MB SID with patchbay -> semi-modular Korg MS-Style Synth?

[SIB]

thanks

[yogi]

GREAT discussion! Thanks guys, opened my eyes to some interesting ideas. I have been scouring the boards for years( trying to learn all I can, sometimes an uphill task) and gathering parts. The other day I finally ordered a set of MB6582 boards to actually BUILD something, yea!

I have a completed STM32 core but had an IT issue with my laptop( win usb driver prob ); had all the tool chains setup and then the PSU died on the Laptop:( BIG setback!! So now I've setup everything on my HTPC (had to build it from my 'dead PC pile', yes I boarder on hording). So the STM32; going to us it for a Seq, but will wait till I have a working SYNTH!

Back to topic, so if I understand correct, 5 CV ins local to each Core (x4 in MB6582). Not sure how to route in the menus, but I'm thinking of adding some of Electric Druid's LOFs and/or ASDRs (all ready got the parts) to the build, with a patch bay setup. Any ideas and thoughts? I had planed on building them for some analog MFOS projects, but I have been planning an MBSID forever, so the WSG will have to wait!

Not wanting to highjack this thread, just wanted to say Thanks, didn't know what I didn't know, Yogi

[NorthernLightX]

I'm thinking of adding some of Electric Druid's LOFs and/or ASDRs

What would they do that the 6 internal LFO's and 2 internal EG's(DADSR with separate curve parameter for the A D and R steps) can't do?

I can understand that you want to go the modular route, but after much research I've not found a reason to add CV in to the MBSID.

[punkdISCO]

What would they do that the 6 internal LFO's and 2 internal EG's(DADSR with separate curve parameter for the A D and R steps) can't do?

I can understand that you want to go the modular route, but after much research I've not found a reason to add CV in to the MBSID.

I just read the thread but I also cannot see the attraction of adding CV. I have only had a MB6582 for about 2 months but, I always learn my synths inside and out – whilst the MB6582 is fantastic, if I had to criticize it would be that it is too complicated. I own a truck load of synths (and have sold even more) and not many synths can compare to the flexibility of a MB6582 – it’s power is literally mind boggling!!

So, why would we need more flexibility?! Maybe this is just an exercise that a DIYer finds attractive rather than being particularly musically useful?

But either way, its an interesting conversation..

Paul

[yogi]

Is it needed, no. Is it complicated to add, no. So for me, its a fun feature of the great software that I might use every once and awhile. I can see using a pedal; adding drum pads, maybe (but i'm not much of a drummer); being able to add a unique analog controller would be a option. Maybe an interface for my Theremin. Don't know how many LFOs or ASDRs I need/want but that's kind of the point of all this; messing around and finding something new ;) The demo TK did with a 3 LFO summer into a CV, was cool (I wish I could find the link to that page again), can it be done with the standard CS, I guess. How hard is it to add a tap tempo to the sid's LFO? Just seems that the AINs add to the range of options.

True analog control is very hard to replicate in the digital domain. This interface is limited to 10 bits so it will never capture the the full range from 0-5v but a range of 1024 can be expressive.

I'm not planning for a PAC TEN case so I can afford more panel space. Guys, for me DIY is the point. It's fun to fire up the DAW and load a VST to have any instrument I want, just not as satisfying. Isn't that why we're all here burning PICs and breathing solder fumes?

[NorthernLightX]

Is it needed, no. Is it complicated to add, no. So for me, its a fun feature of the great software that I might use every once and awhile.

Sure, it's fun! That alone is enough reason to build it.

I can see using a pedal; adding drum pads, maybe (but i'm not much of a drummer); being able to add a unique analog controller would be a option. Maybe an interface for my Theremin. Don't know how many LFOs or ASDRs I need/want but that's kind of the point of all this; messing around and finding something new ;) The demo TK did with a 3 LFO summer into a CV, was cool (I wish I could find the link to that page again), can it be done with the standard CS, I guess.

All that can be done using MIDI, no CV needed. Adding external MIDI control is a very logical expansion (for example adding another midibox based module to the equation).

How hard is it to add a tap tempo to the sid's LFO? Just seems that the AINs add to the range of options. True analog control is very hard to replicate in the digital domain. This interface is limited to 10 bits so it will never capture the the full range from 0-5v but a range of 1024 can be expressive.

You do realise that the CV in will actually be translated into the digital domain, and thus be no more precise than the internal control messages?

I'm not planning for a PAC TEN case so I can afford more panel space. Guys, for me DIY is the point. It's fun to fire up the DAW and load a VST to have any instrument I want, just not as satisfying. Isn't that why we're all here burning PICs and breathing solder fumes?

