Impedance compensation circuit design

Status
Not open for further replies.
Jump to Latest
globalplayer

globalplayer

Member
Joined 2014
#1
I want to get a flat speaker impedance curve.

So I will use a zobel for the top end, R and C in series, parallel to speaker.
(This one is no problem)

For the lower impedance peak I will use R, L, C in series, parallel to speaker.
(Not that sure how to calculate that one yet.)

Maybe it is a stupid question, but can I get away with just one R?
 
#2
You mean an LC series network (let's call it L1 C1) in parallel with a C (C2) and then one R in series with the whole thing? That will result in an extra resonance of L1 with C1 and C2 together. The impedance of your correction network ideally goes to infinity at f = 1/(2 pi sqrt(L1 C1 C2/(C1 + C2))), but I have no idea whether that would be a problem.
 
#4
You can design such a correction using XSim. With a little help from BYRTT I did something similar for my arrays.

http://www.diyaudio.com/forums/full-range/242171-towers-25-driver-range-line-array-116.html#post4323229

impedance-nw.jpg

The XSim schematics

It changed the impedance plot from the blue into the purple curve:
finalimpandcorrected.jpg

Comparison of phase before and after correction

I later removed the network correcting the top end. I liked the sound better without the top end correction.
Based on subjective listening impressions I kept the correction at the impedance peak. Its still there after years.

correction2.jpg

This plot is showing the impedance and it's electrical phase

With my use, using EQ below Fs and being an array etc, might not be a good example to compare to a single driver though.
Just evaluate your own preference of such a network.
The technique will be similar. The impedance plot of the arrays looks like one from a single driver.

This might give you something to start with. In X-Sim you can see how much energy goes trough each component, so it's easy to see what you need, spec wise.
 
Last edited:
#6
globalplayer said:
Maybe it is a stupid question, but can I get away with just one R?
Click to expand...

Well, if you think about it; A resistor does not change its resistance with frequency so it will be impossible to make a frequency specific compensation with a resistor only. Now, impedance compensation for a tweeter is common place but doing the same for woofers are rare and thinking about it I never saw that. It's probably due to that it will take really big inductors as the peak is usually at such a low frequency (30-60Hz?). Low frequency => big inductors and big capacitors => expansive...

Why you want to make a tweeter impedance linear is to ensure that the high-pass filter will work the best and not to make the load linear to the amplifier really... just so you know.

//
 
#8
TNT said:
Well, if you think about it; A resistor does not change its resistance with frequency so it will be impossible to make a frequency specific compensation with a resistor only. Now, impedance compensation for a tweeter is common place but doing the same for woofers are rare and thinking about it I never saw that. It's probably due to that it will take really big inductors as the peak is usually at such a low frequency (30-60Hz?). Low frequency => big inductors and big capacitors => expansive...

Why you want to make a tweeter impedance linear is to ensure that the high-pass filter will work the best and not to make the load linear to the amplifier really... just so you know.

//
Click to expand...

Thanks.
Yes I know.
But I want/need to hear it.
I don`t care about cost.
 
Status
Not open for further replies.
Top Bottom