Bias in Vacuum Tube Guitar Amplifiers
JC Maillet (c) 2004
posted 9/02/04
Active Device Biasing
In linear electronics the term 'Bias' is short for "set of bias conditions', it normally refers to a state described by the terminal voltages around AND current flow through an active device when zero-signal is applied at its input terminals and when it is placed in it's so-called linear mode of operation (... which is actually a highly non-linear realm in relation to dynamic amplitudes) ... there is some confusion about this for tube amp users because of the historical appearance of bias voltage figures in schematics - these we're in relation to the tubes of the day but now, especially with the advent of Russian hybrid tubes, these numbers should be taken as "loose" reference - not that they are invalid per se ...
But fact is, in signal amplifying devices the mathematical quantities from which transfer specs like signal Gain and Load/Drive Impedances are directly derived from the current flowing through the device acting as main variable ... engineers who study the behavior of gain stages under different bias conditions typically play with equations that have output current as the primary equation variable for all small-signal parameters used in first order amplification transfer calculations ... when biasing tubes in guitar amps amp techs typically shoot for this theoretically supported Idling Current - their numbers are learnt emperically, through measurement and listening to a broad range of tube amps over time ...
From a AC signal point of view the job of a vacuum tube is the same as a transistor in the common-cathode/emitter/source circuits ... FET, Bipolar or otherwise their job in the common-CMS circuits is to convert input AC voltage signal to output AC current signal, where the rate of voltage-to-current conversion is known as the small-signal Transconductance - typically abbreviated 'gm' in academic engineering ... the AC current that flows through any vaccum tube can then be used to drop an AC voltage across a resistor (in order to develope an AC voltage version of the output signal) or it can be used to set the AC current in a coil, as in an output stage (single-ended or push-pull) or a reverb coil driver circuit ... either way, the rate of transconductance increases with DC (static) bias current and consequently their associated voltage gains do so as well ... the driving point impedances at the device terminals also change in response to bias variation as well, always with idling current playing the role of dominant reference parameter ...
Biasing Tube Amps
The 'Biasing' of tube amps commonly refers to "... setting the idling current in the Power Tubes in PushPull output stages" - the No-Signal or Static (=DC) Current Levels are the target spec and the surrounding voltage conditions are adjusted accordingly when possible ... here lies a BIG CRUX thingy : the multi-variable voltage-to-current bias relationship that exists for a power tube under a given bias set will respond differently from tube to tube (between brands especially) AND, more-so, over time of use (!) ... some amplifiers have mechanisms for automatically setting the bias of these push-pull circuit tubes - they typically employ some sort of Low-Pass servo-circuit to set the current levels ... this servo circuit can be passive as in the case of Cathode-biased circuits (these tubes self-bias the same way the smaller triode tubes of the preamp and phase-driver circuits bias themselves) or by electronic means as in the case of some modern tube amps ... In order for active servo arrangements to work right there has to be absolutely no signal feedthrough going through the servo circuit and back into the signal chain again or else transient response can be compromised - in practice this is impossible to achieve perefectly since the low-pass circuits are limited (no filter acts as a brick wall) and so some amount of transient focus is lost, especially noticeable at medium to high dynamic levels ...
In fixed or variable bias push-pull amplifiers a separate power supply circuits is employed to generate a negative voltage with respect to chassis ground - this voltage is fed to the grid terminal of the power tubes through bias feed resistors (typically 47k to 220k ohms) ... this is the negative Bias voltage that is typical written in amplifier schematics, in variable bias amplifiers this voltage can be "swept" through a potentiometer.
In Principle As the bias voltage in the grid is brought more negative with respect to the chassis ground (say, from -42volts to -52volts) then you'll typically find that the gain of the output stage will drop along with the drop in idling currents ... if the bias voltage is made less negative (opposite) then you'll find the gain of the output stage going the other way along with the increase in idling currents ... Note: some players mistakingly see the bias control as merely another volume control - the Red-Knob Fender Twins are notorious for this because of their availability of external control ... guitarists should tell each other to be cautious playing with that control alone unless they plan on making a proper science of it (!)
The Bias Limits
There are two primary biasing LIMITS on the network variable sets that need to be observed when biasing power tubes in a clas-AB push-pull stage ... the Lower Biasing Limit, when the amp is underbiased, produces a cross-over notch when observing AC waveforms on an oscilloscope ... shown here is a waveform of an Under-biased set of mismatched tubes in a Fender Super-Reverb amplifier ... notice the diverging turn-on/turn-off levels between the two half-cycles ... when a push-pull output stage is underbiased (bias voltage is too negative) it often sounds fuzzy and weak ...
 |
The Upper Biasing Limit leads to a situation where the dissipation rating of the power tubes is exceeded during normal operation. When the power tubes are running too hot the amplifier will sound very grainy and you might even hear some background crackling when not playing. ĘThis is a sign the power tubes are over biased (too much current at idle). Another tell-tale sign that your power tubes are biased way too hot is if the outer plate inside the tube turns red when you play. As well, bias modulating vibrato circuits set at full intensity can send the power tubes into thermal run-away when biased too hot. When the power tube turns red, it means you have lots of current flowing through the tube and thus the output transformer. Aside from frying the tube this situation can potentially damage the output transformer, so if you ever see this happening while playing it's best to shut the amp down right away and have it serviced. This is an extreme situation that shouldn't happen under proper operating conditions.
