Katana breaks rules of physics?

HAL9000

Yet another Boss/Roland product where, according to the specs, output > input...

Katana 50   - power consumption 47W
Katana 100 - power consumption 77W

JayGee

HAL9000 said:

Yet another Boss/Roland product where, according to the specs, output > input...

Katana 50   - power consumption 47W
Katana 100 - power consumption 77W

The main takeaway from this is that the published specifications are so vague as to be meaningless. Rated power output for my K100 is given as 100 Watts by Boss/Roland with no qualification as to how it's measured and under what conditions. Same for the 77 Watts power consumption. The figures aren't necessarily nonsensical (my knowledge of both theoretical and practical  electronics is at the "just enough to be able to sometimes put together or fix simple things which then more or less work while still potentially being really, really hilariously wrong about what's actually going on inside" level so I'm happy to be contradicted by the people hanging around here who actually do this stuff for a living and have the theoretical and practical chops I don't ) though, and I could envisage an efficient genuine 100 Watts RMS capable SS  amp actually consuming less than 100 Watts if measured over a reasonable period of time with somebody playing a guitar through it in a realistic manner rather than just feeding a test signal in to push it continuously to maximum output. Without any qualification of the figures and test conditions it's impossible to say though.

Having used it at a band rehearsal in a decent size room (as big as some of the venues we play in) I can confirm that the 2x12 version I've bought is plenty loud enough (more than loud enough in fact!) for anything up to a decent sized pub with a loud drummer and un-miced backline with a subjectively clean sound wit7a bit of headroom. Not as stupidly, ear bleedingly. un-necessarily, over the top loud as (say) a Fender Twin or my Blackstar Series One head into a 2x12 cab has the potential for but still very, very loud indeed.

It would still be kind of nice to have enough information to know what I was really buying before I opened the box and plugged it in though...

stratman3142

I first noticed this when looking into the Yamaha THR100 and also noticed this for for Boss products. There was some discussion of this topic on this forum a while back in relation to the THR100.

Despite web searches and discussions with colleagues at work (I was an engineer before I recently retired), I never got to a clear explanation. People offered vague opinions, but I'd like to know the precise methodology (i.e. the exact maths/equations) that the ratings are based on.

olafgarten

stratman3142 said:

I first noticed this when looking into the Yamaha THR100 and also noticed this for for Boss products. There was some discussion of this topic on this forum a while back in relation to the THR100.

Despite web searches and discussions with colleagues at work (I was an engineer before I recently retired), I never got to a clear explanation. People offered vague opinions, but I'd like to know the precise methodology (i.e. the exact maths/equations) that the ratings are based on.

The issue is that we don't know what they are based on, every Manufacturer will have their own system. Most tube amps are rated as an average output but these are probably measured as a peak. This is possible if they have capacitors that can release a burst of extra power when needed. 

ICBM

JayGee said:

I could envisage an efficient genuine 100 Watts RMS capable SS  amp actually consuming less than 100 Watts if measured over a reasonable period of time with somebody playing a guitar through it in a realistic manner rather than just feeding a test signal in to push it continuously to maximum output.

No - that isn't physically possible.'RMS power' (which is a technical nonsense in fact) is actually average power, and the average power output cannot ever exceed the average power input.

What is possible is for the Music Programme power to exceed the input power, if it's reproducing a signal, like a full mix, where the transients - which must remain clean - are several times higher than the average power. In that case as long as the power supply can withstand the current surges for long enough to keep the peaks clean, it's possible for the programme power to be higher than the input power, although the average power must still be less.

But a guitar signal isn't like that - it can be completely distorted, and depending on how you play it can get very close to a continuous full output. Therefore 'RMS' is the correct power measurement.

My guess from the figures is that the Katanas are roughly 30W and 60W amps with a 17W consumption for the preamp and whatever other constant losses there are - that would add up. If so, then the remarkable thing is that they appear to perform nearly as well as a valve amp in terms of dynamics - which is largely what gives valve amps their apparently much higher volume relative to actual power output. That wouldn't surprise me in fact, since it would also explain why they sound so good - and it's been the problem that has been the most difficult to crack so far for solid-state amp designers.

