Oh... okay then!nguideau wrote:Andy was just discussing this in the Technical forum on a thread about the TT (comparing it to a 2061x, and how different factors result in the amps having different ratings). It would be awesome if he could chime in.
(This is all your fault mind... Hope its worth it!)
(Quick OT...)
Right, well I had a pleasant long chat with my friend John Phillips the other day, and we did touch on this stuff for a couple of minutes. Wasn't why I was phoning him though...
As I've mentioned I've just bought a nice '74 Hiwatt DR103 'Custom 100' head. It is in Scotland right now - John has it safely tucked up at home for me. He's keeping it warm for me, powering it up now and again and planning to use it in the studio for some recording next month. Thats fine by me - powering them up now and then is actually good for valve amps, especially for older ones with older filter capacitors in them - about the only components in a valve amp with a pretty finite life expectancy. Hitting them with working voltage fairly regularly actually helps preserve their performance, and the filter caps in any amps over about 15 years old can be suspect, especially if the amp has lain unused for a while at any time. If they have degraded badly, apart from going way off spec and affecting amp performance, they may actually pop, (or even explode), leaking (/explosively spraying!) corrosive gunk around inside your amp - not good. So it is best getting them checked in any older amp - especially if it has been sitting unused for a couple of years or more at any point!
At 35 years and counting this DR103 still has its original caps. They test and perform fine too which means that this amp has been used regularly ever since being built. And in fact apart from a replacement mains socket (to comply with legal H+S regs for gigging) and replacement power valves this amp is totally original - including all the preamp valves. ( !)
Clearly if these amps can be used regularly for so long with no failures or repairs needed Hiwatt were doing something right. And despite what a few might tell you (Matamp! ) they - along with things like old Marshalls and Oranges - do have the reputation of being about as reliable in actual regular use as any valve instrument amps ever made.
John measured its clean power output. It's apparently putting out 135W RMS before distortion. Pretty healthy really especially when you consider how bomb proof reliable these amps are...
(We seem to be right back on topic now...)
How though? 135W clean with only 4 X EL34 valves? And evidently without running the thing ragged either! Heres how...
Okay, technically speaking the amp - just like a 100W 'Plexi' Marshall, or my OR120, or an AD140, or many other 4 x EL34 amps that may or may not actually say 100 watts on the label - is a 'hundred watter'. Not all 4xEL34 amps are 'hundred watters' but most are, and thats usually a pretty realistic name for them. So whats up with people talking about getting 130+ watts clean out of (some of) these amps? Well it CAN happen - it really depends on just two things: 1. the amps themselves and 2. the power valves you put in them.
Not all 100W rated 4 x EL34 amps will be able to sustain 100+ watts of clean output. Limiting factors may be the power supply (power transformer, rectifier, filter caps) as well as the output transformer, or various other circuit components / design features. (More power generally means bigger components, bigger chassis, more ventilation...) But when all these things are capable of creating and handling higher power levels than the 100W 'spec', the limiting factor is simply...
THE POWER VALVES.
So whats happening with them? Okay, couple of different situations:
First how they run in a true class A situation. (Not the TT / AD30 / AC30 so called 'class A', but true class A running like the AD5, THD Univalve or Bivalve amps!)
Genuine CLASS A running is actually very simple!
In class A running all power valves will run their full set current at all times. (Whatever the configuration, a 'single ended' single output valve amp, or an amp with one or more valve pairs in either 'parallel single ended' or 'push-pull' configuration - it doesn't matter!) Regardless of whether the amp is amplifying anything or not the total power they are putting out won't change at all, not between no signal and maximum clean power output anyway. The only thing that will change is where that power is going to.
When you aren't playing and the amp is silent the power is all going up as heat!
At full clean output 50% of that power will still go up as heat.
The other 50% of the power the valves are dissipating will, (minus some inevitable losses), go via the output transformer to your speaker(s). So thats a (theoretical) 50% maximum of the power being dissipated by the valve plates available as useful output power. (Class A is very inefficient!!!)
