How to find the peak decibel gain of the highest gain channel of a high gain amp?

[thinkingchicken]

Marshall JVM410 generated 115 dB as the peak dB gain through its highest gain channel with the gain at max, and the power supply which can be seen at the back panel is 250 V. So, I tried to calculate how JVM410 gain reach the dB reading.

marshalljvm410dbgain.jpg

The input voltage is so low, is that correct? Does that low microvolt voltage come from the guitar itself?

[wpaulvogel]

Yes it comes from the guitar and the guitar produces usually at least .050 volts more like .100 volts with typical output vintage pickups. Once again, it’s impossible for the amp to have 115 dB gain because it causes clipping and the gain is reduced. Why can you not understand

[thinkingchicken]

Yes it comes from the guitar and the guitar produces usually at least .050 volts more like .100 volts with typical output vintage pickups. Once again, it’s impossible for the amp to have 115 dB gain because it causes clipping and the gain is reduced. Why can you not understand

These dB figures actually come from an another amp builder from Australia, if you read the convos in my first link which is here https://www.thegearpage.net/board/index ... t-30886030. Are you saying these dB figures are just plain wrong?

[LOUDthud]

Never thought that Trainwreck Express has so much gain. Did the tremendous gain mainly come from the power tubes? Because I heard when it comes to these Trainwreck Express amps, it is the power amp that is the first to distort then the preamp.

This is pure Internet nonsense. That's the way an amp without a Master Volume has to work. Even if the power amp has a gain of 1. If it doesn't work that way, it's not working.

This B+ thing is the power supply, right? Seems like the more the power supply, the more the sound distorted when only raw gain is used without the clipping diodes. So, you just measured the power supply but not the decibel gain of the Express?

That amp has layout problems because the wires for the output transformer come through the chassis between the eyelet board and the front panel pots. Then the speaker jacks are on the back panel right next to the preamp tubes.

Both of the other amps you mentioned are around the same at 100 dB. Did they started to oscillating too when you drive the gain knob to the max?

Nope. A good layout, short leads to the pots, no nonsense grounding, works as designed. I think the only coax is between the Master Volume and the Effects jacks, then to the power amp.

[teemuk]

Once again, it’s impossible for the amp to have 115 dB gain because it causes clipping and the gain is reduced.

The gain is not reduced, only the output voltage is limited by distortion. All the "gain" - gain being mere amplification ratio - is still there and 115 dB not limited by clipping is possible with low-enough input. This is, of course, purely theoretical speculation because with such figures low-enough input is likely just background noise. Nevertheless, do not mix up concepts of amplification factor/ratio and clipping threshold.

[thinkingchicken]

Once again, it’s impossible for the amp to have 115 dB gain because it causes clipping and the gain is reduced.

The gain is not reduced, only the output voltage is limited by distortion. All the "gain", gaib being mere amplification ratio, is still there and 115 dB not limited by clipping is possible with low-enough input. This is, of course, purely theoretical speculation because with such figures low-enough input is likely just background noise. Nevertheless, do not mix up concepts of amplification factor/ratio and clipping threshold.

You see, these disagreements and different views are what's confusing me.

[thinkingchicken]

Never thought that Trainwreck Express has so much gain. Did the tremendous gain mainly come from the power tubes? Because I heard when it comes to these Trainwreck Express amps, it is the power amp that is the first to distort then the preamp.

This is pure Internet nonsense. That's the way an amp without a Master Volume has to work. Even if the power amp has a gain of 1. If it doesn't work that way, it's not working.

This B+ thing is the power supply, right? Seems like the more the power supply, the more the sound distorted when only raw gain is used without the clipping diodes. So, you just measured the power supply but not the decibel gain of the Express?

That amp has layout problems because the wires for the output transformer come through the chassis between the eyelet board and the front panel pots. Then the speaker jacks are on the back panel right next to the preamp tubes.

Both of the other amps you mentioned are around the same at 100 dB. Did they started to oscillating too when you drive the gain knob to the max?

Nope. A good layout, short leads to the pots, no nonsense grounding, works as designed. I think the only coax is between the Master Volume and the Effects jacks, then to the power amp.

So, if I already understand how NVM amps work, the high gain of Trainwreck Express mainly come from the power amp which is interesting. Somehow I believe the gain of that Express is beyond 100dB because of the occurrences of oscillations.

We can safely say 100dB is the highest gain where distortion is still usable.

[teemuk]

No. Gain of something like an amp is sum of all its gain stages. Just because the power amp stage clips first won't mean its the largest contributor to gain ratio.

