ignorant crossover questions: looking for explanation

[Bob Simons]

I had some semi-custom speakers redone badly by a speaker builder who basically made the upper drivers much too loud in relation to the woofer. The crossover basically has two stages in parallel: the woofer circuit and the mid/tweet/supertweet stage. I'm looking to attenuate the upper range without changing the crossovert frequency of the two stages. I presume I need to know the actual impedence of this circuit at the crossover frequency, and then need to understand the formula for determining what the shunt resistance needs to be for any given series resistance to the upper end circuit. Is this correct, and if so, can anyone point me to anything that might make sense to me? What kind of equipment is needed for measuring impedance at a specific frequency (other than the signal generator to create the frequency in question?

The speakers started out as old Infinity Quantum 5s, with the dual voice coil Watkins woofers and with a dome tweet and an EMIT. The drivers were replaced and the crossover components were replaced with mugh higher grade parts and the new drivers were much more efficient than the former parts. In effect, they now have the upper end of the older Infinitys with the Walsh tweets, but with the woofer of the Q5s (except with better mids and tweets.) My initial hope is to knock the upper end down about 3 or 4 dB and then look into it closer.

Thanks in advance for any assistance.

 

[PeteMcK]

To measure impedance curve: Download ARTA: ARTA ARTA Download, Make this simple jig: ARTA Jig - http://zobsky.blogspot.com/2008/01/s...t-jig-for.html , use ARTA's LIMP module

Calculate L pad values: L pad calculator - attenuation dB damping impedance decibel loudspeaker speaker - sengpielaudio Sengpiel Berlin

 

[oldmanaudio]

There is many ways to do this, one is the L-PAD method as PeteMck described. Or you can keep it simple and add resistance before the tweeter leg of the crossover circuit. When you do it this way it has no effect on the crossover slope. The value shouldn't be more than a few ohms. You can buy a few sizes and add one and listen. If it's too low go down with the resistance of vise versa. Good Luck

 

[Bob Simons]

follow-up

If I remember correctly, the mids are Dynaudio AF-52s, which are at least nominally 8 ohms at the 600hz and up that they run at. Is there any value in trying to measure the actual impedance they have at 600hz, which may not be anywhere near the 8 ohms they read with a meter? I've long presumed that the impedance varies all over the place, not referring to the impedance rise at resonance frequency. According to the calculator PeteMck just referred me to, a 3db drop to the midrange would be (not surprisingly) a 50% difference between 4 and 8 ohm drivers.

Speaker Impedance Z: ohms Wanted Attenuation A: dB ↓ L-padSingle resistor R1 = Rserial: ohmsR: ohmsR2 = Rparallel: ohms Ztotal: ohms Ztotal: ohms A
(dB)Z
(ohms)R1(ohms)R2(ohms) −341.179.70 −642.004.02 −942.582.20−1243.001.34−1543.290.87 −382.3419.39 −683.998.04 −985.164.40−1285.992.68−1586.581.73
Is it sufficiently difficult to determine impedance at a given frequency (by testing) that it is beyond the reach of someone with a meter and a signal generator? Or is it sufficiently not worth the trouble so I should just try the 8 ohm numbers above, try it and then measure the results?

Thanks to both of you for your help.

 

[kbgl]

Start by checking for padding resistors that are in the circuit now, and compare with a chart or spreadsheet. Next make an estimate of how much you want to reduce the level, and refer to the formulas or spreadsheet. If the current padding looks like 2 dB per the chart, then look for values that would pad 5 dB and use them.

 

[speakerdoctor]

Besides the L-pad approach, there's two additional ways. One is a line stage BSC (Baffle Step Compensation) circuit. The main intent of this one is to act like an equalizer by tuning down the lower midrange to mid-high range of the audio spectrum. Changing of values of key components in the circuit can adjust the level of attenuation. The other is a simpler to implement shelving network installed on the positive side ahead of the passive crossover. Both are explained and the results of each are shown in a thread over at classicspeakerpages.net Advent discussion forum.
BSC vs Shelving - The Classic Speaker Pages Discussion Forums