Power output question...

Goldie Glocks

Member
Messages
1,684
Sorry if this is in the wrong place.

I am recording with a Two Notes Torpedo Captor rated at 100 watts. I want to record my 150 watt Triple Rectifier. Can anyone definitively say that this amp will not produce 150 watts at say half output? i.e. master volume only half way up? I don't want to pull tubes and I would rather not buy ew equipment but I obviously don't want to break anything either. It is to my understanding that it is not easy to simply measure the wattage coming from the amp so I decided against that path.

Two Notes website casually mentions that you can do it if you don't crank the amp all the way. But this is an old Mesa and I don't want to risk it over heresay. Anyone have any advice?
 

glpg80

Member
Messages
297
RMS output power is strictly determined by bias setting on the amplifier which determines how much peak power can be utilized from the power tubes at full volume. Without getting into too much detail, at half volume running the power tubes at 70% maximum plate dissipation, there’s headroom for the load to be safe. The most power that’s dissipated is during low frequencies <80 Hz or so, and as such, if your music style is tuned down, or involves lots of palm muted notes, I’d suggest running at less than half volume accordingly.
 

Goldie Glocks

Member
Messages
1,684
RMS output power is strictly determined by bias setting on the amplifier which determines how much peak power can be utilized from the power tubes at full volume. Without getting into too much detail, at half volume running the power tubes at 70% maximum plate dissipation, there’s headroom for the load to be safe. The most power that’s dissipated is during low frequencies <80 Hz or so, and as such, if your music style is tuned down, or involves lots of palm muted notes, I’d suggest running at less than half volume accordingly.

Thank you. That is very helpful. I do use lots of palm muting. I don't have to have it at half volume so I'll just plan on using it at a quarter volume. That will still give me the goods and keep the wattage in the safe zone for the Torpedo Captor.
 

pdf64

Member
Messages
8,914
Can anyone definitively say that this amp will not produce 150 watts at say half output? i.e. master volume only half way up?
No, other than 'all at zero' generally resulting in zero power output, it's not feasible to equate any particular control setting to a certain max power output. The controls act to restrict gain / signal level at various points along the course of the amp's signal path, they don't 'regulate', per se, the amp's power output.
It is to my understanding that it is not easy to simply measure the wattage coming from the amp so I decided against that path
Have you got a meter and are able to measure voltage? Power is directly related to voltage, ie power = voltage squared divided by resistance. Hence it take 28VAC to generate 100W in an 8 ohm load.
That can be measured at the speaker output of the amp.
A cheap meter is unlikely to measure audio accurately, and a reactive load's impedance should rise well over 8ohms at bass resonance / above 1kHz, but if the voltage output of the amp was under 20VAV (equates to 50W), it's pretty safe to say that the full audio power will be nowhere near 100W.
Note that with 16 ohm loads, voltages are higher, and at high outputs at 16ohms, your amp can put out dangerous voltages, ie >50V. So use clip on connectors for the probe tips and keep 'hot' connections insulated.

RMS output power is strictly determined by bias setting on the amplifier which determines how much peak power can be utilized from the power tubes at full volume
That runs counter to my understanding, can you explain your thinking?
As I see it, a tube's plate dissipation is just a limit not a characteristic. For a given set of suitable operating conditions, a 6L6G will give the same performance as a 6L6GB, 5881 and 6L6GC, as regardless of the plate dissipation limit of each being different, their characteristics and plate curves are the same.
So with an AB1 amp, providing the bias isn't too close to class A and the p-p signal drive voltage adequate, its idle plate dissipation is unrelated per se to peak plate current and hence power output. Its bias could be adjusted to class B (ie 0% idle dissipation) without that being affected.
Rather peak plate current is solely determined by the load impedance, (loaded) HT voltage, and plate curve for Vg1=0 at the relevant (full load) g2 voltage.
The idle plate current (again, not dissipation per se) mainly affects the output's linearity.
The only relevance of idle plate dissipation is that it be checked as being appropriate to avoid the time averaged plate dissipation exceeding its limit over a range of signal conditions. That’s the point of the ‘idle at up to 70%’ guideline. Of equal importance is that the HT idle is appropriate for the amp / circuit, eg doesn’t overly tax the power supply / move the operating point too far from that which the designer intended.
 
