CEM 3320 + SID

[Sasha]

What do you think of CEM 3320 + SID comparing to CEM 3378/79? I`m far from expert for tese questions.

Here is datasheets of both:

http://www.synthtech.com/cem/c3320pdf.pdf

http://www.synthtech.com/cem/c3378pdf.pdf

I also have possibility to buy these chips:

CEM 3320 VCF 18 EUR

CEM 3372 VCF VCA 20 EUR

SSM 2056 EG 24 EUR

SSM 2044 VCF 15 EUR (I`ve seen this one much cheaper on ebay)

all suggestion are wellcome

[Jaicen]

My suggestion would be to get the VCA/VCF combo chip. That way you can have proper ADSR on both and the SID bug won't be a problem. It also means that you can get the proper SID sound (sort of) from any SID, regardless of its filter properties.

You could also look at getting a CEM3396, which is a waveshaper & VCA+VCF. Effectively a synth on a chip. If you could feed it a square wave from the SID, you could use the onboard waveshapers, or just use the filter and amp. May be a cheaper option.

[seppoman]

Hi,

the 3320 is a classic and good filter, as is the 3372. Can´t say much about the sonic differences as I didn´t hear them yet. If you can source enough CEMs, a design around them would be surely rewarding, as it would with any other classic analog filter.

My main reason to post is that at the moment there´s a lot of CEM hype in this forum some people who repair vintage gear won´t be happy about. These ICs are mostly really rare, and a lot of vintage 70ies/80ies analog synths could be saved instead of building a 8x MBSID with CEMs. This is not to stop you personally from doing this, just when there are a lot of people doing projects with multiple filters, it starts to influence the market/availability of these ICs and at some point it could get impossible to save the real classic gems like e.g. the Prophet 5. Midibox has become big enough that it can cause serious problems. If even a quarter of the Wilba PCB guys would want to have analog filters and all wanted CEMs, this would draw 200 ICs from the market - the price would go through the roof and availability drop to zero.

You already mentioned the alternative: SSM2044. Seemingly there´s a lot of old new stock available. If you check IC databases, you can find that there are asian companies which have thousands on stock. I recently got 10 pieces on eBay from hongkongsuperseller for 75$/55 Euros incl. shipping. I´ve already built a prototype and it surely does sound fat, warm, analog, really cool. Right now I´m designing a PCB for pairs of these.

As I said, I don´t want to discourage you - if a certain CEM is your dream filter, just do it :) - only if you just want to have *some* analog filter, think about the SSM. Apart from the moral, it´s also way cheaper and easily available. These are the reasons why I´m doing this PCB right now (plus I want this PCB too ;) ), and I´ll publish it when it´s ready :)

Sorry to distract your topic, Sasha  ::)

Seppoman

P.S. a SSM2164 VCA board is also in the queue, but I didn´t start yet, so don´t hold your breath ;)

[NorthernLightX]

To hijack this thread even further (I'm sorry Sasha ;) ): Seppoman, are you also going to do a PCB with your new AOUT and the SSM filters together?  ;D ;D

[ilmenator]

Please not, keep it modular! I want to use the original AOUT module with the filters...

Best regards, ilmenator

[Sasha]

Hi Seppoman, I understand your worries. They makes pretty much sense. It would be shame some old synth cannot be fixed because no more chips left, or they are very expensive. But at the other hand... those synts are already very expensive and only richest people can afford it. I personally hardly can afford any modern gear, not to talk of some vintage stuff. Only solution for me is to make my own MB based hardware. It is hard to feel for those rich people from my position. ...Tons of stuf laying around, and they mostly dont use it, but hey, looks nice in their studio. I recently saw some kind of Orbital interview, and my heart hurts when I saw guy dig a 303 full of dust somewhere from the floor  :'(

Anyway, I`m sorry for gear very much, but I am not sorry, at all, for most of the owners. You know, when you are rich, you don`t really apreciate gear you have. If you brake your Prophet, oh, shit... you`ll buy another one, but If you are poor, you`ll love your few CEMs more than a mother. Not really, but you get my point... Other thing is, you need to kill 8 commodores for one Willba`s synth. Commodore lovers often tells me... you kill C64 for this??? And I reply ashamed... I killed 4.  :-[ It is just a matter what vintage gear you love. We all here have more love for synths than for computers, and that is why we are killing so many C64`s.

