Fun with Impedance
March 6, 2014 11:34 AM   Subscribe

I have some fancy new stereo speakers. I'd like to be as correct as possible when wiring them up, which has led to some possibly-unusual questions about speaker impedance. Halp! (Inside: Series or parallel?)

Here's my scenario:

I was recently gifted a pair of extremely nice B&W loudspeakers (way nicer than I deserve). They've been babied throughout their two decades of existence and I'd like to continue that trend. Their rated impedance is 4 ohms and they have dual inputs (normally wired together in parallel with copper links).

My receiver is a 70s-vintage unit from "Modular Component Systems" (model 3253). According to the internets, this thing rates 53 watts per channel. I don't quite know what the nominal output impedance is since I don't have a manual, but the back panel says "CAUTION: WHEN TWO SPEAKER SYSTEMS ARE USED, THE IMPEDANCE OF EACH OF THE SPEAKER SYSTEMS SHOULD BE 8 OHM OR MORE." The VU meters also say 8 ohms, so I'm guessing nominal impedance is 8 ohms.

So if I'm correct, I have a 4-ohm load on an 8-ohm output. From what I've read this effectively halves the effective power output of the amplifier. So instead of a theoretical maximum of 0dB/50W on the VU meter, my operational ceiling is now -3dB/25W, as far as I can tell.

I know a thing or two about electronics but I've never had to deal with audio balancing (my previous MO: Hook it up! Does it work? Great!). So here are my questions:
  1. This is the main one: Is is plausible/recommended to connect each of the two inputs on a speaker in series in order to increase the impedance? I'm not sure what the resulting impedance would be. I tried hooking up one speaker like this but it didn't seem to result in a difference in apparent volume (although I'm not sure if it should)...
  2. How much can I believe these VU meters? They seem to work. As long as I keep the needles below the maximum power output I should be okay, right?
  3. Am I getting any of this wrong? As usual info on the internet seems to be either too general or too specific.
  4. As long as I'm asking, what's the actual mechanism for speaker damage (from use)? I've always been unclear on this one. I know that it's possible to burn out tweeters with too much energy, and I suppose it's plausible to physically damage speaker cones with too much amplitude, but how do you know what "too much" is?
posted by neckro23 to Technology (11 answers total)
From the warning on the back panel, it sounds like your amp can deal with a 4 ohm load (two 8-ohm speaker systems in parallel = 4 ohms). Since you only have a single speaker system either 4 or 8 ohms should be fine. Just hook it up, it'll be fine.

Not that you'd notice 3 dB anyway.
posted by ryanrs at 11:51 AM on March 6, 2014

You know that speakers must be in phase for best performance, right?
The wires should be connected symmetrically, so that both speakers are pumping out at the same time, rather than one out and the other in.
You can reverse one speaker's wires and listen to the difference in sound quality if you are not sure.
Speaker damage can result from age, too. 20 year-old speakers may have deteriorated, especially around that outside rubber ring that lets them flex. Take it easy.
posted by weapons-grade pandemonium at 12:01 PM on March 6, 2014

1. No this will not work at all. The two inputs are likely for bi wiring or bi amping, which you are not doing. Bi wiring is a bit of bs anyway. Really, don't try this.
2. On a consumer amp the vu meters are not going to be remotely accurate. Pretty though.
3. You are getting quite a bit of it wrong. The most notable bit is output power, which doubles not halves in your situation.
4. The most common mechanism of speaker damage is actually an underpowered amp driven into distortion, not overpowering the speakers. Stay distortion free and you will be completely fine.

Also, look up some typical impedance curves of speakers, you'll notice that that 4 ohm is very much nominal and the amp will see all kinds of impedances, depending on frequency. I wouldn't worry about this at all.

Enjoy your speakers.
posted by deadwax at 12:02 PM on March 6, 2014 [2 favorites]

Isn't it the amp, not the speakers, that can be damaged by impedance mismatching? An amp rated for a 4 ohm minimum load, trying to drive a 2 ohm load, for example, will still try to put out power against 4 ohms, and burn itself out?
posted by thelonius at 12:18 PM on March 6, 2014

Thelonius, roughly ish yes kind of. With a load that wants more current (lower impedance) the amp will try to deliver that current. With really wild mismatch the amp can run too hot and cook itself, if it has no protection. You are likely to hear that something is amiss and there's a fair chance of damaging the speakers too as the amp may deliver dc to them.

