SSM2044 LP Filter designs

This page is a look at how various synths implemented the SSM2044 filter, with the datasheet design as a reference standard. I’ve redrawn all the schematics to make comparisons very easy.

First, the datasheet design.



The datasheet suggests using the two differential inputs (+audio and -audio) as individual oscillator inputs for a two-oscillator synth (or synth voice). It also suggests that these two inputs should be provided with slightly different resistor values (68K Vs 100K) so that the oscillators don’t cancel completely when beating against each other.

Korg PolySix


Korg used the SSM 2044 in various synths. The Polysix is quite an interesting case, as it runs the chip on +/-5V. This probably entails some of the other component changes. The Polysix only has a single oscillator, so it uses the second input to cancel any DC offset on the output.

Korg MonoPoly


Although the MonoPoly has four oscillators, it also only uses one of the signal inputs on the chip, instead using an op-amp mixer (not shown) to mix the oscillator signals. The circuit is very similar to the Poly6, except the MonoPoly uses the more common +/-15V supply. Note the 100nF supply decoupling around the chip, too.

Korg Trident


I read a forum recently where someone claimed that SSM2044 filters sound better or worse depending what circuit they’re in, and said the filter from the Trident was much better than the one in the Polysix. “Hum! How interesting!“, I thought, and went off to find the circuit diagram. Here it is, and it’s virtually identical to the Polysix design. There is no way on God’s earth that this sounds markedly different from the Polysix filter. Moral of the story: Don’t believe every “expert” opinion you read on a web forum! (but you knew that already, right?)

PPG Wave 2.3


I don’t know whether the earlier versions of the PPG Wave used this same filter circuit or not. This one was copied from the service notes for the 2.3. Note that it includes an apparently optional trimmer for output offset voltage, called ‘Klick A’. Nope! See Till Kopper’s comment below for the real reason it’s there.

Siel Opera 6



Siel’s design uses a slightly different arrangement for the Frequency CV input (pin 13). I haven’t shown the CV mixer. It also uses pin 1 (+signal input) as a mixer for the oscillator and noise signals. The design also uses supply decoupling, like the MonoPoly.

9 thoughts on “SSM2044 LP Filter designs

  1. is a nice informative page.

    The PPG wave 2.2 and 2.3 are identically in the filter section. And “Klick A” is a click feed from the sequencer section into just one of the eight voice as the metronome click.

    And I wonder what the sound difference for different capacitors around the SSM 2044 is. One day I should do this on a breadboard to get all other parts of the sound identically.

    Keep on turning these knobs

    Till Kopper

  2. No radical differences, otherwise adding the range of control voltages and signal levels would help correlated the differences in component values?

    1. Yes, certainly for some of the component value changes, different signal levels and control voltages account for the differences. As far as “radical” differences go, everyone seems to have used the SSM2044 as it was intended – as a simple lowpass filter – so the circuit doesn’t vary much. That’s not true of the CEM3320, where there are many different filters built with the chip, or even the CEM3372 which was the chip used for the famous Xpander/Matrix12 multimode filter.

  3. This is a great page, I’m tempted to breadboard them all, to see if I can hear a difference!

    I son’t suppose you know if there is a similar resourse for the SSM2045 chip? I’d love to see some of the variations that have been made with that too

      1. I would love for you to do a similar page for the SSM2045. Mainly because I have 8 NOS 2045s with ’83 date codes. The datasheet shows some significant differences in the ‘typical examples’, and I would like to see an example or two if any manufacturers used the additional VCA.

        Thanks, Tom, for all the excellent information in these pages.

  4. Very nice page, Tom. It’s one that I have referred to many times over the years. One small correction on the PolySix, the FREQ CV resistor should be 43.2K.

  5. Just wanted to point out to anyone building these circuits with the SSM2144 clone that the pinouts have been completely rearranged from the original chip! I just spent the past two days scratching my head and double and triple checking all my work (not finding any mistakes) only to finally pull out the datasheet and read the sentence “pin connections were revised for PCB layout ease”. Oooof.

    Anyhow, thanks for this page! A few jumper leads and we’re hopefully back in business.

Leave a Reply to Greg D. Cancel reply

Your email address will not be published.