The subject of output tube biasing has many otherwise rational men arguing amongst themselves like little school children. It seems that the modern 'guru', never at a lack for self-esteem, loves to bully people into feeling inferior with incomprehensible techno-babble if they should dare disagree with his repair methods. It also seems that all of the vintage amplifiers that we heard on vinyl records (remember them?) growing up had left the factory biased 'wrong'. Those classic tweed-covered tones could have sounded even better if only the manufacturer had heard of (and used) a concept called the current draw method to bias any amplifier before it left the factory. Forget that oscilloscope; your amplifier will be biased very 'cold', and sound weak and thin. Apparently, the 'current draw' method is the only way to bias your amplifier if you care about your tone, and you're a fool if you disagree. Most of this offered 'best' method does have at least a grain of 'truth' behind it. Unfortunately, the current draw method also has a lot of BS behind it as well. There are 7 (un)commonly held 'myths' about bias itself and the proper way to bias output tubes; and they are as follows.
#1: ALL 6L6's SHOULD BE BIASED AT 40mA. Or 35 mA, or biasing strictly to any specific 'magic' number. It's all wrong. You are not taking into account the vast differences in output transformers and tubes. You are also neglecting the poor quality control of tubes available today. Thinner plate material combined with misaligned screen grids yield an erratic parameter regarding how much plate current the tube 'should' be drawing. Lazy people who don't fully understand how a tube or an oscilloscope works arbitrarily chose these 'safe' numbers, that in some instances may not be all that 'safe'. If the sine wave looks good on an oscilloscope (the top and bottom half clipping at the same time, with a slight crossover notch when the wave is 'just' clipping), and you have the expected output power without the plates glowing red, who cares how much plate current the tube is drawing? Not I. The number will vary. Because higher plate voltages tend to decrease the plate resistance1, the 'proper' bias voltage for a 6L6 (or any output tube) will be different for various plate voltages. Try biasing some Chinese 6L6's with 500VDC on the plates at 40mA. Good luck! Also, too, keep in mind that the 'proper' bias point will also depend on the output transformer primary impedance. There is no magic number for any tube type! You can have a 'common' plate current draw for specific tube brands in specific amplifiers, but nothing is written in stone. The wannabe technicians who write 6L6 'Shootout Reviews' or 'Taste Tests' stating all tube samples were biased to 35mA or 40mA are NOT doing fair assessments. Finally, I'd like you to consider the photograph below.
The transformer above left is a genuine 'Silverface' Fender Super Reverb output transformer. The transformer above right is from an old Harmony amplifier. Both amplifiers use a pair of 6L6 output tubes, although the Fender amplifier runs the tubes 'hotter'. Would you try to bias up both amplifiers at 40mA of idle plate current? A voltage discrepancy also exists between say an Ampeg SVT and American Marshall amplifiers. Both may have 6550 output tubes, but while the Ampeg has a B+ of approximately 600VDC, the Marshall will have only about 450VDC for a plate supply. You would not bias up both amplifiers with the same 'magic number'. The moral to this story is that if you only ever own one Deluxe Reverb your entire life, and only retube it with NOS American 6V6's, you can bias by the current draw method to the same 'number' every time and you shouldn't have any problems. Service technicians and people who own different makes of amplifiers need an established scientific approach, and not hocus-pocus. Although, to be fair, the photograph above showed an extreme discrepancy. Typically, most amplifiers with similar circuitry will have the bias fall within a 'range', and in fact this 'proper' bias voltage range can be determined by using an oscilloscope, or simply using numbers such as your shoe size multiplied by your wedding anniversary, divided by how much you paid for your 'Bias ProbeTM'. How you derive at this number is not relevant, although the 'magic number' proponents suggest if you get the very same number they get, and you used an oscilloscope, your answer is somehow 'wrong'.