A four ohm load on an amp stickered for eight is not a big deal with solid state amps though, just don't expect to get off scott free if you get drunk and bury the volume knob full up for three hours. You'll hear that before you smell it though.
posted by deadwax at 12:37 PM on March 6, 2014 [2 favorites]

I should add that what happens with a lower impedance is that the amp hits its maximum power out earlier. Where with an eight ohm load it should deliver full power at roughly the end of the volume knobs range on a four ohm load this will be earlier. Where exactly is impossible for us to predict. With this combo that's likely to be at a pretty damn loud listening level though so you will be pretty right.
posted by deadwax at 12:52 PM on March 6, 2014

Response by poster: The most common mechanism of speaker damage is actually an underpowered amp driven into distortion, not overpowering the speakers. Stay distortion free and you will be completely fine.

Well this is exactly what I'm trying to avoid. I'm trying to figure out how hard I can get away with driving these things on this receiver.

What I don't understand is what the real distinction is between a distorted program signal and a distorted output from the amp. It's all just square(ish) waves, isn't it?
posted by neckro23 at 1:04 PM on March 6, 2014

How loud do you want to go? Fifty watts into these speakers would be enough to give me tinnitus in most loungerooms.

Dc on the input side will be filtered out before it gets to the power amp. I'm running a similar setup to you with a theoretically underpowered amp and if I crank it is pretty obvious when the distortion is on the part of the amp. A bit hard to explain but kind of farty undefined bass yuck.
posted by deadwax at 1:29 PM on March 6, 2014

Keep those copper links in place and you will be fine. The two separate inputs on a speaker are likely for bi-amping, i.e. using different amps for the low and high frequencies. Don't ever connect them in series.

Like weapons-grade pandemonium said, the physical condition of the drivers in twenty-year old speakers is a concern. The flexible part at the perimeter of the bass driver is usually the first to go; it's called a surround and if it's deteriorating you will be in the market for a "reconing."
posted by werkzeuger at 2:41 PM on March 6, 2014

To expand on the comments about speaker surrounds, google the model you have. Some have butyl or cloth surrounds that will essentially never deteriorate. A lot of older high quality speakers are in this camp, but there is also a small-but-vocal-minority of certain brands/models that have foam surrounds, which will deteriorate. B&Ws are more than nice enough to pay the ~$50 to have a local hifi shop redo the surrounds with their jigs and years of experience to really do a nice job. I have a set of the original advent loudspeakers, and i had both this done and the crossovers rebuilt. Yea, i put $100 in to what's essentially a $150 set of speakers max... but now i have an amazing sounding fresh set of speakers.

As for the ohms/load issue, going lower impedance simply generates more heat. The warnings on the amp are likely saying, in a coded way, "you can overheat this amp by having a low ohm load". This is the reason that commercial power amps have huge heatsinks and fans, and that high quality pre/power or large integrated hifi amps just have HUGE heatsinks and weigh 50lbs+.

In reality you will have to be listening to some serious dubstep/house type music with it cranked, or at reference levels which is ear bleedingly loud to cause problems. I destroyed a fairly nice amp this way, but i got it for almost nothing from a thrift store and used to run everything from subwoofers with my PA system to 4 large speakers in my office(and one had a blown crossover) which is in a warehouse... and i had it turned up so i could hear it loudly at the other end of the warehouse while i pulled cable.

My main amp used to be a 60 watt harmon/kardon. Yea, that's probably nicer and likely has more headroom than what you have... but i had it connected to fairly sensitive 8 ohm speakers(both the advents, and some infinitys) and i rarely turned it up beyond like "3" on the knob, and NEVER above 6. Same story with the weaker 60 watt kenwood i had before*. "program" volume, as they rate PA speakers, will probably be around 25 watts for you anyways if you're rocking out.

