paraphase PI. And if I'm right, its way way wackier design than an old Fender Paraphase.

The inverting triode is normal, but the input for the non-inverting (the lower triode in the schematic) is coming from the feedback loop. That sounds like a bizarre version of the "Einstein Chicken and Egg conundrum" to me. How can you feedback what has not yet been created, so as to create it by running it through a triode, then the output tubes, then the primary, then secondary stage of the O/P transformer and then feed a portion of it back to the non-inverting triode of the PI so you can see what you have just fed back? Surely the signal coming into the non-inverting triode would bear little resemblance to what's coming out f the inverting triode.

I thought in any PI you ideally wanted the 2 phases to be identical but 180 deg out of phase. In this design the 2 phases will be 5th cousins...at best. Then there's a presence control draining some of that high frequency signal to ground. I just don't get it.

Have I totally misunderstood the design? Opinions and mockery welcomed. (Enzo: This is where you would insert a comment about the phases actually being 180deg the other way to what I've stated because I'm from the Land Down Under! )

[Enzo]

Hey, don't stomp on my joke, it's all I have...


Everything must be in context. Yes the one side takes its signal "later" than the other. But remember, electricity moves at the speed of light. (Please no arguments about electrons moving) And that can go all the way around the earth seven and a half times in a second. That means the delay for the lower side signal from the first is on the order of a billionth of a second. And at audio, that is about as close to exactly the same time as it gets. So your signal has to go through the inverting triode, then from that plate, it must go through the power tubes, the transformer and the NFB to get to the other triode. 15-20cm maybe?

US NAval Admiral Grace Hopper is famous for her nanoseconds. She used to demonstrate nanoseconds to her classes. She handed out a bunch of 30cm pieces of wire, one for each student. "Here are your nanoseconds." She went on "this is how far electricity goes in a nanosecond. One billionth of a second.


Hopper worked in the world of computing, and in that world, that nanosecond was a limiting factor in computer speed. If the memory was two feet away from the CPU, that computer would be slower than one with the memory only 6 inches from the CPU.

Grace Hopper was an interesting old bird, and sharp as a tack. I wish she was still with us, she could give a much livelier explanation of all this than I could.

https://en.wikipedia.org/wiki/Grace_Hopper

We work in the world of guitar amps, and a nanosecond? Well, not so much of an issue.

[alexradium]

the first triode passes on the signal to the second one through the cathode connection,which comes out in phase to the plate,that's your inversion of the first triode,then you have the feedback insert in the grid,only difference from the fender is that it doesn't pass also through the tail resistor.
i don't know what's your problem,i guess you don't have a scope...

[Dave H]

It's looks like a LTP to me but with a negative supply to the tail and grids referenced to ground. The negative feedback isn't bootstrapped to the tail which makes the outputs unbalanced. They have compensated for this by having different value plate resistors.

[Helmholtz]

Yes, it's definitely a differential amplifier (LTPI). Note coupled cathodes.
With a scope you should be able to see equal but out-of phase signals at both plates - even if you disconnect the NFB.

[Chuck H]

FWIW this was a popular PI for Mesa. They actually use it in a lot of amp models designed over a span of more than a decade. So it must be working for them and their customers. In my experience it works very well. In fact, with the dozens of changes I made to a couple of Mesa amps with this inverter, the inverter was one thing I didn't change.

[g1]

The question reminds me of that EL34 100/100 Marshall power amp we were discussing the other day. I had the same 'chicken/egg conundrum' reaction.
V2a grid gets signal only from OT secondary tap.

[Helmholtz]

V2a grid gets signal only from OT secondary tap.

But that's typical for the non-inverting triode in a LTPI. However the tube has 2 inputs, cathode (non-inverting) and grid (inverting).

[g1]

But that's typical for the non-inverting triode in a LTPI. However the tube has 2 inputs, cathode (non-inverting) and grid (inverting).

