The REAL output of a Mark V/90

[twintriode]

Hello. I had an incident in November where I was playing my Mark V in Crunch mode and had the volume at about 3/4 and two things happened; 1) my speaker blew and caused one of the relays to fail, and 2) I didn't feel it was THAT loud to cause that problem. Every since that issue I have wanted to test the output of the amp to see....is this really a 90w amp? Is it possible the amp it outputting MORE than 90 watts to cause my 90 watt speaker to blow?

To add some context to my testing let me explain my testing method. I removed the speaker from the amp and placed a 100 watt 8 ohm load resister on the speaker input. I put a meter inline with that "load resistor" to measured AC voltage as I increased the volume. In parallel to the meter I also connected my scope to monitor the waveform, and check headroom of the amp. That step is not applicable to this test. Finally, I ran a 1khz signal input to the amp for reference.

For the test I used the RMS formula of Vac squared divided by 8 ohms times .707 to get the full output of each channel RMS. Below are the results;



Channel 1
Mode Tweed

90w mode
full output 42.77 watts

45w mode
full output 10.48 watts

10w mode full
output 2.61 watts


Channel 2
Crunch Mode

90w mode
full output 84.93 watts

45w mode
full output 19.99 watts

10w mode
full output 4.49 watts

Channel 3
Mode Mark IV

90w mode
full output 97.52 watts

45w mode
full output 34.57 watts

10w mode
full output 4.53 watts


In conclusion these findings sort of make sense. With all the modes and different input tubes driving the input signal, the output can NOT be static. You can't expect each channel to output the wattage represented on the wattage switch. There are just to many variables and circuits involved to provide consistent operation in each channel. Secondly, it should be noted that my input signal is 1khz whereas a guitar signal is going to represent more signals than just 1khz. Overall an actual guitar input would represent a small change in the overall output calculated.

My two questions that got answered in this test were 1) (Q) How did the 90 watt speaker fail? (A) the speaker is rated lower than 90 watts and Mesa never intended for the amp to be cranked as loud as I had it, and 2) (Q) Why does this amp seem to not be as loud as some Marshall and Fender amps I've owned in the past? (A) Because its not.

 

[KRichter]

Wow, you did alot of cool checking! Thank you, did the speaker blow cause of the
90w mode,
full output 97.52 watts? Thx!

 

[meesa_jarjar]

I would guess that most speakers are conservatively rated in terms of power handling, and with the thousands of Mark V's with 90W speakers, there haven't been widespread issues. I would guess that pushing the power section into distortion caused some other transients that spiked output enough to blow the speaker. Also as you stated, using a 1k test signal is standard of course, but guitars exist mostly well below that, and with the response of the speaker as one of many more variables, the wattage in actual usage could be quite a bit different.

Also the PI and NFB designs of the Mark series are pretty unique, and I don't think they are as 'clean' sounding as Fender / Marshall designs which are a lot more 'standard' and maybe even vanilla, that could account for the perceived volume differences.

 

[Jay Omega]

Usually amps are rated for average power at a chosen THD based on a sine wave input. As distortion and clipping increase, so does power output. The rule of thumb is make sure your speakers can handle 2-3x the nameplate power if you’re going to push it hard

 

[woodbutcher65]

Those measured power ratings tell me that the power tubes are never reaching the point of clipping. If they were, power ratings would comfortably exceed the published numbers.

Marshall's own documentation says that a "100 watt" model 1959 Superlead will put out 170 watts at 10 percent output stage distortion. In fact a particularly healthy example can put out as much as 220 at full tilt. We're past RMS measurements here.

But Mesas and Marshalls are very different animals. An old non master volume Marshall is a low gain amp and is meant to be running with the power section in saturation in order to get that rich Marshall grind. So it's stupid loud or not much fun. Meanwhile a Mesa is almost always meant to be run with the master volume controlling output level and with the majority, in fact nearly all, of the overdrive coming from the preamp. Actually turning a Mark amp up enough to get a lot of power stage grind is pretty near terrifying, and if you have much gain dialled in it's nearly uncontrollable due to all that feedback.

I've met a few Mesa players who've never had the fun of letting loose with an old Marshall superlead, and a few Marshall players who've never had the chance to play around with a good Mark amp. Both have to learn that the other is very different and you approach them differently. I've got a fair amount of experience with both and appreciate their very different but equally valid ways to get some great tones.

