horn loaded driver Z

[AndrewT]

Hi,
if a horn loaded driver is 50% efficient, then half the input power is converted to heat in the voice coil and the other half is converted to acoustic energy.

Let's suppose we have used an 8ohm driver with an Re=6r0 in this near perfect horn.

What is the effective load impedance seen by the amplifier when operating in the passband of the horn loaded driver?
Could it be 8ohm or 12ohm or 16ohm?
or some other value?

[Iain McNeill]

Interesting question.
I always thought the horn was a "transformer" matching the acoustic impedance of free-space to the acoustic source impedance of the motor. So rather than the resonant peak(s) of a box enclosure I'd expect a step up in impedance at the low frequency limit of the horn to a parallel combination of transformed free space impedance and the motor impedance. Never measured one though.

Can't wait to see a graph. anyone?

[phase_accurate]

The measured impedance will certainly rise. Programs like hornresponse can model this.

Regards

Charles

[GM]

Greets!

In theory Re doubles, or 12 ohms in this example.

GM

[AndrewT]

Thanks GM,
why does horn resp not show this? It does not seem to show on the impedance graph.

Since most of our modern amplifiers are constant voltage, then a 100W into 8ohm amplifier will deliver that same 40Vpk into 12ohm rather than the 8ohms normally assumed.
Does that mean we have 33W of heat and 33W of acoustic power?

A bit better than 98W of heat and 2W of acoustic power.

Is this why Tom Danley says you can use a 2kW amplifier to drive the Labhorn?

[GM]

You're welcome!

Dunno, my experience with Hornresp is very limited since I find it such an infuriating/frustrating experience to use and I feel my quality time is better 'spent' on more productive 'projects'. My mind just doesn't 'process'? like a programmer's?/how it must be programmed?/whatever. That said, most folks don't know how to design a 'perfect' horn (including some 'pro' horn designers if they're not intentionally playing 'dumb'), so even if Hornresp will accurately predict one, it's only as 'smart' as the person doing the inputting. Hopefully you'll pardon me for not going into detailed theory, but FWIW, attached is the closest I could come using MJK's original horn Mathcad WS. I haven't taken the time yet to research his latest one, so don't know if I can improve on it or not.

Since I had no use for the LABhorn I didn't follow its thread, etc., to any great extent, but being a TD design I assume it's a Leach math optimized alignment and it's my understanding that its dual drivers are basically 'bullet-proofed' single coil Avatar/Adire Shivas, so yeah, it should be able to handle plenty of power if wired in parallel and driven with a typical vanishingly low output impedance SS amp.

GM

[catapult]

It's more complicated than a simple rise. Most horns, even relatively benign "waveguides," introduce a number of lumps/peaks in the impedance curve. Borrowing some pics from JonMarsh at HTGuide.....

DDS ENG1 waveguide and BMS4540 driver:



Conjugate network to flatten impedance (no crossover yet):



Impedance with network:

[Jack Hidley]

As already stated the minimum impedance is going to double to 12ohms. Since you are talking about a 50% efficient system, by definition half the power is going into the voice coil as heat, the other half goes into the air as acoustic power. This means there can't be any other losses in the system. Some of the resonances in the system will therefore have infinite Qs, since there is no damping on them. The impedance curve will have peaks into the hundreds of ohms range.

[Iain McNeill]

can someone remind me what the advantage is here????

[David McBean]

Originally posted by AndrewT
Thanks GM,
why does horn resp not show this? It does not seem to show on the impedance graph.

Hi AndrewT,

As far as I can see, Hornresp does show exactly this :-). By way of illustration, calculate the results for default record number 1 and sample the SPL response at 200 hertz. The calculated system efficiency is 44.99 percent (near enough to 50% for the purpose of the exercise). Now sample the electrical impedance at 200 hertz - the calculated magnitude is 11.9551 ohms (near enough to 12 ohms). The electrical impedance is therefore twice the 6 ohm driver voice coil dc resistance under these conditions, as indicated by GM and others.

Kind regards,

David