Mark V 90W combo speaker change

[MarkV User]

Things are more complicated than they looked at first but I am getting a better understanding. I’m still looking only at the speaker outputs and need to repeat observations at the loop out for preamp only given what I’ve learned.

Turning down the presence all the way fixes the problem on every channel and mode. Turning up presence makes it as bad as it can be. Cranking channel gain and volume maximizes the problem.

Extreme mode can still exhibit the problem if the presence is turned all the way up. I thought for a minute that all feedback was disabled in Extreme mode but this is not true. C63 still passes some feedback signal in Extreme mode.

The high frequency spikes are only partly asymmetrical on channel 3. Turn up presence and you see them on both rising and falling edges.

The problem interacts with the global EQ enough that it was confusing me. I turned EQ off on all channels for all the tests above.

I tested with an attenuator between the amp and cab speaker. As I turned down the speaker volume with the attenuator the problem started to go away. In load only mode there was no problem. The problem was greatest in 0dB mode—no attenuation. I want to repeat these observations with more test conditions to be completely sure of this.

So now I need to see if I can reproduce the problem in the preamp only. I don’t think so based on what I saw with the attenuator in load mode.

Here is one thing that could be happening. All those capacitors in the feedback loop add phase shifts at high frequencies. And the speaker also can’t comply as well to high frequencies. So we have more negative feedback up there and it is inaccurate because of those phase shifts. The hypothesis is that is why we see these spikes. The feedback loop is amplifying the error instead of reducing it.

I’ll pick this up again tomorrow.

 

[bandit2013]

Right! I get what you mean about the overshoot not looking like normal ringing or inductive overshoot and I hadn’t noticed but I agree with you.

I saw that apparent asymmetry you pointed out in modes 7 and 8 (ch3). The answer to your question of “why is the top of the upper waveform drooping down to a lower level?” (I‘m sorry for misquoting you) would be negative feedback if we were looking at the power amp output. Do you remember where you had the scope connected in this test? I am starting to think we have the exact same problem.

Here is one thing. Just don‘t play your Mark V again without hearing protection. I’m not making a joke. I think it is pretty bad for you to be exposed to high power ultrasonic and near ultrasonic noise like this.

I was measuring the FX SEND signal level so that is where the oscilloscope was measuring. I did use an FX pedal as a load for the signal so it was not open ended. Ran the JP2C and TC50 much the same way so those waveforms were proper. Only the Mark V CH3 had this effect. Since the rise has more meat and the drop could be a low frequency issue. Since it is in phase with the input signal, that only leaves me with V4B or V6A as it occurs on CH3 only. Does not mean there is no issue with the GEQ. The amp has had many failures over the years, Lost a few JFETS due to the constant power tube failures. It was not until the screen resistors blew out that I decided to change the bias. Also helped to retain the speaker as I was blowing out voice coils too. It is a turd that needs to be retired. I have pleny of other amp in my collection so this one will not be missed unless I forget to remove the EVM12L speaker from it. That is not an ideal choice for the Mark V90 as it is too boxy toned for that speaker. I installed it when I changed over to the STR441 power tubes not too long ago. Since I have the same speaker in an OB112 wide body cab, that actually sounded good with the Mark V90 while I had the chassis out to tune it with preamp tubes. I give up. No point in trying to figure out what is wrong with it when I never use it anymore. I am tired of the torture. I would rather spend my time playing the guitar than tearing it down to change preamp tubes or redial it as it sounds like crap.

 

[MarkV User]

This is good to know. I’ll look more carefully at my preamp and power amp separately next.

I’m getting a pretty good picture now of the end to end behavior. I’m less concerned about ultrasonic energy to the speaker because the really high frequencies are not making it out the stock combo speaker.

This time I looked using a Sure Beta 57 connected directly to the USB scope, no preamp. I would need use an unfiltered preamp and/or or a better scope to see as much detail as I really want, but this is ok for now.

I see spurious high frequency noise coming out up to about 5KHz when the input is a sine wave in the midrange. It doesn’t look good on the scope, doesn’t sound good, and would hurt my ears without protection, but it is not ultrasonic.

