“Let our rigorous testing and reviews be your guidelines to A/V equipment – not marketing slogans”

# Amplifier and Power Supply Basics

By

To gain a better understanding of how amplifiers and their associated power supply work together, we need to do a little math.  Don't worry these are simple equations that even Sarah Palin could comprehend, especially since they can fit in the palm of her hand.

P  =  V * I         (Eq1)

P = V^2 / R       (Eq2)

Let's first discuss amplifier DC voltages to better understand the “V” in Eq1. Amplifiers require a DC (sometimes called rail) voltage to provide a constant power source which the amplifier can in turn use to amplify the input signal.  Assuming the output devices can source (allow to pass through them) the current to the load, (In this case our loudspeakers) the higher the DC rail voltage, the more potential power output power the amplifier can produce.  As long as the input signal increases, if there is available voltage and current from the amplifier, there will be available power to send to the speaker.  If not, and the amplifier attempts to increase the signal beyond the fixed boundaries of the rail voltage you get what is known as amplifier clipping. This also happens if the current demand on the power supply exceeds what is available. That too will cause the DC Voltage rails to sink lower than what is normal and cause amplifier clipping.  These limitations occur when the power supply is asked to supply more power, as either available voltage or current, than it can make available.  If the clipping is caused by drawing too much current,  there will be additional distortion components in the output signal (compared to the voltage clipping scenario) as the variations in the rail voltage will be transferred to the speaker load as AC signal content unrelated to the input signal (i.e, distortion).  Amplifier clipping is bad because it takes the undistorted waveform and essentially clips off the top and bottom of it.  This creates high frequencies not present in the original signal, in a word, distortion.  When done to extremes, the original signal can be converted from a sine into a square wave.  The square wave includes high frequencies not originally present in the signal, which represent a real hazard to the loudspeakers systems high frequency device(s) voice coil(s).  Clipping is therefore especially dangerous to small light and delicate high frequency loudspeaker components in a system as the square wave contains far more high frequency energy than was found in the original music signal.  It can take an unclipped musical signal sending 5% of the energy to the tweeter, and suddenly send 25% of the increased power available to the tweeter.  Clipping, despite what you have read on the blogs, is rarely damaging to woofers by comparison.

Top Left pic (clipped sinewave) ; Top right pic: amplifier power vs distortion

Bottom Pic: FFT Analysis of Amplifier in Clipping

As you can see in the left image above, the middle peak and dip of the sine wave is clipped or flattened once the rail voltage of the amplifier is exceeded.  The pic to the right shows an amplifier power vs distortion measurement I conducted.  As the amplifier starts to visibly clip, it approaches 1% THD. This is clearly an audible amount of distortion that can also be seen with an oscilloscope. (The oscilloscope shows us the voltage versus time). If you do an FFT analysis of the spectrum (showing us the frequency components in the time signal displayed by an oscilloscope), you can see all of the nasty harmonic byproducts (Multiples of the original frequency or harmonics for short) that are added to the original 1kHz fundamental.  This system is producing severe distortion due to clipping.  This not only sounds bad, but it's bad for the speakers too.

According to Yamaha receiver failure rates haven't changed with the inclusion of the impedance selector switch...

If a receiver's power supply has sufficient current supplying capacity, and the amplifiers output impedance is low enough, then it can operate fairly closely to an ideal voltage source doubling available output power as the load impedance halves and the voltage output remains the same. (see Eq2) (Remember, in a good amp the same voltage is present at the loudspeaker terminals for an 8 ohm, 4 ohm, or a 1,000,000 ohm loudspeaker is present, and is dependant on the signal input amplitude and gain of the amplifier., and NOT the load attached) So when a 100 Watt per channel receiver is driving an 8-ohm load, it delivers  28.3Vrms. (28.32/8 = 100) which into a 4-ohm resistive load would produce 200 watts. (23.82/4 = 200)   Most receivers don't operate so ideally however, so while they all will try to deliver more power into a lower impedance, they may first run out of voltage, or worse yet, available current.  This lack of available current will cause the voltage rails to shrink, clipping the signal, and thus generating more heat as losses in the process.

