Marshall Amps

Marshall Amp Myths

Marshall guitar amplifiers have attained a cult following, and their own set of inbred myths and legends, that needs only their one name to identify them. If I told you I owned a vintage Fender amplifier, your first question might be… ‘Which model?’ or ‘Is it brown or black faced?’ When I tell you that I own a vintage Marshall, that is all I need to say. You just know that it is a 50-watt, 4-input, non-master volume model 1987 (only a select few can handle the awesome power of the 100-watt model 1959), even though the earliest Marshall ‘catalogs’ contained combos and up to a 200-watt head. Other buzzwords can include ‘plexi-panel’ and whether it has a ‘small logo’ and a ‘removable impedance selector’. Another catch-phrase is whether or not your Marshall has the ‘lay-down mains transformer’, which is a misnomer. This is not a style of transformer that any manufacturer understands, but the vintage dealers do. These and other myths and legends all came about because guitarists abhor changes to their gear (‘They don’t make ’em like they used to!’) compounded with the fact that Marshall hasn’t made a good amplifier since the JCM800.

First Things First

The first thing that needs to be realized is the ‘classic’ Marshall amplifier is a damn close replica of the 1959 tweed-covered Fender Bassman, or the model 5F6-A. (You can read all about this in the History of Marshall book; see the end of this ‘Lesson’.) What makes the Marshall ‘version’ of the Bassman circuit sound like night-and-day in comparison? A combination of a few deviations.

• There is very bad math going on when someone advertises they are selling a 50 watt (2XEL34) Marshall head or a 100 watt (4XEL34) Marshall head. Technically speaking, the EL34 is a true Pentode with a much higher transconductance (11,000 micromhos) than the Beam-Power tetrode 6L6 (6,000 micromhos). I have never measured a 2XEL34 Marshall that put out less than 60+ watts RMS before clipping (and who plays a Marshall clean?). Many Traynor amplifiers get 90 watts RMS from two EL34’s. A good 4XEL34 amplifier can easily squeeze out closer to 150 watts RMS! To be fair, this fallacy has been going on for a long time, and even the vintage dealers are guilty. How else can two EL34 output tubes leave two 6L6 output tubes in their wake? Read on.
• At similar plate voltages, a pair of EL34’s require only about 2/3rds of the drive voltages of 6L6’s to achieve similar output or even much more output. 6L6’s require about 72VAC grid-to-grid to get 55 watts of output power, as compared to the 47VAC grid-to-grid required to drive two EL34’s to 100 watts of output! (800VDC Plate voltage/400VDC Screen Grid voltage.) The typical Marshall preamplifier will furbish much more than this. This can lead to copious amounts of ‘preamp overdrive’. Once again, the EL34 is a true pentode output tube, as compared to the beam-power tetrode 6L6/5881. Pentodes have a higher plate resistance, and subsequently a lower damping factor with a similar feedback circuit. This yields more output distortion (all else being ‘equal’). Remember, the Marshall feedback circuit is ‘borrowed’ from the Bassman.
• Early Marshall circuits have a fair bit more filtering, and also more negative feedback. Late 60’s and onward Marshall circuits have a lot more filtering. This means the amplifier will have a much tighter bass response, and a lower damping factor.
• An interesting theory recently written has it that because the EL34 was a low-cost tube, it was manufactured with less of a ‘hard’ vacuum and a less rigid internal structure (this is a great example of someone doing intuitive reasoning, something I am trying to ‘teach’ here). Since the signal voltage can reach excessive level at times, the piezoelectric and mechanical behavior of the tube can come into effect and subtly alter the overall sound. These were ideas and questions that I had asked years ago, but had no way of investigating further. This could well be a big part of the equation.
• The Jensen P10R’s in the Fender Bassman have a ‘sensitivity’ rating of about 95dB (@1W/1m). Not too shabby, but four of these 10″ speakers in an open-back cabinet design can’t really ‘keep up’ with four (even more efficient) 12″ Celestion G12M25 ‘Greenbacks’ in a closed cabinet design.

The original Marshall amplifiers used American 5881 output tubes.