Also, in the clean channel at low master volume levels I can’t always make the problem frequencies (the spikes) go away completely by turning down presence as I thought earlier. It is still mostly true and I think it is true on the other channels.

With real speakers in the loop the problem is worse, measured electrically at the speaker output. I do see the problem even straight into the load box though. As I put more output into the load box the magnitude of the problem goes down some.

Summarizing:

1. The problem frequencies are in the single digit kHz for the most part. They can be generated in more than one way, at least with the GEQ and with the presence control.

2. There is no point in outputting really high frequencies because the speaker can’t comply, and if we roll off some more presence that would be good in this amp. It looks like we could use more low pass filtering after the phase splitter. The interaction with the feedback loop makes me wonder if a lower gain tube in the phase splitter might help.

3. Bandit sees this problem on Ch3 and maybe I will too. I’ll also try FX ret to speaker out.

I was measuring the FX SEND signal level so that is where the oscilloscope was measuring. I did use an FX pedal as a load for the signal so it was not open ended. Ran the JP2C and TC50 much the same way so those waveforms were proper. Only the Mark V CH3 had this effect. Since the rise has more meat and the drop could be a low frequency issue. Since it is in phase with the input signal, that only leaves me with V4B or V6A as it occurs on CH3 only. Does not mean there is no issue with the GEQ. The amp has had many failures over the years, Lost a few JFETS due to the constant power tube failures. It was not until the screen resistors blew out that I decided to change the bias. Also helped to retain the speaker as I was blowing out voice coils too. It is a turd that needs to be retired. I have pleny of other amp in my collection so this one will not be missed unless I forget to remove the EVM12L speaker from it. That is not an ideal choice for the Mark V90 as it is too boxy toned for that speaker. I installed it when I changed over to the STR441 power tubes not too long ago. Since I have the same speaker in an OB112 wide body cab, that actually sounded good with the Mark V90 while I had the chassis out to tune it with preamp tubes. I give up. No point in trying to figure out what is wrong with it when I never use it anymore. I am tired of the torture. I would rather spend my time playing the guitar than tearing it down to change preamp tubes or redial it as it sounds like crap.

 

[bandit2013]

Most speakers have a HF drop off around 5KHz. However, some of the speakers do have peaks in the upper boundary past 10KHz that is just enough to become heard.

My Mark V is an exception, I am not sure others have had the same sour experience with theirs as I have had with mine. What I mean by ice pick character, it is that sound you get in your ears or skull that rattles your brain with the sound of glass breaking inside your ear canal. Not only is this sound irritable it hurts the ear drums.

I think I held onto the Mark V90 for experimentation purposes. Money already spent and it would be difficult to sell it considering it sounds like . If I did sell it, I would not be giving up my NOS tubes I have loaded into it. It would have to work with stock Mesa tubes, which seem to lend to the ice pick sounds unless you dial back on the treble presence and gain to 9AM or lower. CH1 with the stock tubes has a 60Hz hum that is louder than the guitar signal. CH2 edge is useless even with a 7 string guitar. Sure I was able to curb most of the tonal issues with NOS tubes (1990 Mesa Beijing square foil getter tubes). Since they are not easy to find anymore, not giving them away with this amp as I did with the Mark III. Person I sold the Mark III to got a handful of those tubes in really good condition, probably like 30 of them. At that point I gave up on playing the guitar and all I had left was the Mark IVB and two guitars. I will not say what got me back into playing again, but I am glad I did pick it up along with other instruments like drums and bass.

All of this round and round on the Mark V90. There are some that are brittle and boxy and some that sound really good with just a hint of boxy tones.

Mark V is not for everyone so if that is the case, move on. I am not trying to be an *** here.

