Skip to main content

I've been internalizing this for a while, but I just want to clear the air. I had quite a bit of work done on a vintage Marshall a few months ago and I can't get one or two things out of my head. The amp came to me biased for KT88s and it had a PPIMV installed. I told the guy remove the PPIMV, bias it for these EL34 tubes which I provided, and check over the entire circuit to make sure it's vintage stock. After receiving the amp back, I opened it and found that at least a few components are still in need of change. The power amp grid bleed resistors look to be the same as they were when the amp was biased for KT88s, 150k value. Stock Marshall value is 220k. The bias ranging resistor wasn't reverted to stock either, it's still the value that was "modified" to fit KT88s. The guy said this didn't "need" to be reverted, but it still bothers me. So there are at least the three components, if not more that I am missing. Do I have reason to be upset? Should I go back and see if he can complete the changes?

Comments

Boswell Fri, 11/22/2013 - 09:48

Going on the account you give, I don't think your Tech did a bad job. He did what he thought was needed to get it back working with EL34s. Maybe you weren't specific enough about what you obviously wanted him to do.

The EL34s will be perfectly happy with grid leak resistors of 150K, and, provided you are not driving them into grid conduction, the only effect using a lower value will be to reduce the time constant (and hence raise the low-frequency roll-off point) if the coupling capacitors from the phase-splitter stage remain the original value.

I don't know what you mean by the "bias ranging resistor". Are you talking about the cathode resistor in the output stage? An incorrect value there would have a much greater effect than slightly lower grid leak resistors.

JakeAC5253 Fri, 11/22/2013 - 10:30

Boswell, post: 408497 wrote: Going on the account you give, I don't think your Tech did a bad job. He did what he thought was needed to get it back working with EL34s. Maybe you weren't specific enough about what you obviously wanted him to do.

The EL34s will be perfectly happy with grid leak resistors of 150K, and, provided you are not driving them into grid conduction, the only effect using a lower value will be to reduce the time constant (and hence raise the low-frequency roll-off point) if the coupling capacitors from the phase-splitter stage remain the original value.

Thank you, that's what I figured, though I do have some issues with the sound of the amp. The amp tends to sound best with the Master on about 2 or 3 because after that it doesn't get any better it just gets muddier. That behavior doesn't sound like what I know to be a vintage "Marshall" amp :confused:

Boswell, post: 408497 wrote: I don't know what you mean by the "bias ranging resistor". Are you talking about the cathode resistor in the output stage? An incorrect value there would have a much greater effect than slightly lower grid leak resistors.

By Bias Ranging resistor, I mean the one that is in series with the Bias trimpot. I know it is probably a non-issue so long as the power tubes can be biased correctly, but I'm still worried that the wrong value resistor is changing the impedance of the bias to the tubes and is causing a drop in sound quality.

Boswell Fri, 11/22/2013 - 11:17

I can't find schematics for the various bias boards on Marshall amps. However, I do see that most of those amps are driven with grounded cathodes and a separate bias circuit driven from a negative voltage supply. It means that the bias ranging resistor you talk about can only have the effect of changing the upper and lower values of the bias voltage that you can set via the trimport and hence the upper and lower values of the quiescent standing currents though the EL34s.

If the quiescent current is set to a level you are happy with, you don't need to worry overmuch about the actual resistor values that achieve that bias. However, if the bias is still set for the KT88s, that could be where your problem lies.

JakeAC5253 Sat, 11/23/2013 - 13:51

Boswell, post: 408504 wrote: I can't find schematics for the various bias boards on Marshall amps. However, I do see that most of those amps are driven with grounded cathodes and a separate bias circuit driven from a negative voltage supply. It means that the bias ranging resistor you talk about can only have the effect of changing the upper and lower values of the bias voltage that you can set via the trimport and hence the upper and lower values of the quiescent standing currents though the EL34s.

If the quiescent current is set to a level you are happy with, you don't need to worry overmuch about the actual resistor values that achieve that bias. However, if the bias is still set for the KT88s, that could be where your problem lies.

I just measured the bias, and at resting idle, I seem to be measuring about 7.3mA

matthewfreedaudio, post: 408515 wrote: Do you like how the amp sounds? If so, stop complaining. If not, change it.

