Marshall Mercury 2060 Repair – Days 4-6
DISCLAIMER: The following is NOT to be taken as a definitive procedure for repairing ANY electronic device and the author takes NO responsibility for any damage or injury that results from anyone using this guide. It is intended for educational purposes ONLY.
If you have ANY doubt about making modifications or repairs to your own equipment then seek advice from relevant qualified persons.
Valve amplifiers use and can store high AC and/or DC voltages that can KILL. You have been warned!
There is a big effect difference between voltage levels that break through the skin conductivity – which happens to be in the 400 to 500V DC range of our amp circuits! BE CAREFUL! Drain the secondary capacitors when the circuit is off and always check the circuit with a multimeter on both AC and DC sides before touching components or working on it – with the mains UNPLUGGED. If the Mains is plugged in, the plug socket and switch contacts are LIVE still, with the switch in the OFF position.
If you are testing the circuit after a change, make sure all your test leads, probes and croc clips are well insulated from your fingers and chassis contact points (chassis edges can be sharp too). I always connect the croc clips and meter and double check. I then plug in the fused mains side. I use the mains toggle switch to power the secondary with the indicator light connected so I can SEE power also, as well as checking the on/off position of the switch BEFORE I plug in the mains. I switch OFF and unplug the mains, checking the meter voltage or current level has dropped to 0V, (or use a drain lead on the capacitors ASWELL) before I move any croc clips to a new circuit point for another measurement.
For the sake of a few extra seconds, why risk safety by not switching off and draining the circuit?
A 240V AC shock will certainly help you not make that mistake again if you ever forget any of this – if you are lucky enough to not get away with it! My last one left my right arm aching for an hour!
A 500V DC shock may mean you don’t get to worry about any of this ever again…but now you know why.
I have spent a lot of time studying the circuit diagram at various points so will do a brief overview of its basic function later. The sites that were really helpful to me in understanding valve function and biasing in particular are:
For now, day 4 involved writing out a list of resistors R1 – R24, as I did for the capacitors, to see what changes have been made to this circuit by prior owners, as it had been modified a little, as the different capacitors values list showed, and that there was a resistor soldered onto the back of the PCB.
Again, a couple of resistors in the tremolo section (bottom left of diagram – footswitch section) were different values than the circuit diagram. I’ll try to explain a bit of the various sections of the diagram later at a very basic level. The different values were:
R18 – 470k (220k)
R19 – 470k (270k)
R20 – missing (220k)
These only affect the tremolo section – not critical, but as this didn’t work at all…
I checked the continuity of PCB tracks from end to end and all were ok. I still bought a silver PCB track pen just to draw over some really faded bits of track, and for the experience. Not recommended at 25 quid for a pen from Maplins! The nib needs pressing down hard, and silver does not flow well at all on partly corroded tracks. It is more of a liability for accidentally joining different tracks that are close together, if you slip. I don’t know how you would remove it either.
I cleaned the board first with contact cleaner (Servisol Super 10) first to remove dust etc. and wiped gently with a tissue. The contact cleaner is alcohol based so evaporates. I cleaned the volume, tone and tremolo pots also.
Once I was happy that the mains wiring (earth particularly), mains fuse, transformer wiring and PCB tracks were good, and the replacement JJ manufactured EL84 arrived, I thought I may as well see what would happen by powering up, knowing there was sufficient electrical safety in place with the earth connection ok and the 500mA fuse intact on the chassis.
I reconnected the speaker (valve amps can be damaged by running them with incorrect or no speaker load), put in the new valve, and turned the amp on. After a few seconds, the valve glowed bright yellow at the top, and the speaker quietly popped into life with a bit of mains hum. I plugged in my guitar, and heard a pretty awful tone from the speakers. It was very distorted and not very loud, even at full volume. After 5 mins trying different pickup selectors, volume and tone control settings, I shut it off – at least it works at this point!
Assumption is the mother of all fuck ups!
Just out of interest I decided to try the old valve that I assumed would be dead due to the burn ring around the top – and it worked! At least I have a spare now, which all valve amp owners should have anyway. It still sounded as shit as the new one, and was also glowing yellow hot at the top, which I read later is NO good for the valve life, and can also cause distortion of the signal as the cathode current is at or near saturation.
I spent the rest of the night studying the circuit and researching valve pages to get some clues as to the crappy sound, and/or the valves running too hot.
I was pleasantly surprised that after 5 mins turned off, there was only 1V at the rectifying capacitor CT, so this amp is not as inherently dangerous as others can be by storing charge for days or more.
I later read that if you turn a valve amp off and keep playing until no sound is heard from the speakers, then the storage capacitors must have pretty much drained as they have no charge left to power the speakers. This happens after a few seconds with this amp.
