Audio for 914 comparison

Here is audio comparing my 914 to Yves Usson's design. I compared the inductor version. There are a couple of things to remember:

1. Yves uses two cascaded filter cells, each one having the same cutoff frequency but different "Q". Mine has two cascaded filter cells as well but also follows the Moog where the first cell's cutoff is a little lower than the second cell's cutoff. The "Q" of the second cell is essentially fixed with a resistor and the trimmer on the input to the first cell is adjusted to change the overall "Q" to get a "Q" of 3.7 (per Yves' research) which seems consistent with the component values on the Moog schematics (although there are a couple of errors on those schematics, no surprise). One glaring is a missing resistor on the second cell which sets the "Q". Somehow, Jurgen missed this. I caught it when I was unable to get a close visual match to the Moog frequency response curves so I did a detailed study of the only photographs I could find of the Moog 914 and worked through the math.
2. The Moog has a fixed high pass filter with an effective cutoff of about 60Hz (if memory serves me) between the summing node of the filter cells and the input to the output amplifier. The 907 did not have this and I don't think Yves included this on his 914, but I could be wrong. Jurgen included it on his 914. This affects the low pass cell response.
3. Yves changed the 700Hz cell to a center frequency of 750Hz.
4. I think the actual cutoff frequencies for the HP and LP cells are a bit different in my 914 as opposed to Yves.

In the audio, my 914 is in the left channel and Yves' is in the right. There is some kind of funky echo or ray gun sound at the beginning of the audio. I have no clue what artifact it is. I recorded the two using a Tascam DP-24. The output of each filter was fed directly into the DP-24 with no additional EQ, or other modification. I'm driving the filters with the same sawtooth from a Steiner Type "A" VCO whose frequency is driven by a sine wave from a Steiner Type "B" VCO. The resulting audio files were mixed down to a stereo track using the DP-24 then uploaded here.

I start with all cells and let you hear the YU 914 alone, then mine. I then take out each cell one at a time HF down. I then compare each cell one at a time. I end by adding all the cells back in and then letting each filter play on its own again, YU first then mine. Saw tooth with frequency swept by a sine wave.

The most exciting 4 minutes of synthesizer sounds you will ever hear! :)

Here is a link to the uncompressed audio. It's 40 MBytes.

AR 914 v. YUSynth 914

Synthasystem VCF

Ronny just finished a build and here are some photos and excellent modifications he made. Thanks for sharing Ronny.

Ronny wrote:
"- I added a second CV potentiometer.

"- I omitted the switch and added three separate attenuator potentiometers (25k)
  for the three inputs (LP, BP, HP). They sit before the 47k resistors. The 47k resistors
  connect directly to the former "switch outputs" now. This enables to get a kind
  of all-pass characteristic when mixing the same signal to multiple inputs or to get
  a kind of frequency-dependent interpolation with different input signals. I also like
  to add a bit of HP while primarily using the LP. I normailzed the "BP in" and "HP in"
  connectors to "LP in".

"- I used 330nF instead of 1.2uF for C10 at CV in to reduce drifting with steady CV
  and to enable faster modulation. Maybe I reduce it even further.

"- I drilled a hole into the Power PCB to easily adjust the "CV reject" trimmer without
  taking the filter apart. Luckily there was just a ground plane at this position on
  the power PCB that could be drilled without damaging something.

"- I had to omit the nice Steiner wave logo on the frontpanel, but added a little hint
  to you "DI" and me "RV" :-)

"Here are some other build notes that might be of interest for you:

"- feed it with +-12V in a Doepfer Eurorack system, but reduce -12V to -10V with
  your Power PCB, also use the filtering on the Power PCB (V+, V-)

"- used SSM2210 and 2N5172 as specified
 
"- used two 2N2907 with HFE of about 140 instead of the PN5138 (HFE measured
  with a  medium-class multi-meter)
 
"- had to use linear potentiometers for everything, because I did not get logarithmic
  ones of this size, but this still feels fine

"- used a trimmer having 200 Ohm instead of 250 Ohm, works fine

"==> self resonance starts at about 2 o'clock resonance potentiometer position, this is fine
" ==> have no problems with power noise, crosstalk or whatever"

Thanks again Ronny.