Push/Pull Pair Removal Theory & Questions

[nickb]

Yes, that range is only specified (-1dB @ max rated power) down to 150Hz.
The inductance for the tube guitar amp 1750* OT range is noted on each individual info sheet, as per previously linked example.
All except the 18watt and AC15 replacements look to be >50H, -1dB 50 - 70Hz.

I think you have to take the Hammond values with a pinch of salt. It's measured with the secondary open meaning the current waveform is very distorted since it's all magnetizing current. I suspect they just measure the current as an RMS value and calculate the inductance from that.

[bob p]

Not the ones I just looked at. I know there are others, because I've seen them along the the way in life. But a *oogle search reveals a lot of Ceriatone models (not sure if those should count?) and a couple of Mesa schematics. None compensate for impedance.

I think it's safer to say "None compensate for impedance with a switch."

As I mentioned in my previous post, early generations of Mesa amps required the user to do this manually, and later generations of Mesa amps implemented this through relay switching (I think). That is to say I seem to recall Mesa specifically implementing switchable loads through relays in their later generation of switching amps, though I don't have a schematic in front of me that can confirm that. If they did patent the idea, which is the typical Mesa MOO, then that would explain why nobody else is doing it, and why you came up with an empty search result.

Another consideration: I'm not sure that it's logically sound to assume that you'll determine whether a designer contemplated changing load impedance during power selection by looking for the existence of OT switches. While the method of looking for switches will reliably identify a true positive result (where switches exist), it could give you a false negative result (where no switches exist). So a little bit of closer examination would be necessary before we're sure about this.

One thing that the schematics normally fail to mention is the turns ratio in the transformer. One thing that is worth considering is that a designer might have deliberately chosen a non-centered value for the OT primary impedance under one set of operating conditions in order to specifically provide added tolerance for a mismatch under a second set of operating conditions.

In the old days when Fender was designing the classic SF and BF amps, Fender didn't directly embrace the mismatching idea. The OT turns ratios were chosen to provide optimal performance with the speaker array being sold with the amp, using design center values that were typically chosen to optimize power output. As an example, Fender sold a Twin Reverb with an OT that was optimized for 4x6L6 with a 4R 2x12 load. If you needed more volume (4x12) then you had the option to plug in a 2x12 extension cab, but you paid a price by doing that; the problem was that the extension cab would shift the 'operational characteristic' of the tube to the left on the power vs. impedance plot, which typically resulted in decreased power output for the amp as you gained acoustic power output from the extra speakers. It was a trade-off. Fender offered a 'better' solution in the Quad Reverb that used a different OT to drive a 2R 4x12 load, while keeping the primary side of the amp running in the optimal area of the 6L6 'operational characteristic' chart.

The same could be said for the idea of pulling two tubes. You have that option with a 4-tube Fender, but pulling 2 tubes is going to shift you to the right on the operational characteristic. Distortion will increase and power output will diminish. Not what Fender intended, as they were designing loud and clean amps, so the 'better' solution from Fender was to downsize to a Pro Reverb.


But things are different today. Today it can be desirable to move an amp over on the operational characteristic to increase distortion while decreasing power output, and some designers might be deliberately making those choices to obtain a desired effect. What a lot of people tend to overlook though, is that when you do this you're deliberately choosing to run an amp off of the centered op characteristic, which is going to diminish the amp's tolerance for a mismatch in either direction. Like Enzo says, there were never any tight specs in guitar amps... but specs are becoming tighter today, because designers are deliberately moving to off-center op characteristics. The conscious decision to move off-center means that tolerance for bidirectional mismatching must be lost. With off-center selections you can end up with twice as much latitude in one direction, but zero latitude in the other. In today's market it's important to understand that. That's why some of the boutique amps specifically tell you that the amp will tolerate some forms of mismatch but not others, and they DO fail when they get driven outside of their intended operational parameters.

Getting back to your point about looking for switches -- there is one pitfall in that approach. There could be cases where an astute designer deliberately chose a lop-sided operational characteristic just to provide increased one-directional mismatch tolerance when an unswitched load change is made. I've seen some amps that specifically instruct users that it's OK to mismatch in one direction but not the other. Reading between the lines, I think this is what they're doing. So it may not be 100% correct to conclude that designers aren't taking into account load variation when they do power switching. It's entirely possible that they selected an OT during the design stage to preclude the need for the switch that we're looking for. The best way to determine this would be to take measurements on the transformers. There's really no other way to be sure.