And no. We can't say "100dB is the highest gain where distortion is still usable" because such statement does not define input amplitude nor clipping threshold. 100 dB is voltage gain of one million, a high figure, but if our input is just one microvolt and clipping threshold meager 10V then the 1V output (microvolt is one millionth of volt) is yet nowhere near the clipping threshold.

I have told several times that if distortion is concerned then gain ratio is a vague parameter unless you also define input amplitude and clipping threshold. What is so difficult to grasp in that concept?

Stop obsessing about specific values of "gain", rather focus on what a certain value of gain will do in a design with its specific clipping thresholds and nominal input signal amplitudes. Gain of x dB will have different results in circuits with different clipping thresholds and different results with input signals with different amplitudes.

For example, 40 dB gain (100x voltage gain) in a microphone preamp is normal, same 40 dB gain in a power amp is way too excessive.
If we think about a mic preamp it may have clipping threshold around 12 volts but its nominal input signal amplitude is only just few millivolts. One millivolt input results to only 1 mV x 100 = 1V output so the preamp nicely tolerates some 12 millivolts (peak) before clipping. (This is just a generalisation, don't think it applies to each and every mic preamp).
Power amp, on the other hand, is often designed to accept signals with higher, "line level", nominal amplitudes, typically around 1Vpeak or so. Power amps may have much higher clipping threshold than a generic mic preamp, like 40 volts, but 100x gain would still be more than twicefold too much and a 1Vpeak input signal would lead them to overdrive and distort with such high gains. So a power amp, with its higher clipping threshold BUT also higher "input sensitivity" rating needs much less gain to reach its clipping threshold.

See how there's interplay between input signal amplitude and clipping threshold and how effects of "gain" are really defined by those parameters. Gain alone is nearly meaningless parameter.

[thinkingchicken]

No. Gain of something like an amp is sum of all its gain stages. Just because the power amp stage clips first won't mean its the largest contributor to gain ratio.

And no. We can't say "100dB is the highest gain where distortion is still usable" because such statement does not define input amplitude nor clipping threshold. 100 dB is voltage gain of one million, a high figure, but if our input is just one microvolt and clipping threshold meager 10V then the 1V output (microvolt is one millionth of volt) is yet nowhere near the clipping threshold.

I have told several times that if distortion is concerned then gain ratio is a vague parameter unless you also define input amplitude and clipping threshold. What is so difficult to grasp in that concept?

Stop obsessing about specific values of "gain", rather focus on what a certain value of gain will do in a design with its specific clipping thresholds and nominal input signal amplitudes. Gain of x dB will have different results in circuits with different clipping thresholds and different results with input signals with different amplitudes.

For example, 40 dB gain (100x voltage gain) in a microphone preamp is normal, same 40 dB gain in a power amp is way too excessive.
If we think about a mic preamp it may have clipping threshold around 12 volts but its nominal input signal amplitude is only just few millivolts. One millivolt input results to only 1 mV x 100 = 1V output so the preamp nicely tolerates some 12 millivolts (peak) before clipping. (This is just a generalisation, don't think it applies to each and every mic preamp).
Power amp, on the other hand, is often designed to accept signals with higher, "line level", nominal amplitudes, typically around 1Vpeak or so. Power amps may have much higher clipping threshold than a generic mic preamp, like 40 volts, but 100x gain would still be more than twicefold too much and a 1Vpeak input signal would lead them to overdrive and distort with such high gains. So a power amp, with its higher clipping threshold BUT also higher "input sensitivity" rating needs much less gain to reach its clipping threshold.

See how there's interplay between input signal amplitude and clipping threshold and how effects of "gain" are really defined by those parameters. Gain alone is nearly meaningless parameter.

I see. I need to re-read that several times though in order to understand what you're saying before I can respond properly to you to avoid any confusions.

So, a typical distortion pedal have a power supply of 9 v. An input voltage of 0.0000001 v, according to the online calculator that I'm using, will generate a decibel gain measuring at 159 dB by using a typical distortion pedal. That is very high, but can a guitar produce that input voltage that low just like a guitar need to generate an input voltage of 0.00044457 v with the power supply at 250 V to produce a decibel gain of 115 dB through the highest gain channel of Marshall JVM410H?

Because of these low input voltages, eventhough the decibel readings are high which is 159 dB for the typical distortion pedal with a power supply of 9 V and 115 dB for the maxed out Marshall JVM410H with the power supply of 250 V, these high dB figures actually refer to undistorted soundwaves. Correct?

[teemuk]

The figures are pretty much theoretical in a sense.

Voltage amplification ratio / factor, or "gain", merely indicates how much the circuit tries to amplify the signal. E.g. 10x voltage gain tries to amplify 1V input to 10V output and so on.