There are other Marshall ‘quirks’. Have you ever thought about the 16-ohm speaker impedance? This is really a Celestion quirk, but one story I heard is that the British engineers had a problem with the losses in the PA systems of the day. People were using cheaper light gauge wire for the speaker leads. These thin wires, when ran over a considerable distance, offered resistance. You now have a voltage divider between the wires and the speakers. Since engineers couldn’t convince people to use proper speaker wire, they adopted the use of 16-ohm speakers. How does this help? Assuming a rather high 1-ohm resistance in a very long speaker wire, in series with a 4-ohm speaker load. You now have a 20% loss of power output in the wire (1-ohm is 20% of the total resistance of 5-ohms). With an 8-ohm speaker the loss was reduced to about 11%. With a 16-ohm speaker, the loss was down to under 6%! (Of course I will get plenty of Email stating that the speaker impedance is NOT constant, or even a DC measurement. These are just mathematical numbers for the sake of argument; the same as probably used by the engineers.) We don’t use long speaker leads in our guitar amplifiers today, but the thinking has stuck with the engineers. Other quirks and myths surrounding Marshall amplifiers include:

• The very first Marshall amplifiers actually used American 5881 output tubes. When these tubes became prohibitively expensive, a switch was made to the inexpensive British tube KT66, followed by the EL34.
• The rest is history. Marshall has a reputation for making very temperamental amplifiers, even though they are ‘borrowing’ a pretty tame Bassman circuit. This came about because although the original circuit design worked quite well with 5881s, the switch to EL34s did not come with many circuit design changes. In general, Beam Power tubes (such as the 6V6 and 6L6) will exhibit less of a potentially damaging Screen-Grid current rise at maximum signal excursions than Power Pentodes (such as the EL34 and 6BQ5/EL84). Marshall took the standard procedure of using 1K Screen-Grid resistors as opposed to the 470-ohm resistor seen with Beam-Power tubes. However, as the voltages crept up in the amplifier over the years, the output stage can still become unstable, especially with the available poor quality EL34’s. Intuitively, you may consider that fact that Marshall actually switched back to 5881’s in their amplifiers for a brief period as a silent admission of the previous statement. Why Marshall neglected the option of ‘fixing’ the circuit remains a mystery. Adding in the fact that many early Marshall amplifiers came out of the factory without any Screen-Grid resistor certainly doesn’t help today’s players replacing their tubes every other day! A final symptom of unstable Marshall amplifiers is the older examples that intermittently oscillate at very high levels. Check to see if your Marshall has grid stopper resistors on each output tube. Some amplifiers had none at all, while other 100-watt models had only one grid-stopper resistor per ‘side’. Either way, ‘modify’ the amplifier to have one grid-stopper resistor per tube. An engineering switch to 6550 output tubes (to try and increase reliability) didn’t go over too well with guitar players. You can make your Marshall very reliable without drastically changing the tone.
• Speaking of changing the tone, adding any Screen-Grid resistor does just that. If you have really good tubes, in a well-designed amplifier, try it for yourself. When the Screen Grid supply is solid (no voltage drop across any resistance) the sound is a lot ‘tighter’. One modification idea is to place a 20uF/500WVDC filter capacitor right on each Screen-Grid, along with raising the resistor value to5K. This runs the tube as a true pentode, with a solid Screen-Grid supply. You loose a few watts in the deal, but some people like the tone and the extended life expectancy of their EL34s. Garnet, Traynor and Hiwatt (only in some 100-watt models) did it much smarter and their methods are my favorite way of running today’s poor quality EL34s in any Marshall. Does it change the tone? Not at all. These ideas are presented below.
• Many people mistakenly believe that Marshall amplifier run their tubes ‘hotter’. The typical 5881 plate voltages of the Fender Bassman Model 5F6-A are about 432VDC. This compares very equally with the early tube rectifier equipped Marshalls. The JTM-45 MkII amplifiers had a typical 430VDC on their EL34 plates. 100-watt amplifiers barely have 500VDC on their EL34 plates. This fact surprises a lot of technicians who expect the ‘big, bad, ‘Marshall’ to have closer to 600VDC on their output tube plates (which an EL34 can easily take, but remember Marshall’s output tube ‘evolution’). Increased output tube plate voltages yield more output, but the risks from a poorly designed Screen Grid supply makes today’s EL34s shudder.