 

[bandit2013]

Interesting short cut.
The thing about the output and actual watts will be off on your calculations. RMS voltage and DC resistance is not going to provide you the truth. The impedance of the speaker will change with frequency as will the load current. How much of that is reactive or apparent? The best way to get a reliable measurement is to use an oscilloscope that has a power measurement capability. That will provide the power factor (which in this case we do not care about), real power (W), apparent power (VA) and reactive power (VAR) along with the phase angle. Even with that, you will get a median, high peak and low set of values. The lower the frequency the higher the current. I almost went to that length when I was measuring the FX output levels on several amps. I did not need to go as far as to measure the output power. I looked at the JP2C, Triple Crown, and Mark V 90 for that experiment.

However, I had blown out many speakers with the Mark V 90 after converting it to a combo amp. It was not until I was using a power attenuator used for recording purposes that had a peak power meter than I realized there was something off with the Mark V90. Before I discovered that, I was using a Mesa Roadster and it never peaked or metered anything above 100W. I was pushing the envelope with the amp too. I wanted to get some recording on the Mark V 90W after I loaded up a Celestion Creme 90W ALNICO speaker is when I discovered the issue. The Mark V was pegging the power meter on the power attenuator that had a maximum rating of 150W. That explained many things I have encountered such as a constant issue with red plating the center pair of tubes, did not matter if I was at 90W, 45W or even on variac power setting. The stock STR440 was an instant red-plate problem. Tung Sol 7851A did last much longer than a few months which was much better than 20 minutes (average time to red plate a new set of STR440. The best tubes I have used that survived the abuse were the SED =C= 6L6GC tubes. At the time, I did not realized to check the bias.

The bias for the 6L6 tubes was not where it should be, I could have run EL34 with the switch set to 6L6. I found the 82k resistor in the bias circuit did not measure 82k ohms. It was much lower than that, so I replaced it with a 91k ohm resistor that measured 87k ohms. The original resistor measured around 76k if I recall. Yes, the 82k resistor runs in parallel with a 3.3k resistor. With that change, the maximum peak output was around 110W. Much better than what was going on with the amp before this. I did measure bias voltage before I made the change. At full power 90W, the measurements were all over the place, not stable and this was with no input signal. When using variac power the bias voltage was stable at -51V. After changing the 82k resistor to the 91k resistor I saw the same variable bias voltage at full power but a different bias with variac power -55V. That was the first time I was able to run the Mark V with the stock STR440 power tubes without fear of red plating issue. Also the amp ran much cooler and was not overheating like it was before the change. That was the main reason I converted the head into a combo amp. Thought the speaker would aid in cooling the chassis. What I should have done was return the amp before making any changes. Also, long before the bias correction and with the red-plating issue, I also blew out the screen resistors. Oh Joy. Now after all that with the corrections and fixes the amp sounds great. Unfortunately I lost all interest in the amp.

 

[bandit2013]

I failed to mention that when power ratings are made for amps and speakers, there are several standards used for assessment. One standard varies from another. They do not list the standard to which they declare is the power rating.

 

[woodbutcher65]

If you want to sell that Mark V, some day, send me a message. I'm not at all opposed to picking up a second one....depending on price.


Your talk about converting it to a combo had me thinking evil thoughts of converting a Triple Rectifier Solo head to a 2x12 combo with a pair of EVM-12L speakers. But that's just EVIL, isn't it?

 

[Eevil]

Your talk about converting it to a combo had me thinking evil thoughts of converting a Triple Rectifier Solo head to a 2x12 combo with a pair of EVM-12L speakers. But that's just EVIL, isn't it?

Try lifting Roadster/Road King combo up first that will usually help with these thoughts…

 

[woodbutcher65]

I think I have about as evil (heavy) an amp as was ever made. I own a Fender 400PS. It's only a head. a 90 pound head.
435 watts RMS out of three pairs of 6550s. A Dual or Triple Recto combo wouldn't be able to scare me much.

I suppose I could prove insanity by building a combo out of it. Maybe it'd be a good use for a pair of EVM-15Ls.


I've also owned a Fender Super Twin Reverb. That was only 180 watts.

I spend plenty of time in the gym. I can handle some pretty heavy gear but even better, I know how to put casters on a cabinet.

