amp watts vs gain

[SteinbergerHack]

 

Now for (more) practical terms, we use the term power factor which is the ratio of real vs. apparent power so if for example we are driving a reactive load with a phase angle of say 45 degrees (somewhat typical of a really inductive LF driver), the power factor is I x cos 45 deg = I x .71 This means that the real power is the voltage x the current x .71. So for those of you who are reading between the lines, the whole power thing has a twist... an amp can not ever deliver full rated power into a real world speaker (which is why more speakers do not blow up) due to the reactive part of the load. In this case the amp delivers 71% of rated power but the currents and voltages are as high as if it were trying to deliver 100%.

 

 

The trouble with this is that the current is still at maximum, even though the delivered power is reduced by the power factor. The heating effect in the voice coil is generated by the resistive losses, which are I*I*R. Thus, even if the delivered real power (Watts) is lower, the thermal build-up is still driven by the actual current (Amperes) delivered by the amp. To put it more simply, the phase difference between Volts & Amps reduces output power, but it doesn't reduce thermal heating in the voice coil(s).

[SteinbergerHack]

 

Are you speaking of the back EMF that must be absorbed by the output stage due to the inertial (overshoot) of the cone. Is this called damping factor? I've always wondered how those super small, high powered amps like Powersoft can absorb this (doesn't it have to be dissapated as heat?). Where are the heatsinks? To me it's just FM.

 

 

No - that's a separate issue. P=V*I is only true for a purely resistive load, like a test-bench dummy load. In the real world, most loads are reactive as well as resistive; a speaker voice coil is both a resistor and an inductor. If you look at the published T-S parameters for a driver, you'll see that there is a DC resistance as well as a nominal AC impedance; the difference between the two is the inductive portion of the load.

 

Now, when you drive an inductive load with an AC signal, the current will phase-shift due to the inductance; the current will lag the voltage by some phase angle. Once you are dealing with a reactive load, you have three different "power" measurements: Real Power (Watts), Apparent Power (Volt-Amperes), and Reactive Power (VAr).

 

Real Power is calculated as Preal = V*I*cos(phaseangle)

Reactive Power is calculated as Preac = V*I*sin(phaseangle)

Apparent Power is calculated as Papparent = V*I

 

Real Power is what is actually delivered to the load. Apparent Power is what the source actually has to drive (actual Volts and Amperes). Reactive Power is the "waste" or "imaginary" power that is present in the portion of the wave where V and I are out of phase.

 

[FWIW, it would seem to me that a good designer should be able to absorb the back EMF you mentioned and actually use it to increase efficiency of the amp. We do that in large power systems all the time with capcitor banks.]

[dboomer]

 

Oh, this just brought back some basic physics memories... yikes.

 

 

You bet. I was just stirring the pot to stimulate discussion.

[agedhorse]

 

Are you speaking of the back EMF that must be absorbed by the output stage due to the inertial (overshoot) of the cone. Is this called damping factor? I've always wondered how those super small, high powered amps like Powersoft can absorb this (doesn't it have to be dissapated as heat?). Where are the heatsinks? To me it's just FM.

 

 

Power factor is different, but what you mention is indeed important and is a trick that is used by class D to make them more efficient.

 

What happens is that the output stages are operated in full bridge mode (different term than analog linear amps) and the back EMF is used to drive the opposite half of the bridge by passing through the power supply rather than being dissipated as heat. This is significant.

[agedhorse]

 

The trouble with this is that the current is still at maximum, even though the delivered power is reduced by the power factor. The heating effect in the voice coil is generated by the resistive losses, which are I*I*R. Thus, even if the delivered real power (Watts) is lower, the thermal build-up is still driven by the actual current (Amperes) delivered by the amp. To put it more simply, the phase difference between Volts & Amps reduces output power, but it doesn't reduce thermal heating in the voice coil(s).

 

 

I was specifically addressing the more common failure mode, which would be much worse if the power factor were 1. It's the mechanical failure that's generally responsible for most LF speaker failures. If course, push things far enough and themal also becomes an issue, as does the current desity of the VC wire and connections.

[agedhorse]

 

 

 

This is true and is done with non-linear (switching or class D) amps. Same as with PWM motor controls.

 

For power systems, generally capacitor banks are used for power factor correction, by introducing a leading factor where there is excessive lag, the energy stored locally is released locally. One danger is resonance effects which must be studied to be sure no ill effects are introduced. This is a specialty unto itself.

[SteinbergerHack]

 

For power systems, generally capacitor banks are used for power factor correction, by introducing a leading factor where there is excessive lag, the energy stored locally is released locally. One danger is resonance effects which must be studied to be sure no ill effects are introduced. This is a specialty unto itself.

 

 

Yup. It's what I do for a living (or more specifically, the people who work for me do it - I don't do much direct engineering anymore:cry:).

[agedhorse]

 

Yup. It's what I do for a living (or more specifically, the people who work for me do it - I don't do much direct engineering anymore:cry:).

 

 

I still design some power distribution systems for larger buildings, but not as much as I used to. Just finished one, 2000A, 3 ph, 120/208. Would have been a lot easier if 277/480 was available.