[Enzo]

Check page three of the Mesa Mark 2b owner manual.

https://srl.utu.fi/~tke/MesaBoogie/Manuals/Mark-IIb.pdf

It does explain if you use two 8 ohm speakers, plug them both into the 4 ohm jacks. But it also suggests trying other combinations to see what you prefer, as no damage will result. It goes on to suggest on 60/100 amps you use 60w with an 8 ohm speaker plugged into the 4 ohm jack. But it also suggests you can flip it to 100 watts that way for more power, and then plug into the 8 ohm jack for even MORE power. SO clearly they are explaining that matching speaker to amp impedance is not necessary, even if it is the preferred combination.

[bob p]

Check page three of the Mesa Mark 2b owner manual.

https://srl.utu.fi/~tke/MesaBoogie/Manuals/Mark-IIb.pdf

There's an illogical propensity on this forum to find an isolated case that refutes a specific premise and to proceed as if it proves some sort of universal truth. It doesn't. It's just an example of violation of a rule or a pattern.

Selecting a single case that violates the rule doesn't nullify the rule, just as selecting a single case that supports the rule doesn't justify it. In response to your citing the Mk IIB owners manual to nullify the rule, I'll cite the MK IV manual to justify it. My point isn't to say that you're wrong or that I'm right, it's to illustrate that Mesa was inconsistent in what they published in their owners manuals as time progressed, and that they changed the content of their manuals over time, for a reason. History will show that they after changing the wording of their manuals they even changed the design of their amps to provide programmable load switching.

Did anyone here ever purchase one of the Mk I or Mk II amps when they were new? Not the Mk IIB reissues, but the ORIGINAL Mark Series amps? Back in the day the Mesa manuals were photocopied from a master document that was written in Pica on a typewriter and then Xeroxed. I know that I did, and I know that Gil Ayan did, once upon a time. Back then the original manuals from Mesa were pretty thin. They suffered from a lack of information. They're very unlike what they are today. Looking at the linked Mk IIB manual -- it has obviously been re-vamped using modern desktop publishing tools and is obviously non-original.

... it does explain if you use two 8 ohm speakers, you should plug them both into the 4 ohm jacks. But it also suggests trying other combinations to see what you prefer, as no damage will result. It goes on to suggest on 60/100 amps you use 60w with an 8 ohm speaker plugged into the 4 ohm jack. But it also suggests you can flip it to 100 watts that way for more power, and then plug into the 8 ohm jack for even MORE power. SO clearly they are explaining that matching speaker to amp impedance is not necessary, even if it is the preferred combination.

Yeah, the Mk IIB manual says you can do anything without damaging the amplifier, but we all know better.

I don't see any value in quoting the Mesa manual and saying that everything is OK because they said it is OK, when we all know that just isn't true. Try hooking up a 16R Marshall 4x12 up to the 4R jack on the Mk IIB and running the amp hard... The Mesa manual will tell you that everything will be fine. Or even worse, try running a 2R load into the 8R jack and running it hard. Tube amps may handle a high side mismatch alright, but they sure don't like a severe low-Z mismatch. The reality is that kind of shitty "it's all good" advice led to fixing a lot of broken amps for people who followed the directions and thought everything would be OK. BTDT.

Like I mentioned in a previous post, the early the Mesa manuals weren't really written all that well. They didn't have much of any documentation in them. They were all that you might expect from a guy who was building amps in his garage. As the company grew they tended to have a lot more content and they even had professionally drawn graphics in them. But they read like they were written by a marketing team that's following an engineer's recommendations to the extent that they understood them, but they continued to make all sorts of silly/unrealistic claims about the amps' do-it-all performance. The manuals were often quite ambiguous and you could interpret them to mean whatever you wanted them to mean, especially if you want to be selective about paying attention to what they said, or if you considered that they guy writing the ad copy didn't really know what he was talking about. That level of ambiguity is great when you blow up a Mesa under warranty and claim that you were following the instructions. They didn't hesitate to fix it. But that kind of ambiguity was not so great when you blew up the amp out of warranty and you had to pay to fix it on your own. That's when the wisdom of gross mismatching doesn't look so much like wisdom after all.

Mesa has been a bit inconsistent in their writings about this. I bought my first Mesa amp (what would now be called a Mark I) when Smith was still building in his garage. The manual didn't say squat about mismatching. When I got my new Mk IV years later the manual made contradictory comments about getting away with mismatching. On one hand they'd say that you could run 6L6 or EL34 and 60/100 power switch at will without worrying too much about proper load matching, though they did qualify that you had to make manual adjustments like putting the 8R load into the 4R jack. But they stopped short of saying something grossly irresponsible, like "whatever sounds good is good," like they did in the earlier era of the Mk II, perhaps because by the time of the Mk IV they had repaired a lot of broken amps, and now they knew better.