The Mark V90 is one of few amps that Mesa has created that does respond well to preamp tube changes. Trying to find a pleasing combination is time consuming but sometimes worth the effort. The Mark V90 also has a similar character to its predecessor the Mark IVB, that boxy characteristic. I found the Mark IVB to be more tolerable. In either case, the Mark V90 or the Mark IVB sound terrible with the EVM12L speaker, black label or classic. Those two speakers will bring out and emphasize on the boxy characteristic. The Mesa Vertical 212 cab (which I have used for all of my recordings) is a bit better but you can still hear the boxy tones. There was only one speaker that put a smile on my face with the Mark V90 combo and that was the Celestion Cream 90W ALNICO speaker. Works really good for all of CH3 but takes away some of the qualities one may prefer from CH1. The alternate is the Celestion Redback. However, if you like torture and the long time it takes to break-in, the other good speaker is the Jensen Jet Blackbird 100W, best to remove the bell cover. It will do wonky stuff during the break-in period with any distortion. Over a span of 6 months it will relax and sound really good. Most do not have the patience to deal with that.

Sorry, I am jumping around on this. Just a few observations and some tube experiences condensed from several forums into one post: Based on my experiences which are limited, what I found to be desirable right out of the box with all stock tubes are the following:

Badlander100W, JP2C, MWDR, and the Mark VII. Going back to history, Mark III and Mark IVB.

Triple Crown TC100 and TC50 could be on that short list too. The TC100 needed a preamp tube swap from the Chinese (Sino)12AT7 to a JAN/Phillips 12AT7. TC50 did not need any preamp tube changes. Most of the issue is with the power tubes. The STR447 EL34 sound on the harsh side out of the box. The tubes will mellow out after using the amp for a few weeks and then it will sound great. Whatever Mesa did with the Badlander, the STR447 do not sound harsh. So far the Gold Lion KT77 were epic in the TC50. I did not have enough of them in working order to try in the TC100. These amps do sound really good with the TungSol EL34B, EH 6CA7 and a few other tubes. I found I much prefer the STR447 tubes after they have some time on them. For those that have the TC50 combo, I would highly recommend the Jensen Jet Black bird 100W Alnico. (need to remove the bell cover, and there is the break in period to contend with. This speaker brings out the low end of the TC50.)

Royal Atlantic (RA100), a few NOS tubes for the hi/lo gain channels and a set of NOS SED=C=EL34 (Mesa STR442) are the best I have ever heard in that amp. The NOS tubes: RFT 12AX7 for V1, and Ei CV492 in V2. Mullard long plate in the phase inverter or Ei CV492 will work as well. For me this is the golden standard as it is one of few amps that I get drawn in and lose track of time.

Roadster, this needs a few preamp tube swaps. Not bad with stock tubes but with the old 1990 Mesa 12AX7 (Chinese Beijing Square foil getter tubes) in V1, V2, V3 it becomes a beast that I can actually use a 7-string guitar with. Low end is still there but does not swamp you out.

For that Mark V90 woes, I would recommend the Badlander100W, Mark VII, or JP2C. If you want something more old school, Triple Crown or Royal Atlantic, may as well include the Electra Dyne in this mix. California tweed or the Filmore will also give you the American dream from yesterday.

For fun, the Badlander 100W sounds great with the STR447, almost as if you were pushing the 6L6GC tube but with some EL34 flavor. The STR446 (TAD red base EL34) are even better. The STR448 (TAD red base 6L6GC in gray bias color) are amazing. JP2C will take advantage with the STR448 tube as well. It is as close as it gets to the STR415 Sylvania 6L6GC tube. However, the STR415 are tops in the JP2C which is what I have in that amp now. I have also tried a blend of the STR448 and STR445 in the Mark VII which is great at 90W. I keep the STR445 in the class A sockets as that basically sets the character for the amp on all power modes.

For those extended range guitars like the 7-string: Badlander, Triple Crown, JP2C, Mark VII were all compatible without flubbing out. RA is really good too but needs the preamp tubes described above. Roadster can be tuned for the 7-string with old Mesa 12AX7 (from1990) tubes and the STR440 6L6 tubes.

What amps do not work well with 7-string: MWDR and the Mark V90.

I am done with my rant. Sorry for old news and nothing new here.