Since transformers and output devices have a finite amount of resistance, those losses will occur as I2*R, amperes squared divided by resistance. Transformer windings and output devices do not have fixed resistances either.  As they are burdened with greater and greater current demands, this heating causes their resistance to increase.  Eventually, the system either reaches a static operating temperature or it simply burns up. While loading an amplifier with a lower impedance loudspeaker load is not universally dangerous it does put an extraordinary load on the power supply components, which can overheat if this demand continues for extended periods of time. More on this later.

Editorial Note About the Cost of Power Transformers
If size and cost and weight were not an issue, manufacturers could supply receivers with their own power station, and then the limiting factor would be the output devices and heat sink.  The cost of steel and copper from China has varied dramatically over the last few years, and manufacturers do not want to see their costs suddenly skyrocket or their supply dry up because they negotiated a low price that causes the vendor to lose money when the cost of materials used skyrockets!  This is perhaps the most expensive part of the entire amplifier! (Except of course for marketing)

## What Does the Impedance Selector Switch Do?

It parts the red sea, stops the BP oil leak and reduces our carbon footprint.  It would be easy to draw these conclusions if you spend enough time reading the misinformed consumer comments in the forums on this very topic.  If this was in fact true Obama would have already been taking credit for such a marvelous invention.

Let's examine some power measurements of receivers I measured in the past that featured impedance selector switches to deduce exactly what they are doing.  The receivers range in price from \$500 all the way up to \$5,500.

Low Impedance (Z) Mode - is the receivers "low" setting that the manufacturer recommends using when you attach loudspeakers rated below 8-ohms. (This mode limits the output voltage, and therefore the maximum current any given speaker can demand of it).

High Impedance (Z) Mode - is the receivers "high" setting that the manufacturer recommends using when you attach loudspeakers rated at 6-ohms or higher.  This is usually the default setting the receiver ships in, and the rating which the manufacturer optimizes his parts for, since they only have to advertise ONE power rating before the consumer will reach for his checkbook or credit card.

PLoss - Power Loss (%) determined by comparing the Low Z and High Z power numbers for each receiver using the following equation: (1 - LowZ / HighZ) * 100

 Manufacturer Model Load High Z Low Z Ploss THD Yamaha RX-Z11 8-ohms 190 watts 190 watts 0% 0.10% 4-ohms 300 watts 300 watts 0% 0.10% Yamaha RX-Z7 8-ohms 170 watts 78 watts 54% 0.10% 4-ohms 255 watts 144 watts 44% 0.10% Yamaha RX-V4600 8-ohms 134 watts 95 watts 29% 0.10% 4-ohms 210 watts 180 watts 14% 0.10% Yamaha RX-V2700 8-ohms 144 watts 78 watts 46% 0.10% 4-ohms 272 watts 144 watts 47% 0.10% Yamaha RX-V661 4-ohms 224 watts 63 watts 60% 1% Onkyo TX-NR5007 8-ohms 191 watts Not tested 1% 4-ohms 171 watts 68 watts 72% 1%

Note: All tests were conducted with one channel driven at 0.1% THD + N using a 1kHz test frequency except the 1% THD tests which were conducted by Sound & Vision Magazine.

As you can see in the tabulated data above, 5 of the 6 receivers exhibited significantly less output power when their switches were set in the "low impedance" mode for both 8 and 4-ohm loads.  In fact the lowest priced Yamaha (RX-V661) and Onkyo (TK-NR5007) models exhibited the most power scaling in the low impedance mode (72% reduction, 60% reduction, respectively)  The RX-V2700 power derating of the low impedance mode was interesting in that it maintained their specified 8-ohm output power rating of 140 watts into a 4-ohm load.  I find it noteworthy that the Yamaha RX-Z11 power output did NOT change based on the impedance switch setting.  It's obvious in this case that the designers of this receiver were confident no derating was needed to pass the regulatory testing.  This isn't surprising giving the sheer size and expense of this engineering marvel, hence why they call it their flagship model!

