What is UL / CSA Actually Testing?
Let's first define what UL and CSA stand for.
UL (Underwriters Laboratories, Inc) - is an independent organization that develops product safety standards with a diverse group of individuals representing government, consumer, manufacturer, electrical inspection, supply chain, and general interests. UL provides national safety certification of products to these standards and global certification using country specific standards following a successful evaluation consisting of product construction and testing.
CSA (Canadian Standards Association) - is an independent certification organization that serves business, industry, government and consumers in Canada and the global marketplace. In addition to certification testing, they also develop standards that enhance public safety.
When consumer A/V gear gets certified by either UL or CSA, it is done so in accordance with the U.S. or Canadian national standard for safety, UL 60065 or CAN/CSA C22.2 No. 60065, respectively. These Standards are harmonized with the International Standard for Safety, IEC 60065. I decided to phone UL to ask them exactly what is being tested on A/V receivers and amplifiers per these standards and how it relates to power and heat dissipation.
The short story is the product (or as they refer to it: the A/V apparatus) is evaluated (including testing) to reduce the risk of user injury and ensure protection of property as a result of:
electrical shock hazard;
hazardous radiation emission including laser, x-radiation, UV;
mechanical hazards – including implosion of CRTs, apparatus instability;
heating under abnormal operating conditions and simulated single fault conditions;
heating under normal operation
The last bullet about heating under normal operation is the crux of this article.
How Does UL Measure Power?
The UL engineer I spoke with told me that the subject of heating under normal operating conditions continues to be discussed as the testing methods have essentially remained unchanged since the time of one and two channel, Class A and AB amplifiers with linear power supplies. UL, CSA and International A/V Standards have established 1/8 unclipped output power, all channels driven into rated load as representative of normal operating conditions. In the old days they used to simply feed a 1 kHz sine wave tone into one of the amplifiers channels when connected to a fixed impedance test load and Oscilloscope across the speaker terminals and ramp the gain up until clipping was visually observed. The resolution of an Oscilloscope is about 20dB or 1% THD.
As best as I can tell, UL/CSA test 4-ohm loads as specified to be used with the product by the manufacturer. There is a test process by which they determine how they will load an amplifier to simulate normal operating conditions. We have already discussed the determination of non-clipped audio output power. Determination of non-clipped output power forms part of what they refer to as a Power Input Test. The results of this test serve to provide data from which they can make a determination of the test set-up for their normal temperature test. During their Input Test, UL will operate the amplifier under all output loading conditions as specified in the user manual (i.e. 8-ohms for all channels, or 4-ohms for the front channels, 8-ohms for the other channels) and/or marking on the rear panel. This can include operation of the front main L/R channels while connected to 4 or 8-ohm simulated speaker loads with the impedance selector switch (if any) at its recommended setting. As discussed, a sinusoidal input signal (1000 Hz in the case of a "full frequency range" audio amplifier) is provided to drive the front main L/R channels loaded as previously described. From the voltage measured across the simulated load, the non-clipped output power is calculated. If the configuration of the amplifier is such that the non-clipped power determination of the mains L/R channels cannot represent the remaining channels, this process is repeated. Armed with non-clipped power output data (voltage and resistance), they then calculate 1/8 power and make a set of input power measurements with ACD at 1/8 power into all combinations of simulated speaker loading. The input current and power are monitored under each loading condition. UL's decision for the test configuration/set-up of the heating under normal operation test -- including speaker loading -- is based upon consideration of maximum non-clipped output power, input current, and input wattage. They make every effort to determine (based upon these values) a single test set-up condition that will be a "worst-case" representation. If this is not possible, the normal heating test under normal operating conditions test is performed multiple times, under varying simulated speaker loading conditions. The bottom line is, UL does use a process (Input Test) to make a determination of amplifier power output and simulated speaker loading conditions used to establish normal operating conditions.
