On April 10, 2014 at 13:05, TimmyS said...
An impedance meter won't help unless it plots the impedance curve of the speaker, simple and complex. Your VOM wont help much because it only measures at one frequency with negligible voltage/current probably at 1k.

Ernie says
Let's just be ridiculously clear: your ohmmeter checks a ZERO Hertz, which is to say, DC. There is ZERO signal at any other frequency.

Timmy Says:

Oh I must have been thinking of my B&K LCR meter that uses 1K for resistance and 120hz for inductance measurements,

For speaker impedance measurements I use my trusty MLSSA system

Tim Asked:

Are they "difficult to drive? It is possible that the "6 ohm" nominal speaker (marketing department spec) really is a 4 ohm speaker over most of the band?

Ernie said

That's usually an engineering department spec and in fact its correctness may be controlled by FTC regulations.

Tim Says

Not that I know of. Anybody's guess on how a particular MFG rates a speaker 8 ohms, 4 ohms etc, usually with a dart board or average in the favor of the marketing dept. A speaker does not have one impedance, it is frequency and phase dependent.

Im talking about in the late 70's early to mid 80's (probabably before you were born?) 8>)


Most of the Sansui's JVC's Technics Luxmans even Sony & Pioneers.... Plus they all had discrete outputs instaed of IC outputs.

Now that all of the receivers have gone digital, they don't play well with 4 ohm or lower impedences.

Most multimeters, even the cheap ones can check impedences of speakers. Make sure no signal of course, and you should be able to get a "nominal" impedence reading + or minus 10 to 15% on the average.

Incorrect. All speakers have a NOMINAL impedence rating. Speakers impedences will change due to amount of voice coil in the gap, and the frequencies they are producing.

Speakers measured in the resting phase with no signal however will give you the reading you are looking for.

I've been using Yamaha recently and no issues yet...

Yamaha is solid. Good value too!

A receiver, of course. But when you write about a receiver, a speaker, a resistor, then say "it" without instead of saying "the resistor" or "the speaker," it's hard to tell what you mean.


Okay, but an actual 8 ohm load will make an amp run hotter than an 8 ohm load with a 2 ohm load in series with it, so the amp running cooler is no guarantee that the speaker presents a bad load.


Let's take a moment to remind ourselves that "rocket surgery" is a phrase intended to insult the person who says it, because there is no such thing. Maybe we could refrain from using that phrase?


Maybe I don't have it because I understand the concepts involved, even to the extent of having run hundreds of impedance curves myself using Bruel & Kjaer equipment, and helping design crossovers.


How the heck does it perform a rub test? Or tell you that a speaker is buzzing (which rubbing can cause)?


Hell, I've seen 'em come off an assembly line with the midrange or tweeter wired incorrectly because a systematic error crept into the crossover assembly area! However, that kind of testing or error is irrelevant to this discussion as it doesn't affect the load the amp sees.

Uh, I started selling at a Pioneer & Sony dealer that carried a bit of some other brands and saw ALL of the popular brands, including the stuff from Playback, Team Electronics, Lafayette, Pacific Stereo, Shaack Electronics and everything Consumer Reports heaped praise on, but sucked. Chip sets CAN work fine- they just need to be cooled well and not rode hard, put away wet. Stacking a bunch of stuff on top of the receiver or amp really made a difference to our bottom line because we did service, too. The STK-0050 that was in Pioneer's SX-780 worked very well and since they really meant their receivers to be played through their speakers, 8 Ohms was fine. It's when Polk and a few other brands started making 4 Ohm speakers that we found out which brands of receivers had balls and which didn't. Most didn't, even though several magazines raved about their sound.

If you look inside most of the brands we carry and still complain about, e.g., Denon, Integra, Onkyo, Pioneer, Sony, etc, they have discreet output transistors, not IC amplifiers.

No, cheap multimeters CAN'T check impedance. They check DC resistance. Period. Set one for Resistance and another for DC Voltage, then see what happens when you connect the leads form one to the other- you're going to see DC, not AC, and you can't measure Impedance using DC because by definition, Impedance is frequency-dependent.