• Many people believe that the Marshall output transformer is a key ‘secret ingredient’ to the Marshall sound. It may well be, but the transformer in and of itself is not all that special. Side story #2. Over the years, one of my tube ‘teachers’ has stated he has changed a handful of Marshall output transformers, substituting regular cheap ‘no-name’ output transformers. He received not a single complaint of ‘What did you do to my Marshall? It sounds like shit!’ I have done the same ‘experiment’ (using smaller than what you’d expect transformers) and achieved a similar lack of complaints. In his book ‘The Ultimate Tone’, Kevin O’Connor is very surprised that the Marshall output transformer is about the same physical size as a Fender Deluxe Reverb output transformer. This is just a simple case of education getting in the way of knowledge. A list of ingredients key to the Marshall output transformer isn’t difficult. See ELECTRONIC PARTS for thoughts on this. Marshall has changed their output transformer a few times over the years, and the result is that people are mortgaging their house to buy a ‘Plexi-Panel’ example of when Marshall just bought a cheap and plentiful light-duty transformer (instead of letting engineers tell them what was ‘needed’).
• Many people mistakenly refer to their Marshall as having the ‘Plexi’ circuit. In truth, there was no ‘Plexi’ circuit, but the same amplifier head Marshall had been producing all along with a silkscreened front panel made from PlexiglasTM, as opposed to the ‘normal’ thin plate of metal.
• Many people who own a 100-watt/non-master volume Marshall will remove two of the output tubes to achieve a more manageable volume from the amplifier. This isn’t such a bad idea (other than the fact that you have to keep tabs on a pair of output tubes that aren’t ‘aging’ at the same rate as the ‘in use’ pair), but a popular myth is that you must reset the output impedance selector to one-half of what your cabinet is rated for. This is 99% untrue. An 8-ohm load connected to an 8-ohm output tap should achieve full rated power and full bandwidth. A mismatch ‘up’ or ‘down’ will reduce the power output and the bandwidth. With a ‘good’ transformer utilizing high-temperature wire and ‘properly’ insulated between the coil stacks, no special care is needed to ensure safe limits of current through the tubes or transformer. You will achieve full wattage and the ‘original’ bandwidth by adjusting the impedance, but the ‘alternative’ has interesting possibilities. It’s all up to your ears. If you aren’t 100% comfortable about this procedure, play the amplifier and feel the shell of the output transformer. Did you get a blister? You may want to check Screen Grid voltages and bias currents. As for the output tubes and reliability, you should have learned by now that output transformers were wound for many different primary impedances over the years. Using 6L6’s as our example, the primary impedance recommended has been 10k-ohms, 6600-ohms, and 3800-ohms. The tubes, transformers, and speakers all lived long and fruitful lives. A better idea came later when some Marshalls (and Fender’s The Twin) featured what is usually called by everyone except the manufacturer a Pentode/Triode switch (which doesn’t really switch the output tubes between pentodes and triodes, but that’s a long story). ‘Triodes’ will have a much lower plate-to-plate impedance, and not even the average tube ‘guru’ was brazen enough to state that the impedance selector must be adjusted whenever the amplifier is run in the ‘triode’ mode. As a side note, those Marshall amplifiers with ‘High’ and ‘Low’ power switching (JCM900) have given service technicians very little trouble, while JCM2000 amplifiers (without High/Low switching) are the source of many output transformer headaches. Draw your own conclusions.
• Speaking of ‘triode mode’, there is much confusion about the ‘proper’ way to connect the output tubes in a ‘triode mode’. While you aren’t necessarily really connecting the tube as a triode, my own preferred method is simply to use the Screen-Grid resistor already there. Quick, easy, and requiring only a switch and a little bit of wire, the modification works well. You need a resistance here because of the poor quality of EL34s available today, none of which can handle the considerable current that would be on the Screen Grid with no limiting resistor. Below is a simple diagram for achieving the modification. Amplifiers with four EL34s require a little more thinking, but switching can be accomplished with a simple DPDT setup as well. Of course you could also have only the outside or the inside pair of tubes on the switch shown below with a 100 watt Marshall, and this idea can even be further worked with a DPDT/Center ‘off’ switch for three way switching. An added safety feature here is that if the switch ever screws up, you’ll just ‘shut off’ the output tubes.

