4, 8 or 16 ohm tonal difference?

Deijavoo

Maybe a dumb question but I've always had heads and cabs that run on 8ohm.

I'm considering a 2 heads into a 4x12 Mesa cab which would then both run on 4ohm each. 

Is there any tonal difference or am I making more of it than I should? 

thermionic

I've wondered about this recently... DC resistance (related to impedance) is a measure of the gauge of wire and number of turns in the voice coil, and we're all aware how DC resistance is a good indicator of tone when it comes to pickups. Assuming the wire is the same thickness, there are 4 times more turns in the voice coil of a 16 ohm speaker compared to a 4 ohm one, right? OK it's impedance not resistance, you have to consider the magnet and the properties of the cone etc, but still, does a 4 ohm Celestion sound different to a 16 Celestion of the same model?

ICBM

There are differences, but they're relatively subtle and you'd probably only really be able to tell if there is a direct comparison. Generally, functionality overrides it - there aren't that many amps and cabs where you can run the exact same amp and speakers at 4 and 16 ohms, and often you don't have a choice at all. If you do have a direct choice when you're putting together a rig it can be better to choose one or the other, but it's much less important than the type of speaker.

The differences are caused by the damping factor and inductance of the speakers and transformer windings, which are much more important than the resistance.

ecc83

I am not so sure that they would use the same wire gauge for 16 and 4 Ohm speakers? But I have never  had a sample each of the same unit to compare the resistances.

I would throw the question back to the OP transformer. These, as do all transformer, work best as a rule when the ratio is 1:1 so a 16R speaker will be using a lower ratio than a 4R one. 

All guitar amp OPTs are, AFAIK just a crude "tap" rather than the "proper" way to do it with multiple secondaries. I would suspect that amps are designed and developed for a 16R load and the other impedances (when given!) provided for convenience? Then of course there are amps that do not have 16R outputs. 

I recall reading once that for transistor amps the optimum speaker Z would be around 6 Ohms and in fact there are a few hi-fi models of that impedance.

But! Consider headphones? These are available in a bewildering range of impedances from about 24 Ohms (I think the totally stupid 8R phones of yesteryear have gone?) up to 600R. There must be an optimum one would think? I have asked the question a couple of times on audio forums but no manufacturer has ever commented. Of course "studio" cans were always 600R  and I had a pair of the superb AKG K 60s for many years. 

Dave.

thermionic

The reason I've been thinking about this is that my kit amp has 4, 8 and 16 ohm outputs. I could have gone for an 8-ohm speaker then I could have tried my Vox AC15 with a Greenback and my kit amp with the Vox's 2x10s - just out of curiosity. I went for a 16-ohm speaker  in the end because of the "crude tap" that Dave mentioned (and I think ICBM said something similar a while ago) - better to use all the windings in the transformer than just half of them.

jpfamps

ecc83 said:

I am not so sure that they would use the same wire gauge for 16 and 4 Ohm speakers? But I have never  had a sample each of the same unit to compare the resistances.

I would throw the question back to the OP transformer. These, as do all transformer, work best as a rule when the ratio is 1:1 so a 16R speaker will be using a lower ratio than a 4R one. 

All guitar amp OPTs are, AFAIK just a crude "tap" rather than the "proper" way to do it with multiple secondaries. I would suspect that amps are designed and developed for a 16R load and the other impedances (when given!) provided for convenience? Then of course there are amps that do not have 16R outputs. 

I recall reading once that for transistor amps the optimum speaker Z would be around 6 Ohms and in fact there are a few hi-fi models of that impedance.

But! Consider headphones? These are available in a bewildering range of impedances from about 24 Ohms (I think the totally stupid 8R phones of yesteryear have gone?) up to 600R. There must be an optimum one would think? I have asked the question a couple of times on audio forums but no manufacturer has ever commented. Of course "studio" cans were always 600R  and I had a pair of the superb AKG K 60s for many years. 

Dave.

re speaker impedance there is no impedance that is inherently better.

However with transistor amps there is always a trade off between power delivery and distortion; a lower impedance speaker will allow the amp to deliver more power, but will do so with higher distortion.

The 6 ohm "optimum" is likely to be someone's opinion of the correct compromise between distortion and power.

Headphones don't need to deliver much power so a higher impedance which is easy to drive with low distortion is sensible.

To answer your other question a 4 ohm speaker does use thicker wire than a 16 ohm speaker.

Cirrus

thermionic said:

better to use all the windings in the transformer than just half of them.

