Elementary question

[Scriffer]

I'm a little curious about the resistance ratings of amps and speakers and how to match them up. I understand that the impedance measures how easily a signal flows through a circuit, and obviously you want to match the resitance of your amp to that of your speakers.

My question is what are the potential consequences of mismatching the loads on the amp and speakers? Note that I have no intention of doing so, but I grow curious because I see 6 ohm speakers and I think "Who has a 6 ohm amp?"

Thanks for the info, and I hope I haven't offended anyone who happens to have a 6 ohm amp.

[Clive]

Amps are built to deliver a constant power output. Let's say that an amp rated at 20volts is putting out 50 watts @ 8 ohms to it's channels, if you half that (4 ohms) the amp will have to produce 100 watts, half that (2 ohms) demands 200 watts. The amp will also need to double its supply of power. Remember the amp delivers constant power so at this point the amp will be producing excessive heat that over a period of time will burn up and/or run out of power and shut down (Go into protection).

Hope this helps.

[Glen B]

Ben,

There is nothing complicated about matching amps and speakers. A speaker’s nominal (i.e., rated/published) impedance does not remain constant but varies with frequency. As an example, let us look at a speaker with an 8 ohm nominal impedance. At some frequencies, that speaker’s impedance may rise above 8 ohms. At other frequencies that same speaker’s impedance can drop to as little as 3 or 2 ohms (a.k.a., the speaker’s minimum impedance). Some speaker manufacturers publish the minimum impedance of their speakers. Other do not publish this information. In the latter case, consumers depend on other end users, product reviewers, dealers, etc. to pass the word when a speaker is discovered to be a difficult (i.e. low impedance) load to drive.

Now, let us look at amplifiers. The average amp should be able to drive any (nominal) load between 4 and 16 ohms. However, all amplifiers are not created equal. Some amps are capable of driving loads of as little as 2 or 1 ohms with little or no difficulty. Other less capable amps may be restricted to driving nominal loads of no less than 8, 6 or 4 ohms respectively. Any attempt to use a lesser amp to drive a speaker for which it is not rated, can cause it to overheat, shutdown or in a worst case, self-destruct, not to mention sounding strained and congested. As long as you match a speaker with an amp that is rated to drive that speaker’s nominal and minimum rated impedance, you should be okay.

Glen

[chrispy35]

Amps are built to deliver a constant power output.

You mean constant voltage right? In your example, power is going up as impedance goes down (P varies inversely with R in P=V^2/R).

Chris P.

[Sim reality]

You mean constant voltage right? In your example, power is going up as impedance goes down (P varies inversely with R in P=V^2/R).

Chris P.

No he means power... Amplifier output is a AC wave so the voltage is really never constant... The only limitation is that there will be a upper limit to the voltage an amplifier can supply (which is the input voltage on the transister power rails).

That said the output is actually more complex them ohms law because you have to take into account reactive impedences such as capacitance and inductance on a AC circuit.

[chrispy35]

Sorry, when I said constant voltage, I meant AC voltage (RMS value).

How could he mean that they are designed to output constant power when the example given illustrates what most would consider an idea amp: power doubles when impedance halves. That's not constant power.

The amp is a voltage gain device isn't it? Voltage out = gain*voltage in. A well designed amp will output the correct voltage regardless of the load placed on it, hence the doubling of power when impedance is halved.

[Sim reality]

Well, there are really 2 concepts here with really fuzzy line between the two: Max RMS power and dynamic power (AKA everything else)...

The Max RMS power is really limited by 2 things:
The ability for the power supply to maintain rail voltage at a particular current draw
The ability for the transistors to dissipate enough heat so it doesn't burn out.

If you remove the 2 limitations you will end up with the perfect "double Max power output at half the resistance" because the limiting factor is the Max voltage at the power rails (which in a perfect power supply is constant).

In reality, sound is a wave which has peaks and valleys and if your amp bumped up against the Maximum voltage the amp can supply (which is DC current at that point) you can fry the voice coils in you speakers.