Last edited:

glpg80

Member
Messages
297
This is my thinking, I do apologize for the longwindedness. To begin, we all know power out is by definition power in, minus losses. You break this law, let me know how, and we will all retire comfortably :)

To go further, total power out is further a function of gain, tube plate impedance, primary OT winding impedance, how well matched the power tube’s plate impedances are to the OT impedance, how well matched the tubes are of one another (there’s no power dissipated for any amount of time spent at the zero crossing), and transconductance of the triode/pentode also relative to where the grid is biased. There are simply too many variables to calculate an assumption which would at a stretch still be valid at only a single tone or very limited bandwidth. This is why I estimated with my oversimplification for ease of answering his question.

Additionally, Those tubes may be rated at larger wattages relative to one another but the power supply must be able to supply the additional current the transconductance demands to properly amplify the AC signal. What adds even further complexity is when you begin to discuss frequency dependent performance and efficiency as everything mentioned so far has been for the most part with static CW (continuous waveform) signals.

Continuing down that train of thought, as you combine frequency dependent losses from the OT, since they are horribly limited in bandwidth, bias limitations for total power output, tube limitations of efficiency, power supply inefficiencies, and the fact the grid is driven at a reduced peak to peak amplitude, the RMS output is further reduced to a substantially smaller value than the peak value listed on data sheets.

You are correct and I agree there’s absolutely no way to place a single number on power output strictly by a knob location, but you can use general rules of thumb and knowledge of how the signal is amplified to estimate a safe area to stay within by accounting for the frequency dependent losses and bias of the grids alone.

My goal is not to overcomplicate every thread with theory, but to give answer to the questions based on logical deductions.
 

JJman

Member
Messages
994
1/2 of the human-perceived volume of 150 watts is 15 watts. That's why volume knobs are log taper.
 

Blue Strat

Member
Messages
30,686
1/2 of the human-perceived volume of 150 watts is 15 watts. That's why volume knobs are log taper.
True, but many amps run out of headroom (maximum power point) way before the volume is set to 10 (or 11), so...

That said, I'd wager that a volume setting of 5 in this situation is safe.
 

pdf64

Member
Messages
8,914
...you can use general rules of thumb and knowledge of how the signal is amplified to estimate a safe area to stay within by accounting for the frequency dependent losses and bias of the grids alone
Thanks for your thinking, but unfortunately I still can't make out how the above could be done :huh
A few of specific things that stood out are -
how well matched the power tube’s plate impedances are to the OT impedance
With regular pentode output stages, plate impedance is much higher than OT primary impedance.

Those tubes may be rated at larger wattages relative to one another but the power supply must be able to supply the additional current the transconductance demands to properly amplify the AC signal
As all 6L6 type plates curves are the same, for a given set of electrode voltages and load impedance, the transconductance and HT current should be the same for all 6L6 types.

frequency dependent losses from the OT , since they are horribly limited in bandwidth
I don't recognize the OT bandwidth limitation described, what exactly is being referred to?

bias limitations for total power output
I don't get what this means?
Many thanks, apologies to keep on about it.

1/2 of the human-perceived volume of 150 watts is 15 watts. That's why volume knobs are log taper.
Often, 'log' pots per se aren't used for volume controls, sometimes not even any 'audio' taper type is used.
The taper description of 'log' tends to be confined to European pot manufacturers, and is generally only used for a nominal '10% of total track resistance at 50% rotation' taper (also known as 'law') type, eg http://www.omeg.co.uk/wp-content/uploads/2018/04/PC20BU.pdf
Whereas a wide variety of tapers get used in guitar amp signal level controls, such that the nominal output at 50% rotation can also be 5%, 15%, 35%, even 50% (ie linear) etc.
eg https://el34world.com/charts/Schematics/files/Fender/Fender_hotrod_deluxe.pdf
The use of linear type pots in the above will mean that, given a typical input signal and other control settings, they could be heavily clipping with the (master) volume at halfway.

