Doepfer

Hi Dieter & Dirk,

I don't own an A-113, so I cannot see the exact problem first-hand, 
but I did observe the nasty 'mush' on the falling edge of the A-111 
square wave output (when it is below about 300Hz).

I notice that the CEM3340 datasheet says that the falling edge is 
more rounded than the rising one, and that this can be improved by 
placing a 1meg resistor between pins 4 and 5 (pulse out and PWM 
input). This adds some hysteresis to the internal comparator, and 
should 'square up' the output. So I tried it and indeed it _did_ 
square up the signal, and also removed all the nasty 'noise' that 
Dieter reported. 

Thus this _may_ offer another possible solution, but the datasheet 
warns that adding the resistor might cause the internal comparator to 
oscillate, depending on the board layout etc. It certainly seemed 
well-behaved for the very simply check I made - perhaps Dieter can 
offer a more informed view on how good this might be as a solution 
(e.g. if the VCO is being rapidly switched between different 
frequencies, could this cause any problems?).

Tim

[*Fourth* attempt at posting this, so I really hope multiple copies 
don't suddenly arrive...]


--- In Doepfer_a100@yahoogroups.com, "Dieter Doepfer" <hardware@...> 
wrote:
>
> Dirk,
> 
> it took a while but now we found the reason and a solution ...
> 
> We had problems to understand why the combination of A-110 
(rectangle)/A-113
> is working perfect but not A-111(rectangle)/A-113 at low 
frequencies.
> Strangely enough the combination A-111(sawtooth)/A-113 causes no 
problems.
> 
> The reason is the negative slope of the A-111 rectangle ! Our 
oscilloscope
> showed in high res mode (1us/div) that the negative slope is not a 
single
> slope but has some ups and downs (a bit like a down-going triangle).
> Obviously the waveform converter inside the CEM3340 has this 
behaviour as
> the rectangle output of the A-111 is nothing but the rectangle 
output of the
> CEM3340 (only via 1k protection resistor and AC coupling 
capacitor). In
> normal applications this causes no problems and cannot be heard 
(each VCA or
> VCF or mixer behind the A-111 smoothes this multiple slopes which 
are in the
> microsecond range). But the A-113 is very fast and detects these 
multiple
> slopes at low frequencies. This causes the problems of the A-111/A-
113
> combination you described.
> 
> There are several solutions for this problem:
> 
> 1. Do not use the rectangle but e.g. the sawtooth output of the A-
111 to
> trigger the A-113 (the sawtooth has no multiple slopes)
> 2. Process the rectangle output by any other module (VCA, VCF, 
mixer) before
> it is used to trigger the A-113.
> 3. Add a capacitor to the A-113 (in parallel to diode D3). We tried 
2n2 and
> then the A-113 tracks perfect from about 10Hz. If you contact me 
directly
> (hardware@...) I can send you a picture that shows the position of 
D3
> on the A-113 board. In this case even the rectangle output of the A-
111 can
> be used to trigger the A-113.
> 
> Best wishes
> Dieter Doepfer
> 
> > -----Ursprüngliche Nachricht-----
> > Von: Doepfer_a100@yahoogroups.com
> > [mailto:Doepfer_a100@yahoogroups.com]Im Auftrag von vliesgaard
> > Gesendet: Donnerstag, 27. April 2006 12:35
> > An: Doepfer_a100@yahoogroups.com
> > Betreff: [Doepfer_a100] Re: A-113 input sencetivity/tracking 
problems
> >
> >
> > --- In Doepfer_a100@yahoogroups.com, "Dieter Doepfer" <hardware@>
> > wrote:
> > >
> > > OK, I'll check this with an A-111 and A-113. But it may take a 
few days
> as
> > > we are very busy with testing and modifying the BBD modules as 
the
> customers
> > > are waiting anxiously for their modules.
> > >
> > > Best wishes
> > > Dieter Doepfer
> > >
> >
> > Don't worry, you get the full 100 points for customer service. If 
only
> > other company's were that good.
> >
> >
> > One use the A-113 is creating interesting rhythms.
> > E.g. if you divide the 16th notes of the sequenser, with the A-
113, by
> > 2 3 5 and 7 you get a very complex rhythm patern.
> > Thats why it should track low frequenty's.
> >
> > Greetings Dirk
>