Personal web pages of

Tim Stinchcombe

The Korg35 Chip, the MS-10 & MS-20 Filters, Clones and Links

Study of the Korg MS10 & MS20 Filters

Several years ago I spent much time analysing how the Korg35 chip in the early MS10/MS20 filters works, and how this compares to the later OTA-based filters, resulting in a fairly comprehensive technical paper (2.25Meg file). Since then it has always been my intention to actually build some versions of the filters, to see how the theory turns out in practice, but other projects have been getting in the way. However I'm soon to take delivery of an actual MS-10, so this may provide a fresh incentive to pick all of this back up again!

The rest of this page is a compendium of links to material related to these filters, from the original Korg schematics to those of various clones and copies that I have come across during my wanderings around the net—there are probably others out there, as yet undiscovered by me.

Korg schematics and information

Lots of information on Korg synthesizers can be found at Ben Ward's incredibly useful 'Korg Kornukopia' site, including schematics of the Korg35-based filters in the MS-10 service manual and the MS-20. The 'Music Machines' website includes the schematics of the KLM-307 daughter-board, being the OTA-based filter of later MS-20s. Korg divulged the schematic of the Korg35 chip late in 2000: here is a scan of the reply to a speculative, and I would imagine very polite, request for information (note Korg's assertion that they will consider legal action should they be given sufficient cause).

Other Korg products utilizing the Korg35 chip in their respective filters include the PS-3100/3300 and the PS-3200, and the X-911 has a very Korg35-like arrangement of transistors in some of its filters (and interestingly also has the OTA-based filter in its SynthE section). The Korg M-500 also uses the Korg35, but I haven't managed to find a link to the schematics online.

Korg35-inspired clones/copies

Most copies of these Korg filters are based around the later OTA version and are detailed below—this is probably due to the simple reason that the schematics for that version have been widely available for a longer time. The only one I know of based around the Korg35 (I've 're-classified' the MOTM-420), is the Doepfer A-106, but because of differences in the biasing resulting from the way the chip doing the work of the Korg35 is used, I don't consider it that close a copy.

According to an old post to the Analogue Heaven mailing list made by MOTM's Paul Schreiber, the MOTM-420 is based around the same topology as the Korg35-based filter, only using OTAs instead of the transistors in the Korg35 (how different this turns out from the later OTA-based filter I cannot say, as I've not seen the schematics). As of Nov 2009, I now know that this filter belongs in the following section, and not here!

Before the internals of the Korg35 were widely known, Jürgen Haible made an impressive guess at the Korg35-based filter with his JH 720 VCF (and here is the referring page).

The Korg Monotron created quite a stir when it was announced in early 2010, as it is described as having

"...the same highly acclaimed VCF circuit as KORG's classic MS-10 and MS-20 semi modular synthesizers."

This suggested that the filter was going to be the Korg35-based one, and was backed up by pictures of the innards from early purchasers of the device, here and here, and having now got hold of one, I can confirm this is the case—it follows the Korg35 topology fairly closely, with detail changes for biasing and exponential control etc., as one would expect, due to the limitations of the 5V supply. In the spirit of saving others following having to do the same leg-work, chips identified so far are:

LM324 op amps—obvious!
TPA6111A2 audio amp
SN74AHC14 hex Schmitt trigger
NPN transistors, 'BR', 2SC5658 2SC4617
PNP transistors, 'FR', 2SA2029 2SA1774
(so my original guess for both these: same family, wrong package!)
Large SOT-23 transistors, marked 'KC'—these are N-channel JFETs, now identified as 2SK433
Diodes, 'A' 1SS355
Dual NPN transistors, 'X9', IMX9
DC-DC boost converter, 'AUJ' TPS61070 TPS61071 (power-save mode is disabled in the '61071)

Update 05 Nov 2010: Astonishingly Korg have published the schematic: the pdf file; the referring page (someone at Korg must have gotten wise to the fact that they can probably sell more of these things to people who specifically want to be able to modify them easily!?). This has allowed confirmation/correction to the device guesses above—I have 'struck through' the original guess were it now appears to be wrong. It is also nice to have the capacitor values too!

