Mark V 90W combo speaker change

[MarkV User]

On speakers: I opened my recently purchased Bogner horizontal 2x12 and found one V30 and one G12 Classic Lead 80W speaker, 16 ohms each and wired in parallel. Soldered in, which is nice.

Just for the time being, and making use of what I have on hand, I changed the configuration to an 8 Ohm V30 and 16 Ohm V30 in parallel, for ~5.3 Ohms, and set the amps to 4 Ohms. That makes for less brightness and some mild phase difference between speakers just in the higher frequencies. I’m sure I’ll upgrade this to something more conventional soon, but despite being a great speaker that G12 Classic Lead was too punchy for me in combination with the high frequency problem in the Mark V. It may be that the phase difference is partly cancelling the problem frequency or just that the V30s roll it off. Well, anyway it seems to be a big improvement.

 

[MarkV User]

Bandit,

This is just enormously useful to see. I‘ll try to repeat the distortion experiment and post the results back here. I don't have a working scope right now still can tell a lot by using the outputs from the Mark V into an audio card. That overshoot is an awful kind of brokenness. I mean it looks to me like your amp is just broken, not like an amp characteristic. And in my mind's eye it is what I think might be wrong with my amp.

Bandit, my reading of your scope trace is that the high frequency part of that overshoot could be in the 5-11 KHz range.

Your pointer to the schematics is really helpful and I will pursue it.

On the Mark V after the phase splitter I don’t understand C100, C110, C115, and RYC3G. The amps that I like tend to have a capacitor across those rails about 1/1000 the value of the pass capacitors. Like 100pF in this case. The fact that reamping through Bogner amps worked for me could have to do with the fact that the Shiva (and probably also the solid state Ecstasy 30W I used earlier) has that high cut after the phase splitter so it applies to the effects loop return and amp-direct input. The Roadster like the JP2C+ lacks this high cut.

The value units are not given but I think C110 is 250pF and C115 is 75pF. Because if those were uF no sound would come out and if they were nF it still wouldn’t sound right. Is RYC3G a pair of relay contacts that selects NO only when channel 3 is selected? In that case channel 3 might be too bright and C100 makes things asymmetrical in a way I don’t understand.

It looks like RYM9F for example might be "mode" 9, maybe the Extreme gain mode? And it would be the F=6th contact pair on that relay or gang of relays?

[edit: sorry for confusion and lack of analysis here and possible wrong conclusions above, but I will try to provide a correct analysis in a later post]

There is a bigger problem in the Mark V in the presence circuit. The JP2C+ doesn’t even have the high cut capacitors I was talking about above, but the JP2C+ has a straightforward presence circuit that depends on just three capacitor values. Tellingly one of those is called out as 10% instead of the normal 20%. There are 6-8 capacitor values in the Mark V presence circuit, depending on relay settings. Other things being equal that means there is probably a lot of room for the circuit behavior to change bases on +/- 20% variation of all those capacitor values. The filter at C57 makes a frequency dependent voltage divider with the filters at C59, C111, etc, and what frequencies make it through also have to do with the impedance looking into C58. I’m still not completely understanding the circuit and not tracking what all the relay positions mean yet, but I don’t like the looks of this.

I'm wanting to use a larger value capacitor across the split phases like the Plexis and Bogner has, eliminating the special mode for Channel 3, and I'm wanting to simplify the presence circuit and make it more Plexi-like. I would also eliminate the mode-based (but not the channel-based) logic in both areas. That might make the whole amp mellower and darker and give it a more open feel... My gut is that based on your experiments it won't fix the root cause of the problem with your amp.

Those changes might make my amp tolerable but they won't get rid of high frequency ringing if it is the same as yours. I still need to make my measurements first, then. I would like to see if I have the same problem you do and where in the circuit it comes in.

Thanks to Bandit especially and to everyone for all the thinking and the trouble you’ve gone through to get this far.

When I slaved the Mark V90 into the Roadster, I got much of the same thing, brittle and ice pick tones on CH3 (IV and Extreme were the worse), CH1 tweed and Ch2 edge were also terrible. Not to mention that the Mark V GEQ circuit is used to create the send signal level, it has an impedance issue which can make things difficult with most FX processor buffers. The only products that actually worked with the Mark V were made by Strymon, DIG, BigSky, Brigadier were all good. Line6 DL4 was nasty, compressed the signal considerably, sure it sounded cool but it was not how it was supposed to sound.

When I was looking at the send levels I did take some pictures, I was running a 750 Hz signal if I remember correctly from a function generator. Here is what the signal looked like with the GEQ turned off. All of the controls were placed in their center positions.

View attachment 2742

The image is hard to see on the scope, here is a close up. The green trace is the FX Send output, note the leading edge marked with the arrows. This was CH3. Since the lead drive circuit comprised of the V5A -> V4B creates an asymmetrical waveform as I was led to understand, the sharp edge has more frequency content, If it had a slope or rounded corners like the portion of the waveform in the negative swing it would be a really good sounding amp. The sharp leading edge is ice pick tone.

View attachment 2743

Now to compare that to a Legacy amp, lets just say a modern version of such, the JP2C. Same frequency and the GEQ was turned off (not that it mattered since it follows the FX loop). Since I was looking at the signal amplitude to determine the signal level, I only took a picture with the amp in the view on the clean channel. It had no distortion. The next picture was on CH2 with some moderate gain dialed in.

View attachment 2744

Again with the CH2 controls at their center position. This is the difference in the distorted signal.

View attachment 2745

There is a sharp leading edge but the not a large spike. The Lead drive circuit is asymmetrical so the bottom portion has more of a normal distorted appearance The 2nd harmonic is visible. The top waveform has 4 ripples in it so that would be from the 4th harmonic. The leading edge having that sharp hook would be the higher order harmonics. My company will not fork out the money for a good signal analyzer as they generally start at $100k for an entry level.

