me
New member
For the last couple weeks I've been trying to reverse engineer the Speck EQ16 circuit from the user manual and a PCB photo I found online. It's a 3 band (50-500Hz, 500-5k, 5k-15k) semi-parametric EQ with +/- 15dB of gain. Interestingly, the bandwidth is variable and coupled to the frequency selection control: at full CCW the Q is 3.3; at full CW the Q is 0.8.
After some reading and experimentation I think it's a Wien network-based design. The low and mid bands are the same circuit with different capacitor values (in the Speck SSM-24 the low and mid EQ bands share one switchable control) and the high band is tweaked to reduce the frequency range.
I've found a single-gang variable frequency Wien network-ish design in Small Signal Audio Design (2010) and I've been able to play around with it in LTspice to get a response that's fairly close, but I'm getting stuck:
I'd love a few pointers from folks with a bit more experience if you have ideas about how I could get closer to the real thing. I'm including the SPICE model if anyone wants to play around with it. I got interested in the circuit as a fan of Moritz von Oswald (Basic Channel et. al) who had something like 96 channels of this EQ in the studio around that time (3x 16-channel mixers + 3x 16-channel rackmount EQs). Eventually I'd love to DIY something close enough that I can experiment with it in the studio myself.
After some reading and experimentation I think it's a Wien network-based design. The low and mid bands are the same circuit with different capacitor values (in the Speck SSM-24 the low and mid EQ bands share one switchable control) and the high band is tweaked to reduce the frequency range.
I've found a single-gang variable frequency Wien network-ish design in Small Signal Audio Design (2010) and I've been able to play around with it in LTspice to get a response that's fairly close, but I'm getting stuck:
- I can't match the resonance behavior across the frequency range (I can get closer to a consistent 15 db boost across the range, but only with wider bandwidths)
- When I try to use the component values from the PCB, I can't get anywhere close to the same behavior. I don't understand how the circuit works well enough to change anything much beyond tweaking values and seeing how the model responds.
I'd love a few pointers from folks with a bit more experience if you have ideas about how I could get closer to the real thing. I'm including the SPICE model if anyone wants to play around with it. I got interested in the circuit as a fan of Moritz von Oswald (Basic Channel et. al) who had something like 96 channels of this EQ in the studio around that time (3x 16-channel mixers + 3x 16-channel rackmount EQs). Eventually I'd love to DIY something close enough that I can experiment with it in the studio myself.