They might be wired as a variable resistor rather than a potentiometer, so the effective taper 'in circuit' is unlikely to match that of the pot nominal, see 'master' of https://el34world.com/charts/Schematics/files/Mesa_boogie/Boogie_mkiic_plus.pdf I understand that's a Mesa patented arrangement :rolleyes:
Hence other than measuring it, I can't see that it's feasible to even hazard a best guess of what the max power an amp might be capable of putting out at a 'master volume at halfway up' setting.
 

glpg80

Member
Messages
297
With regular pentode output stages, plate impedance is much higher than OT primary impedance.

Total impedance is reduced when valves are in parallel as in class AB output stages. Total plate impedance of all tubes in parallel, let’s call that Zout, must match OT Zin for maximum power transfer, less you get frequency dependent amplification irregularities.

As all 6L6 type plates curves are the same, for a given set of electrode voltages and load impedance, the transconductance and HT current should be the same for all 6L6 types.

Output power will also be identical, but what is not identical, and varies the maximum wattage rating between each variant, is mechanical design absolute limitations and bias. From the Radiotron 4th Edition, page 13, output power is dependent on bias, Beta (or amplification factor mu), and output impedance. I’d link an image for further clarity but my image host is currently down.

I don't recognize the OT bandwidth limitation described, what exactly is being referred to?
I don't recognize the OT bandwidth limitation described, what exactly is being referred to?

Impedance transformers such as OTs are similar to baluns used in RF in that they are frequency dependent and highly inductive. Their inductive nature means their bandwidth is limited, and exhibits roll off at the edges of the spectrum - usually close to DC and again at an upper frequency limit. On a smith chart they begin to look like an open circuit, reflecting power back onto the plates of the power tubes. This is not to be confused with maximum power transfer - you can be well matched and still have output power limitations due to inductive rolloff. A passive device is only so efficient over frequency.

Many thanks, apologies to keep on about it.

No problem! I love discussing this back and forth so if I am wrong somewhere please do correct me. I feel Jeff and HBP know much more about the intricacies of tubes themselves than I, but I’m stating only what I know and can reference with certainty.
 

wall_of_sleep

Member
Messages
818
If I have a resistive dummy load, I'll square the output voltage divided by the resistance ( V^2/R=P). A couple 100 watt resistors cost a couple bucks. This can be done without a scope.
 

pdf64

Member
Messages
8,914
Total impedance is reduced when valves are in parallel as in class AB output stages. Total plate impedance of all tubes in parallel, let’s call that Zout, must match OT Zin for maximum power transfer
In the 'class A / both p-p sides conducting' area of operation, yes, both p-p sides will effectively be in parallel. But not at high power outputs, when the signal level pushes into the 'class B / one p-p side conducting the other cut off' area of operation.
Regular audio pentode output stages don't fit with the max power transfer thing. That theorem is only really used in RF.

less you get frequency dependent amplification irregularities
I'm not aware of such irregularities affecting regular tube guitar amps. Are you thinking of transmission line theory? Again, that's more of an RF thing, I'm not sure it has much application in audio.

Output power will also be identical, but what is not identical, and varies the maximum wattage rating between each variant, is mechanical design absolute limitations and bias
I don't see see a problem with the bias being identical, hence my noting the given set of electrode voltages? If it's changed then surely the operating points won't be the same?

From the Radiotron 4th Edition, page 13, output power is dependent on bias, Beta (or amplification factor mu), and output impedance
Sorry, I can't find that on p13 of my copy or an online scan? http://www.tubebooks.org/Books/RDH4.pdf
Here's a clip of p13 from that
rdh4%2Bp13.png


Impedance transformers such as OTs are similar to baluns used in RF in that they are frequency dependent and highly inductive....A passive device is only so efficient over frequency
For most adequately specified OTs used in tube guitar amps, ie the beefier ones that use some degree of interleaving between primary and secondary winding sections, their power output and efficiency over full frequency range relevant to our application is fine, it's not a compromising factor. Even with the smaller non-interleaved types (eg as used in Vibrolux etc), their limitations don't create an issue as such.
 