In the Mk IV manual they specifically said you could do some crazy things, like replacing a quad of 6L6 with a quad of 6V6. But they recognized the perils of power switching to a pair of 6V6 and proceeding with a 200% mismatch and they made specific warnings in the manual that you must never engage the 60/100 switch in the 60 position with a mismatched load impedance. They specifically warned that you must keep the simulclass switch in the simulclass position while concurrently plugging an 8R speaker into the 4R jack or risk damage to the tubes. Why? Because a 200% mismatch just isn't a very good idea.

Decades ago, when I first got the Mk IV, I called Mesa support about this and I didn't get the cavalier "do whatever sounds good to you" response like the tone of the Mk IIB manual you cited. I actually got routed to someone in engineering, rather than in the marketing department, and they told me that there was risk associated with load mismatching, and while they wouldn't replace burned out tubes which were the most likely failure mode in a mismatch problem, they would replace the iron if mismatching took out my iron when it took out my tubes. They made it clear that mismatching was risky, that I should expect that it would eat my tubes, and that I could rest assured that they'd fix the amp for me while it was in warranty. Once it was out of warranty, then I'd be on my own. That policy speaks for itself.

Suffice it to say that in the era of the early Mk Series the guys at Mesa started off by not even talking about mismatching (Mark I), then modified their stance to say that you could mismatch will-nilly (Mark II), and they fixed the amps under warranty that got broken by doing that. Later in the era of the Mk Series, by the time that the Mk IV came along, they began to provide some explicit warnings in their owners manuals about being careful to avoid some mismatch conditions. Then, when the next generation of amps came along, the Road King offered programmable switching relays to perform load switching on the fly as power tubes were switched on/off. You can read the old Mk IV manual if you want to verify this (I've already posted the Mk IV manual on this site) and the current version of the Road King manuals (Road King, Road King 2 v.1 and Road King 2 v.2) are still available on the Mesa site. It's unfortunate that even after they designed the automated switching circuits to cure the load problem, the quality of their manuals remained pretty ambiguous about proper loading -- for a reason.

Mesa is aware that mismatching will get you into trouble if it's done willy-nilly, and history has shown that they'll repair the amps that fail under mismatching under warranty, because it's good for business to do that. While they've fixed the amps that have failed, they've changed their documentation to lessen the number of warranty claims, while avoiding coming right out and saying that load mismatching could kill your amp. Why? Isn't it obvious that they make a hell of a lot of money by convincing people that mismatching is safe, when the reality is that it's only kind-of safe, and isn't really totally-safe? By selling the illusion that you can mismatch willy-nilly, they sell a lot of amps that offer fancy switching options. They only have to repair a limited number of amps that fail during the warranty period, and they don't have to worry about any of the ones that fail out of warranty. In terms of simple economics, they make more money selling the illusion of safety than they spend on fixing the failures. In that light, it's no surprise that the manuals are written the way they are. I think that somebody would be silly to put much faith in the Mesa manuals, especially somebody who knows better.

Of course none of this changes the fact that some of today's boutique amps are not design-centered. for example, one famous manufacturer of boutique EL84 amps deliberately runs them at a 100% mismatch to the low side. If anyone is aware of this, he'd have to be pretty stupid to run it at a 100% mismatch to the upside by pulling a couple of tubes and not adjusting the load, assuming that there won't be a problem with what amounts to a 200% mismatch. There's just no sense in assuming that a 100% mismatch by the user is always safe, particularly not in amps that are already mismatched by design. The margin of safety does not extrapolate that far off-center.

[Enzo]

Bob, I was service manager at a Mesa dealer at the time the Mark series (Mk2 included) was current model. The manuals said exactly what I posted. I don't care if someone has cleaned up the presentation. We received the amps in two boxes: one was the amp/cab, the other the speaker. Together they were too heavy for UPS at regular rates. We were to assemble the speaker into the cab, and instructions were as posted. I don't care how thick or thin the manuals were, it only takes a paragraph or two to explain plugging in the speaker. Right, wrong, smart, stupid, Mesa told us to plug the 8 ohm speaker into the 4 ohm jack.


While I appreciate all the time and effort you put into laboriously explaining how the thinking of others is all flawed and wrong and all, but really...

The point of providing contrary examples is to show that some "rules" are not as absolute as some people would make them out to be. It isn't to prove anything, let alone a "universal truth, it is to assure people they have options.