 

[MarkV User]

OK, I can reproduce the problem using only the power amp section (FX RET to 8-Ohm Speaker out) and in no way can reproduce it on any preamp channel or mode. I was pretty thorough in checking. Because it happens whether the FX loop is on or off, it has to be in the last stage. My issue is with the feedback and presence circuit. Next step is circuit analysis, also comparing this amp with other amp schematics, and then more testing.

 

[MarkV User]

I fo

OK, I can reproduce the problem using only the power amp section (FX RET to 8-Ohm Speaker out) and in no way can reproduce it on any preamp channel or mode. I was pretty thorough in checking. Because it happens whether the FX loop is on or off, it has to be in the last stage. My issue is with the feedback and presence circuit. Next step is circuit analysis, also comparing this amp with other amp schematics, and then more testing.

I forgot to mention that I bypassed the GEQ in most of my testing because it can produce some high frequency spikes on its own, though I think they are normal amplitudes not extra high like the feedback ringing. If I solve the power amp high frequency problems then I will look more at the GEQ in that context.

I think the frequency dependent phase shifts in the presence feedback loop are changing the negative feedback into positive feedback at some frequencies. That is causing the spikes in the power amp output. Maybe depending on the particular caps in the loop this problem only happens on some amps. I have to think it worked before shipping and for everyone who loves their Mark V.

I’m trying to learn enough network theory to quantify the behavior. Failing that I could spice model it or just make a change that will reduce the phase shift in the presence frequencies. My first try will be to make a component change instead of changing the presence loop. If I can’t do that I’ll change the circuit to make it look like an amp that is known to work.

I also want to say that changing V7 to a 12AT7 (lowering loop gain) helped a little but not a really significant improvement. That is not too surprising on either count.

 

[mace]

I’m interested in the result but I sure can’t follow the EE path getting there.

 

[MarkV User]

I‘m sorry about that. Here is a important question though.

How useful is presence to you on all the different Mark V modes?

I see that the easiest thing I can do might be to make the presence work the same in all channels and modes and make it look like the 2C+. Would that mess up any use case you have for the amp?

I am self taught in electronics which might make my communication about it hard to bear sometimes. I’ve shamelessly used this forum to keep myself on track for this big job of finding out what is wrong with my amp and fixing it. But it makes me want to build amps. I love this stuff. Thank you for all your patience and support!

 

[rarebitusa]

I fo


I forgot to mention that I bypassed the GEQ in most of my testing because it can produce some high frequency spikes on its own, though I think they are normal amplitudes not extra high like the feedback ringing. If I solve the power amp high frequency problems then I will look more at the GEQ in that context.

I think the frequency dependent phase shifts in the presence feedback loop are changing the negative feedback into positive feedback at some frequencies. That is causing the spikes in the power amp output. Maybe depending on the particular caps in the loop this problem only happens on some amps. I have to think it worked before shipping and for everyone who loves their Mark V.

I’m trying to learn enough network theory to quantify the behavior. Failing that I could spice model it or just make a change that will reduce the phase shift in the presence frequencies. My first try will be to make a component change instead of changing the presence loop. If I can’t do that I’ll change the circuit to make it look like an amp that is known to work.

I also want to say that changing V7 to a 12AT7 (lowering loop gain) helped a little but not a really significant improvement. That is not too surprising on either count.

Following... so I haven't dwell'd on the V presence circuit till now. My limited experience with presence was implementing the Depth mod on the Stiletto. However that does help me to realize that the V implementation is way more complicated. Either my ears are burnt (very possible) or my V doesn't react in the same way as I have no complaints with it. In fact I've always felt the power section was one of the better features of the amp. But with all those components in the circuit I welcome the deep dive here.

 

[MarkV User]

By the way there is something I learned that should Be recorded in case other people read this thread while debugging their amps. There are two zener clipping stages in the MarkV that also use J175 JFET transistors instead of resistors. There is an error in the circuit design though. The absolute maximum 30V channel to gate voltage of the J175 will be exceeded by the time the Zeners clip at around 31.2V.