 

With the introduction of non-linear loads (SMPS, solid state flourescent lighting, HID), there are things I must be aware of that I used to not have to worry about. Harmonic current loads are now a significant portion of lighting loads and I no longer share neutrals across branch phases for example.

 

Add non-linear loads to power factor correction capacitors can open up a whole can of worms. I don't do any of this, so I don't know the tricks of analysis, but I do know that it's something that must be paid attention to.

[Dogoth]

Ah ha! I learned two things here (It's why I like this forum).

 

1) A V.C. is of course an inductor so It acts like a filter, thereby automaticly losing some of the power sent to it (dependen on frequency of course - As stated it's a reactive load). OK not losing the power but not turning it all into mechanical energy but some as heat.

 

2) Back EMF developed by the speaker can be stored & reused, so it's not so much of a loss.

 

This still doesn't explain how the small switching amplifiers handle the aformentiones back EMF. Where's the room for the caps?

[Shaster]

 

Agreed ... of course you could do it the way we did in the good old days and insert a pad in the line, keeping the knob wide open and reducing the drive at the same time. I don't know why you think it makes a difference whether the amp is big or small however.

 

 

You're right, it wouldn't matter, simply a case of TMI.

[agedhorse]

 

Ah ha! I learned two things here (It's why I like this forum). 1) A V.C. is of course an inductor so It acts like a filter, thereby automaticly losing some of the power sent to it (dependen on frequency of course - As stated it's a reactive load). OK not losing the power but not turning it all into mechanical energy but some as heat. 2) Back EMF developed by the speaker can be stored & reused, so it's not so much of a loss. This still doesn't explain how the small switching amplifiers handle the aformentiones back EMF. Where's the room for the caps?

 

 

Back EMF goes back into the power supply (like a generator) and in full bridge mode the rail that is being recharged is in fact discharging for the opposite polarity.

 

In a non-bridge output stage, the voltage goes back into the power supply but since there is no load on that swing the supply voltage can increase. This is called power supply pumping and can be a problem. This is why we use full bridge configuration in our amps.

[agedhorse]

Here's something else to consider...

 

Som of the newer amps, especially those w/ DSP input sections, but also the new PLX-04 series, have a lower input overload point. The new PLX 04 has an input overload of +15dBu, and IIRC, turning down the attenuators doesn't change this as it's the diff amp that's overdriving. So, you could easily overdrive the input without any indication of clipping or limiting on the output. You turn down the sensitivity, you still have the +15dBu threshold for clipping the input. Same nasty effect.

 

IMO, the inputs should always be able to take roughly +26dBu, that eliminates this potential gottcha from the list of possibilities.

[Craig Vecchione]

Engineers:rolleyes:

 

 

 

 

[Dogoth]

 

Here's something else to consider... Som of the newer amps, especially those w/ DSP input sections, but also the new PLX-04 series, have a lower input overload point. The new PLX 04 has an input overload of +15dBu, and IIRC, turning down the attenuators doesn't change this as it's the diff amp that's overdriving. So, you could easily overdrive the input without any indication of clipping or limiting on the output. You turn down the sensitivity, you still have the +15dBu threshold for clipping the input. Same nasty effect. IMO, the inputs should always be able to take roughly +26dBu, that eliminates this potential gottcha from the list of possibilities.

 

 

Great point.

 

We just got a couple of Crown XTi's (I haven't usb'd into them yet. The sound contracter who setup the system made them R&R Bass heavy IMO). There is a couple of extra stages before the actual input to the amp section. I'm guessing that the knobs on the front are NOT attenuating the signal to the first DIGITAL CONVERTER stage (that would be the loudspeaker managment section). Thanks I'll keep that in mind.

[dboomer]

 

Som of the newer amps, especially those w/ DSP input sections, but also the new PLX-04 series, have a lower input overload point. The new PLX 04 has an input overload of +15dBu,

 

 

That sucks! Now I'll have to only use mixers that have RCA outputs:lol:

 

Talk about screw up gain structure ... I guess there will soon be a big market for in-line pads.

[The Chinese]

That sucks! Now I'll have to only use mixers that have RCA outputs:lol: Talk about screw up gain structure ... I guess there will soon be a big market for in-line pads.

 

 

Or Amps Kept wide open and limited down w./ Proper Gain structure :wave:

 

hehehe

[agedhorse]

 

Great point. We just got a couple of Crown XTi's (I haven't usb'd into them yet. The sound contracter who setup the system made them R&R Bass heavy IMO). There is a couple of extra stages before the actual input to the amp section. I'm guessing that the knobs on the front are NOT attenuating the signal to the first DIGITAL CONVERTER stage (that would be the loudspeaker managment section). Thanks I'll keep that in mind.

 

 

That is exactly what the problem is. Proper scaling (in the design) is essential to maintaining headroom, which can be at the expense of S/N ration But that's really not a problem compared with the potential for overload. I do not know on the QSC LX-04's, if the clip/limit light tallys the front end overload.

 

I will also check the data I have on the GX series. Looks like it's conventional: +24 dB (16 Vrms) so no problem there.

[agedhorse]

Or Amps Kept wide open and limited down w./ Proper Gain structure :wave: hehehe

 

This is correct.

 

(I do not like this lower input overload issue though, it goes against generally accepted industry practices.)