L-C EQ problems

[dmp]

So I'm building an EQ that is passive L-C design. It sounds very musical and is mostly working as it should (frequencies are correct) BUT it is different than the original at a few bands in listening tests.
This is a circuit for the section in question:

There is a mid (1kHz) and a mid-high (5kHz). The 5kHz sounds the same between my build and the original, but the 1k Hz sounds different. They use different inductors. Both inductors for my prototype were wound on RM8 cores.
I used RMAA to compare the frequency response (LEFT CHANNEL=my prototype, RIGHT CHANNEL=original).
The 1kHz boost and cut show that the prototype boosts or cuts the higher frequencies beyond the filter area. In a listening test, it sounded like the prototype was affecting more frequencies, with which this is conducive. What in the circuit might cause this?
For both units, the EQ fader was at max. This should give 15dB boost or cut, as labeled on the original. The original has slightly more boost/cut than the prototype as well. With the faders centered, both units sound the same and have a pretty flat response.



The 5kHz boost and cut also show some differences between the original and the prototype, although they sounded pretty much the same.


So I'm hoping some expert can give me some ideas on what experiment with in the circuit to find a controlling factor for this behavior...
Thanks!

 

[ruffrecords]

I am no expert on this type of EQ but here is what I observe.

1. This is not a passive EQ - it uses negative feedback for its effect.

2. The differences you see are almost certainly due to component tolerances most likely in the inductors. The max boost/cut is determined by the 1.5K series resistors and the series resistance of the inductors. I suspect your inductors have higher series resistance than the original.

Cheers

Ian

 

[dmp]

I'm going to try inductors made with a different core. Thanks for the reply.

 

[warpie]

I'm going to try inductors made with a different core. Thanks for the reply.

why don't you try to use a smaller resistor instead of 1.5K?

 

[dmp]

A smaller resistor would narrow  (reduce) Q, yes?
I'll experiment with that also.

 

[ruffrecords]

A smaller resistor would narrow  (reduce) Q, yes?
I'll experiment with that also.

The Q depends on the total series resistance as does the max boost/cut. The total series resistance is the 1.5K resistor plus the winding resistance of your inductors.  Since you see less than the advertised max boost and cut then it is reasonable to assume your total series resistance is too high and so both your max cut/boost AND Q are reduced. Reducing the value of the 1.5K until you get the right max boost/cut should also give you the correct Q.

Cheers

Ian

 

[dmp]

I adjusted the series resistance to smaller values and observed the change in the frequency behavior. As expected, the max boost and Q increase as the series resistance decreases.
The problem with the +4 dB increase on higher frequencies beyond the boost remains.

 

[warpie]

The problem with the +4 dB increase on higher frequencies beyond the boost remains.

I know it's obvious but have you double checked that all four resistors are 270 Ohms?

 

[dmp]

Yes, the resistors are all within +/- 1 ohm. And the response is flat with the pots centered, which leads me to believe the rest of the circuit is fine. And I am getting another set of inductors made by ChrioN that use the same core material as in the original (N30). With this material, less turns of wire are required, and the dcr should be less. Hopefully the odd behavior will go away.

 

[ruffrecords]

I adjusted the series resistance to smaller values and observed the change in the frequency behavior. As expected, the max boost and Q increase as the series resistance decreases.
The problem with the +4 dB increase on higher frequencies beyond the boost remains.

The only thing I can think of that would explain this is the winding leakage capacitance of the inductor. At full boost and high frequencies (well past the resonant frequency) the capacitor is effectively a short circuit but the impedance of the inductor should keep rising so the response remains flat. However, any stray capacitance across the inductor will lead to a rise in frequency response at high frequencies.

Cheers

Ian