That is a design error. Some unintended behaviors will result (which I think I’ve seen in testing). Those J175s may also be damaged or destroyed during normal operation of the amp.

At the moment I don’t think this problem is related to the power amp spikes I see but I will have to find a way to exclude that possibility also.

 

[Eevil]

I‘m sorry about that. Here is a important question though.

How useful is presence to you on all the different Mark V modes?

I see that the easiest thing I can do might be to make the presence work the same in all channels and modes and make it look like the 2C+. Would that mess up any use case you have for the amp?

I am self taught in electronics which might make my communication about it hard to bear sometimes. I’ve shamelessly used this forum to keep myself on track for this big job of finding out what is wrong with my amp and fixing it. But it makes me want to build amps. I love this stuff. Thank you for all your patience and support!

Just an answer to presence usefulness: it’s effective and useful

I’ve settled ch1 and ch3 fat and Mk IV respectively. Presence before noon.

Fiddling around with ch2 modes.. somedays (and depends on guitar) I venture to Mark I -> presence almost dimed and other days to Edge or Crunch -> presence between 9-12 o’clock.

Somedays would be great to have Mark I and Crunch/Edge in own channels.. but not an actual issue. Road king layout with Mark V innards would’ve been killer

 

[MarkV User]

Just an answer to presence usefulness: it’s effective and useful

I’ve settled ch1 and ch3 fat and Mk IV respectively. Presence before noon.

Fiddling around with ch2 modes.. somedays (and depends on guitar) I venture to Mark I -> presence almost dimed and other days to Edge or Crunch -> presence between 9-12 o’clock.

Somedays would be great to have Mark I and Crunch/Edge in own channels.. but not an actual issue. Road king layout with Mark V innards would’ve been killer

I know this is hard to answer, but if presence on all channels affected the tone like MarkV 2C+ mode do you think that would ok? Making the presence/feedback and phase splitter low pass filter work the same in all channels and modes might make the correct operation of the amp less dependent on precise component values. That would reduce flexibility but might improve the sound of the amp on all channels. I mean it would probably make Mark I mode sound less like a real Mark I, etc. But perhaps all channels would sound better.

I am jumping the gun on this because I need to find a component change to fix my problem before attempting a drastic mod like that. But I’m curious about it because I think it could make a big difference.

 

[Eevil]

I know this is hard to answer, but if presence on all channels affected the tone like MarkV 2C+ mode do you think that would ok? Making the presence/feedback and phase splitter low pass filter work the same in all channels and modes might make the correct operation of the amp less dependent on precise component values. That would reduce flexibility but might improve the sound of the amp on all channels. I mean it would probably make Mark I mode sound less like a real Mark I, etc. But perhaps all channels would sound better.

I am jumping the gun on this because I need to find a component change to fix my problem before attempting a drastic mod like that. But I’m curious about it because I think it could make a big difference.

Have to check how it works in MkIIC mode

 

[MarkV User]

I can make the problem go away by putting some additional capacitance across pins 1 and 6 of v7. 560pF is not quite enough and 1nF is about right or a little too much. This is not perfect, but suddenly my amp became usable.

I tried removing and substituting one of the caps in the feedback path. So far I haven’t found a solution there or any smoking gun, but there are a bunch of things that I still need to try.

 

[MarkV User]

I’m pretty happy with 500pF across V7’s outputs now. The 560pF that I didn’t like was a ceramic disc cap, but two 1nF plastic caps in series sound good to me and leave highs controllable by presence. It is not perfect, but now I’ll take that back out and seek the root cause again. Note that this cap is in the loop so it does bear directly on the feedback issue.

