The Sketchy Labs Frequency Shifter is finished:
Also, mixing the output back in with the audio signal gives a very nice phaser effect when the VCO is run in LFO mode.
So, the panel turned out perfectly. This is actually the second panel I made -- the first one had some textual errors -- I tried to fix them by scratching off the offending text and filling the holes with clear nail polish, but then I made the fatal mistake of rebaking the panel to get rid of some roughness, and the nail polish burned and turned black, and actually ruined the aluminum underneath. I ended up throwing the whole thing away and starting over.
Also, this is the first module that I've built in a while that worked perfectly the very first time -- no design mistakes, no layout errors, no solder bridges. This entire build took me about two days. I've become a bit of a module-building fiend.
As you can see from the rear-view picture, this build consists of five PCBs. They are all 2.1" wide and 5.7" long -- I was able to get the four circuit boards from a single 6" x 8" blank. The outer two boards are the Rubiquad circuit (a simplified TZFM dual-core Rubicon VCO set up to generate quadrature waveforms such as sine and cosine) split onto two boards, the third board is the dual four-quadrant multiplier and the input and output amplifiers, and the fourth board is the 90-degree phase displacement network. The board attached to the panel is the panel board, which has all the pots and switches, and also has an analog multiplexer for processing the four different waveforms of the Rubiquad VCO. This is actually kinda cool -- the two VCO WAVE switches actually represent two digital bits driving a DG409 dual four-channel multiplexer. When the switches are both down (0 0), one gets sine and cosine, when the left switch is up and the right is down (1 0) one gets triangle and cotriangle, in the opposite situation (0 1) one gets zigzag and cozigzag (or zig and zag), and when both switches are up (1 1) one gets square and cosquare (in this context, the prefix "co-" just means that that wave is 90 degrees out of phase with the other). The zigzag wave is something I added to this build and it gives very pleasing results, especially in phaser mode.
Before I tried testing the whole module (which is scary), I tested all the individual parts.
-- The VCO worked perfectly but I still have to finish adjusting the sine shaper and tune it for 1V/octave (I'll do that tomorrow).
-- The dual multipliers give what look like perfect four-quadrant multiplication (as they should, since I hand-selected all the gain resistors to within 0.1%). The dual quadrature multiplication is very cool, especially with the zigzag waveform. I'll make sure to include that in tomorrow's video.
-- The 90-degree PDN is especially awesome. I tested it by feeding it a sine wave from my Rubicon and scanning the frequency from 15Hz to 15kHz by playing octaves and turning the octave selector switch, and measuring the two outputs on my oscilloscope in XY mode. If the PDN is working well, one should get a circular Lissajous figure across the entire frequency range. I got a perfect circle and it didn't budge at all as I changed the octaves -- it was only very slightly deformed at the highest frequency of 15kHz. I compared this with my previous PDN, and I think that maybe this one is just a little bit better than that one. I hand-selected the 12 film capacitors (four each of 102, 103 and 104) to get virtually perfect values. The resistors are 1% tolerance, just randomly pulled from the bags (I have a little kit of resistor values just for this PDN design -- 22 very specific values which, when combined with the three cap values, give 12 allpass filters with very precise corner frequencies -- I can build 10 of these before I have to order more resistors).
Stay tuned for a video demonstration of this beast, which I will try to achieve tomorrow afternoon.
This is a Bode frequency shifter. This thing is effectively an analog engine for carrying out a sum-of-products trigonometric identity. It consists of four main parts: a quadrature VCO for generating sine and cosine carrier waves, a 90-degree phase displacement network for converting any audio signal into its component sine and cosine equivalents, a dual four-quadrant multiplier for multiplying the sines and cosines of the carrier wave and audio signal together, and an output amplifier for generating the sum and difference of the two multiplier outputs. Doing all this manifests a trigonometric identity, as shown in the following diagram, and this identity guarantees that the frequency of the audio signal is shifted by the frequency of the carrier wave, either up or down:
You do not have the required permissions to view the files attached to this post.
A dewdrop can exalt us like the music of the sun.