[chadh]

I almost posted this to "The Lab", but then realized I probably wouldn't understand the answers.  So, as it's obviously a real newbie kind of question, I thought this might be an appropriate venue.

This impedance thing has me confused.  When talking about speakers, people seem to refer to low impedance speakers as being hard to drive.  Speakers that have an impedance dropping as low as 2 ohms pose all sorts of problems for amplifiers.  And some amplifiers (like the chip amps I used to run) apparently just can't cope with these really low impedance speakers.

Then you get headphones, and the whole thing seems to be reversed.  Grados have a 30 ohm impedance, and are much easier to drive than Sennheisers with a 300 ohm impedance.

I gather that a speaker's or headphone's impedance really affects the ability of an amplifier to deliver current.  That would be why an amp delivering X watts to an 8 ohm speaker would deliver as many as 2X watts to a 4 ohm speaker.  This helps to understand why low impedance headphones are easier to drive.  But I just don't understand why low impedance speakers become such a problem.

Chad

[IronLion]

I believe the analogy, however correct or incorrect, that got me to sort of wrap my head around this was thinking of impedance with regards to speakers and amps as the diameter of pipe in a plumbing system.  The lower the impedance, the larger the diameter, and the more water (electricity) can flow through it.  This is, of course, assuming your water pump (or amplifier, in this metaphor) can keep up with the demand coming from the faucet (or speaker).  Lower impedances as I understand it allow the amplifier to get closer to its maximum possible output, but there's a safe point where amps can operate to cover all dynamic peaks without being asked to produce more power than it can output- this is I think why low impedance speakers can be so hard to drive, they can demand too much electricity from an amplifier and take it out of its comfort zone.  Anyways, this is how I understand it...   

[Dan_ed]

Impedance is frequency dependent resistance, sometimes called AC resistance. This means that the resistance the amp sees varies constantly with the frequencies. Impedance and reactance (the mathematical component of the impedance) are associated with capacitors and inductors. This is why caps look like more and more like a short as the frequency increases and why an inductor looks more and more like a short as the frequency decreases.

More impedance equates to more resistance which equates to less current demand on the amp. I guess you can figure out the less impedance case.

As to the headphone example, the simple difference is because the output impedance of the amp is seeing a heavier load with the 300 ohm pair. Again, more current draw on the amp. It is actually much more complicated but this is what people are referring to when they talk about load matching.

Hope this helps.

[planet10]

Impedance and reactance (the mathematical component of the impedance)

To be a bit more mathematically correct:

Impedance is a complex curve in 3-space (freq, magnitude, phase angle). The magnitude (resistance) is the real part and reactance (phase) is the imaginary part.

If we consider the magnitude, and that most amplifiers try to be constant voltage amplifiers, then as the resistance drops the amount of current required increases linearly, but as we have an AC circuit it is more complex.

P=V x I x cos(phase angle)

If phase angle = 0 then cos(phase angle) = 1, but rarely does a speaker look like this.

If phase angle = 90 degrees then cos(phase angle) = 0, and no power can be delivered into the load -- the most common speakers that approach this are electrostats.

If impedance drops & the phase angle gets severe it can be seen that a constant voltage amp can quickly run out of the ability to deliver sufficient current.

dave

[Dan_ed]

Yes, I left out the word "imaginary".    But I think if we keep going we're going to lose Chad. 

[HT cOz]

If one would like an even (easy) impedance should they go with a ported or sealed speaker design?

[IronLion]

If one would like an even (easy) impedance should they go with a ported or sealed speaker design?

Sealed should be "easier" I think, though there must be many other variables involved. 

[twitch54]

If one would like an even (easy) impedance should they go with a ported or sealed speaker design?

Sealed should be "easier" I think, though there must be many other variables involved. 

"Normaly" a ported speaker is Easier (higher efficency) for an amplifier to drive.  Porting along with it's 'tunning tendencies' relieves some pressure if you will from the pistonic forces of the bass driver (so this is how it was explainedf to me years ago !)

[Wayner]

Dave and others are closing in on the answer, but you were on your way yourself with your example. Say we have an amplifier that is rated for 100 watts RMS per channel at 8 ohms. Thinking of the speakers as purely resistive (which they are not), then the example of doubled power at 4 ohms would be a close conclusion, so that number would be 200 watts per channel. Now, where you didn't go was the 2 ohm load, which would make the power output be 400 watts per channel. There is your problem. Most amplifiers power-supplies can't deliver enough ump, if you will, to drive the output stage and the amp will clip, heat up and have bad things happen to it. While the amplifier is making pulsating DC, cause the sign wave has been flattened at the top (from exceeding the power supplies capacities), your speakers are obeying the signal. Which means that while the amp is sending this chopped off sign wave, it now looks like direct current and the speaker motor "freezes" during this cycle, causing the voice coil to super-heat and eventually be damaged.

