Speaker coil winding

[nickb]

I agree. I have the same thought. I'd also like to know just how you would design a voice coil for a specific impedance other than matching and old damaged voice coil turn for turn with the same size wire. For instance, when you send a speaker for recone you can ask for an impedance change. That is if it was an 8 ohm speaker you can ask for the reconer to install a 4-ohm voice coil when the speaker is rebuilt. How exactly do they do that?

I'd keep the coil length the same and use sqrt(4/8) = 0.707 time the original number of turns. The wire would have to be thicker to keep the winding tight but that could be estimated: wire dia = coil length/ number of turns in on one layer.

[anchor]

Same current but fewer Ohms. Power = I^2 * R so fewer Ohms for the same current means lower power rating.

The power dissipated also has a voltage perspective where it does not change. If the amp used with the original voice coil is 100W rated at 4R, reducing the voice coil resistance does not change its power rating. Note p=v2/r, when you halve the resistance, more power is dissipated. The post by nosaj seems only half true and does not tell the whole story.

There´s more parameters involved:
* coil diameter
* winding length
* actual wire diameter (which may or may not match what raw Gauge # indicates)
* number of turns
* insulation thickness
* copper or aluminum
* number of layers.

Of all those listed above, the "winding length" appears to be the only factor relevant to the case above. The number of turns only changes the impedance, not the resistance.

[nosaj]

The power dissipated also has a voltage perspective where it does not change. If the amp used with the original voice coil is 100W rated at 4R, reducing the voice coil resistance does not change its power rating. Note p=v2/r, when you halve the resistance, more power is dissipated. The post by nosaj seems only half true and does not tell the whole story.



Of all those listed above, the "winding length" appears to be the only factor relevant to the case above. The number of turns only changes the impedance, not the resistance.

If wire diameters are the same the 8ohm coil will have more wire to dissapate more heat than the 4 ohm which will be shorter. Shorter wire of same value means more heat in my book.
nosaj

[Chuck H]

The number of turns only changes the impedance, not the resistance.

If wire didn't have resistance. But it does and the number of turns around a fixed size of coil changes the length and therefor the resistance and since the resistance is in parallel with the impedance @ frequency imposed by inductance there has to be an affect on impedance due to wire length if the same gauge is used. Further, the inductance affect on the impedance will be largely affected by the coils capacitance and that related to a large degree on insulation thickness and number of layers. You can see where this is going... Juan is a smart guy and probably the only person posting on any diy amp forum that makes his own speakers. If he says something is relevant then it probably is.

[Helmholtz]

My layman's approach would be like this:

Speaker power rating is determined by its max. voice coil temperature. The voice coil gets heated by the power dissipated in its DCR and is cooled via heat radiation to the surrounding structure and by convection.

An example: An 8 Ohm/50W speaker can take 20Vrms. This means a current of 2.5Arms. It DCR is typically 6.4 Ohm (80% of 8 Ohm), so the max. dissipated thermal power is 40W.
Everything else being equal, the 4 Ohm version of the speaker needs to have a DCR of 3.2 Ohm for a thermal power of 40W and a rated apparent power of 50W.

Everything else being equal means that both voice coils should have the same surface area and (thermal) mass. Realization (including achieving the necessary inductance to get a Z of 4 Ohm) will require a heavier wire as well as less turns for the 4 Ohm version. The optimization process probably requires some iteration.

I am aware that I ignored radiated acoustical power, but this is assumed to be the same for both versions.

Surely JMF knows better.

[nosaj]

My layman's approach would be like this:

Speaker power rating is determined by its max. voice coil temperature. The voice coil gets heated by the power dissipated in its DCR and is cooled via heat radiation to the surrounding structure and by convection.

An example: An 8 Ohm/50W speaker can take 20Vrms. This means a current of 2.5Arms. It DCR is typically 6.4 Ohm (80% of 8 Ohm), so the max. dissipated thermal power is 40W.
Everything else being equal, the 4 Ohm version of the speaker needs to have a DCR of 3.2 Ohm for a thermal power of 40W and a rated apparent power of 50W.

Everything else being equal means that both voice coils should have the same surface area and (thermal) mass. Realization (including achieving the necessary inductance to get a Z of 4 Ohm) will require a heavier wire as well as less turns for the 4 Ohm version. The optimization process probably requires some iteration.

I am aware that I ignored radiated acoustical power, but this is assumed to be the same for both versions.

Surely JMF knows better.

So in simplest terms what I said is correct "Same gauge fewer turns means less power handling." Which anchor took offense to.

nosaj

[Reader]

- - -. The voice coil gets heated by the power dissipated in its DCR and is cooled via heat radiation to the surrounding structure and by convection- - -

Yes. And in addition some speakers are designed to pump a little air thru the gap as the voice coil moves.

[Reader]

- - -. The voice coil gets heated by the power dissipated in its DCR and is cooled via heat radiation to the surrounding structure and by convection- - -

Yes. And in addition some speakers are designed to pump a little air thru the gap as the voice coil moves.

[Helmholtz]

So in simplest terms what I said is correct "Same gauge fewer turns means less power handling." Which anchor took offense to.

nosaj

Things are complicated because parameters are interdependent.
Max. speaker RMS current at middle frequencies (ca. 300..500Hz) is determined by its power and impedance ratings. The thermal power is given by I² times DCR. So in principle less DCR means less voice coil heating. Same gauge half turns number would meet the 3.2 Ohm DCR requirement for equal thermal power. But this would decrease voice coil volume and surface causing excessive VC temperature, in other words: less power handling.
Also to meet the 50% inductance requirement, turns number must be decreased by only 30%.
This means that wire gauge must be increased to get a DCR of 3.2 Ohm with 70% turns and to get enough voice coil volume and surface.

Most surprisingly my measurement on 16 Ohm and 8 Ohm Celestions confirmed the 2:1 ratio of DCRs but the inductance ratio was only 1.6:1 (@1kHz). This shows that there is some leeway regarding actual number of turns.

[anchor]

I just wanted to let everyone know it took a few tries but I learned how to repair speakers. The posts in here were a help, thanks to all.