And if you read my post for what it was rather than with your tiring defensiveness, you'd see that what I really did was expand a little on what you said. It wasn't done with a switch, they did in fact explain it, as you said. The only thing I added by so posting was they they said it was OK to run the "mismatch". Amps do not explode by plugging an 8 ohm speaker into the 4 ohm tap.

I don't believe I mentioned 2 ohm or 16 ohm speakers at all, so I don't know why you bring them up.

[bob p]

And if you read my post for what it was rather than with your tiring defensiveness, you'd see that what I really did was expand a little on what you said. It wasn't done with a switch, they did in fact explain it, as you said. The only thing I added by so posting was they they said it was OK to run the "mismatch". Amps do not explode by plugging an 8 ohm speaker into the 4 ohm tap.

Oh, your agreeing with me wasn't at all lost on me. I saw it for what it was. The problem that needed to be addressed was the reference to Mesa's documentation. It's not a good reference from which to draw conclusions because at times the Mesa docs have been omissive, at times they've been more complete, and at times they've been just wrong. The evolution of how Mesa addressed the topic through the years is evident in how the manuals were written from one amp generation to the next. Hopefully some casual user won't read one of the old manuals and get burned by it.

WRT to "tiring defensiveness" I understand that you may find it tiring when someone disagrees with you, but I don't really care about being defensive because I don't care if you think I'm wrong. I'm not trying to prove myself to you. I'm more concerned about someone who does not have decades of experience as the know it all repair manager, so that they don't make the foolish decision to look at the mismatching question through the safety of the retrospectoscope of mismatching vintage Fender amps, as that would involve unsafe extrapolation between the amps of yore and the amps of today. Back then Fender used design center values, and today some boutique builders purposefully design far off-center. That has implications that are foolhardy to view with a blind eye; anyone who insists on doing that is going to end up paying for a repair bill that should have been avoided.

And in responding to Chuck's comment about not seeing switches, it should be evident to anyone who reverse engineers the OT specs on some modern boutique amps that there are designers who deliberately choose off-center values to make power switching more safe than it is with vintage Fenders, while there are other designers who paid no attention to that, and chose FAR off-center values to impart a specific tonal and distortion characteristic while sacrificing mismatch safety to get it. For this reason it's absolutely fallacious to assume one or the other, or to assume that a 100% mismatch is as safe today as it was 50 years ago; you have to blueprint the transformer specs to be sure.

Some designers are designing the amp to be >100% off center-spec for tone. In that case a mismatch in the right direction would be innocuous but a mismatch in the wrong direction could be particularly dangerous. (Pull a pair of tubes in that scenario without correcting for Z and you have the extreme mismatch condition that you scoffed at as being irrelevant.) Other designers design to straddle a centered value for safety by choosing OT spec that's intermediate between the 2x6L6 and 4x6L6 condition, such that the OT's two operating modes straddle a desirable midpoint, making mismatch safer in both directions than it would be with a vintage amp optimized for the 4-tube Z. Today some smart designers take power switching into account, and they're not ignoring the problem, as the absence of a Z switch might lead some people to believe, because they implement the safety factor by changing the transformer spec rather than by using a switch. To some people reading a schematic, that conscious choice would be invisible, and they could draw the wrong conclusion by looking for a switch that is not there, when the reality of the situation is that the designer chose a different OT turns ratio to eliminate the need for the switch.

My point was to commend Chuck for his investigation into the subject of switching on schematics, to illustrate why the absence of a switch might not be 100% meaningful, and to illustrate that anyone who might make an assumption about mismatching could end up shooting themselves in the foot by making the wrong assumption. Don't take my post as a personal response as if I typed all of that in response to you, I honestly wasn't responding to you, although my response did include a comment in reply to something that you said. Realistically speaking, I don't worry about you, Enzo. I'm sure that you know everything that you need to know to be safe. But some people reading this thread, which is after all based on the OP's question about mismatching theory, may not have your level of expertise. I typed my response for them, in response to Chuck's switch comment, which I hoped would not lead some readers astray. Please don't assume this is all about me and you. I don't view our relationship as being worthy of all of the type that I'm committing. I'm doing that for the benefit of others.

[Gingertube]

I take a short circuit on the arguement:

When you compare the reflected impedance of the speaker into the primary with the output impedance of the output tubes themselves you find the output tubes impedance to be much higher.

That means the output tubes act as current sources - swinging signal CURRENT through the output tranny primary.
2 tubes each side swing twice the current giving double the power.

So for a 1/2 power switch I am quite happy using a screen voltage switch or cathode switch to disable 2 of the 4 output tubes (1 each side) without bothering with changing reflected impedance by messing with speaker taps.