Also when you are listening to music at a particular volume you need the amp to provide pretty consistant power across all the frequecies (regardless of the dips in resistance) otherwise you would end up with frequencies that are over-empathsized (which you would noticed because it does not sound linear).

Anyways... Getting back to the question... Most amps provide Max Power ratings for 4 and 8 ohms and therefore can be used to drive anything in between... (if the amp is rated only for 4 or 8 ohms you are probablly looking at an amp with a low voltage power rail or undersized heat sinks, respectively)

[chrispy35]

Also when you are listening to music at a particular volume you need the amp to provide pretty consistant power across all the frequecies (regardless of the dips in resistance) otherwise you would end up with frequencies that are over-empathsized (which you would noticed because it does not sound linear).

Can you point to a circuit topology that accomplishes this? The output stage of the amp would have to know two things in order to ensure constant power: the output voltage and the output current. The feedback mechanism used in output stages is voltage feedback, isn't it? If the amp somehow could produce constant power, how would the power consumption in the crossover be compensated for since that is not constant over frequency? Also, what about different efficiency levels of the individual drivers in the speaker?

It is the responsibility of the speaker/crossover designer to ensure that, given a constant AC voltage frequency sweep across the audible range, that the output SPL is also constant.

An amp is a linear voltage gain device, not a linear power gain device.

[Sim reality]

The transistor itself is current limited (the semiconductor will only allow a certain amount of current to jump the gap at a applied voltage) so I don't know what your point is on the "circuit topology". Both current and voltage are related and not just at the end load. (I mean if the input voltage on the transister determines the "size" of the "gate"... conceptually we are regulating the amount of current that can pass and the voltage is a function of that... because they are related, are we not regulating "power"?)

I will admit that the speaker/crossover designer "should" be a constant load across frequecies but because we are moving mechinical devices they all produce a dynamic load based on how the cone reacts to a given frequency... The crossovers are not designed to eliminate that but to transition the frequecies between one driver to another.

Don't get me wrong... I think you are correct on a conceptual "infinite load" or a load that is significantly larger then the internal load of the device providing the power. But I don't think it's quite correct wrt what happens between an amp and a speaker.

Anyways it's just my 2 cents... I have been wrong many times before and I did not like using 1st and 2nd order differencials in my circuit design class either.

[Clive]

Thank you gentlemen. It is discussions like this that is most welcomed here. Everybody learns.

The question was about amps and impedance matching could you now then summerize for the poster - reactions, pros and cons between amp and speaker.

[Scriffer]

Hey all,

I didn't mean to start a knock-down drag-out fight, but I'm sort of glad I did. I've been taking notes, and I've learned more in the last couple of days than I have in years, even though some of the terms/equations are still a little foreign to me.

Thanks for the insight, and by all means, keep it up. :T

Ben

[chrispy35]

I don't know what your point is on the "circuit topology".

I'm not 100% sure either, I just don't remember any circuits from school where we were P was constant given a varying load.

Don't get me wrong... I think you are correct on a conceptual "infinite load" or a load that is significantly larger then the internal load of the device providing the power. But I don't think it's quite correct wrt what happens between an amp and a speaker.

Internal load? Now I get to say 'not sure what you mean'. At a glance I'd guess you're referring to output impedance (making a comparison between high-impedance load with low-impedance source). I think in an amp we do have this case...amp damping factor specs roughly indicate the source impedance usually coming in less than 1% of the load impedance (DF=100).

...and I did not like using 1st and 2nd order differencials in my circuit design class either.

We definitely agree on that.

A few more bits after some thinking, the statement 'constant power' implies the following to me:

Say an amp with an easy 10 ohm load, generates 10 W with a 1 kHz, 50 mVrms input, therefore voltage across the load is 10 Vrms and current is 1 Arms.

If I connect a 5 ohm load to the same amp with the same input signal, being designed for constant power, the amp would still generate 10 W corresponding to 7.07 Vrms and 1.414 Arms.