#2: A 1-OHM RESISTOR BETWEEN CATHODE AND GROUND WILL TELL ME HOW MUCH PLATE CURRENT I HAVE. Wrong again. This will tell you the combined plate AND screen grid current. Granted the screen grid current is very little compared to the plate current, but it is there, none the less. You may have 35mA plate current with 5mA screen current. Your meter on a 1-ohm cathode resistor will tell you that you have 40mA, which are the combined currents. You must place your ammeter in series between the output transformer primary lead and the plate connection to find out actual plate current by itself. And, lastly, if you use a 1/2 watt resistor you will have 'drift'. Try the resistor experiment over at ELECTRONIC PARTS (through 'Vintage Advertising') and you won't use 1/2 watt resistors for anything. If you must use the resistor-in-the-cathode placebo, make it a higher wattage rating. And really think about what you are doing if you put a resistor in the cathode of your Marshall. Then think about it some more.
#3: THE 'PROPER' BIAS WILL GIVE ME 'THE TONE'. I wish it could. Less bias (without the plates glowing red) will cause the output tubes to 'drive' easier and have the output tubes biased further towards Class AB2. Increased bias (not at cutoff) will make the output tubes harder to 'drive' and be very much into Class AB1, even approaching class B in extreme cases2. Bias beyond cutoff will 'sound' terrible, but that's the huge crossover 'notch' you are hearing. The rest is the preamplifier section working the output section either harder or not as hard. The sonic signature of the particular tube should remain consistent. Granted, biasing the tubes 'colder' (harder to drive into clipping) will usually sound thinner/colder/etc. while tubes biased 'hotter' (easier to drive into clipping) will tend to sound 'fatter', but if you don't really know the amplifier and the tube(s) you are using how do know when you've reached either point just by using a number? Lastly, to repeat, there is no 'proper' bias point. Between cutoff and saturation, anything is 'fair game'. Audiophiles who only play their tube stereo systems 'clean' and at quiet listening levels really bias their output tubes 'hot', and started many of the biasing 'myths'.
#4: I NEED TO (RE)SET MY BIAS EVERY TIME I CHANGE TUBES. Not true. You may very well feel a 'need' to check and/or reset the bias, but if you are using the same make of tube each time you change, the bias should be close enough. A volt or two either way won't make a difference, assuming the bias wasn't on the edge to begin with. As long as you didn't purchase a bad tube, you will be fine. One prominent Fender warranty repair shop technician I know doesn't check the bias at all when he replaces output tubes as part of the repair. Another repair center, a Mesa Boogie warranty depot, will replace only ONE output tube if that is all that is needed to complete the repair. So much for matching! The warranty shop is told to replace only the bad tube! And the bias isn't checked, because Mesa Boogie doesn't have a bias adjust! As long as you really know what you are doing, everything will work out just fine. Go back to ELECTRONIC PARTS to remind yourself about tubes.
#5: NOS/'MATCHED' OUTPUT TUBES WILL GIVE ME 'THE TONE'. No, they won't. The 'matched' output tube shtick was explained in ELECTRONIC PARTS. To reiterate, the circuit was designed around a particular tube and its non-linearities. Biasing 'fake' 5881's (as one example) to the same magic number as you'd bias 'real' 5881's can lead the uneducated into believing that only NOS 6L6 types sound good. 'Matching' is unnecessary because the output transformer will not be 'balanced'. Nor will the phase splitter (I'll tell you how to check later in this lesson). If everything (phase inverter, output transformer, output tubes) was perfectly matched and balanced, some wonderful harmonics would be cancelled! Even 'matched' output tubes have a high likelihood of aging differently, so why bother3? Also, with supposedly matched tubes, different levels of distortion exist at different output power levels. The fancy 'matched' tube sales outfits seem to spend more money on re-screening their Chinese/Russian tubes than they spent on buying the tubes in the first place! Find a good tube jockey who can make you amplifier sound the way you want to, with available tubes. And forget the hyperbole, OK? Admittedly, there are simply 'dog' tubes out there, and no tube jockey can make them sound 'good'. These would have to include the original (smoked glass) Russian-made 6V6, which has thankfully been recently discontinued.