Anywas, the reason the JP2C sound great, it does not push the limits on the upper harmonics that leads to ice pick tone. May as well call it shot-noise. Yes, that term goes way back to the tube days when the masters were creating their theory on design, mostly for the small signal analysis, Miller effect, and all that jazz. If you ever had any engineering courses on electronics and amplifiers in general (limited to the transistor), it all came from tube circuits. Some say the tube amp has a natural sound and will not generate odd order harmonics. That is BS. Class A/B is not as bad as a Class B amp. They may look similar (yeah, Class B does exist but not used) Class A/B will generate some nasty harmonics of the odd order but is corrected with negative feedback often called presence. The Simul-Class (combination of Class A/B and extended Class A) will create more harmonic content than a Class A/B, again corrected with the use of negative feedback.

I did have all the gear I needed to use to measure the output power except for hearing protection. No ear plugs or sound blocking headphones. Running a single frequency through an amp at elevated levels is pure punishment. 750 Hz was bad enough, 1KHz would have been much worse. Also would have had to use the clean channel as the power test is done without distortion. No thanks. Still tempted though.

 

[MarkV User]

[Edit: this is redundant with one of my earlier messages that disappeared and then reappeared when this message posted]

I’m sorry, I posted a couple of replies earlier that didn’t work for some reason. I have to thank everyone because the last few posts have shed a lot of light on the problems I've been struggling with. I'm going to respond to the speaker help first.

Thank you KRichter and thanks to everyone else who helped with examples of good speaker combos. Thanks to many good suggestions I have a temporary solution, and it might make some people cringe but it works for now and uses the speakers I have on hand. I found I had a V30 and a 90W G12 Classic Lead in my Bogner cab, both 16 Ohms. And the V30 seemed not to be too bright but because of my buggy amp the Classic Lead was too bright. I had an 8 Ohm V30 available and put that in instead in parallel with the 16 Ohm V30. That gives ~5.3 Ohms and I run it from the 4 Ohm output of the amps. It is not really distorted like running an 8 Ohm cab at 4 Ohms. It sounds really good, and it's not even too dark or distorted. But it rolls off the high presence a lot more than running the 4x12 cab at a correct 4 Ohms. This is really a dumb and temporary solution I think, but I need to fix the amp before I can choose speakers.

I'm not really happy with my MarkV tone this way, but it is still the best it has ever been, and it is not hurting my ears at low volumes like the old setup was.

So thank you! This is great. I also listened to many different speakers on YouTube, promoted by the many excellent suggestions. Now there are a half dozen speakers I'm excited about and want to try at some point.

I'll respond to the amp electronics discoveries in the next message.

 

[bandit2013]

Just keep in mind that the phase inverter circuit is a bit different than the designs you would encounter for Class A/B or traditional Simul-Class amps. Considering the 10W power mode is Class A only, C115 and C100 are more than likely used with the 10W power. The capacitor values are missing their units, they would be in pF. When the amp is at 45W or 90W, RYC3G connects the two phases with the 250pF cap, this will also be when RYSE1 is in its normally closed state to connect the bias to the control grids. In 10W mode, the bias circuit is disconnected from the one pair of tubes as this will be running in cathode bias power mode. I would not worry about the 250pF C110 as this is a basic plate-to-plate snubber. Since the phase inverter plates are 180° out of phase with each other, the capacitor serves to remove high ice pick frequencies. It also adds to stability. Typical range is usually 75pF to 100pF, making it larger will cut more ice from the tone. So in 10W mode, half of the phase inverter is not used. So when RyC2G is in the NO position, that is more of a filter and part signal coupler from V7A to V7B. this will all be in the AC signal as the caps will block the DC voltage.

It looks like RYM9F for example might be "mode" 9, maybe the Extreme gain mode? And it would be the F=6th contact pair on that relay or gang of relays?

RYM9F, RYM8RL, and RY-FB I would assume are linked to the Extreme mode. Since the negative feedback circuit gets rewired from the 8 ohm tap to the 4 ohm tap, RYM9F un-shorts the
C63 cap that will cut some of the feed back signal. Not sure how RYM8RL is used. I have not seen the rest of the Mark V90 schematic as it is missing the relay control circuits. Probably why Mesa did not demand its removal as it is incomplete and not accurate past the initial release of the amp. they made changes in 2010 shortly after its release in 2009. It is unclear how many assembly revisions were made (change in components) or if there were other revisions to the main PCB.

If you really want to experiment with the power section of the Mark V, just make use of the FX loop, turn it on, and slave from another amp into the return of the Mark V90. That will disconnect the preamp completely as well as the GEQ circuit. At least that way you will be able to hear how the power portion of the amp will behave from another signal source. It may be preferable to make use of another tube amp instead of a solid state amp. If you did use a solid state amp for this trick you would need an isolation transformer typically used with signal splitters to decouple the two amp's power supplies. Lehel P-Split, or even the ebtech line level shifter would do the trick. If you go direct, not sure I would use a solid state amp for that. The last thing you would want is to sink DC current from the tube amp into the solid-state amp as the one with the larger power supply will always win. I found out the hard way what happens with the input of two different types of amps, I used the Line6 DL4 as a signal splitter from the guitar into a tube amp and solid state amp. That was a short test and did not work for long as the copper foil on the DL4 sort of vaporized. The circuit connected to the solid state amp (bass amp) wound up sinking current from the Mark V90. I was able to repair it with some wire jumpers. At least it did not fry any parts. I did get it working again but if it was dead, not a huge loss.

