Monacor SPH-225C - advanced current drive

[steph_tsf]

hello,

I've just completed a LTspiceIV toolbox for simulating some elaborate active loudspeakers. It is amazing to see the bandwith extension one can get using some kind of managed current drive.

The current drive annihilates the 1st order lowpass at high frequency caused by the coil inductance. This is the reason why the bandwith now extends beyond 10 kHz. I definitely know this is a pure theoretical curve. In a real application, a high frequency roll-off may be desirable for compensating some real cone resonances above 2 kHz.

The deep bass voltage feedback annihilates the closed box resonances. You can see the effect of the deep bass voltage feedback when varying the feedback capacitor. There is an optimal value around 220nF where the bass is at -3dB at 30 Hz, an no closed box resonance anymore.

And we have not yet used motional feedback (coil back EMF extraction or piezo acceleration sensor). Actually, there is a piezo sensor model and schematic in the toolkit (Thiele-Small Reloaded with LTspiceIV.zip). Use this !

You may also use the toolkit for modelling the effect of a passive filter inserted between a power amplifier and a closed box loudspeaker.
Later on I'll release a 4th order bass-reflex toolkit and possibly also a 4th order bandpass (double chamber) toolkit.

See attached files.
Any suggestion or question welcome.

Steph

[infinia]

Hi Steph
Yes sadly this does not happen in real life.
I've seen interestingly, that modeling Le and Re using a first approximation ie 1st order roll-off and the actual cones response off axis have some good correlation. Also we know that cones off-axis frequency roll-off is related to the of departure of the cones simple pistonic frequency range.
I think your work is good so far as I have toyed a little with it using LTspice. but I think most folks in this forum can't relate too much to these particular sims ie current or motional feedback, Concentrating on the driver only model first might be a good idea to this audience? Can you show that your driver model is true, before feedback is applied. Probably the best thing here is to perfect the driver model first using a real driver in a closed box. Seems you should be able to get good correlation to the accepted software modeling simulation programs and even possibly verifying them with open loop measurements using current and voltage drive. What do you think?

[steph_tsf]

Originally Posted by infinia

Can you show that your driver model is true, before feedback is applied ? Probably the best thing here is to perfect the driver model first using a real driver in a closed box. Seems you should be able to get good correlation to the accepted software modeling simulation programs and even possibly verifying them with open loop measurements using current and voltage drive. What do you think ?

Excellent suggestion ! Please post the names of three different accepted software modeling simulation programs, please post the names of three different drivers to be simulated, and I'll compare the simulation results.
Steph

[infinia]

T/S simulation software (driver data base may or maynot be accurate)
Unibox
WinISD
BassBox Pro available here

I have used Unibox and WinISD some/ not tried BassBox yet.

We could use your Monocor SPH-225C and modest 8" Silver Flute W20RC38-08 Silver Flute W20RC38-08, 8" Wool Cone from Madisound

[speaker dave]

Originally Posted by steph_tsf

hello,


The current drive annihilates the 1st order lowpass at high frequency caused by the coil inductance. This is the reason why the bandwith now extends beyond 10 kHz. I definitely know this is a pure theoretical curve. In a real application, a high frequency roll-off may be desirable for compensating some real cone resonances above 2 kHz.


Steph

I have noticed the same when looking at a large number of full range driver response and impedance curves. The response generally tracks the inverse of the impedance curve. Current drive, at least at the top end, will generally give a flatter response.

Makes sense from and F = BLi = ma point of view. Constant current give constant force or acceleration. Very generally, a curve off axis, say 45 degrees off, becomes flattest on a lot of drivers, in current drive.

I think Nelson Pass published something on this.

David

[infinia]

Unibox Monacor SPH-225C
closed box 12 liters no fill/leaks

[infinia]

added lossy BF inductor in series ~ note attenuation from above

[steph_tsf]

just for fun, here is a bass-reflex with the Monacor SPH-225C

[steph_tsf]

just for fun, here is a bass-reflex with another driver, the Monacor SPH-255

[infinia]

That 10" looks nice for the money
How do we relate the model to physical parameters?