Editorial Note on Specified Test Signal of 60065
Back in the "old days" UL used a sine wave signal input for both maximum power determination and normal temperature testing. Today, the 60065 Standard permits the use of either the "standard signal" [band limited pink noise] or a sine wave [1000 Hz or geometric mean frequency] to be used as a signal source for testing. The Standard further states that if testing with the sine wave does not comply with the standard,test results with the standard signal are considered decisive.
So, if a receiver measured 100 watts into 8-ohms unclipped (<1% THD), they would drive all channels at 12.5 watts into 8-ohms continuously while monitoring temperature rise using thermocouples. In addition to monitoring the temperature of internal components, wiring, and mechanical parts, a safety laboratory would evaluate steady state temperature rises of accessible parts to ensure that:
accessible, graspable (e.g., handles, knobs) metallic parts do not rise more than 30 deg C;
accessible, graspable non-metallic parts do not rise more than 50 deg C;
parts of metallic enclosures do not rise more than 40 deg C (except the UL Standard specifies that parts not likely to be touched during intended use, i.e. heatsinks may rise not more than 65 deg C when provided with a hot surface marking);
parts of non-metallic enclosures do not rise more than 60 deg C;
critical electrical insulation temperatures do not exceed prescribed maximums.
Note: UL uses an ambient room temperature of 35 deg C (95 degrees F) unless the equipment is rated for use in tropical climates, i.e. 45 deg C, in which case the permissible temperature rise is decreased by 10 deg C. That’s one hot room!
So what is happening when the impedance switch is flipped to the low setting is that it's bringing the baseline power level down proportionally to what the manufacturer thinks is safe enough to pass UL/CSA heat dissipation tests. Since the manufacturer doesn't specify how much power the product is rated to deliver into 4-ohm loads, it can now tell UL or CSA what power level to test to and de-rate accordingly. Let's take the Yamaha RX-V2700 as example and run some power numbers based on the 1/8th UL/CSA heating tests to see what is happening.
For High Z:
Front Two Channels (4-ohm load): => 1/8*(272) * 2 = 68 watts / (0.5) (estimated efficiency) = 136 watts
Remaining 5 channels (8-ohm load): => 1/8*(144)*5 = 90 watts / (0.5) = 180 watts
Total Power dissipation = 136 watts + 180 watts = 316 watts
For Low Z:
Front Two Channels (4-ohm load): => 1/8*(144) * 2 = 36 watts / (0.5) (estimated efficiency) = 72 watts
Remaining 5 channels (8-ohm load): => 1/8*(78)*5 = 49 watts / (0.5) = 98 watts
Total Power dissipation = 72 watts + 98 watts = 170 watts
Note: the power dissipation calculations only factor in amplifier power consumption so you can expect another 25-50 watts of power dissipation for processing and preamp functions.
As you can see the Yamaha RX-V2700's power consumption for this test is now less than 50% less for the low impedance mode and hence the receiver is dissipating 50% ore less heat during the UL or CSA test certification process.
Editorial Note About Transformers & Power Limiting:
The impedance switch doesn't guarantee 1/2 power as you can see in the variable measurements of the above four receivers we tested. It really depends on the transformer VA rating and what load impedance it was optimized to drive. IF you have a small transformer, and IF the predominant heat was I^2 * R losses in the primary and secondary, then you get a transformer which is less efficient at higher currents which has to be accounted for in the design. IF you have a transformer with not enough steel, and the higher voltages cause the waveform to clip because of the voltage creating saturation in the core (usually it is excess current and high voltage together doing this), the wire size will have to be changed on the transformer to avoid this problem. If the transformer has to work on both 4 and 8-ohms, it will likely work best on a 6-ohm load OR be optimized for one, and the other is an afterthought. It is this action being taken by the manufacturers which are playing "specsmanship" with the numbers instead of making the best engineering compromise and publishing both sets of numbers (power into 8 AND 4-ohm loads) for their customers. Until we as consumers demand more complete information from amplifier manufacturers, they will count on our ignorance and optimize their amplifiers for one load (8-ohms for example) and by virtue of the cost/performance analysis decide that the lower power setting needed to secure UL/CSA certification won't preclude their selling the products if the consumer is never going to ask the question, "and what about power into a 4-ohm load?".