 

[Eevil]

When I slaved the Mark V90 into the Roadster, I got much of the same thing, brittle and ice pick tones on CH3 (IV and Extreme were the worse), CH1 tweed and Ch2 edge were also terrible. Not to mention that the Mark V GEQ circuit is used to create the send signal level, it has an impedance issue which can make things difficult with most FX processor buffers. The only products that actually worked with the Mark V were made by Strymon, DIG, BigSky, Brigadier were all good. Line6 DL4 was nasty, compressed the signal considerably, sure it sounded cool but it was not how it was supposed to sound.

When I was looking at the send levels I did take some pictures, I was running a 750 Hz signal if I remember correctly from a function generator. Here is what the signal looked like with the GEQ turned off. All of the controls were placed in their center positions.

View attachment 2742

The image is hard to see on the scope, here is a close up. The green trace is the FX Send output, note the leading edge marked with the arrows. This was CH3. Since the lead drive circuit comprised of the V5A -> V4B creates an asymmetrical waveform as I was led to understand, the sharp edge has more frequency content, If it had a slope or rounded corners like the portion of the waveform in the negative swing it would be a really good sounding amp. The sharp leading edge is ice pick tone.

View attachment 2743

Now to compare that to a Legacy amp, lets just say a modern version of such, the JP2C. Same frequency and the GEQ was turned off (not that it mattered since it follows the FX loop). Since I was looking at the signal amplitude to determine the signal level, I only took a picture with the amp in the view on the clean channel. It had no distortion. The next picture was on CH2 with some moderate gain dialed in.

View attachment 2744

Again with the CH2 controls at their center position. This is the difference in the distorted signal.

View attachment 2745

There is a sharp leading edge but the not a large spike. The Lead drive circuit is asymmetrical so the bottom portion has more of a normal distorted appearance The 2nd harmonic is visible. The top waveform has 4 ripples in it so that would be from the 4th harmonic. The leading edge having that sharp hook would be the higher order harmonics. My company will not fork out the money for a good signal analyzer as they generally start at $100k for an entry level.

Anywas, the reason the JP2C sound great, it does not push the limits on the upper harmonics that leads to ice pick tone. May as well call it shot-noise. Yes, that term goes way back to the tube days when the masters were creating their theory on design, mostly for the small signal analysis, Miller effect, and all that jazz. If you ever had any engineering courses on electronics and amplifiers in general (limited to the transistor), it all came from tube circuits. Some say the tube amp has a natural sound and will not generate odd order harmonics. That is BS. Class A/B is not as bad as a Class B amp. They may look similar (yeah, Class B does exist but not used) Class A/B will generate some nasty harmonics of the odd order but is corrected with negative feedback often called presence. The Simul-Class (combination of Class A/B and extended Class A) will create more harmonic content than a Class A/B, again corrected with the use of negative feedback.

I did have all the gear I needed to use to measure the output power except for hearing protection. No ear plugs or sound blocking headphones. Running a single frequency through an amp at elevated levels is pure punishment. 750 Hz was bad enough, 1KHz would have been much worse. Also would have had to use the clean channel as the power test is done without distortion. No thanks. Still tempted though.

Just a tangent.. did you measure/record the waveforms with 12AT7s in proper places (never going to remember the sweet slots )

 

[mace]

When you say ‘slave into the return’ do you mean that you connect SEND from one amp into the FX RETURN of another? Is there any risk to the amps when doing this?

 

[MarkV User]

Thank you for your analysis and corrections! Since I am making some mistakes in understanding things I'm going to say what my assumptions are about the relays. This might be wrong in part, and incomplete, and if you notice errors or want to add more or make corrections that is awesome.

Relay Functions

RYznw means relay for feature type z, variant n, circuit w. When the feature is enabled NO will be connected. Otherwise NC will be connected.

Channel relays:

RYC1w means "relay for channel 1, circuit w"
RYC2w means "relay for channel 2, circuit w"
RYC3w means "relay for channel e, circuit w"
RY3 is an additional mode relay for channel 3, used in the reverb

Mode relays

RYM1w Relay, mode 1, ch1/pos1, clean
RYM1 is an additional mode 1 relay
RYM2w Relay, mode 2, ch1/pos2, fat
RYM3w Relay, mode 3, ch1/pos3, tweed
RYM4w Relay, mose 4, ch2/pos1, edge
RYM5w Relay, mode 5, ch2/pos2, crunch
RYM6w Relay, mode 6, ch2/pos3, MarkI
RYM7w Relay, mode 7, ch3/pos1, MKIIC+
RYM8w Relay, mode 8, ch3/pos2, MKIV
RYM9w Relay, mode 9, ch3/pos3, Extreme

EQ

RYEQc Relay for EQ on, channel c

RYEQ1 eq on, ch1
RYEQ2 eq on, ch2
RYEQ3 eq on, ch3

Power mode

RYSEw single-ended power mode
RY100 90W ("100") power mode enables inner tube pair when not in single-ended mode

Other Relays

RYTPw outer tube screen control (?)
RYFXc Relay, FX loop enabled
RY-R Tube rectifier mode, patent# 5168438
RYMU mute mode, also lights the mute LED (?)
RYFB get pres. feedback from 8 vs 4 ohm tap
RYSM shorts 8 Ohm speaker output to GND?

Just keep in mind that the phase inverter circuit is a bit different than the designs you would encounter for Class A/B or traditional Simul-Class amps. Considering the 10W power mode is Class A only, C115 and C100 are more than likely used with the 10W power.

They relay name makes it look like C115 and C100 are used when channel 3 is active.

Didn't you say the icepick sound is worse when channel 3 is active?

