[sdiy] Any SSM2164 pspice model available on the net ??

Olivier Gillet ol.gillet at gmail.com
Wed Mar 19 21:25:26 CET 2014


The LTSpice "model" posted here is a behavioural model, it works like
a perfect current sink, and a perfect current source with the 10^(-1.5
CV) gain. The stability network is not needed to get this behavioural
model to work; and it won't even impact the simulation at audio
frequencies.

Here is a list of errors I can think of, that would make a breadboard
implementation of a 2164 circuit perform differently from the sim:

* Missing stability network. Don't be lazy, this is absolutely needed!
Symptoms: DC offsets and/or very high frequency oscillations. This can
be easily spotted on VCAs, but this can also cause circuits with
feedback like SVFs to crash on the rails or behave in strange ways.
* CV input going below -0.6V. The 2164 behaves strangely when the gain
exceeds 20dB. Symptoms: clipping and/or HF trash. A diode between the
CV input and ground does the job.
* Voltage on the control port coming from a voltage source whose
output impedance is lower or in the same range than 5k. The input
impedance is 5k so this can load excessively the voltage source and
skew the response curve of your parameter. This is an easy mistake to
do when wiring a voltage divider with a 10k pot straight to the 2164
input... Symptoms: the range of the parameter controlled by the 2164
is off.
* Voltage on the output terminal exceeding 100mV in absolutely value.
The "model" allows that, but the 2164 does not like it. Always send
the 2164 output to a virtual ground (op-amp in "transimpedance"
configuration", another 2164 input...). Symptoms: clipping, DC
offsets.
* Of course, the 1.5 coefficient in the "model" is an approximation.
The actual value is temperature-dependent. If your circuit relies on
specific gain ratios to work you'll have a few surprises...

I've made each of them at least once :)

Olivier

On Wed, Mar 19, 2014 at 8:41 PM, Chris McDowell <declareupdate at gmail.com> wrote:
>> The stability network is a not really necessary for low frequency simulations:
>> - the -3dB point is at 507kHz so it has negligible effect on behaviour
>> in the audio band
>
> I seem to recall when I first played with the 2164, I tried a simple VCA without the stability network and it didn't work at all. Was "flipping out", you could say. Adding the network settled things down, so I've always included it from then on. I did also notice recently that in an SVF based on the 2164 on a breadboard, I had taken the 560p caps out (and forgotten) and the thing still worked great. I wish I could remember where, but I remember reading that the 2164 has high gain at high frequencies and that's why that ~500kHz cutoff helps so much. Neil, (or anyone), can you comment on this?
>
> Chris
> On Mar 19, 2014, at 1:00 PM, Mattias Rickardsson <mr at analogue.org> wrote:
>
>> On 19 March 2014 18:13, Tom Wiltshire <tom at electricdruid.net> wrote:
>>> You can comment it out if you want. I've certainly never seen any sim where it had an effect.
>>
>> It's kind of odd to have the stability network *within* the 2164 model though.
>>
>> I usually include it in the simulation (but outside the model) not
>> because it matters there, but in order not to forget it in the final
>> design. *That* would have an effect. I guess. But I haven't tried,
>> does anybody know? :-)
>>
>> /mr
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at dropmix.xs4all.nl
>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy



More information about the Synth-diy mailing list