In practice, within specific margin the circuit works as intented, but overdriven above the margin the results are different and operation is confined to limits of e.g. how high the output voltage can actually rise. For example, with gain of 10x and output voltage limit of 10 volts we can't get 20 volt output with a 2V input. On the other hand, we no longer even get an undistorted replication of the input signal either.

So yes, we can indeed plot a gain curve as a function of input amplitude. This is where we get to discussions about waveform compression, distortion and things like soft, hard, symmetric and asymmetric clipping, but we also dive into the deep end where one must first understand what "gain" conceptually is and what are the pitfalls of using just a single nominal value to indicate "gain".

[thinkingchicken]

The figures are pretty much theoretical in a sense.

Voltage amplification ratio / factor, or "gain", merely indicates how much the circuit tries to amplify the signal. E.g. 10x voltage gain tries to amplify 1V input to 10V output and so on.

In practice, within specific margin the circuit works as intented, but overdriven above the margin the results are different and operation is confined to limits of e.g. how high the output voltage can actually rise. For example, with gain of 10x and output voltage limit of 10 volts we can't get 20 volt output with a 2V input. On the other hand, we no longer even get an undistorted replication of the input signal either.

So yes, we can indeed plot a gain curve as a function of input amplitude. This is where we get to discussions about waveform compression, distortion and things like soft, hard, symmetric and asymmetric clipping, but we also dive into the deep end where one must first understand what "gain" conceptually is and what are the pitfalls of using just a single nominal value to indicate "gain".

Ok so let's look if I get what you're saying. So, the 115 dB figure from the maxed out Marshall JVM410H on the highest gain channel which is 0.00044457 V as the input voltage with a power supply of 250 V, and the previously stated 159 dB generated from a typical distortion pedal with the input voltage at 0.0000001 V with a power supply of 9 V are nothing more than an undistorted sine wave where we don't even started to play the guitar. But the moment we play the guitar, the resulting waveforms will become less gainier in terms of dB gain and instead become clipped (distorted). Also, power supply is like the limit where the sound waves will remain undistorted as long as the soundwaves do not hit the ceiling or limit which is the power supply. Correct me if I'm wrong.

[teemuk]

Forget for a moment what the waveform looks like. Those are theoretical figures. That much the circuit tries to amplify the signal. By that factor / ratio.

If the input signal amplitude is low enough then it can be amplified that much before running into limitations how high output amplitude the circuit can achieve, i.e. "clipping threshold". If that output amplitude limitation is high enough then the signal can be amplified that much without running into limitation.

You see, until we fit input signal amplitude and clipping threshold to equation "gain" indicates nothing except but how much the circuit tries to amplify the input signal. "Gain" alone doesn't even indicate if the circuit can do it or not. Gain is just amplification factor. If we state nothing but gain then amount of clipping is undefined because we don't know at which level the circuit clips, not to mention, we don't know to which level it tries to amplify its output because we don't even know what the input is.

It's like a math test conundrum,

Voutput = G * Vinput, where G is gain, Vinput is input amplitude and Voutput is output amplitude
If Voutput > x, x being the clipping threshold then circuit distorts by clipping and Voutput = x.

G=10. Now define Voutput and maximum voltage output equal to clipping threshold x.

Oh, you can't because you only know value of G and this alone does not help you to solve the equation. Knowing G you can calculate Voutput, but only if you know Vinput, and yet that still doesn't help you to solve x.

So you see, don't obsess about gain; focus on those other parameters too. Analyse those circuits. Find out how they behave, focusing on ALL those variables because they interact. What are expected input voltages at various input points of gain stages? What are their clipping thresholds? It's not as simple as just "gain". That's just a theoretical value what the circuit can achieve if all planets align correctly.

[thinkingchicken]

Forget for a moment what the waveform looks like. Those are theoretical figures. That much the circuit tries to amplify the signal. By that factor / ratio.

If the input signal amplitude is low enough then it can be amplified that much before running into limitations how high output amplitude the circuit can achieve, i.e. "clipping threshold". If that output amplitude limitation is high enough then the signal can be amplified that much without running into limitation.

You see, until we fit input signal amplitude and clipping threshold to equation "gain" indicates nothing except but how much the circuit tries to amplify the input signal. "Gain" alone doesn't even indicate if the circuit can do it or not. Gain is just amplification factor. If we state nothing but gain then amount of clipping is undefined because we don't know at which level the circuit clips, not to mention, we don't know to which level it tries to amplify its output because we don't even know what the input is.