The capacitor values are missing their units, they would be in pF. When the amp is at 45W or 90W, RYC3G connects the two phases with the 250pF cap, this will also be when RYSE1 is in its normally closed state to connect the bias to the control grids. In 10W mode, the bias circuit is disconnected from the one pair of tubes as this will be running in cathode bias power mode. I would not worry about the 250pF C110 as this is a basic plate-to-plate snubber. Since the phase inverter plates are 180° out of phase with each other, the capacitor serves to remove high ice pick frequencies. It also adds to stability. Typical range is usually 75pF to 100pF, making it larger will cut more ice from the tone.

I agree that C110 is a good thing and that 250pF might be an ok value. It is similar to what we see in other amps. If we read each side of the phase splitter as a voltage divider between the short circuit cap at 250pF and the pass cap at 0.1uF then it sort of seems to make sense.

I'm getting a little confused when I try to analyze the circuit though. Hopefully I'll sort this out for myself, but I see the frequency term cancelling out of the expression I have for the voltage, as though we are going to attenuate the same ar all frequencies. That has to do with the fact that I'm not analyzing the output impedance of the triode. If I can sort this out I'll plot the frequency response.

The 0.1uF pass cap is larger than we have on other amps. That affects the size we need C110 to be. My thought before was that we could be more aggressive here and either use smaller pass caps or a larger cap ar C110. And I don't see the need for this part of the circuit to have a different behavior for Channel 3. I think they got rid of the high cut to make the voicing more like the JPIIC+, which doesn't have the fearure. But let's see if we have the same understanding about the relay names and it will be easier to talk about.

So in 10W mode, half of the phase inverter is not used. So when RyC2G is in the NO position, that is more of a filter and part signal coupler from V7A to V7B. this will all be in the AC signal as the caps will block the DC voltage.

It looks like RYM9F for example might be "mode" 9, maybe the Extreme gain mode? And it would be the F=6th contact pair on that relay or gang of relays?

RYM9F, RYM8RL, and RY-FB I would assume are linked to the Extreme mode. Since the negative feedback circuit gets rewired from the 8 ohm tap to the 4 ohm tap, RYM9F un-shorts the
C63 cap that will cut some of the feed back signal. Not sure how RYM8RL is used. I have not seen the rest of the Mark V90 schematic as it is missing the relay control circuits. Probably why Mesa did not demand its removal as it is incomplete and not accurate past the initial release of the amp. they made changes in 2010 shortly after its release in 2009. It is unclear how many assembly revisions were made (change in components) or if there were other revisions to the main PCB.

If you really want to experiment with the power section of the Mark V, just make use of the FX loop, turn it on, and slave from another amp into the return of the Mark V90. That will disconnect the preamp completely as well as the GEQ circuit. At least that way you will be able to hear how the power portion of the amp will behave from another signal source. It may be preferable to make use of another tube amp instead of a solid state amp. If you did use a solid state amp for this trick you would need an isolation transformer typically used with signal splitters to decouple the two amp's power supplies. Lehel P-Split, or even the ebtech line level shifter would do the trick. If you go direct, not sure I would use a solid state amp for that. The last thing you would want is to sink DC current from the tube amp into the solid-state amp as the one with the larger power supply will always win. I found out the hard way what happens with the input of two different types of amps, I used the Line6 DL4 as a signal splitter from the guitar into a tube amp and solid state amp. That was a short test and did not work for long as the copper foil on the DL4 sort of vaporized. The circuit connected to the solid state amp (bass amp) wound up sinking current from the Mark V90. I was able to repair it with some wire jumpers. At least it did not fry any parts. I did get it working again but if it was dead, not a huge loss.

Thank you! I would not have expected that to happen either, and it is a good warning. It is surprising to me that there could be so much ground current between the two amps. If ground lift is off (as normal) then they should be connected to the same electrical ground or neutral, so noise can happen and small currents can flow--ugly enough--but no big currents. If ground lift is enabled on one amp then the other amp should be able to pull the isolated ground. Again there can be noise but things should work. So that is my idea but I know it is wrong because both you and others have reported large currents like that. Do you understand how this is possible?

Anyway I definitely agree with you that more tests are needed before changes make sense. The big one anyway is that thing you showed on the scope, so I'll try to observe something like that.

 

[MarkV User]

When you say ‘slave into the return’ do you mean that you connect SEND from one amp into the FX RETURN of another? Is there any risk to the amps when doing this?

I’ve had success with Slave out to Loop Return of another amp and with Loop Out to Loop Return. And maybe also loop out to instrument in is useful sometimes. These are all sometimes confusing because the Loop Levels don’t always match up and because the effects loop is managed differently on different amps. It seems I’m often trying to figure out what went wrong.

I didnt think there should be a risk until I just read Bandit’s experience. That is not supposed to happen, but from now on I think I’ll be checking the schematics before plugging things together this way. The worst problem I’ve had is ground loop hum and I have a AC splitter cable I use to minimize that.

 

[MarkV User]

By the way, as I remember it I tried to listen to the Mark V power amp in isolation. I have to try it again because I don’t remember what happened. I might not have tested it successfully.

So we will have these tests with a 750Hz signal generator sine wave output normalized to the right level, both low level and clipping, and maybe 600 ohms source impedance. I should build a 50:600 audio transformer isolation box to use the signal generator for this kind of test, and can probably cobble something together from parts on hand. I think the function generator is already internally isolated though, so maybe not necessary to go so far.

1. Mark V preamp from the instrument input to the effects send
2. Mark V power amp from the effects return to slave out

 

[bandit2013]

Just a tangent.. did you measure/record the waveforms with 12AT7s in proper places (never going to remember the sweet slots )

Where did I write that? No, when I ran the test, it was with stock tubes. If I had the 12AT7 in use it would only be in V4. That would be a good trick. Would have to borrow all the gear from work to test it with the 12AT7 but doubt it would make much of a difference. I tried the 12AT7 after I switched to the STR441 tubes, did not make any difference. Still bright. Went back to the 440 tubes, same thing. I must have broken something in the amp as the 12AT7 trick does nothing, sounds no different than the 12AX7. I may have added a resistor in series to one of the ice caps in the tone stack. It is either that or something else has failed in the amp. Hard to say really.