Yes, this is the case for both the Mesa and the Marshall, correct? My big problem is that Marshall, in their usual modern schematic 'brilliance' has decided to go against all schematic and even normal reading convention; signal flow is from right to left instead of the well established norm (left to right).
Their modern 'service hostile' schematics always make me wish they would go broke.

Anyway, I guess my point is that for both myself and the OP, the difficulty is probably more due to the non-conventional drawing style, rather than the concepts. Ironically, both companies were established based on Fender designs, which they always seem so keen to avoid acknowledging in their drawings.

[Enzo]

Fair is fair, Fender got their designs a lot straight from RCA.

Seems to me Mesa drew schematics going the wrong way too.

MArshall draws what I call (perhaps incorrectly) engineering drawings. They seem to draw a schematic for just a board. So we wind up with things like orphaned volume controls or tubes lacking loads...or whatever. I had a long conversation with an engineer at PV once about this. I was asking if there existed system circuit drawings for certain products, like the 5150. The 5150 is one of those disjointed things, but somewhere SOMEONE must have had a drawing of the whole circuit to put the existing stuff together. It took me a long while to get him to understand my point. He was telling me they reuse snippets of circuits and all that. I finally told him to look at page 2 of the 5150 file. Upper left is a single triode cathode follower with the signal NEXT4 coming in and NEXT% leaving. I finally got across that I believe they will reuse that circuit, but I doubt the 5150 team started out with "I think I will stick a CF stage somewhere." We left with him sat least understanding my beef. We both agreed it might not change.

But then the 5150-2 came along et voila the schematic was entire...except for the power tube board.

[christarak]

Hey, don't stomp on my joke, it's all I have....

NO!! It's never boring when you say it, Enzo.

And yes, I may have got a bit wound up in the time delay....

[christarak]

the first triode passes on the signal to the second one through the cathode connection,which comes out in phase to the plate,that's your inversion of the first triode,then you have the feedback insert in the grid,only difference from the fender is that it doesn't pass also through the tail resistor.
i don't know what's your problem,i guess you don't have a scope...

My first thought was a LTP, and then I went into the rabbit hole. On a LTPI, I would expect to see a much higher voltage than 2V at the common cathode junction. I have a scope, and I confess to not using it to see what's on the cathodes. If we add the -16v to the 15K "tail resistor" R112, it still only is 18V compared to the grid. I should really find a way of disconnecting the NFB loop and see if the non-inverting lower triode still functions.

[christarak]

FWIW this was a popular PI for Mesa. They actually use it in a lot of amp models designed over a span of more than a decade. So it must be working for them and their customers. In my experience it works very well. In fact, with the dozens of changes I made to a couple of Mesa amps with this inverter, the inverter was one thing I didn't change.

I know Randall would prefer us to buy Mesa valves than biasing his amps.

Why is there a whacking big 2.2M resistor (R202 and R201) to ground at the grids of the EL84s? Is it to suppress any oscillations? Those 2M2's set up a huge potential divider with micro-amps. Any idea on how to effect a bias change? The 33K R111 would be ineffectual compared to the 2M2 for a global bias change, and it may upset the PI cathode +2vDC level. Have you ever rebiased one of these PI & Bias configurations, Chuck? Is there a better way?

[christarak]

My big problem is that Marshall, in their usual modern schematic 'brilliance' has decided to go against all schematic and even normal reading convention; signal flow is from right to left instead of the well established norm (left to right).

You gotta love consistency, G1. For this Studio amp, Mesa go right to left on the preamp schematic, and then go Left to Right for the PI and power stage.

[christarak]

Yes, it's definitely a differential amplifier (LTPI). Note coupled cathodes.
With a scope you should be able to see equal but out-of phase signals at both plates - even if you disconnect the NFB.

Thank you, Helmholtz. Yes those coupled cathode originally made me think LTPI, but the +2v stumped me. Seeing the grids referenced to earth instead of the bias/tail resistor then set me onto my Paraphase theory. I will disconnect the NFB. It may help get it through my skull.