Production Sound Mixing for Television, Film, and Commercials.
http://www.matthewfreed.com

That's not very helpful advice, but thanks for linking your website anyway.

matthewfreedaudio Sat, 11/23/2013 - 13:58

My point is that many people stress about esoteric differences that have zero effect on the overall sound of an amp. If you like the sound of the amp as it is then don't change it. If you want to change the sound to achieve something else, do that. Whether or not the tech screwed you is up to you and the tech. None of us were present for your conversations so how can anyone here possibly give you advice on what to do? If you change the smallest difference in your amp will it help you write and record a hit song? No.

Production Sound Mixing for Television, Film, and Commercials.
http://www.matthewfreed.com

Boswell Sat, 11/23/2013 - 15:33

JakeAC5253, post: 408516 wrote: I just measured the bias, and at resting idle, I seem to be measuring about 7.3mA.

You don't give the model number of the amp, so I have no way of knowing what the standing current should be, but 7.3mA sounds way too little. I would have expected a value about four times that figure. If the 7.3mA is accurate, it could account for the reported behaviour as you increase the volume above the 2 or 3 setting.

Have you measured the range of currents achievable with the trimpot? Did your Tech say anything about having re-set the trimpot after changing back to the EL34s?

JakeAC5253 Sat, 11/23/2013 - 16:10

Boswell, post: 408520 wrote: You don't give the model number of the amp, so I have no way of knowing what the standing current should be, but 7.3mA sounds way too little. I would have expected a value about four times that figure. If the 7.3mA is accurate, it could account for the reported behaviour as you increase the volume above the 2 or 3 setting.

Have you measured the range of currents achievable with the trimpot? Did your Tech say anything about having re-set the trimpot after changing back to the EL34s?

It's a JCM800 100w 2203 from '83, and it's outfitted with a quad of matched Ruby Winged C EL34s.

I have not measured the available sweep on the bias trim, but this has piqued my interest, so I will do so asap.

JakeAC5253 Sun, 11/24/2013 - 14:50

Ok, I just measured the sweep of the bias trim. Today I am only getting a bias of 6.6mA for whatever reason. The trim sweeps from .14mA to 6.6mA. I'm guessing I'll need to drop the ranging resistor by a factor of 3 or 4 to get a usable bias sweep on this. I think we've found our issue... I guess the tech really did do a bad job. Probably got the thing back together and realized the trim didn't go to a usable setting and didn't want to disassemble it again to change one resistor so he sent it off thinking no one would notice.

Boswell Sun, 11/24/2013 - 15:20

Just before you go and change the resistor, can you say how you were measuring the standing current? If it was on the cathodes, did you insert a milliameter into one cathode lead at a time? One problem with that approach can be that the additional volt drop across the meter has the effect of transferring a large fraction of the standing current into the other valve (tube) of the pair of each phase, thus giving an artificially low reading on the metered cathode. An anode (plate) current measurement does not suffer the effect to the same degree.

On amplifiers that do not have cathode current-balancing resistors, I have seen the bias setting done by removing one valve in each phase when performing the current measurement. This is done to avoid the effect of current shift into the non-metered valve, but you are getting into quite advanced territory here.

The reason I'm a little suspicious is that, while there is a difference in bias settings between EL34s and KT88s, it should not be big enough to cause a factor of 4 drop in standing current, all other things remaining the same. Do you know for certain that the amplifier produced its full 100W output with the KT88s fitted, or was it deliberately under-ranged, for example as a recording amp but with the similar tone colours that you would get at full power?

JakeAC5253 Sun, 11/24/2013 - 15:31

Boswell, post: 408549 wrote: Just before you go and change the resistor, can you say how you were measuring the standing current? If it was on the cathodes, did you insert a milliameter into one cathode lead at a time? One problem with that approach can be that the additional volt drop across the meter has the effect of transferring a large fraction of the standing current into the other valve (tube) of the pair of each phase, thus giving an artificially low reading on the metered cathode. An anode (plate) current measurement does not suffer the effect to the same degree.

Boswell, post: 408549 wrote: On amplifiers that do not have cathode current-balancing resistors, I have seen the bias setting done by removing one valve in each phase when performing the current measurement. This is done to avoid the effect of current shift into the non-metered valve, but you are getting into quite advanced territory here.