Fixed Bias Amps & Aging Power Tubes
There are a few manufacturers that have routinely designed amplifiers to run on specific types and grades of power tubes - the main reason for doing this is for avoiding the nightmare of unqualified people mucking with their amps and having to send it in for repairs afterwards - this is perfectly understandable ... when I visited Gar Gillies a few years ago I wanted to know why he insisted on using fixed bias in his Garnet amplifiers even though manufacturers like Fender and Marshall had trimpots inside theirs around the same time ... His answer was "... I would get boxes of 300 tubes from Mullard and they would all be bang on !! So why bother ?!" ...
Aging Tube Conjecture #1: Because tube specs suffer an offset drift from usage this means they can be brought back into full Bias (Current wise) by adjusting (offsetting) the bias voltage along the way - this is especially effective in the first year or two of (typical) playing - to a much lesser degree just before they totally run out of goose ... typically, the bias voltage has to be made less negative to restore the DC idling currents (and their related transconductance factors and driving point impedance) to optimized bias conditions ... this re-biasing of the power tubes helps maintain the dynamics of the output stage as the power tubes age ... of course there's always the issue of diverging drift patterns between tubes as they age, so re-biasing may not always return the power tubes to their original state completely, especially on the second or third rebiasing ... which brings me to Aging Tube Conjecture #2: even though AC audio signals are assumed to be vertically symmetric there is evidence that suggests to me that power tubes don't operate in a perfectly equal manner inside the output stage, especially during periods of high transient activity ... since spec drift is obviously more so dependent on time of hard duty and in high transients there is typically quite a difference in peak current spikes between both sides of the Push-Pull stage ... this transient differential would consequently causes a proportinal differential in rate of spec drift on one side over the other ... so you may want to swap your tubes from side to side of the push-pull circuit after every few months of use to slow down the drift differential and help maintain spec matching over time ...
How Circuit Specs Change with Bias
Aside from the trasconductance gain of the power tubes their output plate resistance and input grid resistance, which are both highly non-linear network functions, will also vary directly with device bias current levels ... the driving point impedance at the plate of a pentode or tetrode drops inversely to increasing bias cuurents and so does the non-linear average figure of the diode-like input resistance curve ... both have implications in terms of overall signal chain wave-shapping effects and production of harmonic distortion (richness) in the output stage ...
Since the harmonic content of the output signal of a push-pull stage is proportional to the ratio of plate-to-reflected impedances, this means that a lower device plate resistance results in a harmonically richer output signal ... simultaneously, a lower input inpedance at the inut grid node of the power tubes also means an earlier rise of loading currents on positive signal swings, which means more signal loading (or squashing) on that part of the signal cycle - which means more distortion ... so there exists a temptation to overbias the power tubes of a guitar amp for sake of makiing it sound sweeter and raunchier ... this comes at the cost of a shorter tube life since the cathode can only supply a maximum number of electrons in its life time before it becomes stripped of its surface structure - and a higher bias (current = electrons per second) roughly means a faster expenditure of total lifetime electron count ... in some amps, over-biased tubes can sound cool while in others they don't sound so good ... in general, class-AB output stage circuits have a Bias pocket where the dynamic response is most smooth and solid sounding ... this is typically where the amp should be biased for the sake of maximizing tube life as well ...
Should You Bias Your Own Tube Amps ??
In general, relying on yourself to do this is a good way to go except basic knowledge of electronics is strongly suggested before approaching any high voltage gear ... these conditions are necessary for establishing a strong likelyhood of safety and operational reliability - this is no joke ... by the same token, learning about bias can also be a way into studying electronics - so don't be shy about reading up on this ... one way to guarantee that your amp will work reliably and that you will get good life from your power tubes is to purchase tubes that are specifically graded for your amp (like Mesa-Boogie does) - assuming your amp is working fine - or to hire a tech with a good reputation for biasing amps to verify that the tubes you've bought, manufacturer recommended or not, are biased well and will provide best tone for your amp ... if the tech is kind he can show you how to safely bias your amp yourself ... with different amps the approach will change slightly ...
If you decide to go the DIY route hopefully it won't cost you any of your precious output transformers - it's unlikely you can fry it but still it's not a good idea to have tubes run hot through it for too long ... it's a good idea never to leave an amp unattended for too long, especially when experimenting with higher than normal Bias settings ... if your amp is blowing fuses and you keep on replacing the fuse without realising one (or more) of your power tubes goes red hot on the indide while blasting (very common when tubes get old) you may end up damaging your precious output transformer, it's rare but it does happen ... it's a good idea to periodically give your power tubes the eye when you're playing, do so especially if the amp suddenly starts acting funny ... when you see power tubes turn red on the inside it's either time to check your amp out and/or change your power tubes or there's something wrong with the biasing - possibly due to failing components in the circuit ...