The ‘Pentode-Triode’ switch, while not named correctly, is still very useful in high-powered amplifiers.

• Using a ‘Variac’TM is a terrible idea. I truly don’t care who else says it isn’t. Don’t do it. Period. Many Marshall owners are confused by all the hoopla surrounding the variable AC transformer, or Variac TM. This is a tool that should only be used by qualified repair technicians. I’ve owned mine for about 12 years now, and still have only used it to fix TV sets. Misguided ‘wannabe’ technicians claim from one side of their mouths that not adjusting your impedance selector should you ‘pull’ output tubes from your 100-watter is bad, while claiming from the other side of their mouths that reducing the AC input voltage to your Marshall will safely give you a great overdrive sound at a more manageable volume. There are a few good reasons for not using a Variac to lower the AC input in an attempt to achieve a so-called ‘brown’ sound.

If you think you ‘need’ one of these, lay down until the thought goes away.

• Reducing the AC input will reduce the B+ to be sure, but you will also reduce the heater voltage by the same margin. This can cause ‘cathode stripping’ of the tubes1 (if the heater voltage is far too low).
• You are also reducing the bias voltage as an added ‘bonus’. Unfortunately, the bias voltage does not track the plate voltage proportionately. And we all know how anal the Internet ‘guru’ is about the ‘proper’ bias voltage.
• Filter capacitors tend to have a ‘memory’. Running a 500-volt capacitor for any extended length of time at 400 volts will make the capacitor think it is a 400-volt capacitor. The next time you plug your Marshall straight into the wall AC and put 500VDC on a capacitor that has been brainwashed into believing it is a 400VDC capacitor could be trouble (if the capacitor fails to ‘reform’). I have been told this is a remote chance, but I chance I don’t want to take.
• Using an ‘ungrounded’ Variac can cause serious bodily injury, and even death.
• There are two very good reasons not to use a Variac to raise the AC input voltage to your Marshall. You will ruin your tubes2 and possibly your output transformer.

I would like you to study the diagram below carefully. It is from the book Audio Cyclopedia, by Howard M. Tremaine. It deals with the prospect of running the filament voltage on any tube at other than 100% rated value. The illustration shows that tubes will fail prematurely if the filament voltage is either higher or lower. For maximum tube life, the voltage must be ‘right’.

Mr. Tremaine thought about this shit years before the VariacTM was ‘invented’.

From the chart above, you should be able to determine that if a tube has a usable life of approximately 6,000 hours with its filament operated at 100% of the rated voltage, the tube’s life expectancy is reduced to 1,000 hours just by lowering the filament voltage by 10%. This translates to a 6.3VAC tube being run at approximately 5.67VAC, or the input voltage lowered from 120VAC to approximately 108VAC. Ask your local tube guru if they have read Audio Cyclopedia. However, even ‘experts’ can disagree on anything and everything. GE published data confirming tube life is decreased substantially by raising the filament voltage, although tube life can be increased by lowering the filament voltage very slightly. The performance of the tube was purported to be more ‘even’ over its useful life. However, it should be pointed out that the slightly lowered filament voltage GE ‘recommends’ proportioned to the plate voltage (via your VariacTM) won’t make a difference to our ears. People who set their VariacTM to 100VAC are only asking to replace tubes every other month. The GE report also stated that the transconductance of the tube was reduced dramatically, and an audible deterioration in performance was noted when reduced filament voltages were employed. So why bother?

If the filament is operated at higher voltages, some of the cathode coating material is evaporated and permanently lost.