It does NOT mean that you don’t drain and check the DC cap voltages every time before you work on the circuit or move the chassis!
I decided to revert this amp back to the original circuit as much as possible as I wasn’t sure what effect the changed resistor and capacitor values were having on the amp, done by prior owner(s) etc. which I thought, may be part of the incorrect bias settings. It turns out that valve specs can be WAY different off the assembly line, so they should ALL be checked for correct biasing when changed. (Thanks to Lord Valve at DuncanAmps.com for a lot of myth busting info on valves).
At least I would have a baseline to start from if it still sounded crap – and it is all experience – soldering and pondering my long lost basic electronics theory from almost 30 years back. I just kept telling myself there is no rush for this, so take my time and think carefully about what the best is for getting this amp sounding better.
I read in various places that any old amp would probably due for a change of the electrolytic capacitors at least, so went with my capacitor list to Maplins, and managed to get all of them except the main large rectifying 32nF dual anode.
A few are larger values than in the circuit, but in an audio circuit, this would mean more bass should pass generally which is desirable with what I’ve read on this amp, which is that it’s way too trebly on the ears – with the tone control up full – for sure .
All new caps are much smaller these days, and some types don’t even have long legs to solder. I decided to skip the ceramics as I couldn’t get most of those anyway and newer ones are in ridiculously small packages.
I also got a 220k resistor to make a cap drain lead for when I next power it up.
I changed all four electrolytic caps, added those missing, and changed the three resistors that had higher values than the diagram.
The missing or incorrect capacitor values were:
C4 – missing (220pF)
C23 – missing (220pF)
C20 – 680nF (47nF)
C25 – 32uF 350V (10uF) Larger the better for a reservoir cap
The electrolytics that were changed for new ones were:
C12 – 22u (25u)
C15 – 100u
C24 – 10u (22u)
C25 – 10u (47uF) 350V
I couldn’t get a 32uF 450V dual anode to replace the CT (C17/18) so decided to leave it.
You get more familiar with a circuit the more you study it obviously, and I could then break it down into 4 distinct sections;
Power supply rectifying (bottom right) and high voltage (275V) section
Tremolo section (bottom left) with T3 BC184 transistor, oscillator circuit
Input section (low/hi) and transistor gain stages (2 x BC184), as input for the valve (top left)
Output section (valve and 4 Ohm speaker – top right)
The way to see these is to re-draw on paper the DC resistor chains that set the static DC bias voltage for the 3 transistors by following where the 275V rectified output from the mains transformer (X) joins the speaker transformer (X) and goes via R17 (1k), R16 (27k), point A (R18, 220k), R21 (6M8) and R22 (220k) to earth.
There are 2 other equal chains for T1 and T2 – a 22k, a 6M8 and a 6M8 in series for these.
If you do your Ohm’s Law (V=IR) sums for any of these chains you get the voltages at each resistor and the current through each chain e.g for T3 chain:
275V/1k+27K+22k+6.8M+6.8M = 275V/13.848MOhms = 20uAmps.
Now you can work out the voltage drop across each resistor, e.g. 6M8 x 20uAmps = 136V
For JUST these resistor chains, as a learning exercise, (not a real working circuit at this stage) this means that T1 and T2 would have base bias voltages of about 135V, and T3 base has 4.4V above its 220k resistor (220k x 20uAmps = 4400 x 1/1000V)
The next stage to understanding how the rest of the components sit in parallel with R4, R9 and R21 require an understanding of resistor/capacitor circuits for AC and DC behaviour, which goes beyond my knowledge to explain in detail here. Once the capacitors have charged up, they will settle at voltage values that are dictated by the combined resistances of the other resistors that are in parallel with the theoretical “main chain” resistors above. For example, T1 section has R2 (100k) in parallel with R4(6M8) so their combined resistance becomes less than the lowest 100k value as the current now has two paths to take to earth.
As 1/R Total = 1/R1 + 1/R2, the value for the two in parallel becomes 1/R = 1/100k + 1/6M8 = 98.5k Ohms. This now gives a much lower voltage at the base of T1 than the “main chain” alone would give of 136V, which would kill the transistor T1 in reality.
Basically, it a case of trying to break a circuit down into its functional sections – input and output stages, or FX sections etc. – rather than try to understand the circuit as one complex system in one go. It’s the way any complex system is built up – from more basic modular sections that can function independently – then added together.
I hope that helps beginners a little.