Re: "Now that all of the receivers have gone digital, they don't play well with 4 ohm or lower impedences."

What are you talking about? Digital what? Wall wart power supplies can't handle much load but a properly designed digital power supply (PWM) can handle whatever you throw at it. Are you referring to "digital amps"? Again, if it's well-designed, it will handle whatever you throw at it. The DSP circuitry never sees the effects from roller-coaster impedance curves and the real problem with amplifiers not liking "difficult" (and I don't mean low) impedance is the inductive and capacitive reactance. There's a saying- "an amplifier is only a few component values from being an oscillator" and this is where the impedance curve comes into play- if the amp starts to oscillate, is biased too hot, tries to develop high power into a low impedance load or doesn't like seeing back EMF, the protection is going to do its job. Often, the problems with these amplifiers has to do with some of the regulators, but the AVR-991 I sent in had several channels with bad outputs, which I think was due to one of their kids disconnecting speakers and trying to reconnect them when the receiver was still ON. Not really a design or manufacturing problem, right?

The problem is that these aren't designed to pass significant current, the heat sinks are inadequate, the power supplies aren't stiff enough, the regulators can't handle the loads, or a combination of these. OTOH, there's no reason more speaker manufacturers can't do a little more work and design a crossover that tames the impedance curve, other than the fact that they want to hit a price point and that means they don't want to use more, or better, parts.

This is the second new thread I've read tonight where this would be a wise response (screw Integra too).

I just wish my local competitors were as slow to come to this realization...

Nominal impedance is NOT usually what the amplifier will see and from Merriam-Webster, one definition is "of, being, or relating to a designated or theoretical size that may vary from the actual : approximate ". They have many other definitions but they all state that 'nominal' isn't a real world number, it's just something to throw out there.

[Link: stereos.about.com]

Impedance has resistance to current flow, inductive reactance and capacitive reactance. When you look at the impedance curve rising as the frequency increases, it's part of the capacitive phase and if the impedance slope falls as the frequency increases, it's in the inductive phase. The amp has to deal with anything ranging from fairly flat (the speaker manufacturer decided that they didn't want a roller coaster, so they chose their components well and may have used impedance compensation in the crossover) to what looks like the best amusement park rides. Many amps will at least protest a bit, and some will puke on their shoes.

FWIW, I had a Denon AVR-990 and I now have an AVR-2312CI, which I use exclusively. I sold most of my other equipment because I had collected enough, for long enough. The only old piece I have left is a 1980s Sony ES integrated amp. Neither has a problem with the speakers I built and the impedance curve isn't particularly flat, but it's also not particularly low. I'll have to dig out the files with the curve and post it. I have never had a problem with any of the amplifiers in these three pieces and I don't baby them. When I want loud, I play loud. Never cuts out, never overheats, never sounds bad.

Do we remember that the bench tech told the guy to check the speakers and the speaker wiring to see if they're good? That is WAY below the level of detail and information we've been talking about for a while now.  However, this is fun:


"Nominal" means "that's what we call it." "Nom" as in nom de plume, meaning "name." It's usually the minimum impedance in the frequency range above the woofer's bass resonance where the impedance curve has its lowest value before the rise in impedance of the mid, if used, or the tweeter.  But then the number loses all specificity because it is ROUNDED OFF to 4, 6, 8 or 16 ohms.  The nominal rating is an approximation and does not represent the entire speaker performance.


"Change due to amount of voice coil in the gap" is a strange thing to refer to. A speaker with a different amount of voice coil in the gap than another will have a different efficiency and different distortion figures and so will be a different model. If you mean that the impedance changes as the speaker moves, which puts different amounts of voice coil in the gap, then that's true, but you're talking about impedance variations during from one millisecond measuring interval to another.  That's not what anyone is looking for.

Speaker impedances also do not change as such. Instead, when you look at the impedance of a speaker, you will find that it is not the same at one frequency as it is at another. We speak of impedances varying, but this gives the impression that if you put a sine wave in, the impedance won't be the same as it will vary. But no, it's different at different frequencies. "Vary" is not a good word to use for this, but it's what we say.


