JH-4 answers and general desription

Haible Juergen Juergen.Haible at nbgm.siemens.de
Wed Jun 11 14:09:05 CEST 1997



>I have a few questions regarding this -

>1. What gauge were the ribbon cable wires?

Hmm, "gauge" that's the thickness of the wires, isn't it? Honestly, I
don't know.
I used this standard 50-pole ribbon cable that fits into these standard
pcb connectors
(not the high density types). I think the same one is used in the OB-8
for example.
Very tiny stuff.

>2. Do you notice crosstalk between the wires?
>3. Did you put ground wires between each signal wire on the cables to
>cross talk?

No, but there isn't much to crosstalk. The only fast signals that go
over the connectors
are the 2 envelopes, the mixed VCO signals, and the LFO which can go up
to audio
rate. The rest is mostly DC voltages from the potentiometer board. Each
of these
CVs is low pass filtered on its destination (right before the buffer
opamp), so there
is no problem with crosstalk.

>4. Did you mount your jacks on the PCB as well?

No. The idea is to have a building block "a la SEM", so the whole thing
can be used 
separately or as a n-Voice polysynth. I used some trick for the output
signals, though:
You can directly connect many voices together in parallel *without* an
extra summing
amplifier, so in my "4-voice" (Working title (;->) ) I will maybe have
individual outputs
with normalized jacks. If only one pair of output jacks is used, the
outputs of all voices
are simply shorted together.
This is possible with this Minimoog-VCA architecture, which I used.
(High current differential
pnp pair with output at the collector). It even can drive 100 Ohm
Headphones at a decent

>5. I assume the pot board is mounted in a plane parallel to the

Yes. All boards are parallel on top of each other, and will be parallel
to the faceplate.
Think of it as a large Bic Mac (if the SEM is a Hamburger).

>  then you do have a very little amount of pots for everything, don't
you?  Or is
*	everything hardwired, as opposed to being modular?

Both is true. Little amount of pots, and not modular:
It is modular in a sense of one module including knobs for each voice of
a polyphonic,
but not modular in a sense of a Modular synth or a "modularized" (sp)
SEM with all
internal connections brought out.
As the polyphonic synth will have 4 complete sets of knobs, and you
still want to be able
to tweak them all in a resonable time, knob economy was a main design
goal. My philosophy
was: Don't give it much more knobs than a SEM, but stuff as much
functions and electronics
into the engine as possible. The result was some overkill on the
electronics side, sometimes.
Take the LFO, for example:
It would have been easy to switch waveforms with a rotary switch, and
have an extra knob 
for LFO frequency modulation, but I chose a different way. There is one
which selects the Sine Waveform when in 12-o-clock position. Turn the
pot counter clock
wise, and the LFO frequency (with sine waveform) is modulated by
envelope 1. Turn the
same knob clockwise, and you fade thru different waveforms. No frequency
modulation, in that 
case, but well, that was a compromise I have deliberately chosen.
Waveform fades from Sine to Square to Sample & Hold to ... well
something special (see below).
Especially the position between square and S&H is very nice, as it
sounds like a S&H with altered
distribution function for the random levels - very useful for VCF
The full clock wise position (still speaking of the same knob !) chooses
the saw wave of VCO 1, so
I can even have audio rate modulations (on filter, or VCO 2, or panorama
...) without the need
of an extra knob. You get the idea.
Many of the knobs work in a similar way. Think of the original SEM's VCO
modulation: Turning
the knob right performs PWM, turning it left performs FM.

> I'm also curious about the Morphing portion of the LFO.  Is this only
using VC
> over the amplitudes of the different LFO waveshapes coming into a
> mixer?  That one would be easy, I think.  But it's the only way to do
it, that
*	comes to mind at the moment.

Yes, I've done it that way, and yes, it's rather easy (summing up OTA
outputs), but it needs
a large amount of components. The VCLFO core is only a small part of the
LFO. Most
components are used for mixing, and generating the triangle functions
for mixing and FM from
the CV of one single pot.

>Another reason that I'm firing off email though, is the VCLFO.  :)
I've got a
>couple of different versions in Ernie, that I've built.  (Both of these
>adding VCs to existing LFO designs).  I still have a change in the
waveshape of
>my LFOs, when I VC them.  I've been wanting to get away from that,
having a
>constant waveshape, no matter what voltage is coming in from the VC
>Have you succeeded in something like this?

It's a simple triangle VCO design: One 3080 as integrator, one 3080 as
switch. Simple (not temperature compensated) 2-transistor exponential
current source.