Suffice to say, from the BC184 spec. sheet, the actual voltages between the base and collector of each transistor cannot exceed 30V max for this BC184, or 45V max between the collector to base, or 6V between the base and emitter. The main thing to understand here for beginners and novices like me, is to be able to identify and separate the main sections of the circuit, and realise that changing components in these sections affect the operation of these sections. This might be to increase or decrease the gain of each of the gain stages that might give rise to distortion of the transistors, or change the rate that the tremolo can oscillate etc.
Maplins did not have BC184 transistors, but they gave me a spec sheet for this transistor.
These all work luckily, but I am not sure if some of the distortion is not due to the T1 gain stage, as the guitar can overdrive the amp when the guitar volume is full up, but as I don’t have an oscilloscope I can’t check this. The distortion could also be from the T2 stage or the valve.
For the T3 tremolo circuit, you can see that an oscillating output is fed via the variable 100k pot (R4) to the 2k7 (R12) gate input of the valve, which would give an overall slow positive and negative swelling gain (amplitude modulation) to the amplified guitar signal (from T1 and T2 transistor gain stages) to the valve input.
To get a basic understanding of oscillator RC circuits as used in the tremolo section, read here:
Looking at this circuit now and comparing it to the Marshall, I’m confused as to why C23 was missing from this amp? It seems C23 is also helping negative feedback from T3 output back to the base input, assuming C20, 21, and 22 are the phase inverters. Or, would C23 also be 60 degrees out of phase – as C20 is – which is also modulating the 180 degree signal at C22?
Now you can see that the 100k pot VR3 must be the tremolo speed control, as a reduction in resistance here means the charge on C20-24 will flow on and off more quickly. VR4 is the intensity knob, as it dictates how much of the tremolo output is added to the input of the valve gate at R12 to give a slow modulation to the amplified guitar signal.
The calculation of frequency in RC circuits:
If all the resistors, R and the capacitors, C in the phase shift network are equal in value, then the frequency of oscillations produced by the RC oscillator is given as:
f = 1/2(Pi) x RC x (2N)^0.5
• ƒr is the Output Frequency in Hertz
• R is the Resistance in Ohms
• C is the Capacitance in Farads
• N is the number of RC stages. (N = 3)
This would mean that the tremolo in this circuit is a 3 stage type using the three 47nF capacitors, and the three 220k resistors (I’ll miss out the 100k pot for ease of calculation).
Using the formula above, I make this a tremolo frequency of:
1 / 2 x 3.14159 x 220,000 x (47 x 10^-9) x [(2 x 3)^0.5] = 1/0.159138 = 6.28Hz when the Speed pot is on minimum.
Note that the hi and low guitar jack inputs are actually closed (shorted) to earth (I checked this with a multimeter) until a guitar is plugged in, so there is no AC input when nothing is plugged in, so you don’t get noise picked up by the input, fed to the speaker.
I ended up back at Maplins today for the resistors I decided to change, and to change C20 in the tremolo section from a 680nF to its correct 47nF, along with the wrong R18 (was 440k), 19 (was 440k), and 20 (was 470k), as the tremolo didn’t work yesterday. From the oscillator circuit above, it seems you would want all the phase inverting capacitors C20-22 to be all the same value – 47nF, else the charges would flow at different rates and you would not get a clean phase inversion I guess? That may be why I didn’t hear the tremolo working before? I don’t know enough about RC circuits to be sure.
I also got a mains neon light, which was almost identical in look and size to the Marshall 6V, 40mA, SGF99 bulb. I just had to file the hole a fraction bigger to fit it, and solder it from the mains switch neutral wire to earth, instead of across the 6V mains secondary transformer for the old 6V bulb.
The most important bit today was getting a 200 Ohm variable pot so I could try raising the resistance at the valve cathode to try and lower its bias current so it might stop the distortion, and make the valves run cooler. I glued this to the back of the PCB so that there is some extra resistance leeway when valves are changed in future. I soldered it in series with the 100 Ohm (R15) resistor to PCB earth. This will also make it a lot easier for a fully equipped tech to put a multimeter in series with the cathode resistor to measure and set the exact cathode current in future.
As this circuit was already tampered with, it was not important anymore to keep original parts, but just get a half decent sound out of this awful sounding amp.
When all the new components and mods were done I was ready to fire it up again.
Much to my surprise and delight, it warmed up after 15 secs or so, and the speaker popped into life with the normal mains hum, but a lot less than before. I checked the valve colour and it seemed a lot less yellow. I plugged in the guitar, and there was a bit of an improvement in sound and volume. I increased the pot resistance a bit while strumming the guitar and the distortion lessened a bit. I ended up turning this 200 Ohm pot full up before the distortion got minimal. I kept checking to see the colour of the valve, which now is a more normal orangey yellow colour, so it is now running cooler, and the sound is much better. Still very crap overall as an amp, and there is little volume still.