The reason I want analog filter is clear, SID filter sucks! Blame Thorsten. ;) I heard how beautiful SID+CEM sounds.

I wanted to make only one CEM filter as they are "still" very expensive, and I just cant afford to have 4. Other alternative for me is to make SSM2044 filter. If you make PCB for SSM, I would rather make this version than CEM. Thanks In advance.

PS. I hope you didn`t get me wrong.

[seppoman]

NLX:

don´t think so - 1. it wouldn´t all fit into the 80x100mm of Eagle light and I don´t want to redraw everything in crap like Kikad... 2. It´s not bad to keep things a bit modular. 3. This would be not instead of the normal AOUT or the normal (non-existant) SSM board, but in addition. So already three boards to maintain, document etc.

After all, it´s only 4 CVs and power to connect between the two.

Seppoman

[gavgomad]

Well, I pack up half way through my lengthy message and fire up at work this morning with three new posts covering some of what I was going to say! Then I edit my post, and again, new posts already covering stuff! I guess I'm losing my originality! ;-P

I've played with a few of these chips in non-SID situations, and will cut and paste my subjective thoughts below.... ;-)

Seppo, GREAT to hear that you're doing a 2164 board. Right now, for a 4x2 SID system, my plan is to go with an AOUT per voice (I won't say hurry up.... but....  :P) an SSM2044 each for left and right channel, and a 2164 to handle final VCA duties left and right, as well as feedback VCA left and right (to the SID input, with an on/off switch). Unwieldly to do 4 stereo voices on one single board, but was going to do two voices per board and stack them in my case. Soooo.... NLX's combo AOUT and 2164 board would be SWEET, and only leave me with quad or octal 2044s to design!  ;D

Now, thoughts on filters.... YMMV!  ;)

The 3320 is a solid chip, and I personally would prefer it over a '78 or '79 by reputation (again, I haven't personally played with either the '78 or '79, but they were in many later instruments which I found to be more "clean" sounding).

That being said, the 3320 is often accused of being "neutral" sounding.... It is THE non-Roland sound of the early to mid-eighties (same chip in most of the american poly synths of the eighties). BUT is also responsible for the multi-mode filter on the Synthex and for the pure BEEF that is in my beloved Oberheim.... If used right, the 3320 sounds quite good indeed. There are some nice schematics for these, and tips for their use on the net. Still figuring our what to do with the couple I have!  ;)

The 3372 is not as "warm" a filter as the 3320 (IMHO remember!  ) but has built in VCA and is readily configured into a multimode filter (think OB Xpander etc.). These are becoming fairly scarce though, as is noted in earlier posts, and Xpander users are in need of replacements so unless you are using any of the special features of this chip, there are others out there better suited, more "dinosaur" friendly, and probably the '78 and '79 are fairly close in sound (speculation on my part from comments of others and minor playing time on an ESQ-1, which I liked very much!).

My preference right now (already given away above!), even though it is ONLY a 4-pole lowpass (no other modes possible) is the SSM-2044. Not as good as a 2040 (which I own, but not enough of them for my SID box and ULTRA rare!), but pretty beefy and I would say subjectively "warmer" and more "organic" than a 3320. 2044s are also pretty plentiful right now, again, as noted, and are reasonably priced.

So your initial question.... I like the 3320 - Google up the "emulator archive" diy section for the choice of value for the caps in the circuit to sound more Oberheim vs. Prophet vs. Digisound. There is also a nice schematic on the net for a switchable low pass / band pass using the 3320 (look up OG2 - Scott Bernardi's site).

My impression is that later CEM's ('78 and '79) sacrificed a bit of quality (ie. analog "girth") for accuracy and features that made them a better match with digital controllers. If you are running multiple voices and want each voice to sound exactly the same, this may be a positive feature to you.

Overall, which you choose will depend on your goal. All of those mentioned made it to the production stage because they offered something different. For fatter and warmer, the SSMs are tough to beat. You sacrifice a bit of stability, though. For solid operation, multiple modes and stability, the CEMs are a great choice. The 3320 may be the "right up the middle" choice if you want a bit of both - then again, that's what Sequential, Oberheim and others did in the eighties, and while there are those who applaud them for it, there are still others who curse and swear 

To HEAR the difference.... SSM 2044s were used in the Korg Polysix and Trident (among others; early Fairlights and PPGs used them too). I know of a few sites with Trident samples if you google for mp3s, and that will give you an idea of that chip. The 3320 is easier, but depending on which synth it was used in, can be a chameleon - Best and widest range of examples are Oberheim OB-8 or Xa, Sequential Prophet-5, Elka Synthex. The '78 and '79 were, if I recall, used in some of the Ensoniq synths (ESQ-1 / SQ-80), and some late model Sequential stuff (VS, perhaps the 2000/2002 series samplers?).