>It's kind of late now, but I just thought that you may have been
>VCO wise, in my split wave technique that I came up with for the most
>version of my VCOs.  This allows the user to select the Positive and
>voltage halves of waveforms, that are then mixed together at the
outputs.  (I
>have 2 sections of this in my latest VCO version.)  This allows mixing
>halves of each waveform:  Sine, Tri, Saw, Pulse Width #1 and Pulse
Width #2. 
>It's produced some really nice sounding waveforms.  If you've already
done your
>VCOs though, this may not do much for you.  I thought it might be worth
>mention though.  :)

Good idea! another way to make interesting new  waveforms !!

I'll try to give an overwiew of the JH-4 module's knobs and its
(P = Pot, S = Switch with 3 positions, L = LED.
cw = clock wise, ccw = counter clock wise)

The Layout is similar to the SEM:
lower row contains two ADS envelopes,
two left columns contain VCO controls
next two columns contain VCF controls
added is a fifth column with LFO and VCA controls.

A) Oscillator section:

P1:   Initial PULSE WIDTH (both VCOs)
P2:   INTERVALL (coarse detune of VCO2)
P3:   TUNE (both VCOs)
P4:   DETUNE (fine detune of VCO2)
L1:    BEAT (optical control of beating between VCOs)
S1:   SYNC ( 1->2, Off, 2->1 )
P5:   VCO 1 MOD AMOUNT / DESTINATON (12 o-clock = off, ccw = FM, cw =
PWM. PWM part has
        linear function, FM has nonlinear function)

B) Filter Section

P9:   VCF MORPHING (from 2pole LP to 2pole NOTCH to 2pole HP to 2pole BP
to 4pole LP.
        2pole filter is original SEM clone, 4pole is discrete SSM2040
S4 :  VCF MORPHING SOURCE ( -ENV1, Manual, +ENV1 )
P10:  VCF ENV Modulation (ENV 2; 12-o-clock = 0, ccw = negative, cw =
positive modulation)
S5:   VCF KEYBOARD TRACKING ( off, half, full )
P11: VCO 1 WAVEFORM / LEVEL (12-o-clock = 0, ccw = SAW, cw = PULSE)
P12: VCO 2 WAVEFORM / LEVEL(see P11)

C) LFO and VCA section

P14:  LFO FM / WAVESHAPE (ccw: FM of Sine LFO by ENV1, cw: Crossfade
from Sine to
         Square to Sample&Hold to VCO1-SAW)
P15: LFO MODULATION AMOUNT / DESTINATION ( 12-o-clock =0, ccw=VCF Mod,
cw = Pan Mod)
P16: VCA LEVEL / ENV SELECT (12-o-clock =0, ccw = ENV1, cw = ENV2 )

D) ENV section

P17:   ATTACK 1
P19:   SUSTAIN 1

S5:    Voice OFF, Normal, Short Release (The OFF position sends a logic
0 to the keyboard scanner,
so the voice is not just muted, but taken from the voice allocation
cycle as well.)

P20:   ATTACK 2

Every Potentiometer provides a 0V ... +10V output.
Every 3-Way switch provides a 0V / +5V / +10V output.
So it should be easy to make eveything storable.
Someone would like to devellop a storage module ?

Connections on the PCB:

+/- 18V suplly voltage input (regulated to +/-15V on board). Current
consumtion is +170mA / -110mA

GATE / CV inputs. CV is a balanced input to ovoid GND loop problems

LFO Amount CV input (The whole LFO output runs thru an extra VCA which
is controlled by this CV.
    good for Mod wheel, Aftertouch etc. )

VCO 2 CV:  There's an extra CV input that is routed only to VCO 2. Good
for sync sweeps from a pitchbend
    wheel or Aftertouch ...)

Signal Outputs : Balanced, Stereo outputs. Mono Plugs give unbalanced
outputs. Shorting the outputs of
      several voices together will do a mix of all voices (no summing
amp required). Using only one of the stereo
      outputs tursn PAN MOD (in the LFO section) into a simple TREMOLO

Control Output to keyboard scanner: Sends a logic 0 to remove a
switched-off voice from the voice allocation

I hope I find the time to do pictures / schematics etc. to Synthfool one
day - first I have to redraw everything,
however, so don't expect it too soon.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/ms-tnef
Size: 9288 bytes
Desc: not available
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/19970611/d18e06f4/attachment.bin>

More information about the Synth-diy mailing list