This now takes the effective value of R15 to 300 Ohms, and I still think it could have done with a bit more. I played the amp for about ½ an hour and it seemed to improve in sound a bit once warmed up. Maybe that was just me getting used to it? There are particular guitar and amp volume settings that have to be found to retain a clean sound when chords are played and a lot depends on the strength of the pickups overdriving the transistor/valve stages. Overdriven transistors hardly sound great at the best of times, but certainly not when they a NOT designed to be overdriven. There is a serious lack of bass with this amp, and the speaker can’t tolerate much volume without distorting anyway, as I checked it separately through my Technics stereo (see Summary).
The tremolo section works but is subtle and only really noticeable as a slow swell when both speed and intensity controls are up full. Again, I suspect age related component drift, and the odd wrong value capacitors. I guess the last persons mod attempt was to get a more definite tremolo, but changing R18, 19 and 20 to 470k or so should have made it worse I think, because the capacitors would have taken longer to charge then discharge periodically through these than the diagram values, and it is a slow and subtle tremolo anyway, with the right values.
After the amp cooled down I swapped the old valve for the new, and immediately found a slight improvement in sound and volume, but not much. The JJ looks more orange than yellow also, so would appear to be running cooler and closer to circuit spec. I now get a slightly higher volume without distortion, so am leaving it at that.
I put an 8 Ohm 50W Mordant Short in place of the 30W Celestion 10″ to see if that was duff (which it must be – see Summary), but got the same sort of overall distorted sound as the Celestion, so even if the Celestion is duff, it doesn’t change the fact that the amp circuit is not right anyway.
I would have liked to confirm better performance by raising the pot/R15 value to say, 500 Ohms but don’t have the component, and am not driving 8 miles to Maplins then spending another ½ day messing with soldering etc. Quit while I’m ahead. At least I have a working collectors amp, and now the cab is all cleaned with soap and nail brush, it can be sold as a working, classic bit of Marshall history.
Summary and observations
This was a near perfect first project for me – I learned a lot but without taking on more than I could reasonably handle without having a decent workshop and specialist test equipment– from re-capping on my own very basic electronics knowledge, such as amp gain stages, and output impedance matching of power stages, to the practicality of soldering and electrical safety regarding DC storage capacitors. It gave me a lot of confidence for future handling of higher voltage amps up to say 500V DC, and teaches patience. Rushing things may kill you.
It would have been nice to have more electronics knowledge and a signal generator and oscilloscope to actually find where the distortion is happening, and be able to change the correct components to cure that. I’m sure that a more knowledgeable and well equipped tech would be able to clean up the sound and get a higher volume out of the amp – even at 5W it should be much louder than it is (where Marshall got the 25W rating on the badge from I don’t know! ). My 2yr old 10W Marshall practise amp is loud at half volume and deafening at full volume, in the bedroom.
Here is some info on JJ tubes EL84 – power output curves and tests:
There is a very high 100Hz bass frequency spike here, so I can only assume the Celestion G10 Greenback speaker model used is not working properly either, as there is no bass component to the sound of the speaker.
It had NO bass response when I tested it separately through my stereo and could not stand going past volume 1 of my Techics 50W amp without distortion, which for a 30W speaker is not right either. I’ll have to pick up a 4 Ohm speaker from the boot sale and test it with that.
Though I got this to a “working state” from what it was as a repair or spares sale on Ebay, I would like to be able to sell it as FULLY functional, and not just a “working but in need of technical tweaks”, rare Collector’s amp.
I assume that many of these “fixed bias” amps were designed deliberately to create after sales repair fees, as the truth about valves is that they are all different from new, so HAVE to be set up correctly, else may sound shit, or just shorten the valve life, unless you are lucky and get one that matches the circuit exactly, to stay within its optimal parameters. A valve amp with no bias pot is not good as a practical workhorse it seems, hence the addition of these by some techs such as Lord Valve:
I can be so stupid at times its unbelievable!
I didn’t do the obvious things first which were staring me in the face on the circuit diagram, which states what all the T1-3 collector, base and emitter voltages should be, so I checked today and have some differences that seem to point to the T1 and T3 transistors being kaput, or T1’s emitter capacitor C5 being shorted to earth.
collector = 0.7V
base = 0.3V
emitter = 0V
T2 is too high:
collector = 32.6V
base = 11.1V
emitter = 10V
collector = 0.6V (well wrong)
base = 0.6V (about right)
emoitter = 0V (correct – at earth)
Looks like T1 and T3 are shot at least.
On to the web now to order 3 x BC184 transistors…
Marshall Mercury 2060 Repair – Days 4-6