Last edited:

glpg80

Member
Messages
297
The book page you outlined is a bit different than my version but states exactly the same thing. It states what is said involves power out of a stage. You’re then limited only by electrical absolute maximums from mechanical design, and notably bias as I’ve said before and the book outlines as well.

Max power transfer is applicable under any appreciation of AC signals - there’s no free energy. Power out is power in minus losses. It’s a law for a reason, and as such, audio signals are just as applicable as RF. You do have to remember audio frequencies are fundamentally made up of infinite Nth number of harmonics well into the RF spectrum. Just because you ignore it as a second order effect doesn’t mean the law does as well.

OTs are subject to exactly what I said - section 5.3.iii specifically states exactly what I mentioned starting on page 210. Low frequency rolloff due to inductance and high frequency rolloff due to stray capacitances as well as reactive limitations (open circuit or self resonance worst case). OTs are very similar to interstage transformers and you will experience this inefficiency for both class A and class B. You’ll find that page 213 outlines it explicitly - power is not a function of frequency response as it is a large signal excitation variable, whereas frequency response is considered a small signal excitation domain similar to S parameters and 4 port network analysis, such as power transfer theorems.
 

pdf64

Member
Messages
8,914
With regard to RDH4 p13, might voltage amplification be getting conflated with power output?
The max power transfer theorem, whilst of course being applicable, does not fit with pentode output stages, because that is not the goal. Rather it’s to achieve good power output of acceptable linearity without causing undue stress to the output devices; overall system efficiency (ie the ratio of power out to power in) not being a concern.
If the max power transfer thing was being applied with pentode AB1 amps, then wouldn’t plate dissipation increase with power output? Whereas the opposite is the case, eg see ‘operation characteristics’ chart at the end of https://tubedata.altanatubes.com.br/sheets/093/6/6L6GC.pdf
The max power transfer theorem is really only put to use in scenarios where the high losses to heat it imposes can’t be avoided, which pretty much limits it to RF and transmission lines.
Regarding OT bandwidth limitations, of course they exist, with the types used for tube guitar amps being unsuitable for hifi. But for our application, their performance seems fine and I don’t recognise the horrible losses / inefficiency / frequency effects mentioned. eg see the Hammond info for their BF Bassman OT, at the full 50W rating it’s +/-1dB 70Hz-15kHz
My understanding is that within their intended bandwidth, it’s unusual for OTs to lose more than 10% of their input power. Hence they don’t tend to warm up much in use, even if running high power continuous waveforms.
On the test bench with a square wave in the range 100Hz - 1kHz fed into the phase splitter, power amp performance into a resistive load is typically very good. Scoping the output shows good reproduction of the waveform all the way up to (and beyond!) full power output. Surely that just wouldn’t be possible if the OT etc had bandwidth limitations that were of any significance?
 
Last edited:

Blue Strat

Member
Messages
30,686
If I have a resistive dummy load, I'll square the output voltage divided by the resistance ( V^2/R=P). A couple 100 watt resistors cost a couple bucks. This can be done without a scope.
Wouldn't you want a scope to tell you the waveform is just below clipping?
 