If the Katanas are as loud as 30W and 60W valve amps (through the same speakers obviously) that would be a big achievement. But they still shouldn't really call them 50W and 100W amps.

Sporky

JayGee said:

I could envisage an efficient genuine 100 Watts RMS capable SS  amp actually consuming less than 100 Watts if measured over a reasonable period of time with somebody playing a guitar through it in a realistic manner rather than just feeding a test signal in to push it continuously to maximum output. Without any qualification of the figures and test conditions it's impossible to say though.

Correct on all counts.

Remember that the power input doesn't come straight out of the speakers; it charges a bank of capacitors so the amp has a reserve of power to draw on.

That does mean that if you feed it a full-on square wave from the moment you turn it on, and run it flat out, you could deplete that, but who actually does that?

A good class D amp is around 90-95% efficient (more so with a square wave input) so for the moment let's ignore losses there.

With a power output of 100W and an input of 77W, and let's assume 7W for the preamp and other bits you've got 70W going in for 100W coming out. What that translates to is that you can be playing a pure square wave in 70% of the time that the amp is on at absolute full output power. Or you can play a sine wave in 99% of the time.

Add in excess storage in the power supply bank and you can go beyond those ratings, and actual guitar use is between sine and square waves, so if you're not actually thrashing out 100% clipped power chords for several hours without stopping it is entirely feasible to meet the figures they've stated. If you look at the big power amps they use in touring PA rigs they all have higher output power than input power, and they all achieve it night after night after night. For example:

http://labgruppen.com/view-model/fpseries/fp-10000q?page=spec

10,000W output, just over 6,000W input.

I do agree that there needs to be a measuring standard for all of this.

monquixote

Much wisdom for @Sporky ;

This subject comes up again and again. 

JayGee

ICBM said:

JayGee said:

I could envisage an efficient genuine 100 Watts RMS capable SS  amp actually consuming less than 100 Watts if measured over a reasonable period of time with somebody playing a guitar through it in a realistic manner rather than just feeding a test signal in to push it continuously to maximum output.

No - that isn't physically possible.'RMS power' (which is a technical nonsense in fact) is actually average power, and the average power output cannot ever exceed the average power input.

What is possible is for the Music Programme power to exceed the input power, if it's reproducing a signal, like a full mix, where the transients - which must remain clean - are several times higher than the average power. In that case as long as the power supply can withstand the current surges for long enough to keep the peaks clean, it's possible for the programme power to be higher than the input power, although the average power must still be less.

But a guitar signal isn't like that - it can be completely distorted, and depending on how you play it can get very close to a continuous full output. Therefore 'RMS' is the correct power measurement.

My guess from the figures is that the Katanas are roughly 30W and 60W amps with a 17W consumption for the preamp and whatever other constant losses there are - that would add up. If so, then the remarkable thing is that they appear to perform nearly as well as a valve amp in terms of dynamics - which is largely what gives valve amps their apparently much higher volume relative to actual power output. That wouldn't surprise me in fact, since it would also explain why they sound so good - and it's been the problem that has been the most difficult to crack so far for solid-state amp designers.

If the Katanas are as loud as 30W and 60W valve amps (through the same speakers obviously) that would be a big achievement. But they still shouldn't really call them 50W and 100W amps.

What I had in mind was something along the lines of somebody playing, say, "Stairway to Heaven" - You've got a couple of minutes of gentle, quiet fingerpicking with your 100 watt amp barely ticking over, you've got a couple of minutes of louder clean strummy stuff with the amp working a little harder but still not really breaking into a sweat, then at the end you've got a build up to a minute or so of big dirty chords and flat out wailing where the amp is hitting its peak output on big sustained notes in a few places but still with gaps inbetween. Average the power output over that time and in spite of the output hitting 100 watts (measured however) for a few  seconds in a couple of places the measurement over the whole piece would be a lot lower, and hence so could the average power consumption over that period for what could reasonably be described as a genuine 100 watt amp. Obviously that falls down horribly if you do the measuring with a guitarist from a Motorhead tribute act it's not terribly useful to decide how much stuff you can hang off a single socket or run from a generator, and unless everybody else starts rating stuff in StH average Watts it's totally useless for comparitive purposes but then so are Boss/Roland's bare numbers...