The reason I'm a little suspicious is that, while there is a difference in bias settings between EL34s and KT88s, it should not be big enough to cause a factor of 4 drop in standing current, all other things remaining the same. Do you know for certain that the amplifier produced its full 100W output with the KT88s fitted, or was it deliberately under-ranged, for example as a recording amp but with the similar tone colours that you would get at full power?

I am quite perplexed as well, something doesn't seem to add up. Though I do find it suspicious that the bias trim was at its min setting rather than somewhere in the middle of the sweep that would be expected on an amp that was biased correctly. It seems to support the claim that there is something wrong with the way the amp biases, or the components that make it up.

I do not know for sure that the amp worked properly with KT88s. The amp came to me with only 3 working KT88s and I did not want to buy a fourth one so I bought a quartet of El34s instead. I only ever used the amp with two KT88s in it and it performed quite like it does now, not really as loud as I'd expect a 100w Marshall (though still not a quiet amp!), and doesn't really have much edge to its sound, and no dynamics.

Before I do any changing of components or pointing of fingers, I am going to socket a different tube and see what it meters at. Maybe it's just a bum tube.

EDIT: Just metered a different tube, same story: 6.6mA bias.

Boswell Mon, 11/25/2013 - 03:17

OK, those Eurotubes socket adaptors bring out the anode (plate) connections, so my worries aired above regarding problems of cathode current shifting from the metered valve to the unmetered one of a pair are unfounded.

You mentioned earlier that at least one resistor in the bias generator section had been changed from its original value. Do you know the before and after values of this resistor and which resistor in the circuit it was? What voltage do you measure at the output grid common bias point (the common end of the output stage grid leak resistors)? With the conventional screen grid connections, it should be somewhere near -45V, and if it is significantly greater (more negative) than this, it would account for the low standing currents in the EL34s.

If, on the other hand, the KT88 modifications included using higher screen grid potentials (or even triode connection!), that would also have the effect of reducing the standing currents for the same bias voltages. See this curve set for 450V screens and this for 200V screens.

JakeAC5253 Mon, 11/25/2013 - 11:14

Boswell, post: 408569 wrote: You mentioned earlier that at least one resistor in the bias generator section had been changed from its original value. Do you know the before and after values of this resistor and which resistor in the circuit it was?

Now that I look back at this part I am even more dumbfounded. Unless the predicted effect here is inversed due to having a negative biased power section? Unlikely. The resistor in question seems to now be a 15k resistor, and stock it would have been 27k. The trimpot itself is 22k.

I am using this as a major point of reference:
http://www.marshallheads.com/viewtopic.php?f=32&t=346

Boswell, post: 408569 wrote: What voltage do you measure at the output grid common bias point (the common end of the output stage grid leak resistors)? With the conventional screen grid connections, it should be somewhere near -45V, and if it is significantly greater (more negative) than this, it would account for the low standing currents in the EL34s.

I measured a voltage of -49.5V there.

Boswell, post: 408569 wrote: If, on the other hand, the KT88 modifications included using higher screen grid potentials (or even triode connection!), that would also have the effect of reducing the standing currents for the same bias voltages. See this curve set for 450V screens and this for 200V screens.

How would I analyze this?

Boswell Thu, 11/28/2013 - 05:36

JakeAC5253, post: 408584 wrote: How would I analyze this?

Visually inspect whether there have been any changes to how pin 4 (the screen grid) of the EL34s/KT88s has been connected. As I indicated earlier, some amp techs like to experiment with KT88s in triode mode, connecting pin 4 to pin 3 (the anode/plate), either directly or through a resistor. Triode operation would change both the biassing and the maximum output power.

You could try measuring the voltage on pin 4. I would expect it to be around 350V for normal operation (EL34-pentode or KT88-tetrode), somewhat less than the quiescent anode (plate) potential.

JakeAC5253 Thu, 11/28/2013 - 12:19

Boswell, post: 408647 wrote: Visually inspect whether there have been any changes to how pin 4 (the screen grid) of the EL34s/KT88s has been connected. As I indicated earlier, some amp techs like to experiment with KT88s in triode mode, connecting pin 4 to pin 3 (the anode/plate), either directly or through a resistor. Triode operation would change both the biassing and the maximum output power.

