“some amps run out of current first and some run out of voltage first” get me confused.

[duyu389]

Hello everyone! In nwavguy's blog, I came across this passage:

When a headphone source runs out of power with typical compressed pop music you’ll typically either hear distortion or it simply won’t be loud enough with the volume all the way up. With more dynamic music, like classical or well recorded jazz, just the peaks may be clipped so it might not be as obvious there’s a power shortage. Power is a function of voltage and current. And some amps, with a given headphone, run out of current first and some run out of voltage first.

However, there is no mathematical calculation process to substantiate this statement, leading to my confusion about when distortion occurs and when the sound is simply too low. I would simply think that if the amplifier power is insufficient, the headphone's sound would be relatively low, and I couldn't imagine the possibility of clipping.
I would be very happy and grateful if I could receive guidance from everyone. Thank you.

 

[JIW]

Hello everyone! In nwavguy's blog, I came across this passage:



However, there is no mathematical calculation process to substantiate this statement, leading to my confusion about when distortion occurs and when the sound is simply too low. I would simply think that if the amplifier power is insufficient, the headphone's sound would be relatively low, and I couldn't imagine the possibility of clipping.
I would be very happy and grateful if I could receive guidance from everyone. Thank you.

Electrical power P is the product of voltage V and current I, i.e. P = V x I. By Ohm's law, for a load with impedance Z, the voltage is the product of the impedance and the current, i.e. V = Z x I. Inserting this into the power equation gives that P = Z x I^2. Since by Ohm's law I = V/Z, power can also be calculated as P = V^2/Z. Thus, power can be calculated in several ways involving voltage and current but some require the impedance of the load to be known.

All amplifiers are limited in both the voltage and the current they can provide - some more in one way and others more in the other way. By Ohm's law, a lower impedance load requires more current for the same voltage and thus the amplifier is more likely to be limited by its ability to supply current whereas for a higher impedance load, the amplifier can likely provide enough current but not enough voltage.

Suppose an amplifier can provide at most a voltage of V_max and a current of I_max. By Ohm's law, if the impedance is greater than V_max/I_max, the amplifier's output will be limited by voltage whereas if the impedance is less than V_max/I_max, the amplifier's output will be limited by current. Since different amplifiers have different maximum voltages and currents, one amplifier may be limited by voltage but another may be limited by current for a given load.

 

[Philbo King]

Hello everyone! In nwavguy's blog, I came across this passage:



However, there is no mathematical calculation process to substantiate this statement, leading to my confusion about when distortion occurs and when the sound is simply too low. I would simply think that if the amplifier power is insufficient, the headphone's sound would be relatively low, and I couldn't imagine the possibility of clipping.
I would be very happy and grateful if I could receive guidance from everyone. Thank you.

It strikes me as nonsense. Voltage is what causes current flow.
If the supply voltage drops the current drops. If the current demand exceeds the supply capability, the voltage sags.

Whoever came up with this is unable to explain clearly what they mean, assuming meaning is present.

 

[AwesomeSauce2015]

To answer your question, distortion occurs through numerous sources. One way it can occur is by not having enough output power to reproduce the input signal (ie, clipping).
Some devices clip in a cleaner manner than others. For example, class A/B (non-switching based amplifiers) and tube amps tend to have nicer clip performance than class D (switching) amps.

To answer your second question, requires diving into a bit of electrical theory, which I will try to explain.
Audio amplifiers are voltage sources with gain. This means that (assuming an infinite load impedance), my output voltage is a multiple of my input voltage.

When I attach a load to an amplifier, the amplifier’s output will have a voltage and a current value. Say I have 1 volt of output and a load with a 10 ohm impedance. My current (given by ohm’s law, V=ir) is 1/10 amps.

The issue you originally were interested in comes when the input signal increases, and the amplifier either hits its voltage or current limit first.

Say my amplifier’s power supply can only provide 10 volts and 10 amps. If my input signal increases, with that same 10 ohm load as before, I will hit my voltage limit first (10=i*10 -> i = 1 A). Therefore, my amplifier will clip due to “running out of” voltage.

However, if I keep the same amplifier, but now have a 0.1 ohm load, then I will hit a current limit first (V=10*0.1, -> V = 1 volt).

Therefore, my output voltage will not be able to increase above 1 volt with this load. This is because my amplifier is not capable of supplying more than 10 amps of current or 10 volts of output voltage. In the first case, I hit my voltage limit, and in this case, I hit my current limit.
This is why the quote you have specifies that it is per headphone and per amplifier dependent.

In both of these cases, clipping would occur as the amplifier is not capable of increasing the output any further. This would lead to distortion.
The original quote is a bit confusing in that it says that you will either get distortion or it won't be loud enough. If it isn't getting loud enough, then you probably have distortion, just at a low enough level to not hear it. -- [Then again, some systems are designed to run with headroom (no clipping distortion) at max output, so you would just have it "not be loud enough".]