You can't measure a speaker in the resting phase with no signal except with a scale and a tape measure. And maybe a gauss meter for external magnetic field. If you want impedance, you need an AC signal. If you only have a VOM or other ohmmeter that uses DC, you'll get a number that, applied with your intelligence and understanding of speakers, will help you out quite a bit... but still won't tell you a thing about the mids and tweeters.


I also used that nice Bruel & Kjaer equipment to make audio response measurements of speakers in series and I found that if all were the same model, all was okay; if they were different models, the speaker with the higher bass resonance (usually the cheaper speaker) sucked the hell out of the bass of the better speaker by presenting a high source impedance in its mid-bass region.  The better speaker similarly affected the sound of the cheaper speaker via its impedance curve.

And I don't think any receiver manufacturer ever said "be sure to use the same model speakers for the best results."  We all counted on the main speakers to be better than the second pair.


Going back to the subject of this thread, that's way more than is needed for these issues.  And an impedance meter that does not plot will give you numbers that you can write down and plot out yourself, so it's not as convenient, but cannot be said not to help.


I'm surprised to hear of a VOM that uses a 1 kHz signal (you mean 1 kHz, right?) for measurements.  All sorts of circuits will yield incorrect values of resistance with an AC measurement.  "Resistance" completely implies no reactance.  Measure a 100 ohm resistor with a 10 mH coil in series with it with AC and with DC and you'll get very different results.  If you put a 33,000 microfarad capacitor across those probes you'll probably get nearly zero ohms evn though the capacitor does not have a short circuit.  A meter presenting DC would give you the resistance measurement, which would be infinite after a short charging period.


Maybe.  You oughta write to those folks and ask them how they would verify the resistance of a 33,000 mfd capacitor in parallel with a 10K ohm resistor.

What model is that?  I'd like to look at the manual.


Not familiar.  And again, way beyond what we need for the issues here.

Let me know if you want a circuit for an impedance meter.  It won't plot and it's not exact, but it's damn close and only involves one electronic raw component.

I really didn't think there was a reason to ask 'Which?" and nobody else asked that, either.

If you add resistance and it doesn't have a problem, it indicates that the amp didn't like the original load. While the problem could be caused by bad adjustment of the bias, leaky transistors or some other defect, it is a quick way to narrow down the list of faults, when an impedance meter isn't available.

Maybe you should stop nit-picking everyone's grammar and general use of English. This isn't grade school and, while I'm not a big fan of bad spelling and grammar, I don't let it bunch up my undies the way you do. Let it go- the World will keep spinning and the language will keep devolving whether you try to save it one person at a time, or not.

Buzz can occur without rub and if the voice coil isn't wrapped/cemented well, the spurious resonances will be "seen" by the program. If it rubs for long, the impedance test can show it because the coil may scrape on the magnet assembly. I haven't asked how the program finds these and I'm not doing high volume speaker building, so I haven't used it. It requires a known good sample of each type of driver, then compares all others to those.

Ever seen a small speaker builder's products? I don't mean <10K pieces, I mean REALLY small. When they hire stoners to wire the speakers and some show up with the + wires on the woofer and - wires on the tweeter on a second order crossover, you can be sure that the impedance will be different.

Re: the first point, the complexity of the total circuit that results from connecting two separate circuits in series- I saw a lot of people come back for a second pair of what they had bought because of this- some of us tried to package two pairs of the same whenever someone considered using more than one pair.

As far as manufacturers and same models, I meant that if someone bought other electronics, they marketed their speakers along side so they could sell "package systems". I don't remember any reps who mentioned selling two pairs of the same, unless it was a flippant "Then sell TWO pairs" comment at a sales meeting. Other than that, most weren't very technically-oriented.

MLSSA is a software suite, used for speaker/crossover testing/design. If you look at the specs from a lot of speaker driver manufacturers, you may see an impedance/phase chart with 'LEAP' at one of the bottom corners and this is part of that suite.

[Link: mlssa.com]

Remove the suspect speaker and throw it in a deep well. If it floats, it's bad. If it sinks, it's good.

This method has worked for centuries.

Could be a duck.