You do need to be aware of what circuit your amp has before just pulling 2 of 4 output tubes - you can't just do that on cathode biased amps unless each output tube has its own cathode bias resistor.
For fixed biased amps it will generally be OK.

I know that KOC (London Power) argued exactly the same over many years.

Cheers,
Ian

[loudthud]

P=R*I^2. Half the current means ONE FOURTH the power. KOC doesn't understand this and gets it wrong every time!

Now half the current into twice the load impedance, that's half the power.

[Gingertube]

I don't agree, the P=R*I^2 is correct
BUT
- that is the thing about output tubes operating as a current source. The voltage swing does not change. Each of the 2 tubes on each of the push pull sides swings the same voltage. From B+ down to the tubes saturation voltage (typically 40 to 50 V).
The second parallel tube just doubles the current.
Same voltage and X2 the current is X2 the power. The Voltage swing is NOT dependent on the load R. So use P=V*I the derived formulae above where you sub in V=I*R just confuses the issue.
Cheers,
Ian

[loudthud]

If the Voltage is the same and the load is the same, then the current is the same. If half the number of tubes are conducting the same current, they won't last very long.

[Gingertube]

That doesn't happen because the tubes are acting as a current sources with current = Vg1-k * gm.
The "composite tube" made up of a parallel pair simply has 2 x the gm.

The P=R*I^2 does not apply as contains an incorrect assumption.
You start with P=V * I and then ASSUME that V= I * R, in the is case it doesn't, it equals the B+ less the saturation voltage of the tube (x2 for push pull across the full Raa).
So that substitution of V = I * R into P=V * R to arrive at P = R * I^2 is in this case invalid.

Perhaps it might be easier to simply think of this as one tube has a gm value and a parallel pair has 2 x that gm.

Cheers,
Ian

[nickb]

That doesn't happen because the tubes are acting as a current sources with current = Vg1-k * gm.
The "composite tube" made up of a parallel pair simply has 2 x the gm.

The P=R*I^2 does not apply as contains an incorrect assumption.
You start with P=V * I and then ASSUME that V= I * R, in the is case it doesn't, it equals the B+ less the saturation voltage of the tube (x2 for push pull across the full Raa).
So that substitution of V = I * R into P=V * R to arrive at P = R * I^2 is in this case invalid.

Perhaps it might be easier to simply think of this as one tube has a gm value and a parallel pair has 2 x that gm.

Cheers,
Ian

The truth is somewhere in between. The tubes are not perfect current sources and i =gm.V is just valid(*) for small signals . Due to these factors, with constant load resistance the peak current in the single pair will be about 25% higher than when a pair are working. This means the power out is around 40%. The single pair of tubes are working very hard and may over-dissipate.

(*) well it is valid but gm is not a constant
PS: The 25% I got a above was with an unusually low output impedance. If I go for something a bit more realistic the current increase is more like 50% i.e the power output is 56%.

[Malcolm Irving]

By looking at the plate characteristics for a particular pentode, with a given bias, with appropriate load lines, we can see exactly what happens when one pair of a push-pull quad is pulled (but the load resistance and transformer tap are unchanged).

The plate characteristics for a parallel pair of pentodes is exactly like one pentode, but with the figures on the vertical (plate current) axis all doubled.

Or putting that another way, the load resistance ‘seen’ by each pentode of a parallel pair is doubled, compared to that seen by a single pentode, if everything else is unchanged.

[Mick Bailey]

Now suppose that you had an quartet type amp, like a Twin Reverb. It might have an OT with a turns ratio of about 700:1, so that when you place a 4R load on the amp (two 8R speakers in parallel), it reflects a primary Z of 2800 to the quartet of 6L6: 4R x 700 = 2800R. That puts us right in the area of the design center values for running the quartet.

Hey Bob, you've mixed up turns ratio with impedance ratio - the turns ratio for the example given would be 26.46:1.

[loudthud]

The P=R*I^2 does not apply as contains an incorrect assumption.
You start with P=V * I and then ASSUME that V= I * R, in the is case it doesn't, it equals the B+ less the saturation voltage of the tube (x2 for push pull across the full Raa).
So that substitution of V = I * R into P=V * R to arrive at P = R * I^2 is in this case invalid.

There is a reason it's called Ohm's LAW, not Ohm's Vague Suggestion. If the current is half, then the Voltage can't be the same,,, unless somehow the load is doubled. Like Kevin, I suspect you don't have the 100X probe needed to look at the plate Voltage in a power amp stage.