#6: CATHODE BIAS ON THE OUTPUT TUBES MEANS CLASS 'A'. Not at all. The myth that a bias voltage placed on the output tube grids is equal to operating in Class 'AB1' or 'AB2' is equally wrong. In a nutshell, if the plate current doesn't increase more than a few milliampres between idle and full (unclipped) output, you probably have as close to Class 'A' as is possible with vintage guitar amplifiers. If the plate current increases about 25% between idle and full (unclipped) output, you have either Class 'AB1' or 'AB2' (you need more information to determine what sub-class you are operating in). Turning to our tube manual, two 6BQ5's in class 'AB' are listed as having a maximum of 36mA plate current (each tube) at idle and 46mA plate current (each tube) at full output (17 watts/4% THD). We are then made aware of the parameters; 20VAC grid-to-grid driving voltage, 300VDC-plate supply, 300VDC-screen supply, and a 130-ohm common cathode resistor! These numbers don't vary from RCA tube manuals to Amperex tube manuals to Mullard tube manuals. Class A amplifiers have their own myths as to why they sound 'different' or 'better', but that's a whole other 'Lesson'.
#7: I USE THE ONLY 'PROPER' WAY TO BIAS. Not at all. I use a very good method that produces very good results for me. Coincidentally, many other service technicians and full-scale manufacturers use the very same method(s) while others have their own preferred alternative method(s). It seems the new-age 'gurus' know something the major manufacturers were never taught in school. While repeating myself; some big name warranty shops I know don't even 'bother' with checking/resetting the bias. Other people make their money selling tools designed to help you measure current draw in the output tubes. Who's right/wrong? Is it a matter of right or wrong? I have gotten in over my head by stating my preferences for biasing on BBS 'threads' and really showed what little theory I know compared to everyone else. But the bottom line isn't that simple; while I understand the hard-on everyone has for biasing strictly by current measurement (who needs an expensive oscilloscope? Or a signal generator?), the moral of this story is that if it makes you feel better about the tone of your amplifier and the life expectancy of your tubes, who am I to argue? I promise you though that any amplifier that left the Fender/Marshall/Gibson/et. al. factory did so without the current draw being measured. I wonder why they didn't think it was important enough to measure. To repeat my answers to the few critics of the oscilloscope method, and hopefully remove any doubts you may have yourself (be sure too that you read the bit on improper oscilloscope usage in Articles That Didn't Quite Make the Cut);
A) The oscilloscope method is inaccurate. No, it's viewing a representation of the definite location of the DC operating point (Q point) of the load line where one tube is being driven out of cutoff while the other tube is biased into cutoff (due to phase inverter action). It ain't the equipment, it's the operator. There is no ambiguity there, while the current draw proponents can't even make up their minds what 'magic number' to use. Was that 35mA or 40mA of idle plate current? Do I bias to 60% plate dissipation, or 70%? One early book (How To Service Your Own Tube Amp/Tom Mitchell) wrote that to 'properly' bias EL34's, we set the idle plate current to 60mA. Of course, that number has since been lowered somewhat in reprints since many a Marshall started 'eating' modern EL34's. In the book 'Desktop Reference of Hip Vintage Guitar Amps', Gerald Weber states... 'any number (between 10mA and 40mA) that gives you the tone you like is correct.' Now there's Accuracy with a capital 'A'. As an added 'bonus' you can also check for the top half of the wave form clipping at the same time as the bottom half using the oscilloscope, meaning your output tubes are 'matched' as well as need be (assuming the phase inverter is designed properly). I will concede that the measurement can be inaccurate (due to poor oscilloscope operation/operating conditions) but the 'meaning' is not. With the current draw method, the measurement is accurate, but the 'meaning' is anything but. Even the instruction sheet accompanying the Bias ProbeTM is a puzzle. We are given a list of 'estimates' along with a chart of plate voltages and output wattages. After cross-referencing the chart, we derive at a mere guess about what current the tube 'could be' drawing. At least they don't commit to one end-all-be-all number, instead giving 'nominal' and 'maximum' values.