Why Do Some Receivers NOT Have the Impedance Selector Switch?
There are three reasons actually, with one being the most common with respect to consumer A/V gear as we discuss below:
The receiver in question has a large enough power transformer, output devices and heatsink area to properly dissipate heat under full 4-ohm power tests (very uncommon, I can count the number of multi channel A/V receivers on one hand that do and still have fingers left over)
The receiver is NOT rated for 4-ohm loads at all (most common)
The receiver was actually designed for a low impedance load, (like in the Professional Audio arena) and the user is willing to accept a lower available power number into the high impedance load than he is going to see advertised for the minimum impedance it can drive. This is the standard in the Pro Audio business as can be seen in the QSC data table below which was taken from this link.
QSC CX 2-channel Amplifiers
Watts per channel
Model 70 V* 8Ω** 4Ω** 2Ω†
CX302 − 200 325 600
CX502 − 300 500 800
CX702 − 425 700 1200
CX902 440 550 900 1500
CX1102 1000 700 1100 1700
CX302V 250 − − −
CX602V 440 550 − −
1000 700 1100
*1 kHz, 0.05% THD**20 Hz – 20 kHz, 0.05% THD †1 kHz, 1% THD
Note: 70 Volt applications are for commercial distributed systems like malls or supermarkets
Let's now explore the most common reason in consumer A/V regarding the lack of impedance switches on many of the A/V receivers on the market today.
Yamaha RX-Z7 (left); Denon AVR-5308CI (right)
The Yamaha RX-Z7 receiver has an internal impedance selector switch while the Denon AVR-5308CI receiver does not. Can you guess which receiver has a 4-ohm rating for its speaker terminals?
If you guessed the Yamaha, then you guessed correctly. Despite both receivers being similarly sized and having the same power rating, the Yamaha is the only one with the 4-ohm marking on the back panel. If you look closer, there are two caveats to this rating:
the receivers impedance selector switch must be set to the low setting
the 4-ohm rating is only applicable to the two front channels
So What Should I Do?
The wisest thing to do is to be prepared to ask the right
questions when shopping for your AV receiver prior to your visit to the store.
Read reviews from reputable review publications that test and measure amplifier
receiver power into 8 and 4-ohm loads. Know if your speakers, the ones you own, or
the ones you intend to buy for use with this receiver are rated at 8-ohms or 4-ohms, or something
else. If your preferred speakers are 4-ohms, or a highly reactive load, you may well do better
off with a separate power amplifier with a well-built hefty power supply. If economics prevent you from choosing that
option, find a review that actually measured the speakers you plan to use. As we know from the impedance magnitude graphs previously shown here,
using a simple number to identify them is not a reliable method, and in some
cases, nearly useless. Even when the
speaker is rated at 4 or 8-ohms, without an
actual measurement we have little idea of how it can or will behave with any
given amplifier or receiver. A load
which changes rapidly with frequency is also a difficult load to drive. Having a large phase angle means the
amplifier has to deliver current before or after the voltage changes. That is a difficult thing to do with any
amplifier, much less one with a modestly designed power supply.
Alternatively, you can stay uninformed, throw in the towel and switch back to a portable MP3 player listening on earbuds. You can even blindly use the low setting and have a false sense of security that you are doing the right thing. As Tommy Boy used to say "you can get a good look at a T-bone if you stick your head ups a cows..., but I'd just take a butchers word on it." Frankly if you can't trust your butcher, then you better get ready to grease your ears and head.
In this case however, I'd advise a healthy dose of common sense and ignore that switch all together when using 8-ohm speakers and think hard about your listening habits and how well ventilated your receiver is before flipping to the low setting. Remember the low setting may severely power limit your system which can degrade sound quality and potentially damage your speakers.