Some amps have a separate fuse on the High Tension circuit to protect the output transformer and power supply (the Mesa-Boogie 400 Bass amp is a good example of a tube amp which should have the HT fuse but doesn't) ... if you're still interested in learning the technicalities behind guitar amp biasing I would recommend putting your hands on anything that Ken Fischer has written on the subjext (see early 90's issues of Vintage Guitar and Aspen Pittman's earlier books) ... also, have a look at my tube amp book IF&MTA which goes into this in more detail ...
When an amp comes to me, whether it's a class-A or class-AB push-pull circuit, fixed bias or whatever I go in there and measure the current flowing through each power tubes without altering the circuit in any way (I disagree with cathode resistor methods) ... in fixed bias amps the bias is tested two power tubes at a time whether they are quad or sextet amps ... in cathode biased circuits like the VOX AC-30 all four power tubes have to be present in order for the bias conditions to be appropriate - the only way to roughly infer individual bias of tubes is to measure total current in each leg of the output tranny, then swap the inner two tubes and make the same two measurements - you can write down two sets of linear equations and isolate for each of the four terms to infer per-tube current ... even when I stick manufacturer recommended tubes I will double check that the bias measures proper values for each power tube - I've learnt to never take things for granted in any tube amp ... there are many ways of measuring bias but the technique that I use is most accurate and also most dangerous and so cannot not talk about it openly ... trust me, THIS IS DANGEROUS STUFF AND IT SHOULD NOT BE ATTEMPTED BY NOVICES !!
Is Power Tube Matching Crucial ??
Because of the forgiving nature of push-pull circuits, output stages will work fine and sound pretty ok even though the bias of the power tubes might lie all over the map - i.e., one tube might be biased screaming hot, one almost off, and two sort of ok and the thing will sound acceptable - that's because the output stages are fairly forgiving to tube mismatches by nature ... regardless, in most push-pull tube amps, whether for Instrument or HiFi, it is now well known that the proper biasing of a set of well matched power tubes usually leads to premium results ... I have found that it is impossible to evaluate the usefulness of a preamp mod if the output stage isn't in such shape - something to consider when modifying guitar amps ...
Making Fender Silverface Amps Variable Bias
When CBS bought Fender in '65 one of the things they did to almost all their push-pull tube amps was introduce a bias circuit that had an innefective differential adjustment on the power tubes ... this circuit attempts to rectify bias mismatch between power tubes by introducing variable resistance between the bias-feed circuits to the power tubes in hopes of skewing grid voltages (through grid leakage currents) as a means of rectifying plate spec mismatch - it's complete non-sense and does absolutely nothing - it's likely one of the worse flops in all of tube amp history ... in the early 90's I decided to test theories behind Differential and Separate (double-source) biasing on sets of matched and mismatched tubes and found both techniques did little to nothing to re-balance things out ... the conclusion of these trials convinced me that the only way to make the push-pull stage of a tube amp sound really good (sometimes people use the word "glassy" to describve this depending on the output tranny) is by using well-matched tubes and the single common-mode bias adjustment circuit found in the AB763 issue Fender amps ...
Standard work done on Silverface amps ... whether they come in for repairs or re-tubing ... I always convert the bias circuit in these amps to the true-bias '63 circuit - I also augment the circuit slightly with an extra stage of filtering on the "sweeper" for less hum (see my Fender amp page) ... double checking bias and matching is the only way for me to make sure that the power tubes are sitting in the right bias pocket for these amps - with the larger Silverface amps it's crucial in order to attain the nice smooth tone many players expect from them ... if you want to try hot-biasing your output stage you're on your own - there's no choice but to tabulate bias settings for a given tube brand and see how long they last at these idle current settings ... in general you should get plenty of good tone at nominal bias levels and over-biasing the output stage of your tube amplifier is rarely recommended ...
Scoping un-Matched Power Tubes
 |
With the bias set appropriately the notches disappear but notice the side "walls" are not perfectly symmetric ... this is characteristic display for un-Matched power tube activities ...
 |
Waveform of biased set of same mismatched tubes running at full blast ... notice sharp spikes at turnOn/turnOff boaundaries and the difference in corner sharpness between the two half-cycles ... half-cycle shape differences gets exaggerated in both large-signal transient and large-signal steady-state conditions ...
In any tube amp, however old it may be, if you've upgraded the power supply with new and proper valued filter caps and prediodically measure/adjust bias on a good set of power tubes - you've just covered the two main sources of failure ... following this practice on a routine basis brings the reliability and tone of your tube amp to its peak.
viva Analog /// jc -> lynx.net