 

[bandit2013]

I’ve had success with Slave out to Loop Return of another amp and with Loop Out to Loop Return. And maybe also loop out to instrument in is useful sometimes. These are all sometimes confusing because the Loop Levels don’t always match up and because the effects loop is managed differently on different amps. It seems I’m often trying to figure out what went wrong.

I didnt think there should be a risk until I just read Bandit’s experience. That is not supposed to happen, but from now on I think I’ll be checking the schematics before plugging things together this way. The worst problem I’ve had is ground loop hum and I have a AC splitter cable I use to minimize that.

Slaving one amp into another, you can only do this with amps that have FX loops. It is a method to use a preamp of one to drive the power section of the other.

(AMP Master, FX send)->>> to ->>>(AMP Slave, FX return).

So if you tried connecting the Mark V90 send to a return of another amp, just do the opposite and take that other amp send and connect it to the return of the Mark V90. This will bypass the Mark V90 preamp so you can experiment with the power section. (note that the presence controls will remain active, if you run CH1, you can adjust the presence of CH2 to affect the output, CH2 and CH3 will behave the same way. Since CH3 has two presence modes, one for Extreme and the other for the other two modes, it will also affect the output characteristics since the signal is only from the low side of the OT feeding back to the phase inverter). The power modes will also work too.

A tube amp master send output to a tube amp master return should be ok since the chassis ground will be at the same potential. If you have concerns you can use a power strip but disconnect it from the wall outlet. Plug in both amps to that power strip, turn the power switch to on (nothing will happen as there is no power) and then use a multi-meter to test for resistance between chassis ground of one amp to chassis ground of the other. If it appears as a short circuit, safe to slave one to the other. If it does not, not a good choice unless you have a signal isolation transformer commonly used in guitar signal splitters (most should have a ground lift to uncouple the circuit grounds between the two amps). Solid state amps may not be using earth ground as the circuit ground for the FX send/return. You can measure the resistance from the ground prong to the sleeve of a cable plugged into the send or return. If it is earth grounded, the chance of one power supply back-feeding into the other is reduced considerably. If you cannot get a low resistance measurement between the cable sleeve to the chassis ground or the ground prong on the plug, it would not be advised to use that amp as a slave or master if the other amp is a tube amp.

 

[bandit2013]

By the way, as I remember it I tried to listen to the Mark V power amp in isolation. I have to try it again because I don’t remember what happened. I might not have tested it successfully.

So we will have these tests with a 750Hz signal generator sine wave output normalized to the right level, both low level and clipping, and maybe 600 ohms source impedance. I should build a 50:600 audio transformer isolation box to use the signal generator for this kind of test, and can probably cobble something together from parts on hand. I think the function generator is already internally isolated though, so maybe not necessary to go so far.

1. Mark V preamp from the instrument input to the effects send
2. Mark V power amp from the effects return to slave out

If you have a function generator, you would need to set the AC signal level to 750mV peak to peak. That is a typical guitar signal strength from passive pickups. It can vary but that is what I set it for.
I used a Lehel P-Split II as it has an isolation transformer which is what I needed to decouple the function generator from the ground of the amp. Did not want DC in the input signal either. A signal transformer with balanced coils should be fine, 1:1 ratio if you have an adjustable signal generator with an AC level. I only did this test to see what the send levels of three amps were at. TC was +4dBu if not a bit higher, JP2C was much closer to instrument level and the Mark V90 sort of fell in between.
Here is the link to what I was trying to do:

https://boogieforum.com/threads/mark-v-90w-fx-send-levels-measured-others.74195/

I was not intending to make a science project of out this. It was just for curiosity to see why the Mark V90 FX loop is terrible. I believe it is more of an impedance matching issue that a signal level problem as the FX send source impedance is too high for a line level signal driver. Also, it is unclear if the leading edge is a factor of the GEQ circuit. The driving transistor Q1 (NPN-MPSA20) and the two transistors that form the differential pair Q2 and Q3 (NPN-MPSA20) are still active as well as the output driver transistor Q4 (PNP-ECG159). This is the source for the signal to the Send jack. When the GEQ is set to off, the ground is lifted by a 1Meg resistor. That source is then passed through some large caps (10uF all modes except IIC+, or 0.22uF for IIC+ mode). The Send level circuit will be in parallel to a 100K resistor (IIC+ mode) or a 10K resistor. Send level pot is 25K ohms with a 1.5k resistor to ground. That is a bit high for a source impedance for line level. Also I looked at the link, I forgot I ran the test with the GEQ on in a V shape and GEQ turned off. The picture of the Mark V and scope may have been with the Extreme mode, GEQ turned off and the bright switch set to normal.

I am done with the Mark V90, not sure why I am participating in the first place. Perhaps there may be something I want to discover to make the amp worthy of ownership. Right now, it just takes up floor space.

 

[Eevil]

Where did I write that? No, when I ran the test, it was with stock tubes. If I had the 12AT7 in use it would only be in V4. That would be a good trick. Would have to borrow all the gear from work to test it with the 12AT7 but doubt it would make much of a difference. I tried the 12AT7 after I switched to the STR441 tubes, did not make any difference. Still bright. Went back to the 440 tubes, same thing. I must have broken something in the amp as the 12AT7 trick does nothing, sounds no different than the 12AX7. I may have added a resistor in series to one of the ice caps in the tone stack. It is either that or something else has failed in the amp. Hard to say really.

Sorry about being unclear I meant just that 12AT7 tube trick in V4

Sad thing to hear that your Mark V is still in brightness-limbo..