Radiotron Designer’s Handbook, 4th Edition

• There are so many devices on the market today that replace the poorly designed ‘Power Soak’ TM of the 1980’s that you shouldn’t need to gamble with your tubes or your life in the search of tone. All the rumors began when Eddie Van Halen joked about using a Variac to get his sound. He’ll never admit if his Marshall was ever ‘modded’ and this set all the rumors in motion. You can still find discussions through various bulletin boards arguing about whether Eddie’s Marshall(s) were ‘stock’ or not. The ‘best’ rumor I personally heard was he connected the channel II input to the AC outlet! The most common rumor was actually his own interview where he stated he set his Variac to 140VAC and ‘watched the tubes melt!’ A better (and safer) idea is to reduce the B+ voltages inside your high powered amplifier, as shown below, complete with a sag resistor. This is a Tone Lizard exclusive circuit; you won’t see it anywhere else! The B+ is lowered a small enough amount that your filter capacitors are not in any danger when the resistor is switched in and out of the circuit. Play around with the value of ‘R’ to get the voltage drop you like, keeping in mind you will need a POWER resistor to drop any substantial amount. If you don’t want the sag option, just get a SPDT switch and use the 1/2 of the circuit you need. It involves seriously modifying the power supply, so be damn confident of your capabilities. Be sure to mount the capacitor away from the resistor, and observe polarity. Reducing the plate supply of any preamplifier tube will yield much less headroom, but the amplification factor (gain) of the tube remains pretty well constant. Reducing the output tube voltage will yield less output. It is a win-win situation. Another thing that happens with reduced output tube voltage is the effective plate-to-plate impedance changes, but I wouldn’t worry about it. Just enjoy the ‘new’ amplifier you have. If you decide that you would sleep a little better by also reducing the bias voltage a couple of volts when reduce the B+, use the DPDT and utilize the second 1/2 to add/remove a resistor in series/parallel with the bias circuit to accomplish the task (you won’t have the ‘sag’ option). Otherwise, set the bias ‘in between’ the value for the full and reduced B+ voltages.

Marshall voltage reduction method; far superior to using a VariacTM

• Don’t waste you time and money on ‘gurus’ who are hot to add an extra 12AX7 to your Marshall for more distortion. Instead purchase ‘The Ultimate Tone’ from Kevin O’Connor and ‘discover’ 101 ways to add distortion without adding extra tubes. People who just want to add tubes don’t really understand tube circuits very well. Any Marshall has a multitude of places in the circuit where extra preamp gain is just ‘sitting there’ not being used.
• A friend of mine returned from a trip out East where he met a fellow who apparently owns the Marshall. The amplifier, a model 1985, has been heralded by all who hear it as having the tone. The owner ‘blueprinted’ the amplifier and wrote a ‘report’ for all who were interested. He even gave the amplifier a cute name; Chocolate Cake, and called the ‘report’ Grandma’s Chocolate Cake Recipe. Everybody likes to think they have the Marshall, but compare yours to ‘Grandma’s’. You’ll find it’s probably similar, and nothing special. Other comments include transformer markings, the filter capacitor/decoupling resistor values, and the fact that ‘Pam’ assembled the amplifier. Conspicuous by its absence is any mention of the output transformer primary impedance. The only observation of any physical attribute is… ‘output transformer seems slightly larger than other 50 watts units.’ How astute. It may have more iron, be wound for extended frequency response, or have a higher wattage rating. This would change the tone, but we’ll never be told. Thanks, ‘Grandma’.

 Plate Voltage Other Comments V1 (ECC83) +147VDC No ‘Brite’ cap on 270K mixer resistors No ‘Brite’ cap on volume controls V2 (ECC83) +246/+130VDC All tone controls measure almost +50% V3 (ECC83) +174/+180VDC V4/V5 (EL34) +392VDC +392 VDC Screen-Grid Voltage 42mA Idle Current @-27VDC Bias 27K Feedback Resistor 

‘Grandma’s’ Chocolate Cake recipe isn’t much different from my wife’s!

If the filament is operated for long periods at reduced voltages, the effect is a reduction in emission.

Radiotron Designer’s Handbook, 4th Edition

An excellent idea to promote longer tube life in your Marshall (especially with the poor quality EL34’s available today) is to utilize a few ‘tricks’ seen in collectively in Hiwatt, Traynor, and Garnet amplifiers. Below is a pictorial drawing of the stock output tube socket wiring scheme, simplified and drawn as well as I can. You should be able to make out ‘heads or tails’ though. The ‘grid stopper’ resistor is usually standing up vertically right behind the Screen-Grid resistor on a Marshall, and the 1500-ohm grid-stopper will be laying down across the tube socket on a Fender amplifier. On every Marshall the Suppressor-Grid is tied to the Cathode and then wired to ground.