Actually given that Boss have seen fit not to give the combos extension speaker outputs (given Boss/Rolands insistence that  they shouldn't be used with anything lower than an 8 Ohm load I'm guessing that's because there's a good chance that a lot of people would blow them up by trying to run the internal speakers and an extension cab together) they might have been better off ignoring or at least playing down output power and just specifying the maximum SPL they can generate like some PA speaker manufacturer's seem to be doing for their active boxes these days.

They are properly loud though and certainly seem to have the dynamic range and "touch response" of a decent sized valve amp. Surprised the hell out of me I can tell you, I can't help thinking that I could have saved some money and even more weight/space by getting the 1x12, and that if the "50 Watt" one had the same facilities (specifically number of presets and ability to work with the GA-FC foot controller) I could have saved even more...

ecc83

The first thing you need to do is define 'Power Output' and the only fair way to my mind is, watts into a known load resistance for a specified distortion level ('clipping' is no good as that is in the eye of the scope beholder) and for guitar amps, 10%Total Harmonic Distortion would be a good benchmark (lots of valve OP data uses 10% THD) . The frequency is say 1kHz but test should be done at say 100Hz for a decent report.

But, 'watts'? Can't measure those directly so it becomes  a matter of measuring the voltage across the load and THAT is the 'rms' bit! Now, we can measure the rms voltage at 1kHz fairly easily (but check YOUR DMM is accurate at 1k ref 50Hz!) Bit of simple maths give us P=V x V (rms) over R. But there are snags depending upon how fussy/accurate you want to be..

How then to measure the consumption of the amp whilst delivering X watts ? Mains I times mains V? Yes but IS the current a sine wave? You bet your sweet bippy it ain't! 'True' rms meters should take this into account but they will likely be expensive.

I suspect Boss just ran an amp up to full welly with a guitar signal and measured the mains consumption with either a stock (B&Q?) power meter or, measured the current and pronounced from that..Could even have just been the ball park figure lifted from the P transformer spec! Of course, the music signal will be nothing like a sine wave so the power output is indeterminate.

Don't want to knock but making the amp 8 Ohm min load capable is IMHO a retrograde step and will come to bite their corporate posteriors. Of course they could have some very aggressive internal limiting, or a fuse but neither option is very user friendly.

Dave.

Sporky

ecc83 said:

Don't want to knock but making the amp 8 Ohm min load capable is IMHO a retrograde step and will come to bite their corporate posteriors. 

Does seem odd considering the plethora of modern amp topologies that pretty much don't care about load impedance. Quite a few of the Lab Gruppen range (which don't start at the expensive end of the market) put the same power into anywhere from 2 to 16 ohms, or even 70v/100v line.

JayGee

ecc83 said:

Don't want to knock but making the amp 8 Ohm min load capable is IMHO a retrograde step and will come to bite their corporate posteriors. Of course they could have some very aggressive internal limiting, or a fuse but neither option is very user friendly.

That 8 Ohm loading thing struck me as odd. In my admittedly limited experience it's unusual for a solid state amp to have a minimum load that high, most things seem quite happy into 4 Ohms (and quoted maximum power ratings usually seem to be at 4 Ohms). I wondered if there was something clever and/or unusual going on in the power amp design which was contributing to the perceived "valviness" with the unusually high minimum load impedance being a consequence or contributing factor to this...

ecc83

JayGee said:

ecc83 said:

Don't want to knock but making the amp 8 Ohm min load capable is IMHO a retrograde step and will come to bite their corporate posteriors. Of course they could have some very aggressive internal limiting, or a fuse but neither option is very user friendly.

That 8 Ohm loading thing struck me as odd. In my admittedly limited experience it's unusual for a solid state amp to have a minimum load that high, most things seem quite happy into 4 Ohms (and quoted maximum power ratings usually seem to be at 4 Ohms). I wondered if there was something clever and/or unusual going on in the power amp design which was contributing to the perceived "valviness" with the unusually high minimum load impedance being a consequence or contributing factor to this...