 

[bandit2013]

Ah I see, Eevil, Yeah, the Mark V90 is still in ice pick mode. I doubt that will change. I should check some of the dc blocking caps to see if there are any smaller caps in place of the 0.047uF. The amp is either ice pick or flubby.

I was able to cure the low frequency hum and ice pick character of CH1 (tweed bringing the ice but all three modes had noise). Along with CH2 fix on the edge mode, that no longer sounds shrill and lifeless. The trick with current production tubes or what may be in stock: V1 =Svetlana 12AX7 V2= Tung Sol 12AX7. I did try a Svetlana in V3 but found the unobtanium tube works much better (Mesa 6N4-J Beijing Square foil getter from the 90's). Actually, I loaded the rest of the amp with those, V3-V6. Took care of the flub too. Also, the STR441 tubes helped considerably. V7 is a Mullard 12AX7 long plate (similar to the Sovtek LPS).

When I can get it to sound half decent, I start to dislike it for something else. Boxy or lacking any top end. Just no happy median. However, I can actually enjoy the Edge mode as it sort of reminds me of the TC CH2.

 

[bandit2013]

Thank you for your analysis and corrections! Since I am making some mistakes in understanding things I'm going to say what my assumptions are about the relays. This might be wrong in part, and incomplete, and if you notice errors or want to add more or make corrections that is awesome.

Relay Functions

RYznw means relay for feature type z, variant n, circuit w. When the feature is enabled NO will be connected. Otherwise NC will be connected.

Channel relays:

RYC1w means "relay for channel 1, circuit w"
RYC2w means "relay for channel 2, circuit w"
RYC3w means "relay for channel e, circuit w"
RY3 is an additional mode relay for channel 3, used in the reverb

Mode relays

RYM1w Relay, mode 1, ch1/pos1, clean
RYM1 is an additional mode 1 relay
RYM2w Relay, mode 2, ch1/pos2, fat
RYM3w Relay, mode 3, ch1/pos3, tweed
RYM4w Relay, mose 4, ch2/pos1, edge
RYM5w Relay, mode 5, ch2/pos2, crunch
RYM6w Relay, mode 6, ch2/pos3, MarkI
RYM7w Relay, mode 7, ch3/pos1, MKIIC+
RYM8w Relay, mode 8, ch3/pos2, MKIV
RYM9w Relay, mode 9, ch3/pos3, Extreme

EQ

RYEQc Relay for EQ on, channel c

RYEQ1 eq on, ch1
RYEQ2 eq on, ch2
RYEQ3 eq on, ch3

Power mode

RYSEw single-ended power mode
RY100 90W ("100") power mode enables inner tube pair when not in single-ended mode

Other Relays

RYTPw outer tube screen control (?)
RYFXc Relay, FX loop enabled
RY-R Tube rectifier mode, patent# 5168438
RYMU mute mode, also lights the mute LED (?)
RYFB get pres. feedback from 8 vs 4 ohm tap
RYSM shorts 8 Ohm speaker output to GND?



They relay name makes it look like C115 and C100 are used when channel 3 is active.

Didn't you say the icepick sound is worse when channel 3 is active?



I agree that C110 is a good thing and that 250pF might be an ok value. It is similar to what we see in other amps. If we read each side of the phase splitter as a voltage divider between the short circuit cap at 250pF and the pass cap at 0.1uF then it sort of seems to make sense.

I'm getting a little confused when I try to analyze the circuit though. Hopefully I'll sort this out for myself, but I see the frequency term cancelling out of the expression I have for the voltage, as though we are going to attenuate the same ar all frequencies. That has to do with the fact that I'm not analyzing the output impedance of the triode. If I can sort this out I'll plot the frequency response.

The 0.1uF pass cap is larger than we have on other amps. That affects the size we need C110 to be. My thought before was that we could be more aggressive here and either use smaller pass caps or a larger cap ar C110. And I don't see the need for this part of the circuit to have a different behavior for Channel 3. I think they got rid of the high cut to make the voicing more like the JPIIC+, which doesn't have the fearure. But let's see if we have the same understanding about the relay names and it will be easier to talk about.



Thank you! I would not have expected that to happen either, and it is a good warning. It is surprising to me that there could be so much ground current between the two amps. If ground lift is off (as normal) then they should be connected to the same electrical ground or neutral, so noise can happen and small currents can flow--ugly enough--but no big currents. If ground lift is enabled on one amp then the other amp should be able to pull the isolated ground. Again there can be noise but things should work. So that is my idea but I know it is wrong because both you and others have reported large currents like that. Do you understand how this is possible?

Anyway I definitely agree with you that more tests are needed before changes make sense. The big one anyway is that thing you showed on the scope, so I'll try to observe something like that.

Ice pick is worse with Extreme and using the preset GEQ controls. Ice pick is totally intolerable in CH2 Edge and CH1 Tweed. I was able to manage most of the ice pick on CH3 by dialing out the treble and presence. It can also be fixed with preamp tubes without any circuit modifications.

There is some other matrix of circuits that turns on or off the GEQ on each channel. I assume it is in some of the hidden parts not in the schematic that pass the signal to the JFET. The RYEQx relays are not selected by the channel matrix, they are managed by the mini toggles next to the sliders.

RYEQ1 connects the differential amplifier (or at least that is what it reminds me of, Q2 and Q3) to the preset matrix or slider array.
RYEQ2, and RYEQ3 are also used to channel the preset matrix or the slider array to the associated inductor and capacitor.

The on/off per channel is managed by JFETS. Three J175(F,G,H) used to enable the preset pot based on channel (controlled by the channel bus ). The one below that array enables the preset circuit J175PS, or slider circuit J175SL (this will short out the preset circuit) and at the bottom is J175EQ to source to ground to enable the GEQ or to lift the ground with a 1M ohm resistor to cancel out its effect.