It's like a math test conundrum,

Voutput = G * Vinput, where G is gain, Vinput is input amplitude and Voutput is output amplitude
If Voutput > x, x being the clipping threshold then circuit distorts by clipping and Voutput = x.

G=10. Now define Voutput and maximum voltage output equal to clipping threshold x.

Oh, you can't because you only know value of G and this alone does not help you to solve the equation. Knowing G you can calculate Voutput, but only if you know Vinput, and yet that still doesn't help you to solve x.

So you see, don't obsess about gain; focus on those other parameters too. Analyse those circuits. Find out how they behave, focusing on ALL those variables because they interact. What are expected input voltages at various input points of gain stages? What are their clipping thresholds? It's not as simple as just "gain". That's just a theoretical value what the circuit can achieve if all planets align correctly.

I see. So it is not just gain, but we should know about the input and output voltages and the clipping thresholds. But I'm curious what is your take on my previous statements in which I stated both the 115 dB figure from the maxed out Marshall JVM410H and 159 dB figure from a typical distortion pedal with an input voltage of 0.0000001 V and 9 V power supply are nothing more than undistorted sine waves when someone don't play the guitar. Is that correct?

[teemuk]

I'm not going to waste my time analysing those circuits to find out whether that's true or not. Besides, like I said, how they operate goes way beyond knowing just the nominal gain figures. To find out about inner workings of these effects and amps I'm way more interested in frequency responses, harmonics, waveshapes and clipping thresholds plus dynamic effects at specific circuit sections rather than if the nominal gain figure in total is 105 dB or perhaps 110 dB. In the great picture past a certain point all that has lot more importance than how much "gain" - nominally so - a circuit will have when you dial all its controls to ten, which no one will ever do in practice.

No. I'm convinced that both JVM and generic high gain distortion pedals will have plenty enough of gain for all practical purposes they were designed for. You are the one obsessing about these nominal gain figures so maybe you should be the one also who does the "homework" with these designs. You can then inform rest of us about your discoveries so maybe we can learn something in turn.

--

If the interest is merely about what happens when you amplify the signal with astronomical gain figures then there's the easy solution of cascading a bunch of high gain pedals (preferably to input of amp's distortion channel) and cranking all the gain and volume controls. If one pedal has gain of, say, 100x then two cascaded will give you 10000x and three 1000000x. See, it's easy to get plenty of gain because gain ratios multiply. How much do you need?

[thinkingchicken]

I'm not going to waste my time analysing those circuits to find out whether that's true or not. Besides, like I said, how they operate goes way beyond knowing just the nominal gain figures. To find out about inner workings of these effects and amps I'm way more interested in frequency responses, harmonics, waveshapes and clipping thresholds plus dynamic effects at specific circuit sections rather than if the nominal gain figure in total is 105 dB or perhaps 110 dB. In the great picture past a certain point all that has lot more importance than how much "gain" - nominally so - a circuit will have when you dial all its controls to ten, which no one will ever do in practice.

So, if these high dB gain figures don't refer to an undistorted sine wave do they refer to nothing? I don't understand.

No. I'm convinced that both JVM and generic high gain distortion pedals will have plenty enough of gain for all practical purposes they were designed for. You are the one obsessing about these nominal gain figures so maybe you should be the one also who does the "homework" with these designs. You can then inform rest of us about your discoveries so maybe we can learn something in turn.

I thought they have different level of dB gain, with the JVM being 115 dB (input voltage is 0.00044457 V and the power supply/output voltage is 250 V) with the gain at max on the highest gain channel and a typical distortion pedal being only 36 dB if the input voltage is 0.141 V and the power supply/output voltage is 9 V.

If the interest is merely about what happens when you amplify the signal with astronomical gain figures then there's the easy solution of cascading a bunch of high gain pedals (preferably to input of amp's distortion channel) and cranking all the gain and volume controls. If one pedal has gain of, say, 100x then two cascaded will give you 10000x and three 1000000x. See, it's easy to get plenty of gain because gain ratios multiply. How much do you need?

Remember when HotBluePlates said this previously in our convos,

"The tube amps have "more dB gain" only because they have a power supply of 200-400 volts, and get deliver a larger output signal approaching 100v peak."

Later he said, "No, not even close. I already explained here, "36.1 dB gain is max possible," with this pedal powered by a 9v battery.

But you've waved off internalizing the technical reason why. That 9v battery means you could string 1,000,000 Metal Zones in series and they can still only deliver "36.1dB Gain."

But it doesn't matter, because the Metal Zone gets the distortion from clipping diodes, not raw gain."

What prevent these multiple Metal Zones from increasing their decibel gain? Is it the power supply which is at 9 V or the clipping diodes?