Now for a real example:
My desk system has a Dayton KABD-430 (running in 2 channel mode) driving a pair of KEF Q100s. It has a 14 volt, 16 amp power supply hooked up to it. KEF says that the Q100 is an 8 ohm speaker, so we will use that for the calculations (Impedance is frequency dependent...). Using Ohm's Law, 14 = I*8, We get I = 1.75 A. Multiply that by 2 channels and you get 3.5 amps of current draw. In this case, my output power is limited by my power supply's voltage. However, this isn't an issue as my system gets plenty loud as is.

If you are still interested, looking into "Electrical Engineering (intro) Circuit Theory" may be useful.

 

[dlaloum]

A different question... what are the audibly differences with regards to running out of current vs running out of voltage?

 

[AwesomeSauce2015]

A different question... what are the audibly differences with regards to running out of current vs running out of voltage?

"It depends"...
There really isn't a simple way to put it, because it's not simple.
In some cases, having limited current capability can change the sound of a system (See Truthear's Zero IEMs, due to their "subwoofer", they are sensitive to DAC / Amp output impedance, ie: DAC/Amp current capability).
In voltage cases, usually it will round off the peaks, which introduces harmonics into the signal. This can be audible, or it can not be. It depends on a lot of factors...

 

[dlaloum]

"It depends"...
There really isn't a simple way to put it, because it's not simple.
In some cases, having limited current capability can change the sound of a system (See Truthear's Zero IEMs, due to their "subwoofer", they are sensitive to DAC / Amp output impedance, ie: DAC/Amp current capability).
In voltage cases, usually it will round off the peaks, which introduces harmonics into the signal. This can be audible, or it can not be. It depends on a lot of factors...

My AVR had audible signs that I took to indicate running short on current (or low impedance driven instability?) - mostly because the AVR has ample power for the levels at which I was running it at (circa 4W continuous MAX) - with peaks beyond that obviously... but unlikely to be above 16W - with the AVR rated at 100W@8ohm, I would not expect that a voltage limitation would be reached...
The speakers are 4ohm nominal, down to 3 ohm at woofer x/over, and 1.63ohm (capacitive) on the tweeter...

note headphones, obviously, but same electrical issue/question.

Using pre-outs and external amp capable of handling (rated for) 2ohm speakers resolved the issue.

But the question remains of why it resolved the issue! - at such low levels the AVR should be capable both in current and voltage terms... but audibly that was not the case.

 

[radix]

A different question... what are the audibly differences with regards to running out of current vs running out of voltage?

Running out of current can also cause an amp to shutdown in overload protection. If you are running out of current, you're liking pulling too much power and the amp will go into thermal shutdown. For example, you try to drive 2-ohm speakers on an amp that cannot do it.

 

[dlaloum]

Running out of current can also cause an amp to shutdown in overload protection. If you are running out of current, you're liking pulling too much power and the amp will go into thermal shutdown. For example, you try to drive 2-ohm speakers on an amp that cannot do it.

No shutdown in this case, and on an AVR that has a particularly sensitive "nanny circuit" known to self protect after 30s at 5W into 4ohm...

It sounds OK - just not as good as the system is capable of - and using the alternate amps clear it up.

 

[restorer-john]

It all comes down to the ability of the power supplies to maintain the internal rail voltages in the face of varying current draws, and the output devices in the amplifier stages themselves being able to supply the current without excess voltage drops developing.

There are perfectly valid reasons for having a 'loose' supply with high voltage rails that sag under excess current draws, just as there are good reasons to go the opposite way- have a tight, highly regulated supply (a stiff supply), that can deliver increases in the current as the load demands it, all without losing voltage regulation.

There is no one correct way.

 

[boXem]

Running out of current can also cause an amp to shutdown in overload protection. If you are running out of current, you're liking pulling too much power and the amp will go into thermal shutdown. For example, you try to drive 2-ohm speakers on an amp that cannot do it.

One thing is the short circuit protection (overload), the other is thermal protection (overuse). They may have or not have the same effect from user standpoint.

 

[restorer-john]

A different question... what are the audibly differences with regards to running out of current vs running out of voltage?

Interesting question.

"running out of current" is generally (not always) a situation where the designers have deliberately included current limiting circuitry. That can be anything from benign to quite nasty sounding as the OPT devices are starved of drive due to circuitry detecting excess (to design paramters) current flow in the output devices.

Voltage limiting is simply (in most cases) sagging rails, and that is generally quite benign also, but squashes dynamics and allow a rapid onset of clipping, but almost in a 'soft' way- often accompanied by mains (60/120Hz) hum and noise at high levels modulating the music.

In modern Class Ds, the limiting is a completely different animal. Excess current is a trip situation, resulting in silence. Voltage limiting is perhaps not an issue except with ultra high power amps being run into no load.