pdf64

Member
Messages
8,914
Impedance transformers such as OTs are similar to baluns used in RF in that they are frequency dependent and highly inductive. Their inductive nature means their bandwidth is limited, and exhibits roll off at the edges of the spectrum - usually close to DC and again at an upper frequency limit. On a smith chart they begin to look like an open circuit, reflecting power back onto the plates of the power tubes. This is not to be confused with maximum power transfer - you can be well matched and still have output power limitations due to inductive rolloff. A passive device is only so efficient over frequency
I'm reminded of my transmission line EE module several decades ago :)
I don't think that OT bandwidth limitations act to reflect, per se, power back to the power tube plates, as power pentodes aren't operating using impedance matching, so the transmission line mechanisms don't apply. I concede that plate dissipation will increase when used outside its intended bandwidth, but that will primarily be due to a reduction in load impedance skewing the plate load line clockwise, moving it across to the right of the knee. Primary impedance will fall at low frequencies, as detailed in the RDH4 p213 reference, due to the primary inductance no longer providing sufficient reactance to support the impedance being reflected from the secondary. Above the intended bandwidth, I guess that parasitic capacitances will act to reduce primary impedance, though I'm a bit fuzzy on the mechanisms :huh
FYI a few years ago I measured the output impedance of my 6L6 BF Fender type amp, using a differential change in the output load (8ohms reduced to 7ohms), measuring the corresponding change in output voltage, and from that calculating the output impedance from that. At low power levels it was about 100ohms, dropped to about 50ohms at max unclipped output, then halved again to about 25ohms when fully squarewave overdriven.
The amp used p-p 6L6WGB into 4k2:8 ohm OT @ idle VHT of about 425VDC at 35mA cathode current per 6L6 (GZ34 rectifier, 47uF reservoir cap), no global NFB.
The application of the typical BF degree of global feedback (820 - 47) reduced the unclipped output impedance levels by about 3.
 

wall_of_sleep

Member
Messages
818
Wouldn't you want a scope to tell you the waveform is just below clipping?
I'm sure we agree that a scope always beats no scope. And a twenty can get a person into a basic DSO150 that will perform the lions share of audio diagnostics.

The heart of the question to me is what "wattage" is most important in determining if the output device will be safe. When running an amp full tilt I'm more likely to be throwing musical square waves. What's your rule of thumb here?
 

HotBluePlates

Member
Messages
14,057
If I have a resistive dummy load, I'll square the output voltage divided by the resistance ( V^2/R=P). A couple 100 watt resistors cost a couple bucks. This can be done without a scope.
The heart of the question to me is what "wattage" is most important in determining if the output device will be safe. When running an amp full tilt I'm more likely to be ... What's your rule of thumb here?

Rule of Thumb is if you must be certain and safe, don't run a 150w amp into your 100w load box.

Apart from the discussion above, some amount of drive-voltage from the preamp will make the output section do 150w (maybe more after distortion). That drive-voltage is the result of Pickup Voltage * Preamp Gain / Volume Control Loss.

The player can carefully work out what Volume Knob setting imposes enough loss to keep preamp Drive Voltage small enough to stay under 100w.​
The problem is the next time they get excited/aggressive & whack the strings harder, "Pickup Voltage" got a lot bigger and Volume Control Loss is suddenly not enough to stay under 100w.​

Will it matter? No one knows for sure. The OP asked whether we can "definitively say." I can say "probably" (safe if the Volume or Master is turned below some level), but the more I learn the less-certain I am.
 
Last edited:

rl2020

Member
Messages
2
Was researching for a load box for my triple rectifier and ended up here :D


Found my answer on two-notes, and thought I would come back here to share my findings:

In short :

Yes, you can use an amplifier more powerful than the rated power of a Torpedo product. With a Torpedo Studio or Reload, you can use an amplifier rated for more than 150W. With a Torpedo Live or Captor, you can use an amplifier rated for more than 100W. The only thing is : you just have to keep the output volume of the amp low enough.

...
If more power is fed into the Torpedo, it will overheat, and go into overheating protection. You will be warned about this, by means of a visual indication, ...

... Here at our offices, we've been using a 500W bass amplifier for years on all our Torpedo products without any problems, ...
 
Last edited:

HotBluePlates

Member
Messages
14,057
Was researching for a load box for my triple rectifier and ended up here :D


Found my answer on two-notes, and thought I would come back here to share my findings:

"The only thing is : you just have to keep the output volume of the amp low enough."

That's true. Except you'll never know for sure whether you're "low enough" or not. Well, until the load box is damaged.

The exact same thing happens when people debate speaker ratings. There are "safe beyond doubt" combinations of speaker/amp, and "might be okay; use at your own risk" combinations. And somehow folks come back surprised when they blow their speaker in one of those marginal use-cases.
 




Top Bottom