I see people saying this all the time, and on the face of it the sentiment seems fair enough. But what's the actual logic behind it? Does an extra length of coiled secondary wire add to the tone somehow? Does slicing through a greater volume of magnetic field have some effect?

Not being snarky, am genuinely curious if anyone has any thoughts.

jpfamps

Cirrus said:

thermionic said:

better to use all the windings in the transformer than just half of them.

I see people saying this all the time, and on the face of it the sentiment seems fair enough. But what's the actual logic behind it? Does an extra length of coiled secondary wire add to the tone somehow? Does slicing through a greater volume of magnetic field have some effect?

Not being snarky, am genuinely curious if anyone has any thoughts.

There are definite theoretical reasons for potential differences in performance due to the HF response of the transformer.

The HF response of the transformer is determined by two parameters: leakage inductance and stray capacitance. Ideally the lower the better.

Both these are affected by the number of turns on the transformer (all things being equal which they rarely are!): less turns means less capacitance. Leakage inductance (whilst impossible to calculate with any accuracy) is proportional to the square of the turns ratio.

Additionally splitting up the primary and secondary into sections improves the HF performance of the transformer by reducing leakage inductance.

One of the "classic" designs uses 5 sections: 3 for the primary and 2 for the secondary. For example, the JMI era AC30 transformer uses this pattern.

By employing the 4 ohm tap rather than the 16 ohm tap in a transformer wound like this means that you now only have a 3 section transformer rather than a 5 section.

If the transformer is "straight wound" (which the cheapest way of winding a transformer, so you can imagine this is a popular method!), then there is no sectioning: all the primary is wound then all the secondary. With this type of transformer there will be less of a difference than a sectioned/ tapped transformer, but there is potential for some differences due to changes in the turns ratio.

Of course whether any of this is audible in a guitar amp is debatable.

Regardless, it's' virtually impossible to do an exact comparison, as a 4 ohm load is usually either 4 x 16 ohms in parallel, or 2 x 8 ohms in parallel, whereas a 16 ohm load is usually a 4x16 ohms with either parallel/series wiring or series/parallel wiring, or possibly 2 x 8 ohms in series.

The difference in speaker wiring will probably have more effect on the sound.

ecc83

"Headphones don't need to deliver much power so a higher impedance which is easy to drive with low distortion is sensible."

Back in the day, all decent headphones were 600Ohms but the coming of the iPod changed that and much lower impedance cans appeared. The pods of course not having enough drive voltage for 600R. 

Then there is now the massive home recording market and shedloads of cheap to very expensive mixers and interfaces. There should be no problem driving any sensible phone impedance from a mains powered device, you would think! But many cheap mixers do struggle to put enough volts into the higher Z models. Bus powered usb interfaces are an even worse problem due to the limited power available. 

"We" await the coming of USB3.0 interfaces that can make good use of the extra couple of watts!

Dave.

thermionic

jpfamps said:

Regardless, it's' virtually impossible to do an exact comparison, as a 4 ohm load is usually either 4 x 16 ohms in parallel, or 2 x 8 ohms in parallel, whereas a 16 ohm load is usually a 4x16 ohms with either parallel/series wiring or series/parallel wiring, or possibly 2 x 8 ohms in series.

An exact comparison would be two single speakers, identical models but one 4 ohm and the other 16 ohm, which was almost what I was considering (going for an 8-ohm or a 16-ohm Greenback). But I tend to agree, any difference would probably be negligible to most people's ears. Have a wisdom for that informative post.

jpfamps

thermionic said:

jpfamps said:

Regardless, it's' virtually impossible to do an exact comparison, as a 4 ohm load is usually either 4 x 16 ohms in parallel, or 2 x 8 ohms in parallel, whereas a 16 ohm load is usually a 4x16 ohms with either parallel/series wiring or series/parallel wiring, or possibly 2 x 8 ohms in series.

An exact comparison would be two single speakers, identical models but one 4 ohm and the other 16 ohm, which was almost what I was considering (going for an 8-ohm or a 16-ohm Greenback). But I tend to agree, any difference would probably be negligible to most people's ears. Have a wisdom for that informative post.

That's as close as you will get, but a 4 ohm (if you can find one) and 16 ohm speaker aren't the same.

For starters the 4 ohm will have a heavier voice coil due to the thicker wire used.

Having said this, this is about as close as you will get to a fair contest.

imalone

ecc83 said:

"Headphones don't need to deliver much power so a higher impedance which is easy to drive with low distortion is sensible."

Back in the day, all decent headphones were 600Ohms but the coming of the iPod changed that and much lower impedance cans appeared.