I say the 'full set current' because you don't have to set them to 100% of rated plate dissipation for them to be running in class A! Setting the amp to run them at less than 100% will reduce output power but won't itself take them into class A/B running. Run them at 85% dissipation if you want (and it still sounds good to you) - or run them at 45%. It'll be less loud, but every bit as class A! Doesn't say whether it'll sound good or not, but that is a totally different matter...
Whatever power level you choose to bias the amp to run the valves, they will be switched on and fully working at that set level all the time. Simple really.
I say 'maximum clean power output' because as soon as the amp exceeds that level and goes into distortion it is no longer running in class A - or any other class really! When an amp distorts it goes outwith the parameters defining operating class, and even a true class A amp will then put out more power. But talking about operating class really is meaningless unless you stay within the amps clean output range! So I won't be going there, except to say that yes you'll get more actual power output by fully saturating and then overdriving the amp.
I'm totally sticking to talking about clean amp power# here though!
OT (#Measuring clean amp power? Yes it can be measured and rated in different ways. But measuring true clean output means 'before the onset of clipping distortion'. Many / most valve instrument amp manufacturers will rate for the output at '5% total harmonic distortion' though. The power figure will be a bit higher, but makes at least some sense because (unlike with solid state amps) it'll still sound pretty clean. You'll hear 0.1% distortion from a ss amp at least as easily as several percent distortion from output valves, simply because valve distortion sounds much more natural to our ears! If you try to judge by ear when your valve amp is starting to distort that will seem to happen well after an oscilloscope trace actually starts showing early output distortion.)
Okay, thats class A. Now CLASS B running.
(No, not class A/B, just B!)
Its just a little bit more complicated, but really not so bad. Basically the difference is that in class B running all the power valves will be resting totally for half of the time. The rest of the time they will be running somewhere between zero and maximum set output current.
In class B amps the output valves are always run as 'push-pull' pairs. While in class A you can run 'push-pull' or (parallel) 'single-ended' configurations, all class B amps (and all class A/B ones too) are push-pull configuration. So what is that...?
Imagine the signal being amplified as a squiggly trace going up and down, running from left to right. Imagine a straight mid line (going left to right) which is neither up or down but bang in between. Lets call that a 'zero point'.
In class B running some valves (one of each push - pull pair) only amplify the bits of signal above that line, which we can call the 'positive' (+) signal. Other valves only amplify the bits of signal below the line, the 'negative' signal. (The separate +ve and -ve signals are derived from the 'phase splitter' or 'phase inverter' stage of the amp which is between the preamp and power amp sections. The amplified signals are then re-combined at the output transformer stage to give the full amplified waveforms.)
So all valves running in class B spend half of their time resting, and the rest of the time passing a varying level of signal current. Guess what? Its a bit like the speaker situation I was talking about recently here:
http://forum.orangeamps.com/viewtopic.p ... 71#p465271" onclick="window.open(this.href);return false;
Speakers can handle temporary bursts of power above their power handling rating - as long as those peaks are not TOO high and the speakers also have time to cool off again in between.
Same with valves - they are thermal thingies too, technical name 'thermionic valves'! If they aren't having to put out power all the time then they can temporarily put more than their rated power when they actually are amplifying a signal. Just like speakers the limiting factor is heat buildup, and the important thing is to run them so that the average power produced doesn't cause thermal damage. (Red plating --> actual meltdown of the valves internals.)
As they are only running half the time, and (when running a clean signal anyway) apart from at the very highest signal peaks will still not be putting out full signal, you'll see that when actually being asked to run signal they can put out a lot more power than they could continuously handle. Simply because they get plenty rests in between you can push them much harder when they actually are working. Remember, in class B no signal = no plate current.
Guess what, class B amps are much more efficient then class A ones. If you're wondering just how much power you could get from a class B instrument amp the answer is I don't really know, but well over double that of a class A one anyway. Really doesn't matter though - because as far as I know there aren't any! Class B has its uses, but amps for musical instruments / audio aren't one of them...