Now most people do not play music that loud, but the speaker is not a purely resistive load, either. My MartinLogans can dip down to 1 ohm (which is one ohm from being a short), and it takes a very special kind of amplifier to handle this kind of extreme condition, even at moderate loudness levels.

Now you can also imagine a speaker that is low impedance (like 4 ohms), but highly in-efficient, needing lots of power to drive (like 84-85db), that is a dangerous combination.

This is also the reason why you can damage a speaker easier under-powering it with a smallish type of amplifier. Clipped signal is a death sentence for speakers.

Wayner 

[tekwiz]

.

[Ericus Rex]

Welcome Jeff!

Sounds like you'll be an asset to us all here at AC.

[NewBuyer]

Hello Jeff,

I agree, thanks for that great post.  I am wondering please: Do the concerns you addressed above, have equal relevance for active speaker designs (i.e. in-built amps individually mapped to each driver, active internal crossovers, etc)?

[vett93]

What exactly is an autoformer? If it is a transformer, how can the amp sees a constant impedance when the speakers change impedance at various frequencies? In other words, the ratio between the primary side and secondary side of a transformer is fixed. If the load on the secondary side varies, won't the primary side changes also?

Guess I still don't understand the autoformer. Could you please explain more, Jeff? Thanks.

[Ericus Rex]

I think the definition of an autoformer is that it is a transformer where the primary and secondary actually share coils, whereas your standard transformer keeps pri and sec completely seperate.  Is this correct?  If so, how does this work at all and not act like a short?

[tekwiz]

.

[tekwiz]

.

[tekwiz]

.

[lonewolfny42]

Welcome Jeff....thanks for your fine responses to the questions....

                 Chris

[bosound]

To follow this thread to the practical conclusion for my simple mind:   

I use a Glow Audio amp one to drive headphones and small monitor speakers.  The specifications for that amp list 8 ohm speakers, and the manual indicates that speakers which drop down to low impedences should be avoided.

If I took that Glow Audio amp one and hooked it up to my old a/d/s 1530 speakers with a nominal impedence of 4 ohms and started playing all kinds of music through them at normal listening levels, what would be the result?   

    Would it blow up, taking it and my precious collection of Milli Vanilli records with it? 

    Would the amp get hotter and hotter so that I would come home to find a molten lump of glass
    and metal?

    Would the output tubes fail, leaving the rest of the amp unscathed and requiring nothing
    more  than re-tubing?

    Would multiple parts fail, and reduce my Glow audio amp to a paperweight?

    Would I damage the speakers?  If so, which would go first, the amp or the speakers?

Any guidance would be useful.   Just kidding about Milli Vanilli...


bosound

Unnecessary and unsolicited info about my systems follows: 

Main System:  BPT signature power conditioner, ModWright SWLP pre-amp/phono stage with outboard power supply, FirstWatt F3 J-FET amp, Omega XRS speakers with Fostex drivers, unmodded transporter as digital source, Nottingham Analogue Dais turntable with Grado cartridge as (main) analog source, speaker cables Crimson music link, interconnects Crimson and Zu cable.

Second system:  Glow Audio amp one (original version), Paradigm atom speakers, Logitech(Slim) squeezebox, speaker cables Crimson music link.  The squeezebox is named "Throne Room Orchestra" and the speakers sit next to my shaving brush and soap mug...

Third system:   a/d/s/ PA4 amp fed directly from a Logitech Squeezebox and driving a/d/s M1530 speakers which are more than two decades old  (a ModWright SWL 9.0 SE should be in this system, but is loaned out to a friend who is far from home and family for work reasons and had no sound system)

Headphone system:  Benchmark pre DAC, Glow Audio amp one ver '09, Sennheiser HD650, AudioTechnica ADH W 1000  usign Benchmark as pre and computer DAC and sometimes as
a headphone amp


Average listening time:  three to four hours daily

In the event of fire, I would rescue the BPT power conditioner, ModWright SWLP and FirstWatt F3 first, then ask the kind firefighters to help me lug the Nottingham turntable out of the blaze...

[tekwiz]

.