As always, the mileage of others may vary! Hope my two cents help!

Gavin

[seppoman]

Hey Sasha,

no offense taken, but: Just that a Prophet 5 is expensive doesn´t mean its owner is a rich bastard - this Prophet example came from a co-student of mine who just bought a slightly broken one for 1700 Euros and asked me to check it. He doesn´t earn much money and makes his own music all day and night. He saved his ass off for it.

And you´re partly right about the SIDs, but: millions of C64s existed, and even the ones not having been trashed by now must be hundredthousands. A Prophet 5 was an expensive beast back then and probably was built a few thousand times, remaining functioning units perhaps a few hundred. And so are the necessary spare parts. CEM and SSM were nowhere near as big as CGS/MOS were. I think the big number of 2044s comes from SSM having missed the end of the analog era. They must´ve had produced like 10,000 of it and shortly after nobody wanted to buy them anymore.

As I said, this was not a "don´t do it, you evil !$%&$"!$§!", my point was just: Only use really rare parts when you know and value the difference to the alternative. If you just love the CEM3320 sound and this is exactly what you want, then go for it. If you just want to have any great-sounding analog filter from good old times, then do the Synth Lovers a favour and choose the SSM. Even while there are probably several great machines collecting dust in some rich-ass basement, most people spending money on these machines do appreciate the difference to a Nord Lead :)

Seppoman

[seppoman]

Seppo, GREAT to hear that you're doing a 2164 board. Right now, for a 4x2 SID system, my plan is to go with an AOUT per voice (I won't say hurry up.... but....  :P) an SSM2044 each for left and right channel, and a 2164 to handle final VCA duties left and right, as well as feedback VCA left and right (to the SID input, with an on/off switch). Unwieldly to do 4 stereo voices on one single board, but was going to do two voices per board and stack them in my case. Soooo.... NLX's combo AOUT and 2164 board would be SWEET, and only leave me with quad or octal 2044s to design!  ;D

Hi Gavo, you mixed up the numbers a bit :) NLX wanted a AOUT+SSM2044 board. The first thing I´m doing right now is a double 2044 board and then some time (definitely more than 2-3 months!) a 2164 board. And as I said, I don´t think I would do a combined board of any type - this just makes the number of necessary boards higher. With single modules, you can combine stuff more flexibly. For example, I also want to add some 2044s to my MBFM, so I don´t need 2164s there and it would be annoying if I had only two 2044s together with an AOUT on one board.

Modularity is one of the main concepts of Midibox, and while I really appreciate the availability of Wilba´s board (and have ordered one myself), I think that doing all-in-one boards for every possible configuration just doesn´t make sense for the community. The Wilba board is something different because it provides a good and integrated way to have everything for a full normal MBSID on one board. But stretching this to all combinations of nice filters, nice VCAs and every combination of this would be too specialized IMHO.

Thanks a lot for you comments about the filter sounds, really interesting!

Seppoman

[gavgomad]

Apologies, Seppo (and NLX!!) That's what I get for skimming to catch up - saw your PS and latched on to the 2164.... Excited to hear about the dual 2044 board! Wow, almost missed that! I may need to send you virtual beer for designing half of my synth! (I'd send real, but I think any decent beer would not be healthy by the time it reached you, and you probably have better there!)  ;)

Maybe my task will be to handle the 2164 design? I plan on including some analog bypass switching as well, so perhaps there is somewhere where I can modestly contribute - and more my speed - I am a true rookie at PCB layouts!  ;)

I'm fighting for space in my box - Wilba base board, and trying to get all that other stuff as well, all into a two-space case with enough room for all that silicon and voltage to breathe! UGH!

Thanks for all of your efforts as well, Seppo!

Gavin.