There is also a way that the mute JFETs might turn on very briefly because of gate leakage (marginal breakdown) if the signal goes below -18V anywhere in the signal path. That could create very narrow return to ground that might interact with inductors to make an RF spike or upset the feedback loop because it can end up looking like an inductive reactance at some frequencies maybe. But that hypothetical chain of events would be asymmetrical. So to see the problem on both edges of the signal it would habe to affect different JFETs in different stages. This whole scenario seems really unlikely but still hasn’t been excluded And it looks to me as though the JFET/zener anti-pop circuit has this mistake where the JFET is being run past its absolute max and very well might malfunction or be damaged. It is also possible that the JFETs are ok due to something I’m not understsnding yet. I know JFET properties vary depending on the vendor, the manufacturing run, etc. J175 data sheets from different vendors don‘t even agree on all the parameter specifications.

By the way the GEQ definitely creates its own high frequency spikes if you turn the highest band all the way up, etc. I still need to look at this because the 500pF V7 filter doesn’t remove those spikes completely. I don’t think this is a big problem at normal GEQ settings but will investigate.

The current issues of concern are safely accessing all the components and making the mods without breaking things. From here on out I can put the 500pF back any time, remove the mode dependent filters there, and be on good ground to use or sell the amp. I expect to learn more and come up with a better fix though.

 

[bandit2013]

Have you played through another Mark V90 before? If so, was this characteristic present with that amp?

 

[Eevil]

I know this is hard to answer, but if presence on all channels affected the tone like MarkV 2C+ mode do you think that would ok? Making the presence/feedback and phase splitter low pass filter work the same in all channels and modes might make the correct operation of the amp less dependent on precise component values. That would reduce flexibility but might improve the sound of the amp on all channels. I mean it would probably make Mark I mode sound less like a real Mark I, etc. But perhaps all channels would sound better.

I am jumping the gun on this because I need to find a component change to fix my problem before attempting a drastic mod like that. But I’m curious about it because I think it could make a big difference.

Tested today..

I did not hear a difference.. played with a looper in a fx-loop so the preamp and eq sections would not have effect on tone. Switched between MkIV and MkII while the loop was playing and fiddled with presence from 0 to full

No documentation only hearing

 

[MarkV User]

Tested today..

I did not hear a difference.. played with a looper in a fx-loop so the preamp and eq sections would not have effect on tone. Switched between MkIV and MkII while the loop was playing and fiddled with presence from 0 to full

No documentation only hearing

Oh wow, that is a lot more than I thought you were going to do. Thank you. I'll send an update after another session of looking at the feedback circuit then. I have an idea for a fix.

 

[MarkV User]

Have you played through another Mark V90 before? If so, was this characteristic present with that amp?

Unfortunately I never have played another Mark V:90, so I can’t be sure the feature that bothers me is not in all amps. I don’t think the V:35 I played had the problem. I actually thought it was only in some V:90 amps because I read your experiences and saw those scope traces. But the situation with your amp was so different, and after spending more time looking I’ve seen that we can get the high frequency overshoot in several different ways, and I can make it symmetrical or asymmetrical depending on the control settings. The one thing that makes me think this only happens on some amps is that no one would like the Mark V if they were all like this.

I did more experiments that convinced me that the Mark V loop from the 8 ohm tap to the phase splitter is not the culprit. And this is really obvious in retrospect because turning presence all the way up removes almost all the negative feedback makes the problem worse. Also, bridging C38 makes the Mark V look like the other Boogies and didn’t make the problem go away in any mode. So much for the phase shift idea.

The JFETs are also not going to be involved as you pointed out, because the signal levels are too low to cause a problem when the gate is grounded, and when they are off (gate positive) the source and drain are both near ground.

There is a very strong treble boost from C57, but this part of the circuit Is identical to the other Mark II and later amps. But there could be an interaction with C110/C116, parts that are not present in the other amps. I haven’t tested this yet but it would look something like this: when V7A turns on suddenly its anode initially sees low impedance from the C power supply voltage due to C110/C116 discharging. After that the anode sees voltage C through the 82K Ohms from R107. The impedance change during the edge transition changes the impedance through C57 and C58, but not by the same amount, so the voltage seen by the grid of V7B goes up more during the transition and we get the spike. If we add a bigger capacitor around 500-1000pF across C110/C116 then the spike will be shorted across the differential pair. This is all easy to check with the scope and easy to fix (if it turns out to be the problem) by removing the filters across the phase splitter output. We don’t see those C110/C116 capacitors in other Mark series amps.