So yea, basically i think the condition of the speakers is a lot more of a concern than the load on the amp. You're probably never going to push the amp anyways. Even on my new digital amp i have mixed feelings about, i rarely get more than 20 or 30 watts into it's "70 watts". You want headroom anyways though, the really buttery crisp deep bass and really smooth, liquid highs come from the amp cruising midway through it's power and being able to easily handle spikes.

The one bit of advice i'd give you honestly, is go buy a nicer amp. Spend $100 or less on craigslist on say a nakamichi stasis amp from the 80s, 70s harmon kardon, 70s sansui, maybe even a marantz from that era if you can find a deal. I was BLOWN AWAY by the difference when i connected my nice speakers to a nice amp for the first time(And a nice source! get a USB dac and some lossless audio files, or a decent turntable, or a quality cd player from the 80s when they cared for cheap). I seriously thought i was high or something how different the music sounded on tracks i already knew well. That MCS amp is likely a bit nicer than most of the stuff you'd get at frys nowadays, but it's like the cheapest decent bike you can get at REI compared to the really really nice stuff from the 70s and 80s that's practically being given away out there.

*How was it weaker with the same watt rating? well back in ye olde days, amps used to be rated one of two ways. The bullshit way they are now where the power is "this is the point at which beyond this, it's totally distorted or catches on fire" and sometimes even "This is the max it could ever put out even with tons of distortion and probably wrecking your speakers". No, amps used to be rated like "This is 60 watts with the minimum distortion in the absolute worst case scenario". Some even used to be rated for actually 100 at that, and they just said 60 so that even if you gave it more than you should you'd still have a good experience. Your MCS amp likely falls somewhere in the middle of this, but towards the "This is definitely 60 real watts" end, not so much the "this is totally 100-110 no problem".
posted by emptythought at 5:14 PM on March 6, 2014

So if I'm correct, I have a 4-ohm load on an 8-ohm output. From what I've read this effectively halves the effective power output of the amplifier

This is not correct, and you've been reading the wrong things.

Let's work through a DC example to show why not (the outcome for AC signals and impedances is the same, but the DC maths lets us ignore frequency-dependent effects for the time being).

Ohm's Law says that the current through a load in amps equals the voltage across that load in volts divided by the load's resistance in ohms.

Power dissipated by a load, in watts, is the voltage across the load in volts multiplied by the current through it in amps.

Combining those two formulae: power dissipated by a load, in watts, is the square of the voltage across it in volts, divided by load resistance in ohms.

An amplifier delivering 50W to an 8 ohm load does that by putting √(50W × 8Ω) = 20V across the output terminals.

If you connect a 4 ohm load to that same 20V output, the load will dissipate (20V)2 ÷ 4Ω = 100W.

Now, amplifiers are not perfect voltage sources: they have internal resistances as well, and as a very rough rule of thumb you can estimate the effect of those on a Class B amplifier as meaning that the amplifier itself will also be dissipating about as much power internally as it manages to push out to the speakers. So if the amplifier's cooling system is designed to deal with the effect of supplying at most 50W to the speakers, and you connect speakers that induce it to supply 100W instead, it's going to run hotter than the designers intended and possibly suffer damage as a result.

But as long as you don't make it do that, it will be happy. So if you've got a front panel meter that shows some ballpark approximation of output power, and you're running speakers that have half the nominal impedance your amplifier is built for, you should mentally double what the meter says to get an idea of how hard your amplifier is actually working.

What I don't understand is what the real distinction is between a distorted program signal and a distorted output from the amp. It's all just square(ish) waves, isn't it?

Yes. And if you run a mostly square-wave signal to a nice set of B&W speakers at high volume, regardless of whether that's because your amplifier is being driven into clipping or because you just enjoy listening to square waves, you do indeed risk damaging them.

Also, I once found out the costly way that a nice amp driving nice speakers can actually be running way louder than you think it is, simply because one of the cues we all use to judge subjective loudness is how much distortion is present.

If you turn a rig capable of delivering very loud, very clean signals up to the point where you're feeling the bass in your chest but your ears are still demanding that extra push over the cliff you can only get by turning it up to eleven, you're probably about to cook your speakers.
posted by flabdablet at 6:43 PM on March 6, 2014

« Older Business Taxes--How do they work in smaller towns?   |   A different sort of podcast question Newer »
This thread is closed to new comments.