 

[bandit2013]

Yeah, I did a bad thing with the Mark V. I even sold the head shell and regret doing that after some of the corrections I made to it. Now it sounds great and I am stuck with a combo. I have run it through a traditional cab loaded with Creamback H75 and that was awesome. Bummer. Since I blew out all of the C90s as well as the Celestion Creme 90W Alnico, along with a GWS BlackHawk HP100, all I had left to use were EVM12L black label, EVM12L classic and a pair of Organic Timbre Rhapsody G12F speakers. The OTR is similar to the EV but has a different voice to it and not as bloody bright. Forgot about the redback. At the moment I have the OTR in the combo and one of the EV speakers in a widebody extension cab (after market, not mesa). I also have a widebody open back 112 with EV loaded in it too. It is not the end of the world. Note the EV or OTR will not fit with the amp chassis if mounting the speaker on the inside. Have to mount the speaker on the front of the baffle and modify the grill for clearances. Thought about selling the Mark V, but have not made any decisions yet. Perhaps I am not done with it. I suspect the bias issue may be with one of the capacitors since it is not stable unless variac power is used. Measured without any input signal. Diode is fine. EL34 bias is also correct but should check it. Also I was incorrect on the parallel resistor, it is not 3.3k but 33k ohms. I did try it without the 82k in the circuit just for S&G. That was a bit cold. I may just add a pot to this so I can dial it in. Not the best way to set a bias though. While I am at it, may pick up some bias probes. As it seems the center pair of tubes running extended class A are the first to go when the 82.5k 1% resistor was in the circuit. I did check all of the resistors on the power tube circuits. Nothing out of whack there. I did see that the Mark IV uses a -57V bias. Not that it should matter with the 10W power mode since it runs the two power tubes in cathode bias mode and not grid bias mode.

 

[pmac11]

I failed to mention that when power ratings are made for amps and speakers, there are several standards used for assessment. One standard varies from another. They do not list the standard to which they declare is the power rating.

Agreed. It would be nice if the industry just adopted W = (Vrms)^2 / R, where Vrms = 0.707 Vac, and Vac is measured at the onset of clipping. I've seen that standard in a few audio texts.

 

[pmac11]

I failed to mention that when power ratings are made for amps and speakers, there are several standards used for assessment. One standard varies from another. They do not list the standard to which they declare

Duplicate, sorry

 

[rgx612a]

Nerds!

 

[steve Rhenius01]

Hello. I had an incident in November where I was playing my Mark V in Crunch mode and had the volume at about 3/4 and two things happened; 1) my speaker blew and caused one of the relays to fail, and 2) I didn't feel it was THAT loud to cause that problem. Every since that issue I have wanted to test the output of the amp to see....is this really a 90w amp? Is it possible the amp it outputting MORE than 90 watts to cause my 90 watt speaker to blow?

To add some context to my testing let me explain my testing method. I removed the speaker from the amp and placed a 100 watt 8 ohm load resister on the speaker input. I put a meter inline with that "load resistor" to measured AC voltage as I increased the volume. In parallel to the meter I also connected my scope to monitor the waveform, and check headroom of the amp. That step is not applicable to this test. Finally, I ran a 1khz signal input to the amp for reference.

For the test I used the RMS formula of Vac squared divided by 8 ohms times .707 to get the full output of each channel RMS. Below are the results;



Channel 1
Mode Tweed

90w mode
full output 42.77 watts

45w mode
full output 10.48 watts

10w mode full
output 2.61 watts


Channel 2
Crunch Mode

90w mode
full output 84.93 watts

45w mode
full output 19.99 watts

10w mode
full output 4.49 watts

Channel 3
Mode Mark IV

90w mode
full output 97.52 watts

45w mode
full output 34.57 watts

10w mode
full output 4.53 watts


In conclusion these findings sort of make sense. With all the modes and different input tubes driving the input signal, the output can NOT be static. You can't expect each channel to output the wattage represented on the wattage switch. There are just to many variables and circuits involved to provide consistent operation in each channel. Secondly, it should be noted that my input signal is 1khz whereas a guitar signal is going to represent more signals than just 1khz. Overall an actual guitar input would represent a small change in the overall output calculated.

My two questions that got answered in this test were 1) (Q) How did the 90 watt speaker fail? (A) the speaker is rated lower than 90 watts and Mesa never intended for the amp to be cranked as loud as I had it, and 2) (Q) Why does this amp seem to not be as loud as some Marshall and Fender amps I've owned in the past? (A) Because its not.

 

[steve Rhenius01]

Don't see too much wrong with these figures - except failure of the speaker ! I have a Mk 5 which stood unused for the Pandemic but when switched on afterwards ( unused for 3 years near enough) it didnt come on at all. The 12 V supply had failed - IN4007's in rectifier all blown. Replaced these, checked all valves and 12 supply to heaters and relays restored. Signal appears as far as V3 but nothing at grids of V7 to supply power valves. From posts I've seen this is not an uncommon problem = possibly relay failure or failure of J175 FET's. Puzzled about what could have caused this ? There's is a report from Boogie recommending upgading the IN4007 diodes. I'm curious about what caused this failure in the first place.