 

[IAtaman]

It all comes down to the ability of the power supplies to maintain the internal rail voltages in the face of varying current draws, and the output devices in the amplifier stages themselves being able to supply the current without excess voltage drops developing.

There are perfectly valid reasons for having a 'loose' supply with high voltage rails that sag under excess current draws, just as there are good reasons to go the opposite way- have a tight, highly regulated supply (a stiff supply), that can deliver increases in the current as the load demands it, all without losing voltage regulation.

There is no one correct way.

Point of well regulated supplies is quite straight forward. Loose supply, not so much. Could you please care to elaborate what would be the reasons for wanting a 'loose' supply?

 

[restorer-john]

Point of well regulated supplies is quite straight forward. Loose supply, not so much. Could you please care to elaborate what would be the reasons for wanting a 'loose' supply?

Entire brands have built their amplifier legacy on 'loose' supplies. NAD, Proton, Hitachi with their Class G HMA-8300 (the first), Pioneer in the 80s, Yamaha's 'X' supply and Yamaha pretty much all through the 80s, 90s and well into the present day. Their transformers in Class ABs have always been voltage demons, not current monsters.

It's a perfectly valid design concept. Use a power supply with a higher than what is needed voltage, but a deliberately limited current capability, so it can deliver phenomenal peaks in voltage (and short term current from the caps), but sags under really heavy continuous loads. You get an amplifier that is underrated in real terms, for less money, less weight, runs cooler and with a dynamic capability that belies conservative specs. All things that are desirable.

Take the much lauded Benchmark AHB-2. It has a regulated supply so it has virtually no dynamic headroom over and above its moderate power. When it clips, that's it. No reserve for transients. A 1984 NAD 2200 will output 4 times or more the Benchmark, for the same 100WPC rating- all to do with a switchable 'loose' power supply that deliberately 'collapses' after 300-500mS on the 110V rails and reverts to the low rails.

More than one way to skin a cat as they say.

 

[popej]

However, there is no mathematical calculation process to substantiate this statement, leading to my confusion about when distortion occurs and when the sound is simply too low.

It is a simplification. Whenever you overload amplifier, it will clip and there will be distortion. Depending on music type and amplifier properties, you will hear it at once or not.

As for voltage/current limit, you can think like this: amplifier reach voltage limits, when volume is set too high and expected peak of output signal is higher than internal supply voltage. Reaching current limits means, that some protection circuit has been activated. Like for example a fuse has been blown.

 

[boXem]

a fuse has been blown

In 2024, when a fuse blows, it means that all other mechanisms of current limitation have failed.

 

[Hayabusa]

It strikes me as nonsense. Voltage is what causes current flow.
If the supply voltage drops the current drops. If the current demand exceeds the supply capability, the voltage sags.

Whoever came up with this is unable to explain clearly what they mean, assuming meaning is present.

As explained in earlier post: no nonsense!
I seems you don't understand...
It all depends on the impedance and the signal it self if clipping occurs and if this is clipping(limiting) in current or voltage

 

[IAtaman]

Entire brands have built their amplifier legacy on 'loose' supplies. NAD, Proton, Hitachi with their Class G HMA-8300 (the first), Pioneer in the 80s, Yamaha's 'X' supply and Yamaha pretty much all through the 80s, 90s and well into the present day. Their transformers in Class ABs have always been voltage demons, not current monsters.

It's a perfectly valid design concept. Use a power supply with a higher than what is needed voltage, but a deliberately limited current capability, so it can deliver phenomenal peaks in voltage (and short term current from the caps), but sags under really heavy continuous loads. You get an amplifier that is underrated in real terms, for less money, less weight, runs cooler and with a dynamic capability that belies conservative specs. All things that are desirable.

Take the much lauded Benchmark AHB-2. It has a regulated supply so it has virtually no dynamic headroom over and above its moderate power. When it clips, that's it. No reserve for transients. A 1984 NAD 2200 will output 4 times or more the Benchmark, for the same 100WPC rating- all to do with a switchable 'loose' power supply that deliberately 'collapses' after 300-500mS on the 110V rails and reverts to the low rails.

More than one way to skin a cat as they say.

So instead of a regulated power supply that can deliver say straight 50V up to 4 Amps , you oversize the power supply with extra headroom so that the rails are 100V, and even though they might sag down to 50V supplying 4A continuous, they have the capability to handle transients thanks to extra voltage headroom - did I understand correctly?

That is a nice way of doing it indeed. Also explains the high peak current ratings of NAD amps which did not make a lot of sense before.

Thank you.

 

[popej]

In 2024, when a fuse blows, it means that all other mechanisms of current limitation have failed.

It was an example

Actually my old Rotel stereo amp has a fuse at output. Maybe designers thought it was better solution than current limiter. It has never been blown

 

[Cbdb2]

Fuses increase resistance with current (heat) reducing the output current. This will dynamicaly compress the signal at high levels, unless the fuse is in the feedback loop, which can cause other problems if the fuse blows.