Problem with class B amps is what they call 'crossover distortion'. When the signal passes across the center line between +ve and -ve signal the valve(s) on one side of the push-pull pair(s) take up the signal while the others stop putting out any power. (Resting valves are described as having gone into 'cutoff' = no signal produced. That isn't the same as clipping distortion which happens at above their maximum clean current capabilities BTW - totally different thing!)
Trouble is that it is very difficult to make the transition between signal and cutoff smooth at this crossover point and you get really nasty sounding crossover distortion with audio signals in class B amps. And it really isn't anything like the nice 'soft clipping' saturation distortion we like from our valve guitar amps - it is horrible sounding, and very obvious.
The solution used in most instrument / audio amps is to run the output devices (valves, output transistors, whatever - I'll stick to 'valves') in an intermediate state. Guess what that is called...?
In CLASS A/B running the amps valves never quite go into cutoff.
Unlike in class B running they never completely stop passing signal. But the main thing is that unlike class A they aren't passing full set current all of the time, only when full output is needed. They will be passing some current all of the time though.
What happens is that the '+ve signal' valves of each push pull pair will also amplify a little bit of the '-ve signal' part of the waveform. Not all of it though, they may only just cover the fairly low current area just negative of the 'mid line'. Similarly the -ve signal valves will also amplify a little of the +ve part of the waveform.
Unlike valves operating in single ended class A though none of them will ever go 'full wave' and amplify the whole signal.
(Again the split signals from the 'push-pull' pair halves are recombined at the output transformer stage to form the complete amplified waveforms. In a well sorted and adjusted design the crossover region around the 'zero line' will be handled smoothly without crossover distortion.)
Class A/B amps are less efficient then class B ones then, but more efficient than class A ones. All the valves are putting out some power all the time, but will only pass full current when asked to amplify the highest peaks of the waveform. So though never fully resting they are also hardly ever working at full clean output. This means of course that while getting less rest than valves running in class B they still do have plenty rest from full output. Which means that they can still be pushed quite some way past their rated plate dissipation when actually needing to amplify higher signal levels - because at lower signal levels they can still cool off some. You can't push them as hard as in class B because they get less rest, but you can on average still get double the useful working power output of class A running without causing them to thermally overload and red plate.
This explains why valves running in class A/B, can basically output their rated plate dissipation even though not working at 100% power 100% of the time. If they are working some of the time at say 70% power output they can be pushed to run temporarily to well over 100% at other times. The 'temporarily' bit is important though. They won't be able to sustain that which is why you can't normally simply bias a class A/B amp hotter and expect it to run in class A. You'll very quickly fry the valves by running them on average at above 100% of their power dissipation rating. The average is all important!!!
Which still doesn't explain how four EL34s can sustain 135W RMS CLEAN output in my Hiwatt!
Well sorry, slight apology here. After reading all the above I'm going to have to tell you that it is really quite simply down to the valves you use!
Not all valves are equal, and certainly not all EL34s. The better modern production ones will sustain the EL34 plate dissipation rating of 25W right enough. But not a lot more, and only for a few hundred hours in total before failing.
Some old ones while still rated at 25W RMS could actually sustain considerably more than 25W output without terminally red plating. And could do so for a far longer working life too. (Not hundreds but thousands of hours!) They were much better built with internal components that were like proper heavy engineering compared to the relatively flimsy modern 'to a price' components used in valve manufacturing today. The reason for this was that valves were used much more widely in all sorts of electronics, and also (especially) they were used by the military. Reliability not cost was all important so valves were usually comfortably over spec. The strongest and most consistent performing EL34s ever produced were probably the famous old Mullards. My amp is fitted with some well worn old (OS) Mullard EL34s which even look well used but may well still outlast a couple of sets of brand new modern ones - while still considerably outperforming them!