UL's Official Position
UL’s position would be that the manufacturer’s guidelines/recommendations should be followed: use the high setting when using 8-ohm (nominal) loudspeakers and the lower switch setting when using lower impedance rated loudspeakers. The A/V test standards also require operation of the amplifier under faultl conditions. This includes driving the amplifier to any output power from zero to maximum attainable to the rated output impedance or to the most unfavorable load impedance – including short-circuit and open circuit. Higher temperature rises are permitted under this abnormal operating condition. Under fault conditions, the device must not reach a temperature that would present a danger of fire to the surroundings of the apparatus or otherwise create a condition where safety is impaired by abnormal heat developed in the apparatus.
The Audioholics Position
To play devil's advocate here, it's important to recognize that consumers don't operate their A/V receivers under constant load conditions using a signal generator. Music is dynamic in nature and hence the amplifier is rarely put under continuous full load conditions. Remember, we are not like musicians on tour who run their gear into continuous clipping. The typical audiophile is listening to his music at or near 1 watt average power but requiring plenty of reserves to ensure unclipped dynamics.
Hence why I suggest more practical considerations as bulleted below.
Ensure your A/V receiver has plenty of open ventilation (follow the manufacturers guidelines here)
Select more efficient speakers (>88dB per watt @ 1meter or more if they are 4-ohms or less)
Use the receivers bass management facilities and engage the LFE/Subwoofer channel
- Use external amplification for at least
the front 3 channels when more power is needed or when you're driving low
efficiency 4-ohm speakers
Real World Usage Considerations
As a professional reviewer and President of this website, I am constantly evaluating amplifiers and A/V receivers in my 6,000ft^3 theater room using my 4-ohm reference speakers playing at very high levels (>100dB at the listening area) using the high impedance setting and have never run across a situation that warranted me to flip to low impedance mode. In fact, when I've tried the low impedance mode, I noted significantly more audible distortion as the amplifiers prematurely entered voltage clipping due to the power limiting. Remember the impedance rating is NOT an average, but rather a warning of about how low the speaker's impedance can drop over the whole audio range.
In long term testing of receivers such as the Yamaha RX-V2700 and RX-Z7, I've run both products in a 7.1 speaker configuration where at least 3 of the 7 speakers were 4-ohms rated and never ran into reliability or thermal issues at extreme continuous listening levels utilizing the high impedance setting. My center channel is 4-ohms rated, but I do apply the receiver's bass management at 80Hz to limit the bass output for that channel and redirect it to the dedicated subwoofer/LFE channel.
Use External Amplification
If you like to listen to music at loud levels and are using a speaker with a tough load to drive and find that your A/V receiver is overheating and shutting down, then you probably should purchase external amplification to compliment at least the front channels. Today's modern A/V receiver is loaded with overload and short circuit protection. Chances are if you're over heating the receiver, it will shut down far before any real damage can be done. That said, who wants their musical experience to be shut down during a crescendo?
Amplifier power is cheap these days with very competently designed amplifiers from the likes of Emotiva for as little as $299. External amplification for the front three channels can really increase system dynamic range while the unused front channels can be conveniently redirected for a second zone.
Use Bass Management
With most modern A/V systems, the users are advised to bass manage all of their speakers sending power hungry bass signals to the dedicated powered subwoofer channel. This takes considerable stress of the power supply of your A/V receiver, allowing it to breathe more freely while also providing more uniform bass across all the listening seats. Think more dynamic range, better bass, less heat. Any questions?.
Budget A/V receivers represent a compromise of features vs. performance. They are typically optimized for 8-ohm loads because of limitations in the power supply and/or the output devices. Unless you are considering one of the mid to high end spectrum of their product offerings, chances are you're not getting an amp section that will adequately drive 4-ohm loads in either impedance setting.
Don't think that flipping the impedance switch in your receiver will be your miracle solution, because in this case it certainly is NOT. What this really is in most of the receivers I have tested is an inexpensive way to claim the receiver is 4-ohm compatible without the manufacturer having to actually spend money on better internal parts in the power supply. In virtually all of the receivers we've tested with impedance switches, flipping the switch to the low setting would put your speakers at risk of receiving dangerously clipped AC signals when attempting high output levels. While it may operate more safely for a steady-state UL heat dissipation test, it's important to note the real world tradeoffs once this switch is engaged in its low setting.