[seppoman]

If you want to take a shot on a 2164 board, just do it :) As I said, I probably won´t have time for this anyway in the next few months, so any activities are welcome :)

In a 2U case, space won´t be the problem. Just use two "floors" of PCBs, e.g. put all analog stuff on the bottom of the case, buy some really long screws and spacers and cover everything with the Wilba PCB :)

Seppoman

P.S. virtual beer is ok ;)

[Wilba]

Other thing is, you need to kill 8 commodores for one Willba`s synth. Commodore lovers often tells me... you kill C64 for this??? And I reply ashamed... I killed 4.  :-[  It is just a matter what vintage gear you love. We all here have more love for synths than for computers, and that is why we are killing so many C64`s.

You don't know how hard I tried to buy SID chips before killing a Commodore 64. And now I've gone to extreme lengths to obtain new-old-stock 6582 SIDs and sell them so people don't kill any more C64s, especially just to populate an 8xSID PCB.

Sorry, I just don't want to be seen as the cause of more C64 killings... I've restored the karmic balance as much as possible.  ;D

[Sasha]

As I said, this was not a "don´t do it, you evil !$%&$"!$§!", my point was just: Only use really rare parts when you know and value the difference to the alternative.

Sure, you didnt. ;) But, hardly anybody here knows the difference (AFAIK) as nobody tested both, CEMs and SSMs. I know I like CEMs that much that I found my SIDs filters completely useless. Would I be satisfied with SSMs, probably yes.

If you just love the CEM3320 sound and this is exactly what you want, then go for it.

...If not, the one I`ve haded will gather dust till somebody other buy it, maybe to fix some synth, maybe to make a new one.

The guy is selling it because he wanted to make his own synth once but he gave up due to the lack of time. Time is passing, and If you ask me... it is better to do something with those chips now, than skip it because somebody will need it one day. Maybe it sounds selfish to you, but I look at it different.

It is hard to change global deficits. In less than 30 years want be any oil left on earth, but people still use it same way. Who can afford it drives. I never owned a car or motorcycle (I`m 31), and I`m sure in the time I can afford one, to drive, it would be luxury. There will be electric cars, but there is also going to be vintage cars lovers, who appreciate that (analog) roaring sound. Nothing lasts forever...  :-\

@Willba, It is really nice of you to help us to stop killing!  ;D

[Jaicen]

I agree with Wilba, I hate to see good C64's trashed just for thier SID's. That pic of TK's showing all the boards he'd pulled makes me die a little inside every time I look at it ;)

That's not to say I don't appreciate the mighty MBSID, because I see it as finally realising the true potential of the chip. It's now doing what it was designed to do, no longer handicapped by the hardware.

The way I see it is, the older C64's with the 6581 will be less hard hit by the MBSID. I would never trash a C64 for chips, if I had to pull the SID, i'd hoard the motherboard just in case it was needed again. I look at that as my duty (and that's why there's a broken VIC-20 packed ready to move to my new house with all my other stuff ;) ).

Whew, that was a bit long winded. I was actually posting to remind people that the CEM/SSM chips are not the be all and end all. Some great filters can be built with just a few common components like a CA3080 for example. Maybe we should look at that before buying up huge stocks of rare chips that really should be used for keeping old stuff alive. I think a lot of chips will be re-manufactured in the not too distant future,  for example, I know that the Juno voice chips are being re-made, but they're not yet available.

[gavgomad]

Hey Jaicen!

True, there are some excellent discrete filters out there which would beat out many of the CEM and SSM chips.... Moogs etc. I actually plan on an MB based project to control and switch between stereo sets of discrete filters, actually!

That being said, with discrete, you have a number of problems if you are looking at multiple channels - Most important of which are differing component tolerances (voices WILL differ somewhat!), and for my situation, SIZE. THE most tantalizing discrete filter for me right now is, strangely enough, a discrete recreation of a chip - the SSM2040. Eight of those bad boys would be amazing, but those will be BIG boards, with LOTS of components of different tolerances, not to mention how they'll react individually when they heat up in a closed space together....

Doug Curtis and Dave Rossum were brilliant in getting their designs onto a chip, and eliminating many of these variables. Accordingly, in those nice old polysynths, you don't cycle through eight voices with the same sound, and get: Bop - plonk - doink - bowwww.... I exaggerate a bit, but you get the idea! ;-P

.... Ah, but that's the FUN of analog in my estimation, and why analog filters marry so well with the SID! ;-P

Cheers,

Gavin.

