Passive Mastering EQ

[ruffrecords]

This is a placeholder for a reboot of the Passive Mastering EQ project I began back in 2017 (how time flies). The original thread petered in out in 2019 mainly because I had a lead customer who was paying for the materials and driving the specification. I won't go into details but suffice to say that relationship has ended and I returned their money. Shortly after that another customer appeared and I did build and deliver a special mastering version of the EQ for him. Since then I have designed, built and delivered three different custom tube consoles that last of which I shipped last week. Now, at last I have some free time to restart this.

More later,

Cheers

Ian

[capacitor]

I can't find a thumbs-up button, so doing it with a reply instead Ian, I'm also moving on from the former place and for probably the same reason. Thanks for everything you do. I've learned a lot from you and continue to do so.

[scott2000]

+1

[ruffrecords]

It is great to be here. I feel a lot more relaxed already - but that could partly be because I am watching Creedance Clearwater Revival at the Albert Hall in 1970

Cheers

Ian

[AnalogJoe]

It is great to be here. I feel a lot more relaxed already - but that could partly be because I am watching Creedance Clearwater Revival at the Albert Hall in 1970

Cheers

Ian

John Fogerty is hands down my favorite singer.

[ruffrecords]

OK, here is a summary of the basis of the design:

1. The EQ is based on a single resistor ladder shown on the left.
2. Input is between the top and the bottom of the ladder and the output is taken from the half way point.
3. The total resistance from the top to the half way point is the same as the resistance from the half way point to ground.
4. The insertion lass is therefore 6dB as is the maximum boost.
5. The ladder is wired across to any number a rotary switches, one per band.
6. There is no limit to the number of bands but there are caveats (see later)'
7. Boost is achieved by connecting a frequency selective network between the mid point and one of the upper switch position.
8. Cut is achieved by connecting a frequency selective network between the mid point and one of the lower switch positions.
9. With all switched centred the response is flat.
10. A single CR network can be used for a high frequency shelving response.
11. A single LR network can be used for a low frequency shelving response.
12. An LCR network can be used for a bell shaped response.
13. For the most consistent results, the EQ should be driven by a low impedance source and drive a high impedance load.

Practicalities and Caveats

a). If the top and bottom halves of the resistor ladder each total 600 ohms, then the worst case load occurs at maximum boost. This means the load presented by the EQ will eve be less than 600 ohms although it can vary from 1200 ohms to 600 ohm.
b). The nominal output impedance is 300 ohms (600//600) or less,
c). The above two factors mean it should be possible to operate it as a completely passive device driven by any source capable of driving a 600 ohm load and feeding any 10K inpu
d). With any network including an inductor the amount of boost will be limited by the resistance of the inductor.
e). For a mastering EQ, 6dB of boost/cut is rarely if ever required. 3dB or maybe 4dB is the normal maximum
f). Although any number of bands can be used, there can never be more than 6dB boost at any frequency. This means that where bands overlap there will be interacti0on between them. This may or may not be desirable.
g). Since perhaps only 3dB of boost is required, series resistors can deliberately be added the LCR network to limit the maximum boost to this value. This has the added advantage that it reduces the interaction between adjacent bands by allowing their maximum boosts to add. Again, this may or may not be desirable.
h). Busing any number of 12 way (or greater) rotary switches together does not lend itself to hand wiring. A PCB is a possible way to go but that will fix the number of bands and the spacing between them. A daisy chain ribbon might be an alternative but I would be concerned about stray capacitance affecting the response.
i0: There is no reason why each band should not have a range of selectable frequencies. There are only two wires connecting the band switch to the LCR circuit so it should be straightforward to create ban frequency select PCBs

Cheers

Ian

[mediatechnology]

Ian is this topology similar to LC graphic EQs?

[ruffrecords]

Ian is this topology similar to LC graphic EQs?

Probably not but then again there are lots of LC graphic EQs so who knows? What I will say is that most graphic EQs I have come across use what is sometimes referred to as a "swinging EQ" where there is a potentiometer across the plus and minus inputs of an op amp and the pot wiper is wired though an LC(R) to ground. There is of course a feedback resistor and there is also an equal value resistor in series with the plus input. At one extreme the wiper connects the RLC to the plus input where with the series input resistor it acts as resonant attenuator (max attenuation set by the value of R), and at the other extreme it connects the minus input to ground via a resistor and thus acts as a resonant boost (max boost again set by the value of R). You get various values of boost or cut as the wiper "swings" from one extreme to the other.

My EQ is completely different.

Cheers

ian

[mediatechnology]

What came to mind were the White Instruments 4000 units.

[ruffrecords]

What came to mind were the White Instruments 4000 units.

OK, they are LC EQs alright but active rather than passive. It does not specify a Q other than to say it is optimised for summation with adjacent bands, but given they are one third octave apart it will have to be greater than one. Also the range is plus/minus 10dB which is too much for a mastering EQ. So, overall not like my EQ at all

Cheers

Ian