What the "low impedance" setting accomplished (except for the Yamaha RX-Z11) was to step down the rail voltage fed to the amplifier by the secondary of the power transformer. The unfortunate side effect was clipping at a much lower power level as seen in the tabulated test results.   The low switch setting appears to limit the maximum available current draw on the transformer to about 1/3rd (Onkyo TX-NR5007) as much as the high setting so that it would be able to play continuously (at a significantly reduced power level) during the UL/CSA certification testing while generating significantly less heat.

There is no set derating number that manufacturers use as far as I can tell.  I surmise that they calculate the maximum power their receiver can deliver at a certain distortion level while still maintaining a low enough temperature so that it will  pass UL / CSA certification, deeming it safe to operate into 4-ohm loads.  This value depends greatly on how well the amplifier can dissipate heat which is a function of heatsink area and ventilation and also the VA rating (Volts time Amperes) of the power transformer and how much actual power the amplifier can deliver before failing.  Realize that the power transformer is wrapped up tight and into a ball.  In most cases, manufacturers don't heat sink them but some do add a cooling fan which turns on during high current demands to cool off the power supply.  They do put resettable fuses and heat sensing switches inside to prevent them from burning the enamel off the windings. The "low" setting of the switch is providing some protection to the power transformer from blowing up under a continuous maximum power condition into 4-ohm loads, but so is the overload protection circuitry that causes the receiver to shut down if driven into clipping for more than a few seconds in either impedance mode.  The low impedance setting is also killing amplifier headroom and maximum available power which will send more distorted and clipped signals to your loudspeakers.

most A/V receivers have overdrive protection built in regardless of the impedance switch

### But where are the Failures?

It's equally important to reiterate that most receivers have overdrive protection built in independent of this impedance switch so if the power supply is stressed into exceeding maximum rated power, or full rated power simultaneously into multiple channels, the power gets scaled back significantly or the receiver shuts down.  This can be seen in some models during All Channels Driven (ACD) tests where a limiter kicks in and cuts power down to 1/4 or less rated one channel power.  I've never blown out a receiver during my power torture tests, but I have shut many down when driven hard into 4-ohm loads in both the low and high impedance settings.

In speaking with Yamaha, they told me they could not confirm any failures of receivers directly attributed to running the high impedance mode with 4 ohm speakers.  Their power supply / amplifier failure rate hasn't changed with the inclusion of the impedance selector switch on their A/V receivers. One has to wonder, just how much over-protection does the user need until it starts compromising real world dynamics and headroom while using the product with musical program material and not continuous test tones in a lab environment?  Is it worth potentially damaging your speakers to protect your receiver?  I guess that’s a philosophical question depending on which is more valuable to you.

### Recent Forum Posts:

ntsarb posts on March 22, 2021 20:17
Thank you @PENG. Very interesting and useful information.
PENG posts on March 21, 2021 10:10
ntsarb, post: 1469677, member: 94157
* Why does the manufacturer need to reduce the rail voltage more than required to keep the power rating steady?

I believe they did it on as required basis. The issue is not so much “power”. In my opinion, It was unfortunate that in the beginning, manufacturers started rating their amps by “power” output instead of voltage and current. It is now too late to change.

Loudspeakers are sort of voltage devices, their sensitivities are often rated as X dB/1W/1m but should be more appropriately rated X dB/2.83V/1m. 2.83 V is picked because for the popular 8 ohm loads, it would be equivalent to X dB/1W/1m. The fact is, if you apply a voltage signal to the loudspeaker's terminals, it would make sound in proportion to the applied voltage, not to the “power input” as such. That is, the speaker may only consume only 0.1 watt in one moment but make a very loud sound, yet may consume 0.2 W in another moment and make a much quieter sound.

If an amp is rated 100 W into an 8 ohm resistor, then it can be rated 50 W into a 4 Ohm resistor safely, and that means the output voltage has to be reduced by half in order for the current to be the same. The current would be 4.8 A. (using your 38.47 V, 8 Ohm example).