Here is an annotated photo showing the component layout, as best as I can tell (click for larger image). I make absolutely no guarantees that this is correct, so if you trash your Monotron by pulling off the wrong component through not checking yourself first, then you only have yourself to blame! (But if anybody spots any errors, I shall be glad to hear of them!). Original photo courtesy Gregg Janman:

Monotron filter poles: I have spent some time looking at the pole locations of the filter, and how they are affected by the capacitor in the feedback path and the bandwidth of the op amp used. (Some of the results also read-across to the Korg35-based lowpass filters.)

Clones/copies based on the later OTA-based filter

These are much more plentiful than Korg35-based copies, as alluded to above. I have kind of ordered them as commercial first, then DIY versions second.

One of the first copies of the filter was the Frostwave Resonator, produced by Paul Perry, but which hasn't been made since about 2008 (here is a link to its page archived at The Wayback Machine). Recently (2021) Paul Perry himself alerted me to this old thread at TalkBass (from 2013), which contains a tracing of the circuit from an actual unit, and which Paul believes is faithful to the original. There are also several other links from that thread on matters concerning the unit, which may be of value to those with an interest in it.

Analogue Solutions have incorporated the filter into a number of their products over the years: there is the SY02 module in their 'Concussor' range; it forms the basis of a complete 1U 19" filter effects unit, 'Filtered Coffee'; and the filters used in their 'Vostok' synthesizer are also based on the MS-20 (the SY02 is a slavish copy, even down to an obvious mistake on the Korg schematic!).

I obviously read more than was warranted into the Analogue Heaven post linked above, as on acquiring a Synthesis Technology MOTM-420, I was a little disappointed to discover that the core of the filter is nothing more than a faithful copy of the OTA-based filter.

Several EFM designs (Electronics for Music) include an 'MS-20 style' filter. Over recent years EFM has had a very chequered history, with the website and various forums coming and going with alarming regularity. Fortunately Matthias Herrmann has produced an excellent repository of many of the old EFM schematics, from which we can see that the EFM VCF8E is very MS-20-like, but has an overdriven OTA for distortion in the feedback path rather than the diodes, but that VCF2 in the Wildcat has had the diodes re-instated. A built version of the VCF8E used to be available from MegaOhm Audio—there are still plenty of sound samples there, and I understand they are working towards having a similar filter available in the not-too-distant future.

René Schmitz has a fairly straightforward take on the filter, using CA3080s (plus op amp buffers) and a single LED instead of the three diodes for the distortion. Several others have expanded/modified this circuit: Scott Stites' 'Dual René Schmitz Late MS-20 Filter' has two filters that can be paralleled/cascaded in a number of ways; Jürgen Schmitz has added a few bits and bobs to it to change some characteristics; Ryan Williams has used CMOS inverters instead of op amps in it; and recently I came across this version in a post by user softfin at electro-music.com.

Motohiko Takeda has two MS-20 Clones at his site: the first schematic is referenced from a page which is currently only available in Japanese; the schematic of the second 'clone' is apparently a re-drawing of the Jürgen Haible Korg35 'guess'-circuit mentioned above (though I've not checked this in detail)—the associated page (in English) contains some chit-chat about the Korg35 story, but unfortunately there are quite a few broken links there.

Finally, Tim Escobedo's 9V-powered circuit, ostensibly intended as a guitar effects unit. It is perhaps debatable as to whether this really qualifies as a clone or not, as there are no distortion diodes in the feedback (and it isn't even a Sallen-Key topology as stated, as indeed later MS-20 filters aren't either!).

[Disclaimer: inclusion of a link above in no way constitutes an endorsement by me of any of the designs or products. I have not built nor seen nor heard the majority of them, and my comments are solely based on what I've seen in the schematics or what I've read. If you want to know what any particular filter sounds like, then I suggest you try and find one to listen to, or buy or build it, or try and find a review of it by someone whose aural tastes are known to closely resemble your own!]

[Page last updated: 30 Jan 2021]