[last modified: December 31, 2004] I had a problem with my Odyssey driving a DigiTech 7 second delay, here's how I solved the problem. First, the problem: The Odyssey and other analog synths can put out signals that are rich in high frequency harmonics. That's one of the things I like about analog synths, usually. When driving certain digital sampled effects such as my DigiTech delay, certain notes cause an inordinant amount of scratchy sounding distortion. In tracking down the problem, I devised a simple test where I played a note using a single oscillator and swept the note through the range of the VCO. Certain frequencies generated lots of grunge in the delay. To compound the problem, since I'm using the delay in a mode where I modulate its internal VCO, the offending frequency moves around. The solution: I built a simple LC pi type low pass filter. After a bit of experimentation with fixed values, I decided to build one which can be adjusted via switches. This is a passive design so it doesn't add a lot of hiss to the signal path. The grunge is effectively removed along with some high frequencies :-( but nothing is free anyway, is it? This circuit is useful for many filtering applications. The filter can be placed between the synth and the mixing board or better, right on the input to the digital delay so that the non-delayed signal is not low-passed. Here's a simplified ASCIImatic, It's a basic pi lowpass filter. The 1P5T switch is not shown. >-----------------\_a__ T1 ______ X ---------------< | )||( | | Input/Output C1= b__)||(__d | C2= Output/Input | )||( | | >--------------- c__)||(__e | ---------------< | | | | | |__________| | |_______________________| Capacitors C1 and C2 are both the same value, in my design I have a 2P3T switch set up so that they can be selected to be 0.001uf (1nf), 0.003uf (3nf), and 0.01uf (10nf). T1 is a center-tapped 600 ohm to 600 ohm transformer which was ripped out of an ancient 300 baud modem. T1 is cross-wired to form a tapped variable inductor. Other transformers will also work, the more taps the better. A second switch (1P5T in this case) goes from point X (wiper) to points a,b,c,d, and e on the transformer. I use 1/4" phone jacks for the input/output connectors. I mounted everything in a small metal box to keep the hum out. Note that the circuit is symmetrical so either port can be the input or output. Two of these circuits could be cascaded for a sharper cutoff. Adding a 1 to 10K resistor on the input side of the circuit may improve the low-pass performance, it will also reduce the volume a bit. Use: Connect the filter between the synth and the effect, start with the minimal capacitance and inductance, then tweak the values until the aliasing noise disappears.