Know the facts when purchasing and setting up home theater equipment. An informed consumer makes smarter purchases. Most importantly, step back and evaluate how you are using your system and make adjustments accordingly and NEVER eat at restaurants that charge you full price for 1/2 the meal! Let the boss pay for those.
However, the points that I would like to make plain, and that I believe that Gene obfuscated, are as follows:
1. This may well cause distortion to increase, but ordinarily not unless you are playing so loudly that the amp is caused to run much hotter than it ordinarily runs.
2. The effect of the switch is not necessarily to lower the supply rails. In the case of my Onkyo receiver, it is manifest that the primary effect is to reduce the idle current, i.e., the operating mode of the amplifier is shifted further toward class B, where idle current is reduced and the amp runs cooler. This again may well cause an increase in distortion, but it is not necessarily the case that you will be able to hear the difference, whereas it is necessarily the case that you will use less electricity.
3. Even in the case where the lower impedance setting causes your amp to clip at a point in loudness where otherwise it would not clip, this is not likely to cause any damage to your speakers. This is a myth that started back in the early days of the battle waged by tube aficionados, who claimed that because clipping in transistor amplifiers produced stronger high-order harmonics as compared to tubes, that clipping in transistor amplifiers would be more likely to damage speakers.
On our beloved Internet, you can find several half-baked attempts to prove that clipping distortion will fry your tweeter. I have read several of them, and they all use erroneous, convoluted reasoning. The question that they end up answering is not the correct question. They do this because they are hell bent on finding an answer different from the one they get when they stick to the correct question.
The correct question, stated succinctly, is this: When sound is played at the same perceived volume level through an amplifier that is clipping and an amplifier that is not clipping, is the RMS power sent to the tweeter greater for the amplifier that is clipping than for the amplifier that is not clipping?
I have yet to see one of these so-called proofs stick to this correct question. In all of the analyses that I have seen, that purport to prove that clipping damages tweeters, it is virtually impossible to infer exactly what question they eventually did manage to answer.
Clipping does not increase the total RMS power that the amplifier delivers to the speaker. As such, it is manifest that the only way that it is possible for clipping to increase the power delivered to the tweeter is if the spectral power distribution shifts significantly from lower to higher frequency. For the most part, clipping removes some of the original high-frequency content and replaces it with harmonics of the lower-frequency content. Possibly, there is some shift in spectral content from low to high frequency. However, in the hypothetical case where the spectral power content is ordinarily distributed evenly between the woofer and the tweeter in a two way system, the entire spectral content below the crossover frequency would have to shift above the crossover frequency in order for the RMS power delivered to the tweeter to increase by 3 dB. Granted this is not entirely realistic because the spectral power division between the woofer and the tweeter is not ordinarily 50/50. But the point is that an enormous shift of spectral power would have to occur before the tweeter RMS power would increase by 3 dB, and way, way before that happens, you would hear the difference, not because you would hear the 3 dB increase in high-frequency content, but because when this occurs, the shift in overall tonal quality would be absolutely impossible to miss. It would be practically the same as disconnecting your woofer and turning up the volume by a tad.
This is what the audible effect would have to be in order for the power delivered to the tweeter to increase by a few dB. But even if you were so deaf that you could not hear this happening, this much of a shift in spectral power distribution does not occur. When an amplifier clips, the spectral power distribution takes on a character inherently similar to white noise. All you really need to ask yourself is whether you believe that if you were to play white noise loudly through your amplifier that is guaranteed to never clip, if this would burn out your tweeter. If you think the answer is yes, then you had better pay very close attention to the sort of music you play, because there is plenty of prerecorded commercial music where there are significant periods where the spectral power distribution is shifted more toward high frequency than is the case with white noise.