Standard Marshall output tube socket wiring scheme, albeit simplified.

The standard values are 1K/5W for the Screen-Grid resistor and 5.6K for the ‘grid stopper’, although that value can vary slightly. Most technicians will automatically think they are doing you a favor if they raise the value of the Screen-Grid resistor to help Marshalls that ‘eat’ output tubes, with some mail-order mod ‘kits’ raising the value to 2K or even considerably higher. While that does work quite well, you can certainly do better, as seen below.

Improved Marshall Suppressor Grid wiring scheme.

The above wiring scheme has quite a few benefits, but please be 100% sure you know what you are doing (if not, refer to a Traynor YBA-3 schematic). To give the appearance of factory wiring, I drill two small holes on the circuit board and use similar looking wire. The above circuit will reduce secondary emission of the tube even better than the original wiring scheme. You also get improved linearity. This will end up making the tube’s plate current more efficient to the point that you will have to rebias! The last Marshall I ‘modified’ in this manner had the plate current increase 5mA. So, please recheck your bias after doing this procedure. Where some technicians insert a 1-ohm resistor between pin #8 (cathode) and ground, this arrangement is far ‘safer’ for the tube. Think about it. Instead of having the Suppressor Grid at ground potential or even at a negative voltage, you would have a positive voltage (granted only 50mVDC or so, but positive nonetheless) on that grid, because no one ever separates the Suppressor Grid from the Cathode. Where most inexperienced technicians have trouble with EL34’s is in a ‘poorly’ designed amplifier that was biased by the current draw method via a 1-ohm resistor in the cathode! The idea presented above, in concert with an ‘improved’ Screen-Grid wiring scheme, yields a very efficient environment for any EL34. That wiring scheme is seen ahead in ‘Mods and Odds’, but I’ll give you a sneak preview. Remember that those 100-ohm resistors seen below are the Screen-Grid resistors soldered on the tube socket.

Improved Screen-Grid supply promotes longevity for EL34 service.

This setup is seen in 100-watt Hiwatt amplifiers and Traynor Custom Special amplifiers. Of course the 100-ohm/5-watt resistor in the power supply takes the place of the choke seen in Fender and Marshall amplifiers. Keep in mind Hiwatt ‘got away’ with using 1/2 watt resistors for their Screen-Grid resistors seen above, and the failure rate is surprisingly low. Well, surprising to most ‘gurus’ it would appear. A final tip in weeding out bad EL34’s is to bias up the amplifier and then measure and compare Screen-Grid voltages. If you have a 100-watt Marshall and the voltages you measure are (A)520VDC, (B)518VDC, (C)521VDC, and (D)493VDC, the last EL34 (tube ‘D’) is ‘bad’. It is drawing excessive screen current, and that tube should be thrown in the garbage. A volt or three of difference is nothing to worry about; you are looking for the tube with a considerably lower Screen Grid voltage.

Recommended reading for this lesson is ‘The History of Marshall’. To read my own personal ‘review’ of this book, look here. If you ever get the chance to have a tweed Bassman and a Marshall 50-watt head (Model 1987) near your bench at the same time, make some comparative studies. Measure voltages and scope the output. If you own a Marshall, place a 200 to 250-ohm/25-watt resistor across the output jack (this can be mounted on the chassis). Or at least change the jack to a shorting-type (similar to the input jacks). If you decide to implement the ideas for the Screen Grid and suppressor grid supplies, please give me a ‘report’ on how the ideas worked for you and if you noticed any changes in tone, along with how your tubes are ‘holding up’. 

Try a few mod ideas, in a ‘Lesson’ I humbly refer to as THE ULTIMATE JCM800?, and stop using fuzz boxes, attenuators, and never add 12AX7’s to your Marshall again! After you’ve purchased a ‘new’ set of output tubes for your JCM800, learn how I would bias them up by here.

Sources

1. Radiotron Designer’s Handbook, 4th Edition, F.Langford-Smith.