B) The oscilloscope method leaves the tubes 'cold', and the notch will 'reappear' after a while. That's because potentiometers 'drift'. This is why I don't use adjustable bias, like Mesa Boogie, Hiwatt, and many others won't either. Blackface Fender amplifiers, with their adjustable bias, really started going downhill from the tweed-era designs. And to preserve the 100% originality of your collectable tweed Fender you wouldn't dare put a bias pot in there, would you? Besides, even if you bias by current draw, the potentiometer and resistor will still drift, and the notch will still reappear (unless you 'under bias' enough to compensate). As an intuitive example, many technicians complain about the poor quality bias potentiometer in their Marshall amplifier, and the more expensive multi-turn pots are in vogue right now. These also appear to be the amplifiers with the most tube troubles. Draw your own conclusions. Lastly, according to The Radiotron Designer's Handbook, "During the life of a valve there is always a slow drift which is particularly apparent in the plate and screen currents........... in turn has the effect of reducing the plate current which flows at a fixed bias........... In general, the grid current crossover point tends to drift in the positive direction.......... Most of the drift usually occurs during the first hundred hours of operation. If stability is required it is advisable to age the valve for at least 2 days........". Well, that puts a whole new light on the subject, doesn't it?
C) If done five different times over a time span, the oscilloscope method will yield five different results. Déjà vu; pots (and tubes, apparently) drift. Bias by the current draw method five different times over a time span, and you'll get five different measurements, too. Will biasing by the current draw method 'magically' make the bias potentiometer (or tube) not drift? Read the above quote from The Radiotron Designer's Handbook a few more times until the point sinks in.
I have no hate for measuring plate current draw in your amplifier. I've done it myself, in a learning atmosphere. Until you learn to recognize the common measurements in guitar amplifiers (and get a sixth sense of what's 'right' for that amplifier), I say measure everything that can't outrun you. As for biasing by oscilloscope versus current draw, my question is this; How does the tube or amplifier know how you've arrived at your bias voltage and 'decide' to let the plate current drift if you chose the oscilloscope method? On that note.....
Recommended reading for this lesson is to get your hands on at least one receiving tube manual. Reprints are available from Antique Electronic Supply (see LINKS). RCA are the most common ones around, and for many good reasons. Aside from the fact that they bullied everyone into using their licensed circuits and vacuum tubes, they were simply the best-written manuals, period, with an excellent theory 'primer'. Older tube manuals won't necessarily have all of the tube data listed guitar players are interested in. Up until about RC-24, for example, there is no listing for the 6CA7/EL34, 6550, 5AR4/GZ34, and others. Any tube manual you are about to purchase should be checked out first to see what tubes it has listed and doesn't have. You should also be cognizant of the fact that the 6L6 changed somewhat over the years. The data for a 'plain' 6L6 differs from that of a 6L6G or a 6L6GC. Even the 5U4 changed slightly over the years, with later examples rated for more output current. This led to one confused book writer claiming 5U4's each had their own sound! You would want to look up the data for the exact tube you are using. Owning really old tube manuals with, as an example, the 'metal' 6L6's listed may be fun from a nostalgic view point. However, since you'll never use one in your Super Reverb, why put out good money unless you decide to collect tube manuals (and I'm the only one stupid enough to do just that)? Voltage amplifier tubes, like the 12AX7, didn't change at all in fifty years. Sylvania tube manuals are also popular, and the 'later' ones seemed to list tubes before the RCA manuals did. You could get by with just one Sylvania manual for most (if not all) of your relevant tube queries and one RCA tube manual for your theory course. The 'black binder' Sylvania tube manuals are the older style while the 'paperback book' style are the most recent ones, published after RCA stopped at RC-30.
Once you have a 'relevant' tube manual in your hands, put your favorite amplifier on the bench. Measure anything that can't outrun you. How much plate voltage do you have at quiescent? At full output (unclipped)? How much plate current do you have at quiescent? At full output (unclipped)? How much driving voltage AC do you have at the output grids for full (unclipped) output? Is it the same for each output tube? How do these numbers compare to your tube manual? Study your manual diligently. Learn the common numbers for a few amplifiers you have. This, and one more detailed tube theory book from a vintage book dealer will make you twice as knowledgeable as most repair technicians and tube amplifier 'gurus'.
The following is a gathering of information taken from various websites. Some are quite popular, while others may be a little more unknown. Regardless, let's sample how these various websites advise you to bias your tube guitar amplifier.
I want to remind everyone reading this; Your Plate voltage will decrease as the idle Plate current is increased, and vice versa. Try it and see. This is only one of many 'flaws' to the current-draw method. I will point out a few more pitfalls and traps as necessary.