The walkman surely?

ecc83

imalone said:

ecc83 said:

"Headphones don't need to deliver much power so a higher impedance which is easy to drive with low distortion is sensible."

Back in the day, all decent headphones were 600Ohms but the coming of the iPod changed that and much lower impedance cans appeared.

The walkman surely?

Hmmm? You could be right! Ten minutes of Googling failed to find me the Z of the original phones but I have a sneaking feeling they were 8 Ohms (as were many big "hi-fi" cans of the time) and the Walkman had  wee,wee transformers in it?

BTW! I didn't say the first bit!

Dave.

ICBM

Walkman headphones are either 16 or 32 ohm. Pretty sure the originals were 32, but I have some here (not Sony) which are 16.

Cirrus

Cool, so you could play a guitar amp through them for low volume practice! 8-X

ICBM

Cirrus said:

Cool, so you could play a guitar amp through them for low volume practice! 8-X

I know you're joking, but… if you're extremely careful with the volume control, yes.

Actually true with any headphones - at the sort of volume levels that won't blow the headphones, the impedance won't matter to the amp anyway. But it's unwise to try, since it would be very easy to give it a bit too much volume and blow the headphones. And it will sound crap anyway.

ecc83

ICBM said:

Cirrus said:

Cool, so you could play a guitar amp through them for low volume practice! 8-X

I know you're joking, but… if you're extremely careful with the volume control, yes.

Actually true with any headphones - at the sort of volume levels that won't blow the headphones, the impedance won't matter to the amp anyway. But it's unwise to try, since it would be very easy to give it a bit too much volume and blow the headphones. And it will sound crap anyway.

This is why the first headphones designed (apparently!) for early transistorized hi fi amps were stupid when made in an 8 Ohm impedance. If accidentally connected to even a modest 20W per channel amp they could burn out and even if the punter did keep the wick down the PA produced unacceptable noise in the very sensitive phones. 

Then a switching jack was needed together with an attenuator circuit to cut the level and reduce noise. The jacks produced measurable distortion through their contacts (top amp designers would not use them) . Had high impedance ~600R cans been used some attenuation would have been needed but not so much because the phones need about 1volt rms to produce 100dB SPL . If speakers HAD to be killed a dedicated, high current switch was better. Better still was a proper, separate headphone amplifier.

Dave.

ICBM

ecc83 said:

 If speakers HAD to be killed a dedicated, high current switch was better. Better still was a proper, separate headphone amplifier.

…which (mentioning no names! ) is finally being done in many guitar amps too - although at least as much because they need to be speaker-emulated.

Would you believe, I've come across a valve guitar amp which has a headphone socket done the 'standard transistor amp' way, by breaking the signal path to the speaker and padding the output with a large resistor… and which does *not* switch the output to a parallel dummy load as well. It's the worst-designed and one of the worst-built valve amps I know of.

It still annoys me how many guitar amps have headphone jacks which directly break the speaker circuit - particularly in those where the speakers are connected to the amp section by a jack and plug… there's no reason at all to introduce extra unreliability (which very much does occur) by using the jack switches. If you want to mute the speaker, unplug it!

imalone

Cirrus said:

Cool, so you could play a guitar amp through them for low volume practice! 8-X

Briefly. And then for high volume practice for an even shorter time.

Sessionman

jpfamps said:

Cirrus said:

thermionic said:

better to use all the windings in the transformer than just half of them.

I see people saying this all the time, and on the face of it the sentiment seems fair enough. But what's the actual logic behind it? Does an extra length of coiled secondary wire add to the tone somehow? Does slicing through a greater volume of magnetic field have some effect?

Not being snarky, am genuinely curious if anyone has any thoughts.

The difference in speaker wiring will probably have more effect on the sound.

TFB newbie here.

Interestingly, because the impedance of speakers vary with frequency, and the resonant nodes of the cone are reflected in this too, two different model speakers connected in series can sound quite different to two of the same speakers connected in parallel.  This is due to the two different impedance curves subtracting or adding when the speakers are in series.  Thus creating an entirely new impedance curve, therefore tonal response.

Also relative: Speakers, when driven from a high output impedance such as a valve output transformer or transistor power amp with modern 'constant current drive', will actually enable the addition of harmonics added to the tone that are the result of the cone's instability from being 'under-damped'.  The mentioned newly created series speaker impedance curve 'tone' (I prefer the term sound, as new harmonics are created) is greatly influenced by this too.  Due to the cone under damping, you hear many harmonics in the sound that are NOT created by the amplifier itself.  Behaviour rather like a car wheel with knackered shock absorbers.  Very rock 'n' roll!

Sessionman