tag:blogger.com,1999:blog-1249615655314260587.post4239855631344392943..comments2024-02-21T01:04:06.333-05:00Comments on Synth Hacker: Mystery Circuit - Polysix Post-Effects VCAChiphttp://www.blogger.com/profile/10352943033779293161noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-1249615655314260587.post-85592910272810365842013-02-28T06:52:50.594-05:002013-02-28T06:52:50.594-05:00On the Polysix Yahoo Groups, Johannes posted this ...On the Polysix Yahoo Groups, Johannes posted this very informative reply. I hope that he doesn't mind, but I'm posting it here for my own record keeping (when I want to find this later, it's a lot easier for me to find it on my own blog than to search the Yahoo Groups archive)...<br /><br />>>>>>>>><br /><br />Hi,<br /><br />The part of the schematic you posted in you blog clearly is a voltage<br />controlled low pass filter, 2 stages, therefore 12dB/oct. They had<br />such circuits in the 80s in tape decks and other Hi-Fi equipment. It was<br />then called "Dynamic Noise Reduction" or similar.<br />The analog delay lines used for the effects introduce quite some noise<br />and additionally the clock frequency. Apparently it was necessary to<br />put in substantial amount of circuitry to make the noise acceptable.<br /><br />Another measure for noise reduction is the compressor/expander system<br />built from NE571 (IC7, IC8). The NE571 consists of two identical<br />stage which can act, depending on the external circuitry, as compressor<br />or expander. The topmost circuit on the KLM-368 schematics (one half<br />of IC8) is the compressor. After that follows the input anti-aliasing<br />filter (Q13, Q25). Then the signal is split into three paths and enters<br />a BBD delay line for each phase. The BBDs are clock with a modulated<br />clock, which is built out of CMOS inverters used in linear amplfifier<br />mode (IC22). The modulating LFO is of the multivibraotr type and built<br />around IC18. DEpending on the type of effect the modulation depth,<br />frequency and audio paths are switched (4066 switch IC14). After going<br />through the torture of the BBDs the signals are again anti-alias<br />filtered (same type of filter, two transistors) and fed to the output<br />mixers and swsitches.<br />Then the sum signal goes through the discussed controlled low pass<br />filter and goes to KLM-369. OF course all this filtering and processing<br />will (potentionally) have a big impact on frequency response.<br />--<br />Thre three effect types Chorus, Phaser, and "Ensemble" result from<br />different delay times and different signal paths and modulations.<br />E.g. the third phase is only used in the "Ensemble" setting. It would<br />be interesting to hear how a flanger would sound, if they had provided<br />the possibilty for it. Has anybody on the list done that? The circuit<br />changes should not be too dramatically.<br />--<br />I always wondered why the Resonance control is part of this. Similar<br />to the output LPF there is a controlled (by the Resonance knob) LPF<br />at the input.<br /><br />All in all the KLM-368 offers a lot analog processing circuitry and<br />some unusual circuits (e.g. the modulated oscillator made of 4069<br />inverters). If you look at the board size it is larger than the<br />KLM-367 board; and this is just for the effects. But this was a<br />distinguishing factor at that time.<br />Today this all just a piece of software...<br /><br />Johannes<br />Chiphttps://www.blogger.com/profile/10352943033779293161noreply@blogger.comtag:blogger.com,1999:blog-1249615655314260587.post-89022585371862964782013-02-26T07:49:08.276-05:002013-02-26T07:49:08.276-05:00By examining my circuit, I just found that R107 is...By examining my circuit, I just found that R107 is actually 4.7M, not the 4.7K value shown in the schematic. This makes a pretty big difference! Now, that first gain stage won't rail at 15V. The circuit operates without going into saturation. It makes more sense now!Chiphttps://www.blogger.com/profile/10352943033779293161noreply@blogger.comtag:blogger.com,1999:blog-1249615655314260587.post-10194976975051075442013-02-25T17:06:07.050-05:002013-02-25T17:06:07.050-05:00Looking at the LM13600 data sheet, down through al...Looking at the LM13600 data sheet, down through all the example circuits, this use of the LM13600 does NOT match the circuit shape of a VCA. The negative feedback and the capacitor between the OTA and the darlignton pair actually make this look like a voltage-controlled low-pass filter.<br /><br />The app notes say that each stage of the LM13600 configured as a VCF generates a -6 dB/octave filter. So, in the Polysix, the two stages cascaded together would give a -12 dB/octave filter.<br /><br />Looking at the values for the resistors and capacitors, and knowing the control currect (Iabc), one can calculate the cutoff frequency of the filter. The circuit values are clear from the schematic. It's Iabc that's trickier. <br /><br />For Iabc, there's a tiny barely-legible note in the schematic shows that the IE is 1.2 mA, which is then knocked down to 1/3*IE (which is 0.4 mA) which is then divided between the two halves of the LM13600. As a result, the maximum (?) control current to each 13600 is 0.2 mA. Running this through the F_cutoff equation in the LM13600 datasheet, you get:<br /><br />Fc = (Ra*gm)/((R+Ra)*2*pi*C) <br /> where gm = 19.2*Iabc<br />Fc = (220*19.2*0.2e-3)/((10E3+200)*2*pi*470e-12)<br />Fc = 28 kHz<br /><br />This suggests that if the Polysix really drives this LM13600 to the 0.2 mA level suggested by the note in the schematic, that this element will open up to just beyond the range of human hearing (20 kHz). That's pretty reasonable.<br /><br />With less current drive, the 13600-based VCF will close down, cutting out the high frequencies (often noisy and hissy) first. Since the current drive appears to be based on the magnitude of the input audio signal, it suggests that the filter opens when the input is strong (loud) and closes when weak (quiet).<br /><br />Is this the behavior that the designers intended? Is this the right interpretation for this portion of the "Compressor/Expander" circuit? If so, I'm still troubled by the fact that the current drive circuit (the level detection circuits) seem to be railed at +15V. How are they doing their job?<br /><br />Thoughts?<br /><br />ChipChiphttps://www.blogger.com/profile/10352943033779293161noreply@blogger.com