You could try measuring the voltage on pin 4. I would expect it to be around 350V for normal operation (EL34-pentode or KT88-tetrode), somewhat less than the quiescent anode (plate) potential.

I do not see any joining of the 3rd and 4th pins either through a resistor or bare wire.

At pin4 I seem to be measuring about 452v.

JakeAC5253 Thu, 11/28/2013 - 18:55

Just measured again and the Plate seems to come to rest on about 450 while the Screen now rests a little lower than last time right around 448.

The fact that there are minor variances in measured voltage each time, is that due to differences in facility current usage/city power you think? Or just the nature of tube amps?

Boswell Fri, 11/29/2013 - 02:08

So that's looking good, with pretty standard pentode (EL34) or tetrode (KT88) connections. That amp has no regulation on the V+ supply, so mains input variations will indeed have an effect on the measured potentials.

It seems to come down to the puzzle as to why, with around 450V on the top end and -50V on the grid, you are getting only about a quarter of the expected idle current through the output valves. In this sort of case, the finger of suspicion might normally point towards some sort of measurement error, but you have the under-performing sound of the amp to back up your meter readings.

Maybe it's time to get back in contact with your amp Tech and, in a non-accusatory way, go through the elements of the puzzle with him. There's probably a simple explanation for it all, and I for one would be interested in knowing what that is.

DrGonz Tue, 12/10/2013 - 01:12

Oh my god!! That is a horrible amp tech. Actually that guy does not deserve the Tech in there... Amp "Guy" is good enough. I guess I am just pouring fuel on the fire, but a good tech would have made sure that the amp worked good. Personally, I would not take my amp to that "Guy" ever again. He did not alter the bias circuit and the switching from kt88 to el34 tubes needs just that done.

JakeAC5253 Tue, 12/10/2013 - 22:03

DrGonz, post: 408851 wrote: Oh my god!! That is a horrible amp tech. Actually that guy does not deserve the Tech in there... Amp "Guy" is good enough. I guess I am just pouring fuel on the fire, but a good tech would have made sure that the amp worked good. Personally, I would not take my amp to that "Guy" ever again. He did not alter the bias circuit and the switching from kt88 to el34 tubes needs just that done.

Well, this IS southern Jersey (a.k.a. - the "Deep South") so we can't have our hopes too high for such a simple people, can we?

On a related note, I don't even feel like calling the guy up, I've been avoiding it. He'll probably want me to bring it in so he can look at it. Frankly, I'd rather just perform the modifications myself and see where we are afterward.

DrGonz Wed, 12/11/2013 - 02:25

You live and you learn. I mean you seem to know what your doing and this guy does not. You know how to drain the caps and solder right? That's how it all starts in the first place, and then you will only be able to blame yourself if you mess up lol. (y) Don't drink and work on amps with high voltages! Always ask questions just like you did and it all usually turns out fine. Take care!

JakeAC5253 Wed, 12/11/2013 - 12:12

Alright I have updates. I did the work myself, and the amp is biasing now! I manually changed the 150k grid bleed resistors to 220k resistors, and changed the 15k bias range resistor for a 27k resistor. Before I powered the amp on I put the bias trim from the max position to the middle of the sweep and when I clicked the amp into play mode after a quick warmup... the meter immediately read about 25mA. WOO! Finally I have a usable bias sweep on this. Oh and finally the amp sounds appropriate, or at least I can only assume it does. 2 on the Master right now is no longer a comfortable volume, so I've only tested it at very low volume, but the amp has opened up and I can imagine that it will sound much better when I open up the Master as well. (y)

JakeAC5253 Thu, 12/12/2013 - 16:22

I don't have an easy way to measure both the currents and voltages on the fly, as it would require powering the amp off and warming up again before taking each reading. Instead, I have the power section biased to a stable 33mA currently and the voltage readings for this setting are as follows:

Plate: 436V
Screen: 428V
Grid common: -40V

Boswell Fri, 12/13/2013 - 01:50

JakeAC5253, post: 408897 wrote: I don't have an easy way to measure both the currents and voltages on the fly, as it would require powering the amp off and warming up again before taking each reading. Instead, I have the power section biased to a stable 33mA currently and the voltage readings for this setting are as follows:

Plate: 436V
Screen: 428V
Grid common: -40V

That's looking good. The big test is how it sounds.