John reckons that unused NOS Mullards will happily sustain not the rated 25W but an actual real world 35W RMS usable clean output. Multiply that by 4 and you'll get a real world 140W RMS clean, which in his experience is about the upper power output limit you can expect from a healthy spec '100 watter'.
Don't expect to get that sort of power out of a set of modern EH or JJ EL34s or similar though. If you tried to bias them the same as good NOS valves to allow that much output they would red plate and burn out within minutes.
Theres your answer! (Hope it was worth the wait...)
A couple of last thoughts for you...
Biasing...:
Remember in the class B section I talked about the 'zero line' in the waveform. Well that 'zero line' doesn't mean zero volts! All biasing does is to shift that zero line (and with it the whole waveform) up or down the voltage scale. All you are trying to do with biasing is to fit the waveform (or whichever part of it you are asking the valves to handle) within the valves working voltage (and power handling) range. It is exactly that simple!
The different types of class A/B instrument amps...
Class A/B covers a lot of ground in terms of how amps run the valves. But most class A/B guitar amps tend to be at one or the other extreme. Some amps, like the (Vox based) hot running cathode biased 4 x EL84 30ish watters depend to a large extent on pushing (and punishing) the valves hard to get their sound. That is why I am reluctant to burn up expensive and increasingly rare NOS power valves in them. These are the so called 'hot running' class A/B amps! (Often incorrectly labelled as 'class A'.)
The much more efficient higher output amps like these '100 watters' are at the opposite end of the scale. They actually need to have the valves biased as cool as possible to achieve their maximum sustainable clean power. Bias them hotter and they'll run hotter at every power level and run out of steam sooner. Setting a cooler bias doesn't limit the ultimate output power in the way that many think - rather the opposite!
Setting the bias only sets the available 'power range'. Your volume control sets the actual power!!!
If biased hotter not only will the amps distort earlier, (having less clean headroom), but this will burn the valves out faster too. Old amps like my Orange and Hiwatt were totally designed to run the valves biased as cool as possible, and that is the way to get the best power and reliability with them. This is why you'll read people like John and myself saying we really don't hold with biasing them up near the theoretical 70% maximum idling plate dissipation figure. That really is a maximum guideline figure, beyond which - specially with modern valves - you'll red plate the valves very quickly when you turn the amp up! With these amps, where the limiting factor to power and reliability is the valves themselves, it is better to get them as close to the class B end while still staying above the 'crossover distortion threshold'. That way you get great tone, great output and great reliability.
And IMO (just my opinion!) these more efficient amps are exactly the right place to use NOS or used 'OS' power valves. Thats when you'll get the big dividend returns in terms of sound, power and longer life that these valves can give you.
Bung NOS valves in hot A/B amps (eg. AC30 / AD30 / TT) and you'll not notice nearly as much sonic difference - any valves will get thrashed in there, being run way over their design limits. And don't expect NOS valves to last all that long in them either - yes, they'll last longer than modern ones, but not that much longer - you'll probably be disappointed!
With my own 'hot A/B' amps I prefer just to use the best current production valves which can sound really great too. I also don't feel as bad running them as hard as the amp needs to get them to sound really great, and can live with a few watts less ultimate power. My '60s AC30 head last tested out at around 28W RMS clean with its modern Harma EL84s. Thats set about right IMO - the amps were rated at 33W clean - but that was in the days of really good performing old valves! I wouldn't really expect to be able to run modern ones at that sort of power though - not for any length of time anyway.
Guess what Brian May uses in his AC30s...?
Any of that make any sense then? If so, and you want to know more here are a couple of suggestions for more details:
http://en.wikipedia.org/wiki/Electronic ... er_classes" onclick="window.open(this.href);return false;
Also... If you really want to know about this stuff do yourself a favour and check out the 'Introductory' and 'Advanced' sections by Randall Aiken under 'Tech Info' here:
http://www.aikenamps.com/" onclick="window.open(this.href);return false;
Very informative and reliable site - Aiken totally knows his stuff!!!
Andy.