I may be wrong in that analysis but it is my current best shot at understanding what is happening now. If this idea above is what is really happening then C110/C116 are working very differently than I thought and maybe not how they were intended to. It could also be that the effect changes a lot within the component tolerances. The Marshall plexis and 5F6A Bassman have a smaller 47pF capacitor across the differential pair, but except when the presence knob is at the very top of its range they lack the low impedance shunt to ground that we have with the Mark V. Until I check and report back please keep in mind that this is all speculation.

I was hesitant earlier to probe inside the amp with the USB scope, and I took some time to sort of learn where everything is on the crowded PCB, but my understanding has gotten better thanks to time looking and your help. Another round of careful testing should get this.

 

[bandit2013]

I am not sure I am reading this correctly. Sorry.
C38 is a bypass cap across the 69k resistor which is part of a voltage divider circuit to the control grid of V4B. Bridging that would effectively short out the signal to ground. C39 on the other hand couples the grid to the cathode of that same circuit, this will tune out some midrange and top end before the signal gets its non-linear amplification. You cannot bridge that component, it has to be removed to determine its affect. The same would apply to C38.

As for the JFETs (J175 P-channel), they are in full conduction mode when there is no voltage applied to the gate (open circuit, or pulled to 0 VDC). Since they are mostly used in the mute circuit, they will shunt the control grid signals to ground as long as the strobe mute transistor (NPN Darlington) is turned on [this will bring the voltage on the gates of the mute JFETS to 0VDC]. This should only be a momentary function. The collector side of the circuit is connected to the 24V supply but will also have some capacitors connected to ground. So, when the NPN transistor turns off, there will be a charge effect on the circuit that feeds the JFETs so they do not turn off instantly but gradually. (The JFETS must be turned off with 6V to 12V on the gates to stop them from shorting the signal.) The Mute for the reverb circuit has a larger capacitor so its charging rate will be slower than the other JFET mute circuits. As the voltage on the gate rises with the charging of the caps in the circuit, the channel will get pinched off slowly. In short form it begins to act like a variable resistor controlled by the rise in the gate voltage until it reaches the 6V which is where the pinch off occurs. It may still be conducting some current but will no longer be acting like a short. Once it gets to 12V, the JFET channel is turned off and the signal is permitted to pass onto the control grid of the next stage. The charge rate is quick but not immediate.

The other factor to consider is the amps you have compared the Mark V90 too. I would assume they are either full class A or are Class A/B. I doubt they are a combination of both as that would infringe on the patents for the Simul-Class method used by the Mark V90 and its predecessors like the IIC+ DRG, Mark III DRG, Mark IVA or Mark IVB. The extended class A sockets of the Mark amps is what defines the overall tone of the output. More so with the IV or V than the III or IIC since they ran with a 15W class A for a total power of 75W, the IV class A was pushed a bit to 30W triode, and a bit more in pentode with a total power of 85W. Mark V90 also has the triode/pentode switch to deliver 45W in class A to 90W total power. I am unsure if the intent was to retain or recreate some of that EL34 Class A sound with the use of 6L6GC tubes so some boxy tuning may have been done that was to simulate that effect. I feel they missed the Mark with the IV and V.

The JP2C on the other hand does not have that tuning. Does not sound shrill or boxy. But then again it is based on the IIC+ HGR model. 100W/60W class A/B amp. Not odd overtones or shrill artifacts in the signal at all. Power tube saturation is there but different.

Mark VII is still a Simul Class (DRG) but Mesa must have decided to ditch the boxy tones and ice pick flair for something more useful. It is very close to the character of the JP2C but yet different in its delivery and tone density. I much prefer the Full Class A/B of the JP2C at 100W but that is my preference. Seems that my favorite amps are all Class A/B. the Mark VII is an exception, so it has gained more respect and admiration. As for the power capacity, it is on par with the JP2C or the Badlander 100.