This image can be found by searching on-line. Not difficult to find the old schematic from 2009. It is incomplete.

There is one component, call it the boxy cap. C39 (120pF) cap that couples the grid to cathode on V4B. there is one on V5A and V6A as well. I tried removal of the one on V6A(C44) which was not a good part to remove. Never removed C35 on the V5A triode circuit as that is present on the IIC+ all the way up through Mark III, and Mark IVA, or B. What made the Mark IVB on the boxy side is the cap on the equivalent V4B circuit that is 250pF. I found it cuts some frequencies out and what gets added in is the harmonic content. I am considering removing C39 again, that was a real pain to restore, may also add another cathode resistor bypass cap to R57, one that is smaller in value than the 0.22uF cap that is used in normal or removed for bright. I did try some caps across the 3.3k resistor only to get into the mud. Plate resistor of 270K is on the high side. May even try a parallel resistor across it to reduce the gain a bit. Again, why? I gave up on this amp. Sure, it would be great to be able to correct the flaws in its design. I want to get rid of the amp more than hold onto it.

 

[MarkV User]

Thanks Bandit. You are asking what you're doing here and one thing I know you've been doing is helping me and other people our who might have gone through some of the same struggle with this amp you did.

I think a lot of people bought this amp thinking it would be their first and last really high quality guitar amp. Because of it's flexibility I always assumed I could kind of get any sound from it. When I found out how much I had misunderstood the amp recently, and had somehow come to dislike it, I went through a phase of being kind of pissed at the amp and at Mesa Boogie.

My take on this conversation is that now we are just waiting for me to do some homework and debugging with my amp. You showed me a lot of things, and with your encouragement I've finally gotten over the hurdle of understanding the relay configuration well enough to read the schematics.

I'm hampered at the moment by not having a good scope here, but the tooling we talked about earlier in the conversation should be just as good for this problem. So I need to get that together. More soon I hope.

When we started talking about this in a different thread some weeks ago I didn't have a really good characterization of what was wrong with this amp. I had just grown to hate it. If the problem is high audio or ultrasonic noise from the amp, which is what I've come to believe, then it is fascinating how I reacted to it. The pain from this ultrasonic noise is registered through the parasympathetic nervous system that I don't have conscious access to. So I never knew I was experiencing it and came to have a hugely negative emotional response to the amp itself. Consciously I noticed that my ears would ring afterwards though, so I did have some idea. But mainly my reaction manifested as a big emotional response. Well, that is my current take. It will be easier to talk about this with confidence after I do my test and measurement homework.

I want to thank you again for all your help and for selflessly revisiting your frustrating experience with this amp. Your experience makes sense to me because I went through something similar.

 

[MarkV User]

By the way, at the risk (as always) of speculating incorrectly about things, that EQ circuit might be capable of generating the thing we saw in your traces--a kind of inductive overshoot when the signal got clipped was what it looked like. In its simplest situation the spikes should be symmetrical but maybe if we analyze the circuit later we can figure it out. I started looking at it earlier but have to go out for an errand now.

Another possibility I keep thinking about is the negative feedback loop from the 8 Ohm output of the output transformer. I remain very suspicious of that circuit. There is a mix of positive and negative feedback there. It might be that there is one thing wrong with your amp and a different thing wrong with mine, and they both add up to high frequency resonance.

And like you just said, with whatever happens before that, the low/high filter on the output of the phase splitter should be filtering out ultrasonic frequencies at least. I'm not touching anything in the amp until at a minimum I have my test setup and a bunch of observations recorded, but I'm interested in changing that filter to match the plexi as closely as possible, or maybe the Bogner even. There are plenty of high and low frequencies in the plexi. The Bogner is darker but extremely satisfying. I think even if something is messed up upstream of that, correct filtering at that stage will help. And also I think that fixing the root cause and also adding correct filtering will make our amps sound as good as they can.

I also need to figure out how to calculate the input and output impedance of these triode and pentode stages so I can analyze the circuit. So there are still some hurdles. If I don't make progress by Monday I'm going to seek the help of an amp tech.

 

[bandit2013]

I know my issue resides in the preamp. I would not doubt some of it is with the GEQ circuit. The peak on the waveform and decay rate of the signal to the cut-off section is not inductive. It seems that the positive going portion of the waveform had more of a capacitor discharge or leakage effect. I had moved the center of the input signal so I could see that characteristic of the positive portion of the output waveform. You can see that the after the peak, there appears to be a decay to the zero crossover before the inversion takes place. That large white field just below the waveforms is the statistics window that I had open. It is a bit washed out as the display is on the bright side. Also there are two bright horizontal lines that are narrowing in on the average of the positive and negative sections. The waveform should have 2 divisions in rise and 2 divisions in fall based on the squares on the display. If it was anything like the other images, it would be 200mV per division. That peak is not overshoot as it may appear. Why the drop in the signal though? The negative portion is the amplified V5A signal. This is supposed to be an asymmetrical distortion of sorts the way I understand it.

Say if I were to superimpose the Mark V image over the JP2C, this was hand drawn but is as close as I can get to the peaks on the positive transition. The Mark V waveform decay was drawn in blue over the JP2C distorted trace. Just as a representation to indicate the positive and negative transitions are balanced. What is causing that? May be the key to the CH3 ice pick.

I looked at the phasing of each stage starting at the input. To the positive transition would be in phase with the V4B circuit, it gets inverted by V3A then the inverted again by V6A. First transistor is an emitter follower so no phase shift there. Q2 will be an inversion on the collector to drive Q4 base negative to turn it on.

