Speaker watt rating

[JohnH]

Hey guys,

I'm fiddling around with a couple of small amplifier circuits (like this), but am totally confused by speaker requirements.

I have a circuit which calls for an 8ohm 1W speaker, and another which calls for a 4ohm 5W speaker. Now I have the first speaker already, but for the second one I only have a 4ohm 3W lying around. I'm assuming the watts rating is the maximum the speaker can withstand, and so the 3w speaker will be no good? If I buy another speaker, does it matter how close the speaker rating is to the circuit output? Like if I find a 4ohm 20W speaker, will that be ok?

Sorry for the noob questions. I did a bunch of googling but just confused the hell out of myself!

EDIT: one of the circuits I was looking at uses a BA5406, which I happen to have on hand. It mentions that it should have a heatsink, which I don't have. Does anyone know what kind of heatsink I can use for these? (Again, google proved unhelpful).

[Marcel]

The general rule is to match the power rating of the speaker to the amp. So a 5W amp should have a 5W speaker. ..... And now that we have addressed all the Engineers concerns we can give them a beer and send them to bed and the rest of us can laugh about what goes on in the real world....

Often a larger wattage speaker is connected to a amp, and it's rare to find a smaller speaker on a larger amp. For example in many Marshall guitar amps the speaker rating is 1.2 to double the rated power of the amp. My 100W JVM combo has 2x60W speakers, and a typical 4x12 has 4x60W speakers to hook up to any 50W or 100W head. Even when loaded with 25W vintage Greenbacks a 100W head is a match to the 4x25W of the cabinet. Disco and PA systems often use a similar approach with speakers wattage being equal or up to two times the amp power ratings... Many Bass players frequent the 1.5 to two times the amp ratings, though there are some who like to keep it closer to 1:1 for tonal and response sake...

But back to the 'problem of a 3W speaker on a 5W amp. When we are dealing with the smaller number of watts the chances of burning out the speaker are far less, and it is for two reasons. 1/- If you do crank the volume up of our MP3 player to the full 5W chances are that your ears will pass the message to your brain that "this sounds like S@#t... turn it down" mostly due to the distortion the speaker will make when driven so hard. And 2/- The short time you have the volume at stupid levels for such a tiny speaker will not generate enough heat to do any major damage to the voice coil.

And that's the crunch.... If we put 5W of 440Hz tone into that little 3W speaker then yes it will heat up and after a few hours will eventually burn out. If we put 5W of 440Hz tone into a 20W speaker then it will probably sit there all day annoying the neighbours and not even get warm... If we put 3W of 440Hz into our 3W speaker it will probably last a few months before shaking itself to bits..... Playing a MP3 at reasonable volume levels will probably average 0.3W with peaks of maybe 1.5W to 2W which is well within the 3W (average) maximum that the 3W rated speaker can handle.
(As a side note - I have a very old Realistic brand power monitor that is permanently monitoring the power levels of my 200W+200W Wharfdale studio monitors. It gets pretty damn loud when I push the volume past 5W, and the building is truly shaking when the meters flash out any +100W peaks, but most of the time 'comfortable listening' is less than 1W.)

A huge impact on your speaker size decision will depend on what you put through your amp and speaker.... If it's going to be a MP3 player at comfortable listening levels then a 3W speaker would be fine and would probably last years.... but if you're intending to play endless guitar solo rehearsals at the 'fully cranked' 5W volume then a better than 10W speaker choice would be wise.

[dave.king1]

Words of wisdom from Marcel right there.

I'll add, don't install a lower impedance speaker than the circuit calls for or you'll let the smoke out of the output transistor(s)

A while back I inadvertently wired a quad cabinet in parallel instead of series parallel giving 2 ohms instead of 4, when switched on it let out a deafening thump then silence. The replacement transistors were surprisingly cheap but being from the 1970s were a pain to source.

[JohnH]

Thanks guys, this is really helpful. Better get back to soldering...

[DrNomis_44]

I've currently got a Jensen C10Q 8 Ohm 35 Watt speaker installed in my 5 Watt Legacy guitar amp and have had no major issues so far, in short, you can get away with using that 4 Ohm 20 Watt speaker with no problems, as Marcel has explained.

I'll also add that if you try to use a 4 Ohm speaker with an amp that needs an 8 Ohm speaker, the amp is going to be working about twice as hard to drive the 4 Ohm speaker as it would driving the 8 Ohm speaker, if the amp is a valve type, you must make sure that you are using an 8 Ohm speaker with it if it needs an 8 Ohm speaker, otherwise the amp won't put out the correct amount of power, you also risk burning-out the output transformer, or the output valves suffering an internal arc-over.

[Marcel]

I tried not to include ohms in my post earlier because its two buddies (Volts and current) would also have to become involved, and that makes it a foursome which can get really messy....

All four are inter-related... We all know ohms law being E=IR or Voltage = current times resistance. Well power is a relative too being P=EI or Power = voltage times current... which when we transpose values in the equations means power is also equal to I2R or current squared times resistance.

