SSI2140 Filter Overdrive

[lickspittle]

Hi,

I'm looking for some advice on driving an SSI2140 filter ic.
The data sheet recommends a peak input signal of 2V p-p (with its reference design) but adds "Higher input levels overdrive the input stage, resulting in a distortion that many will find desirable. See [xxx] for further discussion." [xxx] appears to be a placeholder.

Does anyone have any idea of what the maximum input level should be limited to? The overdriven distortion is a good part of what makes this filter desirable but I don't want to break it.

Thanks

[oldenjon]

Nobody answered this? I think you should be fine with any input voltage due to the 10K/200R voltage dividers. Hopefully someone with a bit more expertise will come and confirm that.

[BugBrand]

Hopefully Neil Johnson will chime in, but you won't break it by feeding in a large signal.
Curious to hear too as I've got a simple proto design board coming shortly - didn't have the scope to put an onboard drive control but will experiment with external pre-level control to see the responses.

[lickspittle]

Thanks oldenjon and BugBrand,

Ideally I'd like to see THD/Frequency curves for various amounts of Vin but I don't have the test equipment.
Eventually I'd like to use the SSI2140 for what guitarists call "playable distortion" which they (rightly) consider desirable and which I've never seen on a synth. I'm hoping that velocity controlled drive will do the trick.

[lickspittle]

...or...

Possibly more elegant, change the usual subtractive synth flow from VCO -> VCF -> VCA to VCO -> VCA -> VCF where nothing goes through the VCF until the gate opens. Assuming that the VCA envelope is velocity sensitive, it might be interesting.

Any thoughts?

edit:- Maybe not, the clipping would ruin the dynamics.

[neil.johnson]

What you really mean is the input signal level to the gm cell input, which you would normally keep below +/- 20mV. In the example circuit the 10k/200R attenuator allows that circuit to handle a 2Vp-p signal without distorting. If you have a hotter signal then increase the 10k resistor to suit.

Generally it's not a great idea to put the filter after the VCA otherwise you will get filter noise even when not playing anything.

If you want to overdrive the input then you can do so by pushing up the input signal a bit so that the first gm cell input is overdriven and distorts. Make sure this is to your taste

In a forthcoming update (don't ask when, m'kay) to the data sheet there will be a circuit which allows you to control filter overdrive without changing the output signal level.

Cheers
Neil

[oldenjon]

What you really mean is the input signal level to the gm cell input, which you would normally keep below +/- 20mV. In the example circuit the 10k/200R attenuator allows that circuit to handle a 2Vp-p signal without distorting. If you have a hotter signal then increase the 10k resistor to suit.

But assuming that distortion is desired, you'd want to use hotter signal with the 10K/220R. Isn't that the point? I thought a big part of hype around SSM2040 is the unique asymmetric distortion as described in AN701.

[neil.johnson]

What you really mean is the input signal level to the gm cell input, which you would normally keep below +/- 20mV. In the example circuit the 10k/200R attenuator allows that circuit to handle a 2Vp-p signal without distorting. If you have a hotter signal then increase the 10k resistor to suit.

But assuming that distortion is desired, you'd want to use hotter signal with the 10K/220R. Isn't that the point? I thought a big part of hype around SSM2040 is the unique asymmetric distortion as described in AN701.

Well, you can either use a hotter signal with the 10k/200R input, or if you don't have a hotter signal you can reduce the 10k resistor to drive the input harder. Not every application of the SSI2140 has +/- 12V rails so you have to use what you have - the key point being driving the gm cell input beyond 20mV.

Neil

[BugBrand]

..
In a forthcoming update (don't ask when, m'kay) to the data sheet there will be a circuit which allows you to control filter overdrive without changing the output signal level...

Any brief hints/outlines on how this will be achieved?

[neil.johnson]

Spoilers!

[lickspittle]

I don't understand this, Neil (not uncommon). Isn't the maximum peak to peak set by the onset of clipping?

[BugBrand]

Spoilers!

Tease!

Fingers crossed for the updated datasheet - not that I really need more projects to delve into at the moment..

[neil.johnson]

Here's the helicopter view: SSI2164 on the way in, SSI2164 on the way out, diode-limited inverter on the CV to the second VCA. So when you amplify the input signal you attenuate the output signal to keep the levels the same. The diode limit is needed to counter the clipping action of the gm's input stage.

Neil

[BugBrand]

Cool - thanks Neil.
That won't upend my current simple 1-space test idea which should be able to be set as one of LP/HP/BP/AP easily enough.
But likely a useful idea for a larger future design.

[neil.johnson]

I don't understand this, Neil (not uncommon). Isn't the maximum peak to peak set by the onset of clipping?

No, that's only when the Q is turned up. In this case we're talking about the input stage which you can overdrive for a pleasing overdriven distortion effect.

Neil

[lickspittle]

No, that's only when the Q is turned up. In this case we're talking about the input stage which you can overdrive for a pleasing overdriven distortion effect.

Neil

Is this a different distortion mechanism to the asymmetric soft-clipping outlined in AN701?

[neil.johnson]

No, that's only when the Q is turned up. In this case we're talking about the input stage which you can overdrive for a pleasing overdriven distortion effect.

Neil

Is this a different distortion mechanism to the asymmetric soft-clipping outlined in AN701?

Yes.

Neil

[lickspittle]

Ok, thanks Neil. I look forward to the updated datasheet.