[NorthernLightX]

... That pic of TK's showing all the boards he'd pulled makes me die a little inside every time I look at it ;)

I believe that not TK but Ian Hurlock destroyed that pile of C64. Anyway, I too have killed some C64's for SIDs in the past. The C64's were often partly broken and/or cigarette smoked, so I don't feel guilty at all  ;D

[Sasha]

I feel no guilt for killing commodores, rather pride. We are giving new life at least a part of it, and that is better than gathering dust. C64s was great computer once, but what for would somebody use it nowdays. I know lot of hardcore exC64 users that didnt turned it on for ages, and probably they never will. Maybe some day for short, just to recall the memories. Hey, we should go towards the future. Not all gear from the past are still useful. As Jaicen said:

I see it as finally realising the true potential of the chip. It's now doing what it was designed to do, no longer handicapped by the hardware.

[b-funk]

hey,

i just wanted to mention that fonik of electro-music.com forum is considering to make a second run of his quite brilliant ssm2044 pcb if there is a interest of 40 pcb´s. his first run sold out really fast, and 22 pcb´s are already pre-ordered for the second run.

i breadboarded his circuit some time ago, and, well, it´s GOOD...  ;D

see here: http://www.electro-music.com/forum/viewtopic.php?t=19720

there is link in the thread to the documentation of the project.

best wishes,

tobias.

ps: electro-music.com seems to have technical problems at the moment. so, if the site doesn´t load and you get an error message, retry several times. after a few tries the site does load. or it´s maybe my conecction which is lame...  ::)

[Sasha]

Second run, you say. That is great!

So, you like how it sounds. Can you attach some samples please? It would be great.

Thanks for the info man.

[b-funk]

hi sasha,

i can post some samples at the weekend.

to get a slight idea, watch this:

http://www.mdz.de/www/blog/106#more-106

that´s mainly a demonstration of how easy it is to build an lfo, but the lfo is modulating the breadboarded ssm2044 circuit. a roland mc202 was used as oscillator.

quite poor soundquality and even poorer picturequality. the little clip was shot with my digital camera, that explains the poorness quite good  ;D

but for a first idea it´s ok.

i´ll post some audio samples as soon as possible.

best wishes,

tobias.

[Sasha]

Cool. Can`t wait, but take your time. ;)

Yes, quality of video is pretty bad+youtube degradation, but I get a clue. I already heard how SSM sounds before but not with SID what interested us most.

[seppoman]

Hi,

here´s another link to the Fonik board: http://www.modular.fonik.de/Page27.html

I also took some inspiration from it. The schematic is not very far away from the reference design in the SSM datasheet. The Fonik board is definitely nice for modular synths and completey manual control. One thing all (except one) schematics I found are missing is a linearization for the resonance CV. The datasheet suggests a neg log pot, and Fonik simulates this by the resistor in parallel to the pot. So for manual control this probably works fine. Only for CV control, like with the MBSID/MBFM, the problem is that the response is still exponential. This means that in the lower quarter of the CV range, not much happens. Then it starts to have more and more influence and after about midway the thing goes off like a rocket.

The only implementation I found that deals with this issue is this one:

http://www.emulatorarchive.com/assets/PDF/SSM2033%20SSM2044%20Circuits.pdf

Here, a 3080 is used for linearization. As the 3080 is obsolete by now (still widely available, but nevertheless), I used a LM13700 instead and played with the values until it worked well. For MB applications, this is quite a benefit!

I don´t want to diss the Fonik board at all, but from my board you´ll get two channels on a similar PCB size (left out all the pot and additional CV stuff) and a (kind of) linear response of Q. Also, some values are adjusted for +/-12V instead of +/-15V supply. So it´ll be more suitable for Midiboxing (while being less suitable for modular stuff because of the missing direct control).

Seppoman

[gavgomad]

That virtual beer is waiting for you when you finish off this board as well, Seppo!  ;D

Does your implementation of the linearizing function work comfortably with 13600s as well? I know they're harder to find, but I have a TON of them....  ;)

Thanks for all your work, and looking forward to seeing the results!

Gavin.

[Sasha]

Hmmm, linearity is very important for MB. I think I will test both PCBs. Fonik`s as a standalone manual filter box, and yours for integration with SID. Who will be first?  ;D