From a design point of view it is tempting perhaps to make a transistorized amp with fairly high internal supply rails so that it can drive 'full power' into 8 Ohm speakers and around 75% of that into a 16 Ohm load?

The fly in the emolument of course is that such a design might self destruct into 4 Ohms. The reality is of course, 'it'appen, speaker lines DO get shorted . No sensible amp maker would make such a vulnerable device? "Back in The Day" transistor amps WERE very feeble, even some quite expensive 'hi fi' amps would blow if shorted and even if run at high levels for too long and got hot. Those days SHOULD be firmly behind us!

The REAL answer of course is to do what we do with valves. Tapped OP traff. Trouble is, it is the OPT that causes the destruction when you forget to load a valve amp, but you 'usually' get a minute...With transistors? Micro seconds.

Of course, my comments could come home to roost and the country, nay world, be littered with the smoking remains of ID/TVP power amp modules! To the best of my information to date, that is not happening.

Dave.

darthed1981

The Katana presumably does include some nifty design as it does seem to just behave differently to most SS amps.

Be interested to see how it is different - I do love how its just got a few knobs, the ones with Apps would just have me tinkering with settings and not playing.  As it is I set it to crunch channel, bit of reverb, maybe some delay, and off you go!

ecc83

"Lab Gruppen range (which don't start at the expensive end of the market) put the same power into anywhere from 2 to 16 ohms, or even 70v/100v line."

Sat WHA! Sporks? I have not looked at prices but they are highly sophisticated amplifiers and seem to be under close CPU control?

With THAT sort of technology you could make valve output stages virtually bombproof! I have done 'back of envelope' circuits for OP current sensing and suchwhich.

As I said up the page a bit,  power into various Zs is easily solved for SState amps by using an OPT but that cause more problems than it solves UNLESS you also use the La Gruppen philosophy. 'We' would go broke.

Now then, there was ONE failrly common transistor amp that used an OPT and was pretty unburstable, the Quad 50E. Would like to find one and let a guitarist rip on it?

Dave.

Sporky

ecc83 said:

"Lab Gruppen range (which don't start at the expensive end of the market) put the same power into anywhere from 2 to 16 ohms, or even 70v/100v line."

Sat WHA! Sporks? I have not looked at prices but they are highly sophisticated amplifiers and seem to be under close CPU control?

The E series isn't vast money for the power. I think Lab's explanation is that the two-channel amps are actually four-channel with internal bridging which means they can run channels in series or parallel or singly as appropriate for the load. 

They also power share across the channels, so the E2:12 (for example) has 1200W across 2 channels - split however you want. And the splitting is automatic - as one channel gets more level it gets more power allocated.

Sounds like witchcraft I know, but they work, they're bombproof and they're comparable price-wise to the competition which isn't as clever.

ecc83

Sporky said:

ecc83 said:

"Lab Gruppen range (which don't start at the expensive end of the market) put the same power into anywhere from 2 to 16 ohms, or even 70v/100v line."

Sat WHA! Sporks? I have not looked at prices but they are highly sophisticated amplifiers and seem to be under close CPU control?

The E series isn't vast money for the power. I think Lab's explanation is that the two-channel amps are actually four-channel with internal bridging which means they can run channels in series or parallel or singly as appropriate for the load. 

They also power share across the channels, so the E2:12 (for example) has 1200W across 2 channels - split however you want. And the splitting is automatic - as one channel gets more level it gets more power allocated.

Sounds like witchcraft I know, but they work, they're bombproof and they're comparable price-wise to the competition which isn't as clever.

Yes but is there any 'competition' in the guitar amp market? Top makers, Crown, probably  one of the first, have been making virtually indestructible sstate amps for decades but AFAIK nothing of that sophistication has found its way to the gitamp market?

The IDs get around the reliability problem by  what is essentially a very over powered amp module plus some fairly clever overload detection and protection. The trick with the latter (as you WELL know ,Mr S!) is to save the amp but not make it sound like ***t when the protection kicks in.