 

[twintriode]

Interesting short cut.
The thing about the output and actual watts will be off on your calculations. RMS voltage and DC resistance is not going to provide you the truth. The impedance of the speaker will change with frequency as will the load current. How much of that is reactive or apparent? The best way to get a reliable measurement is to use an oscilloscope that has a power measurement capability. That will provide the power factor (which in this case we do not care about), real power (W), apparent power (VA) and reactive power (VAR) along with the phase angle. Even with that, you will get a median, high peak and low set of values. The lower the frequency the higher the current. I almost went to that length when I was measuring the FX output levels on several amps. I did not need to go as far as to measure the output power. I looked at the JP2C, Triple Crown, and Mark V 90 for that experiment.

However, I had blown out many speakers with the Mark V 90 after converting it to a combo amp. It was not until I was using a power attenuator used for recording purposes that had a peak power meter than I realized there was something off with the Mark V90. Before I discovered that, I was using a Mesa Roadster and it never peaked or metered anything above 100W. I was pushing the envelope with the amp too. I wanted to get some recording on the Mark V 90W after I loaded up a Celestion Creme 90W ALNICO speaker is when I discovered the issue. The Mark V was pegging the power meter on the power attenuator that had a maximum rating of 150W. That explained many things I have encountered such as a constant issue with red plating the center pair of tubes, did not matter if I was at 90W, 45W or even on variac power setting. The stock STR440 was an instant red-plate problem. Tung Sol 7851A did last much longer than a few months which was much better than 20 minutes (average time to red plate a new set of STR440. The best tubes I have used that survived the abuse were the SED =C= 6L6GC tubes. At the time, I did not realized to check the bias.

The bias for the 6L6 tubes was not where it should be, I could have run EL34 with the switch set to 6L6. I found the 82k resistor in the bias circuit did not measure 82k ohms. It was much lower than that, so I replaced it with a 91k ohm resistor that measured 87k ohms. The original resistor measured around 76k if I recall. Yes, the 82k resistor runs in parallel with a 3.3k resistor. With that change, the maximum peak output was around 110W. Much better than what was going on with the amp before this. I did measure bias voltage before I made the change. At full power 90W, the measurements were all over the place, not stable and this was with no input signal. When using variac power the bias voltage was stable at -51V. After changing the 82k resistor to the 91k resistor I saw the same variable bias voltage at full power but a different bias with variac power -55V. That was the first time I was able to run the Mark V with the stock STR440 power tubes without fear of red plating issue. Also the amp ran much cooler and was not overheating like it was before the change. That was the main reason I converted the head into a combo amp. Thought the speaker would aid in cooling the chassis. What I should have done was return the amp before making any changes. Also, long before the bias correction and with the red-plating issue, I also blew out the screen resistors. Oh Joy. Now after all that with the corrections and fixes the amp sounds great. Unfortunately I lost all interest in the amp.

I always get a great laugh when reading your comments. Thanks

 

[twintriode]

Wow, you did alot of cool checking! Thank you, did the speaker blow cause of the
90w mode,
full output 97.52 watts? Thx!

No it blew on ch2 mode 2 at about 2/3 full volume. The Black Shadow is really a 80w spec Celestion with a Mesa Boogie label on it saying its 90w. When I got the replacement speaker I found factory evidence of the 80w, and I also had to apply my own label to the end.

 

[twintriode]

Don't see too much wrong with these figures - except failure of the speaker ! I have a Mk 5 which stood unused for the Pandemic but when switched on afterwards ( unused for 3 years near enough) it didnt come on at all. The 12 V supply had failed - IN4007's in rectifier all blown. Replaced these, checked all valves and 12 supply to heaters and relays restored. Signal appears as far as V3 but nothing at grids of V7 to supply power valves. From posts I've seen this is not an uncommon problem = possibly relay failure or failure of J175 FET's. Puzzled about what could have caused this ? There's is a report from Boogie recommending upgading the IN4007 diodes. I'm curious about what caused this failure in the first place.

Because the quality of tubes is getting/has been pretty bad, Mesa Boogie uses the 12VDC to provide heater voltages to some of the more critical input tubes. These tubes would be more likely to go microphonic in AC heater situations so they use the DC on select tubes to somewhat overcome that. The reason for the 1N4007 failure is because they spec'd them too small on the current side of things by choosing the 1N4007.

Depends on the amp, but I've mainly seen the JFET's used to handle footswitch pop. You start hearing pops when changing channels its time to replace those little guys.