Now if the manufacturer knows their amp can be rated higher than 4.8 A, then they wouldn't have to lower the rail voltage as much for the 4 Ohm setting, may be lowering it by 30% is enough, just an example. That's why the lower rail voltage would vary depending on the specific amp's rated voltage and current capability. So far so good? And you can see why I said amps should have been more appropriately rated for their voltage and current limits, than just “power” that by itself actually makes little sense? By the way, keep in mind, you really don't know how much “power” your speaker would actually consume, all you know is how much current it draws on moment by moment basis, as a good portion of the so called “power” would be consumed, or dissipated in the amp itself!!

Would dropping the rail voltage from 38.47V (in the following example) to 33.3V significantly increase power loss in the form of heat? If so, does this relate to the transformer's efficiency at different voltage?

No, dropping the rail voltage should result in less current so less loss, not more, all else being equal.
Copper loss = I^2 R for an resistor load. For an inductive load such as many loudspeakers, it gets more complicated as much of the loss would be dissipated in the amp, not just in the speaker. It would depend on the phase angle vs frequency characteristics of the speaker.

Here's a good article for you:
Phase Angle Vs. Transistor Dissipation (sound-au.com)

* In case more heat is produced by use of lower impedance speakers at same wattage, would it not make more sense keeping the heat production steady (i.e as if n 8Ohm speaker was driven) by adjusting the voltage accordingly? Why would the manufacturer drop the rail voltage more than that?

Yes, but you are assuming the manufacturers drop the voltage more than necessary, you don't really know that for sure, as you don't know their products current capability. Again, think current, and phase angle, not “power”.

* I have not fully understood the certification process, hence, my next question is: is the certification process biased towards use of higher impedance speakers, to the extent that it forces manufacturers to degrade the amplifier's actual capabilities (practically leading to the production of less heat with lower impedance speakers than the other way around)?

Not sure I fully understand what you are asking, but this is a complicated issue that I don't think there is a right or wrong answer even if you clarify your question.
ntsarb posts on March 20, 2021 21:08
Very interesting article. Thanks for sharing. A few things I don't understand, so here are my questions:

* Why does the manufacturer need to reduce the rail voltage more than required to keep the power rating steady? Would dropping the rail voltage from 38.47V (in the following example) to 33.3V significantly increase power loss in the form of heat? If so, does this relate to the transformer's efficiency at different voltage?

Example: if the system can supply 185W to an 8Ohm speaker, that means 38.47Volts output. At this voltage, a 6Ohms speaker would require (38.47^2 / 6 =) 246W, but if rail voltage was dropped to 33.3Volts (using the switch), the power rating would remain 185W.

* In case more heat is produced by use of lower impedance speakers at same wattage, would it not make more sense keeping the heat production steady (i.e as if n 8Ohm speaker was driven) by adjusting the voltage accordingly? Why would the manufacturer drop the rail voltage more than that?

* I have not fully understood the certification process, hence, my next question is: is the certification process biased towards use of higher impedance speakers, to the extent that it forces manufacturers to degrade the amplifier's actual capabilities (practically leading to the production of less heat with lower impedance speakers than the other way around)?

Many thanks.
PENG posts on December 31, 2019 16:21
hotrabbitsoup, post: 1359291, member: 90421
Thanks for the replies guys. Multiple secondary windings are common for the power supply transformers used in tube amps and is where I'm getting my inspiration from.

I stated that too, yes it is quite common, but for different purposes.
hotrabbitsoup posts on December 31, 2019 16:11
Thanks for the replies guys. Multiple secondary windings are common for the power supply transformers used in tube amps and is where I'm getting my inspiration from.

I think some cheaper amps, that actually have an impedence switch, implement the power restriction by using a tap on a single secondary winding that supplies a lower voltage to the power supply, but in a more expensive design I don't see why they couldn't use a second coil altogether. But, all that said I don't think it would be worth it, as you said, just make the transformer with the secondary coil that's going to work for all intended purposes in the first place.

But all manufacturers don't have acces to the same. Some shops have the size to invest in custom transformer production runs, some even wind themselves, while many others look to off the shelf parts.

I'll try and find some schematics of older 80s receivers with the switch and report back if I notice anything useful. I am EE too.