From HJ2743@denbgm3xm.scnn1.msmgate.m30x.nbg.scn.de Tue Apr 23 09:06:27 1996 From: HJ2743@denbgm3xm.scnn1.msmgate.m30x.nbg.scn.de Date: Tue, 23 Apr 96 12:06:13 -0400 23 Apr 96 12:06:11 +0500 22 Apr 96 12:57:00 +0500 2 Apr 1996 16:12:48 GMT id sma024045; Mon, 22 Apr 96 18:12:21 +0200 id sma001207; Mon Apr 22 18:11:27 1996 id <317C2E3E@MSMGATE.M30X.NBG.SCN.DE>; Mon, 22 Apr 96 18:11:26 PDT From: Haible_Juergen#Tel2743 Subject: Error in OB-8 service manual Date: Mon, 22 Apr 96 18:10:00 PDT Encoding: 35 TEXT Hi! Someone has asked me about this, and I didn't know an answer back then, though I had encountered the same problem as well: It's about the envelope time calibration in the OB-8. In the manual it says that you have to adjust some voltages at some pins to 0.000 volt (in calibration mode). This is wrong. In my OB-8 (as well as in at lest one other one) you cannot even reach 0 volts, so what I did so far was matching the voltages for the whole synth at least. Now I found out that I had ended up with far too slow envelopes (I often asked myself: Why is the attack so tame in spite of hardware envelopes?) You get the fastest envelopes with 0 Volt at the control pins of the CEM3310's (higher voltages are not allowed). With 0 Volt at the control pins you get a *slightly* slower time than in a Prophet for example, but the difference is minimal (the Prophet uses a smaller capacitor). The point is, that 0V in calibration mode do *NOT* corespond to 0V in normal running mode !!! I did some iterations: Setting some value in calibration mode, storing it (important!), then switch to nornal mode, turn attack pot fully ccw, and measure the actual value. After some iterations I found that at -250mV in calibration mode I had the desired 0V in normal mode. This was the right value for *all* time constants in my machine, so probably it's just a software error and applies to all OB-8's. At least, the -250mV would be a good value to start with. Yes, my OB-8 can say "JUMP" now. (;->) JH.