This goes against what most would consider an easy way to tell a good amp from a bad one via the specs alone. With the 5 ohm load I'm pretty sure we would, assuming the amp is of decent quality and operating well within its rated power, get 10 Vrms and 2 Arms == 20 W which is double the power when the load is halved.

Given the same input signal, power just isn't constant with varying loads, whether the load is changing due to impedance change on the same speaker or due to just swapping speakers. Every amp spec sheet indicates this.

Also, to make use of the infinite load idea, when we have no speakers connected to the amp but there is still an input signal, where does the power go assuming the amp is generating constant power for a given input?

As for Scriffer's original question, I think Glen B's response above sums up my feelings. I'm just nitpicking Uncle Clive's choice of words in his post. No offence intended Uncle Clive.

[Sim reality]

I was thinking about it last night and realised I goofed up with the linear statement...

There is some build in control on the transister circuit because when the "gate" on the transister is open when charge migrates across the gate the voltage potential across the gate is reduced which limits the current (where the feedback happens) but I think Chrispy is right and this more of a inherent limitation of the circuit then the ideal design...

Where I was getting caught up what when the resistance dips if there was a increase in power consumption and the power has to go somewhere, but then is stuck me last night that more power is consumed when the voil coil resists movement and maybe the dips are the inefficiencies where driver does not want to move at frequency (it might be at .5 increments of the mutilples of the resonance frequecy so you have power cancelation happening).

ANYWAYS, to summarize:

1) You can run any speaker on any amplifier as long as it's at low volume
2) At high volume running a 8 ohm speaker on an amp rated for 4 ohms could cause the amp to clip because the 4 ohm amp is running up against the maximum voltage it has available which will damage the speaker
3) Running a 4 ohm speaker on a amp rated for 8 ohm could cause the transistors to burn out because of heat management issues because of the amount of power that needs to be dissipated.
5) Speakers are a complex load anyways so look for an amp with really big heat sinks.

[Clive]

Chris, thank you all for the input on this subject and for summarizing :T
The points are well taken.

Thanks!

[Scriffer]

Follow up questions

I'd like to start the first question by saying that I wouldn't do this in reality, but it's just a question of academics.
1) Would it be theoretically possible to add resistance to a speaker to bring it up to a load that would be acceptable to an amplifier? I seem to remember from high school physics that a power source doesn't know or care about the source of the resistances on a circuit. Is this close to accurate, or would it drastically affect the sound quality?

2) What is the benefit sonically of having 4 ohm speakers over 8 ohm or vice versa? Can you get the same sound out of a system with the different drivers, making it purely a question of efficiency with respect to the amplification?

Thanks.

[Sim reality]

I'll try not to over think things this time...

1) The power dissipated by the resistor would reduce the volume of the speaker so you would have to turn the volume up which would increase the power consumed but the amp anyway so you don't gain anything (you just can't get something for nothing). It can drasically affect the sound quality depending on the resistor (a resistor using a low emf resistor material like carbon probably would just reduce the available power, but a wire-wound type would definitily introduce distortion)

2) This is one of those questions where theory and practice really diverge... In theory 4 ohm should have better fidelity but less efficient (Lower indectance, less wire on the voice coil [reduced mass], etc) but in reality you end up with thicker wire (to dissipate heat) more distortion of the magnetic field (less and more course wire loops), proportionaly more difficult load to drive, etc... So it then boils down to how the driver is designed...

I think the general trend of speaker designers now is for low excursion drivers (tweeter, midrange) the preference is to use 8 ohm because you gain efficency without much compromise on performance and 4 ohms on high excursion drivers drivers (woofers, subwoofers) because the stack of the coil is shorter so voice coil can remain within the "uniform" part of the magnetic field of the magnet during the entire throw of the driver.

[Sim reality]

"performance" is a bad word... higher efficiency usually means greater sensitivity and more detail... Maybe "time fidelity" in the last paragraph...