So ... if we have 100W of audio into a 8 ohm cabinet then = Sqrt of 100/8 gives us about 3.53A... and 3.53A x 8ohm = 28.24V
If we have 100W of audio into a 4 ohm cabinet then = Sqrt of 100/4 gives us 5A ... and 5A times 4ohms = 20V
If we have 100W of audio into a 16 ohm cabinet then = sqrt of 100/16 gives us about 2.5A... and 2.5A x 16ohms = 40V

All interesting stuff... but note how by increasing the resistance by 4 times only resulted in a halving of our current, yet our voltage doubled.... That doubling of voltage is reflected back into our amp, back to the last active device, which in a tube amp is our power tubes, and when transformed by the output transformer up to power tube levels in the same way the audio was transformed down to speaker levels it becomes easier to see why sparks can fly if the impedance selector is not set correctly. And when you realise everything is made for a price so they often use the thinnest wire possible for the transformer output winding the extra current on some settings of the impedance selector will quite easily turn your output transformer into a fuse.
Transistor amps don't usually use transformers, and the ratings of output transistors usually far exceed any mis-matched back EMF so they do have a reputation for being much more tolerant of speaker impedance variation.

At lower powers of a few Watts it is more likely that the amp will fail to deliver enough voltage to drive an impedance to a certain volume. In higher Wattage power amps other factors like current and heat dissipation become increasingly important factors. Kick it up to the +1kW range and even the small resistances and ratings of the interconnecting wire and Speakon connectors need to be seriously considered and addressed.

Even the best engineered systems can be undone by an errant operator... The largest system I ever worked on was a 56kW RMS per channel stereo active 4 way crossover rig with multiple 5kW 3way delayed stacks, with lots of dbx limiters and top end 32 band EQ units. One night a guest band came in and their sound guy ran the 40ch FOH desk.... first song and every peak lamp on the desk lit up and stayed on for the entire set (which is the way he ran his own usual rig)... At the end of the set no less that 60% of the 4kHz to 20kHz drivers were blown due to simply too much high frequency (from all the squared off clipped signals in the desk) getting to the PA... The sound guy from the next band up on stage vocalised loudly that the PA was garbage, and we only discovered the real truth why when we checked the system over with a fine toothed comb the next day. We knew the PA itself never went into clipping so the energy required to blow the high frequency drivers could only have come from the clipping in the desk....

[DrNomis_44]

We could also examine the differences between a typical solid state guitar power amp, and a valve guitar power amp, first thing we would notice is that the valve power amp uses less components than the solid-state power amp (assuming they are both made using discrete components), second thing we would notice is that the valve power amp uses less negative feedback than the solid-state power amp, this is mostly due to limitations in the valve power amp's output transformer which limits the amount of negative feedback that can be applied.

Applying negative feedback to a valve power amp does three beneficial things I can think of, firstly it reduces harmonic distortion, secondly it flattens the frequency response, and thirdly it lowers the output impedance of the valve power amp, of course, solid state power amps use negative feedback too but they tend to use a lot more than valve power amps, if you try to increase the amount of negative feedback in a valve power amp, at some stage the valve power amp is going to become unstable in it's behavior at some frequencies in it's frequency response, why does this happen?, it's mostly due to phase-shifts caused by the windings in the output transformer, now, because most solid-state power amps found in guitar amps don't use an output transformer, like valve amps do to isolate the speaker from the HT supply going to the power valves, they don't suffer from the same phase-shifts and so can afford to use higher levels of negative feedback, and it is because solid-state power amps use higher levels of negative feedback that they inherently have much lower output impedances, this means that the solid-state power amp can supply more current to the speaker.


There is one downside to solid state power amps, when the solid-state power amp is run below the point where it starts clipping, everything is great, soon as you start running the solid-state power amp harder into clipping you run the risk of burning-out the voice coil in the speaker because you're effectively putting pure DC across it.

This is where valve power amps are at an advantage, since if the valve power amp is driven into clipping, the output transformer is going to limit the current being delivered to the speaker.


Here's an interesting fact, the impedance of a speaker is actually not constant for all the frequencies that the speaker can reproduce, it actually varies, at some point in the speaker's frequency response the impedance will rise to a peak, this is where the speaker is operating at maximum efficiency due to resonance, therefore the impedance-rating of a speaker is actually a nominal rating.

[Singing Frog]

Be careful about RMS vs peak values of speakers and amps and it doesn't hurt to give your speakers a bit of head room ie over rate them so if your amp is say 50w RMS then use a bigger speaker say 75w RMS. A 1 watt speaker is not going to cut it for a guitar amp...

[DrNomis_44]

Be careful about RMS vs peak values of speakers and amps and it doesn't hurt to give your speakers a bit of head room ie over rate them so if your amp is say 50w RMS then use a bigger speaker say 75w RMS. A 1 watt speaker is not going to cut it for a guitar amp...

That's a good idea, my Marshall amp is rated to put out 100 Watts, and it has two 16 Ohm 12 inch Eminence speakers rated at 75 Watts which are connected in Parallel, giving an 8 Ohm/140 Watt load for the amp.

[Simon Barden]

I have a forum mate who used to work for Blackstar. The only speaker brand he actually trusts for their ratings are Celestion. Anything else he'd match with an amp rated at half the speaker rating.

A lot of valve amps are rated around the point the THD exceeds 5%, i.e. when they still sound clean, but can put out a lot more power when really driven hard. A 100W Marshall amp can put out between 180-200W if really pushed. Transistor amps normally have far less headroom at this point and put out very little extra power if driven hard.

But as Marcel said, the average power level when playing a guitar is a lot lot less. All those gaps between the notes, plus the level dying away after the initial attack really lower the average energy level. I asked Celestion about the 90W rating of the speaker in my 100W Mesa Nomad and they said it was fine.

But until recently there hasn't been a universal standard for testing and rating speaker wattage e.g. at what frequency or mix of frequencies, how long the test went on for and what (if any) level of 'damage' was acceptable. There is now an AES standard, which does give a figure very close to most normal RMS watts figures, but it is a set test with known program material and a set duration.