This same website also mentions the maximum plate dissipation for the 6L6GC as being 23 watts. Oh, oh. Which is it? At least this website give the not so common-sense advice of not worrying too much about how much current your output tubes are idling at. Moving right along, we now come to my personal favorite missive against using an oscilloscope to bias tube guitar amplifiers.
Not that my fecal excrement is any less malodorous than the next person's, but who would paint such a broad statement with such a broad brush? The words 'accurate' and 'acceptable range' go together in the same sentence as well as 'Grandma' and 'naked'. One guru-written tome even accuses tube manufacturers who suggest the 'crossover-notch' method of biasing of giving bad advice! But wait; it gets better! Also from the same website;
First of all, to suggest your method is the only 'accurate' method, and then use the words 'around 50% to 60%' suggests to me that you should really get a new dictionary. Let's do some math; if a 6L6GC has a plate dissipation of 30 watts, and my Super Reverb has a plate voltage of 460VDC, 50% of maximum plate dissipation is 15 watts, for 32.6mA of idle current. Now, 70% of 30 watts is 21 watts, and 45.6mA of idle plate current. This does not take into account that changing the bias will also slightly alter the plate voltage, but we'll ignore that for now. By the above math, this 'guru' seems to claim that any bias point of between 32mA and 45mA is 'correct', but only if you measure the plate current. Again; if I use an oscilloscope, and arrive at some highly accurate number of anywhere between 50% and 70% of maximum plate dissipation, how is using an oscilloscope giving wrong results? The very same website also has the statement......
....... yet if I arrive at any of these 'magic numbers' by using an oscilloscope, my number doesn't seem to have the same 'magic'. My advice? Flush twice; I can accurately predict that the next person will really appreciate it.
Lastly, I did a little eavesdropping on a popular BBS board, and would like to pass along the following 'conversation'. It begins with some confused fellow telling us information he has read 'elsewhere' about biasing does not seem to coincide with his particular amplifier. This is to be expected, as I hope you have learned here that there are no hard and fast rules as to how much idle current any tube 'should' be drawing. Keep in mind too that everyone involved in this conversation has his own experiences to guide him, and there really are no wrong 'answers'. My final thoughts are at the end of this section.
Each piece to the conversation seems like good, sound advice. Everyone is entitled to his preferred 'magic number', and his opinions on how to properly bias a tube guitar amplifier. I do find it strange, never the less, that the following has been conspicuous by its absence.
There are other considerations as well. We must keep in mind too that just as with oscilloscopes, different brands and qualities of DMM's will give different readings! Also, I really doubt anyone will hear the difference between 32mA and 36mA, especially playing the amplifier 'dimed', and using a Tube Screamer.
There is a very interesting 'assignment' for this 'Lesson'. Read up on oscilloscope usage, and biasing common vintage tube guitar amplifiers, by CLICKING HERE. After you have digest all the pertinent information, you can listen carefully to your amplifier, without using any effects pedals. We are after a good clean tone, close to full rated output power of the amplifier. Play chords, and a few single string scales. Now reduce your bias by exactly 1VDC. Can you hear a difference? Now play with all the overdrive you normally use. Can you 'hear' how 1-volt of bias difference changed the tone? Probably not. Repeat the experiment after you've increased the bias by exactly 1VDC. Next, try adjusting the bias by +2VDC. Keep this up until you see the plates start to glow red, and until you 'hear' the tubes biased into cutoff. Write the numbers down, and rebias your amplifier with the current-draw method, and again using an oscilloscope. Both methods will give you a bias point 'somewhere between' the two highly accurate numbers you wrote down earlier. I promise. Now, worry less about bias, and more about how you were so naive in the first place to worry about bias.
1) Basic Theory and Application of Electron Tubes; Department of The Army Technical Manual.
2) RCA Receiving Tube Manual (RC-17, page 23).
3) Radiotron Designer's Handbook, 4th Edition, F.Langford-Smith (page 23)
HOME EMAIL LINKS TABLE OF CONTENTS
MADE IN CANADA CLASSIC ARTICLES