I do not want to go down that path again, it is too long of a story to write about. I brought home a lemon but played a non-lemon at the shop in PA (should have taken the dirty scuffed up demo I tried instead of the nice minty clean one). No places in NC carry Mesa unless you drive to the other side of the state. I have family up in PA so I bought it while on vacation. Seeing that Lamb of God video on the Mark V90 made me realize they all sound like crap.

Badlander100 one of the best amps I have bought that is in current production.
RA100, best Mesa amp I ever played through and that is why I have two.
Mark VII, yeah, that is what a Mark is supposed to sound like.
JP2C, a little different but on par with the Mark VII and Badlander
California tweed, nice clean amp that will get mark like with the Flux drive up front.
Triple Crown, not bad, fun amps to run with.
Mark III, may be an early model but still far superior to the Mark V90.
Mark IVB, honky sound, more boxy like the Mark V but never ripped my eardrums out like the V with its glass shredding tone.

Mark V90. Pure ice, and it is not ice cream. I need a lobotomy after playing through it so I can forget the experience. I gave up on the amp as it may be more work to get it up to suitable usable sounds. My V90 has tone suck without using the FX loop. Call it sucky tone.

I think I may take the amp out to the front yard and set it on fire for Christmas.

 

[MarkV User]

Before you set the amp on fire, I may have some good news.

I’m kind of excited about it right now. I should be able to document it for you tonight but I’ll just tell the short version first.

For me the problem is after the effects loop return and I’ve proven it here using a cheap but serviceable USB scope I had in the closet. All settings in the preamp at 12:00 except for gain and volume which I turned up to get clipping. Scope connected to effects loop return, volume up. No spikes were happening at any settings looking at the clipped waveforms. I used a guitar with hot pickups for that but I’m sure it clips cleanly. I looked at all three channels but not at all modes so-far, but I’m going to repeat the test with function generator and look at all modes.

Then I switched the scope to the speaker out from the attenuator and saw overshoot like you saw at the termination of each rise and fall. I was looking at channel 1.

Then I switched to the function generator at 750Hz sine wave, low 700s mV max. And I moved the scope to the speaker out. I put on hearing protection and used an amp attenuator so I don’t damage the combo speaker. The overshoot seems to scale no matter whether the clipping is from high gain or high channel volume. An it happens the same on every channel and all modes except mode 9, extreme mode. There is no overshoot in extreme mode. Also the overshoot is symmetrical except in modes 7 and 8 where it is asymmetrical, just like you saw in mode 7, 2C+ mode.

I looked at the FFT of the waveforms and there are high frequency peaks way up into the ultrasonic. For human ears it looks pretty bad. In mode 9 with square clipping it looks completely different.

So here are some early conclusions. RYM9F turns off the feedback loop in mode 9, ch3/Extreme. I’m sure now that this problem in my amp is a problem in the feedback path.

In all Channel 3 modes RYC3G is in the NO position with its asymmetrical T-shaped filter cap arrangement. That is why half of the ultrasonic overshoot is truncated in modes 7 and 8.

More soon.

 

[MarkV User]

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 know my issue resides in the preamp. I would not doubt some of it is with the GEQ circuit. The peak on the waveform and decay rate of the signal to the cut-off section is not inductive. It seems that the positive going portion of the waveform had more of a capacitor discharge or leakage effect. I had moved the center of the input signal so I could see that characteristic of the positive portion of the output waveform. You can see that the after the peak, there appears to be a decay to the zero crossover before the inversion takes place. That large white field just below the waveforms is the statistics window that I had open. It is a bit washed out as the display is on the bright side. Also there are two bright horizontal lines that are narrowing in on the average of the positive and negative sections. The waveform should have 2 divisions in rise and 2 divisions in fall based on the squares on the display. If it was anything like the other images, it would be 200mV per division. That peak is not overshoot as it may appear. Why the drop in the signal though? The negative portion is the amplified V5A signal. This is supposed to be an asymmetrical distortion of sorts the way I understand it.

View attachment 2786

Say if I were to superimpose the Mark V image over the JP2C, this was hand drawn but is as close as I can get to the peaks on the positive transition. The Mark V waveform decay was drawn in blue over the JP2C distorted trace. Just as a representation to indicate the positive and negative transitions are balanced. What is causing that? May be the key to the CH3 ice pick.

View attachment 2787

I looked at the phasing of each stage starting at the input. To the positive transition would be in phase with the V4B circuit, it gets inverted by V3A then the inverted again by V6A. First transistor is an emitter follower so no phase shift there. Q2 will be an inversion on the collector to drive Q4 base negative to turn it on.

I do not want to go down that path again, it is too long of a story to write about. I brought home a lemon but played a non-lemon at the shop in PA (should have taken the dirty scuffed up demo I tried instead of the nice minty clean one). No places in NC carry Mesa unless you drive to the other side of the state. I have family up in PA so I bought it while on vacation. Seeing that Lamb of God video on the Mark V90 made me realize they all sound like crap.

Badlander100 one of the best amps I have bought that is in current production.
RA100, best Mesa amp I ever played through and that is why I have two.
Mark VII, yeah, that is what a Mark is supposed to sound like.
JP2C, a little different but on par with the Mark VII and Badlander
California tweed, nice clean amp that will get mark like with the Flux drive up front.
Triple Crown, not bad, fun amps to run with.
Mark III, may be an early model but still far superior to the Mark V90.
Mark IVB, honky sound, more boxy like the Mark V but never ripped my eardrums out like the V with its glass shredding tone.

Mark V90. Pure ice, and it is not ice cream. I need a lobotomy after playing through it so I can forget the experience. I gave up on the amp as it may be more work to get it up to suitable usable sounds. My V90 has tone suck without using the FX loop. Call it sucky tone.

I think I may take the amp out to the front yard and set it on fire for Christmas.