The ID power is still max with a 4 Ohm load, I can see no simple, economic solution to that problem?

Dave.

Sporky

ecc83 said:

Yes but is there any 'competition' in the guitar amp market? Top makers, Crown, probably  one of the first, have been making virtually indestructible sstate amps for decades but AFAIK nothing of that sophistication has found its way to the gitamp market

Not that I know of, but it does surprise me that more modeller users don't go for commercial power amps. Proven reliability, massive power density, convection cooling...

ICBM

ecc83 said:

The ID power is still max with a 4 Ohm load, I can see no simple, economic solution to that problem?

I've actually wondered for some time whether a deliberately very poorly-regulated power supply might be a partial solution to this, as well as providing more of a 'valve feel' to the amp.

The ironic thing about power supply regulation is that the better the regulation - ie the 'better' the power supply - the *worse* the power drop-off at higher impedances since the supply voltage doesn't rise as much.

Voxman

ICBM said:

My guess from the figures is that the Katanas are roughly 30W and 60W amps with a 17W consumption for the preamp and whatever other constant losses there are - that would add up. If so, then the remarkable thing is that they appear to perform nearly as well as a valve amp in terms of dynamics - which is largely what gives valve amps their apparently much higher volume relative to actual power output. That wouldn't surprise me in fact, since it would also explain why they sound so good - and it's been the problem that has been the most difficult to crack so far for solid-state amp designers.  If the Katanas are as loud as 30W and 60W valve amps (through the same speakers obviously) that would be a big achievement. But they still shouldn't really call them 50W and 100W amps.

That is what the Blackstar ID TVP series amps (True Valve Power) were supposed to do. However, I was very disappointed with the  ID60TVP .  Haven't yet heard a Katana and I'd be interested in how they sound & feel as compared to the Blackstar ID's (the bigger versions) and my Vox AD120VTX - obviously the modelling will be more up to date, but its the valve feel and how it changes at volume that I'm interested in. 

ecc83

ICBM said:

ecc83 said:

The ID power is still max with a 4 Ohm load, I can see no simple, economic solution to that problem?

I've actually wondered for some time whether a deliberately very poorly-regulated power supply might be a partial solution to this, as well as providing more of a 'valve feel' to the amp.

The ironic thing about power supply regulation is that the better the regulation - ie the 'better' the power supply - the *worse* the power drop-off at higher impedances since the supply voltage doesn't rise as much.

Ha! One thing I never did with an ID chassis was check the supply 'droop'! I would guess though it would not be much. The Tech director is 't hot on having reliable, well rated mains traffs. The rectifier is pretty chunky and the caps 3,300 and 4,700 per leg IIRC for the 60 and 100 W amps.

But! How about an 'anti' regulator? a power FET in the supply that was 'programmed' by the connected impedance and gave both 'sag' and protection for the amp?

Thing is though, this all assumes MOST of the goodly sound of valve amps is in the OP stage when in actual fact the pre amps play a much larger part than many people realize. Your good self likes big amps but tell us you don't blow the beans off them?  The fact that so many people now seem happy to put a 'modelled' sound or at least an emulated one, through the PA means they must be happy with big, clean amplifiers?

 Dave.

ICBM

Voxman said:

That is what the Blackstar ID TVP series amps (True Valve Power) were supposed to do. However, I was very disappointed with the  ID60TVP .  Haven't yet heard a Katana and I'd be interested in how they sound & feel as compared to the Blackstar ID's (the bigger versions) and my Vox AD120VTX - obviously the modelling will be more up to date, but its the valve feel and how it changes at volume that I'm interested in. 

There's a big difference - as Dave said, the Blackstars were actually far *more* powerful than their claimed outputs, not less - double, if I remember rightly. That gives them the ability to be 'as loud as valve' (as they claimed) by matching the full distorted output of a valve amp while the power section remains clean. And they certainly are loud.

It's rather a over-engineered solution in some ways though - at least compared to making a lower-powered amp more dynamic... I also found them rather stiff and dark, but I don't know if that's to do with the power section (or power supply) or just the Blackstar voicing - I find all their amps sound like that, even the Artisans.