Really love the filter, most analog sounding filter in the rack. Pings really pretty too. I love looking over the technical stuff, math is totally awesome. But the sounds are super attractive and it will be a favorite for sure.
Is this polyphonic as well, seems like I could merge/split a left/right to save a lil cpu?
Super nice work Mr. Simper!
The circuit solver is an ongoing area of R&D. It’s primarily a way to speed and trying different variations of model detail to work out what is actually audible or not, all the while making sure no errors are made calculating that solution. In that regard it’s really modelling specific devices accurately and efficiently that’s the primary driver, and writing all the other bits just enables that, and requires time and effort adding bits or fixing bits all the way along that process.
The circuit diagram is the main thing I work with. For the CF100, which contains components I know well and have modelled in sufficient detail previously, it was a straightforward process and I worked solely off the schematic. I first reproduced the entire circuit diagram using LTSpice, and checked various voltage and current levels for where things would be clipping, and how much drive there is at various points in the circuit to know what I may need to model if full or could use a less cpu expensive approximation. I then re-created this circuit using my circuit modeller, and tried out different levels of model complexity until I was happy with the results.
I usually like to double check against the actual analog circuit, but since I’ve already done this type of OTA based cascade filter multiple times before (Roland Jupiter 8 / Juno 6, SH-101 etc) I fine tuned the final stages by ear to what I liked best instead of matching it to a particular version of the analog circuit, and this also saved some time. I have the LA 67 MACCA SSI2140 OTA filter arriving soon, so I’ll double check everything against that - it’s pretty much the identical circuit, but all inside an IC, that is based off the SSM2040. The MACCA has also implemented the BP4 output response, so it will be great to listen to that. The resonance stage is done slightly differently inside the SSI2140 compared to the JH version, but it will be easy to update my model to match that. From the schematic alone I can see the resonance will be stronger in the JH version, and push into clipping more easily at high resonance, which I like. I’d also like to track down an analog E440 and have a go, since from what I can gather it’s pretty much identical to the JH schematic, and has JH written on the circuit board.
For the A-124 I needed the analog circuit to match the sound, since the CMOS devices are so wildly different in response depending on the particular batch / model.
Yep, it’s poly and computes four voices for the price of one, so you could even do this:
Which is a stereo 8 pole filter for the price of a mono 4 pole filter. Shown in the spectrum is the left signal of the 4 pole output in green, and the left signal of the 8 pole output in orange.
PS: I know the Befaco even is also poly, but I wanted to keep it clear on how to route for stereo without using poly for that bit.
Couldn’t agree more! Andy’s contributions (in both knowledge and insanely detailed analog filter modules) are a massive win for the community. THANK YOU!!! Can’t wait to see what you come up with for the next Cytomic module!!
Is your solver developed in Mathematica or lower level? Some codes in your tech papers look like Mathematica code.
Thanks again for that detailed and fascinating reply. At the beginning of the talk you posted, you mentioned your respect for the hardware designers. It seems to me that you are in a position to design modules yourself if you wanted to.
Good luck with sourcing an E440. There are a couple of US sellers on Modwiggler. I’ve been eyeing the LA67 kit on Thonk for a while. Out of stock at the moment.
Yes, I could design some of my own Eurorack kit, or possibly even chips, but I’m fine with modelling for now. There are lots of amazing designs that haven’t been modelled yet, and I can still lean a lot from them. Like the JH filter, I’ll try and add my own little touches, like the NPN2 output, and VC Drive, and differen PNP forward beta values for fine tuning, but there wasn’t anything else really needed on that one. The A-124 was begging for a self osc mod.
LA 67 MACCA is in stock here (this is where I ordered mine from): LA 67 | Analogue Haven
I’ll just loan an E440 to check out.
Thanks, I’ll wait for a kit though. Speaking of which, if you fancy testing your ability to discern the character of the resonance from the components/schematic then this filter is tremendous fun: Neutral Labs – Scrat – Full DIY Kit – Thonk – DIY Synthesizer Kits & Components
Yeah, that looks great. I’m not sure why they called it a Steiner Parker filter, since it’s not. It’s a Sallen Key filter very similar to an MS20 mk1. I’m not sure it will track 1v/oct very well since they are using opto isolator BJT for the cutoff control, I’ll probably just ditch that and make it have more accurate freq tracking. It’s also a shame they didn’t add a band pass input, which would just need the switch wired to the bottom of C5 shown here:
Looks like fun and something I could definitely do a model of pretty quickly, thanks for showing me this!
I have two smallstone pedals that I absolutely loved to play with, but they are my worst pedals for picking up radio signals and injecting them into my guitar signal. I’d love to have a good digital version of that someday.
Thanks for all this cool work, really looking forward to any future modules you make!
Here is the Korg 35 internals used in the MS20 mk1, two transistors used as variable gain elements, then this goes into a buffering opamp with diode clipping, and is fed back into the base of the first stage capacitor for resonance:
Sorry this stuff is a bit off topic for the CF100, shall we move discussion elsewhere for product suggestions? Also I’m not everyone’s personal analog modeller, I may choose to do something or may not, so please don’t get offended if I never do something suggested. I still do appreciate things being pointed out to me that I haven’t seen before just out of general curiosity.
edit: there is this thread here which seems to already be “stuff I’d like in vcv rack”, so perhaps just post there with suggestions?
Wowww, so clean you could wash your clothes in it!
cooool, looking forward to watching. Also, giving a talk at 4am sounds like an especially cruel form of torture, there should be a circle for this in Dante’s hell!
Yeah, definitely not, I’m a bit embarrassed I tossed in my own ‘suggestion’, haha, very selfish.
This thread was great though, I just caught up on the whole thing all at once. sorry for the barrage of responses! But wonderful to read through everyone’s posts.
No probs with talking about stuff, but I realised we were getting a bit off topic, so for those wanting to read about the CF100 it probably wasn’t as much fun 
Used to just be a 303, now everybody needs a personal analog modeller.
Hi Andy!
Thanks for the amazing filter! It is the grand star in my patch emulating the Wing Pinger of Meng Qi. Only two filters being (cross)pinged as a sound source.
here is a quick testdrive and video: a little lullaby
Hope you like it!
Nice! I’m working on a video too, but there seems to be quite a lot of noise when I’m pinging it. Presumably that’s part of the modelling. Anyone else noticed this?