Of course. I'm just trying to give you feedback so you don't end up spending a lot of time reasearching and building something that will not behave according to your wishes. :flowers:

[yogi]

=> NorthernLighX, just got to say I really like your build so far! That is the direction I want to go!!! Also following your thread on the AOUT redesign, ++, (I'm in line for 2 if you plan on a fab run of the boards).

As to

All that can be done using MIDI, no CV needed.

yes, yes and yes. At the moment i have a serious lust for a mbSEQ, must stay focused, must stay focused...I would love to build a mb64 at some point also or something similar; but for now, here are some 'free' AINs directly connected to the core and mios supports them. With a slight change in the SW each core in a mb6582 can have 5 CVs in.

... no more precise than the internal control messages?

Know exactly your point, I guess I wasn't to clear. My point was, we are not dealing with an temco VCO; but from a HUI view, adding some expressive control is easy and the AD resoulation is high 'enough'. It is atomic to the synth, which to me is a plus. These AINs have the same resolution that a mb64 would offer, so we're eliminating the overhead of the midi connection, but giving up some of the added features another core would have. Of course with MIOS32 and the LPC core many new 12bit doors open( PIC moves to back of bus).

I think in your first post you were asking more about the Electronic Druid designs?

I kind of got off on a tangent, but to that point, http://www.electricdruid.net/index.php?page=projects.taplfo

DCLFO on an 'F684 with PWM D/A 10bit resolution. I had been looking at these design for some analog projects, but when I discovered this 'hidden' mbSID feature, just seemed like the planets aligned. His ASDR looked cool also, but again it's doable within the MIOS framework.

So in your opinion, given a single mb6582 setup, (I've seen your videos, very nice studio space!) is there nothing to be gained by using these AINs? I suffer from a total lack of 'hands-on-it-ness'

On a related point concerning your efforts with the AOUT, do you see a need for singal conditioning to allow for bipolar inputs? I think that as a min, a pull down R and a 5.4V Zener to protect the AD inputs. Maybe an ext. op amp bridge if needed. I'll look closed at the mb64 as an example.

Tnks for your input, Yogi

[NorthernLightX]

So in your opinion, given a single mb6582 setup, (I've seen your videos, very nice studio space!) is there nothing to be gained by using these AINs?

I'm not sure to which videos you're referring, because as far as I know now video's of me or my home exist on the net. :angel:

For my application I won't use CV in on the MBSID. I'll use analog CV modulation for example on the SSM2044 filters; MBSID will send audio out to the filters and CV can be modulated from a different source than the MBSID. I plan to implement a companion board for the SSM2044 filters that will add a second CV input for the frequency, and bring both the frequency and resonance CV input to respond to +/-5V (now frequency responds to +/- 2,8V and resonanve to 0V-10V).

I can't say that there's nothing to be gained; I just haven't imagined a need. If you have a crazy envelope or LFO you want to use, or have other CV equipment that does tricks that the MBSID cannot do, it could be interesting.

I suffer from a total lack of 'hands-on-it-ness'.

See point above. If for you it's easier to grab a patchcable than to press a few buttons and turn a few knobs then go for it. It's your box. Buttons and knobs do introduce recall though, and that's hard to achieve with patchcables.

On a related point concerning your efforts with the AOUT, do you see a need for singal conditioning to allow for bipolar inputs? I think that as a min, a pull down R and a 5.4V Zener to protect the AD inputs. Maybe an ext. op amp bridge if needed. I'll look closed at the mb64 as an example.

This might be one other downside of the Core8 CV in option; it only responds to 0V-5V, which is only half of proper CV resolution. I have little experience with CV in, this is territory you need to navigate yourself I'm afraid. If you have modules that output very hot and/or negative signal it's wise to implement some form of protection on the CV inputs; how you do that you'll need to figure out yourself, I can only give some theoretical pointers:

Imagine you have 2 CV inputs for one CV function; one for bipolar and one for linear. You need to use switching jacks for this idea. The bipolar jack switch is connected to GND, so when no jack is inserted this input is at 0V. The linear jack switch is connected to +5V so the input is at 5V when no jack is inserted. Both inputs are summed with an opamp. The sum is then halved with a voltage devider and fed to the CV input.

When a -5V signal is received on the bipolar input and nothing is in the linear input the +5V from the switch will cause the sum of the signals to be 0V. When this is halved it will stay 0V.

When a +5V signal is received on the biplor input and nothing is in the linear input the sum will be 10V, 5V after halving. a +/-5V signal will thus be converted to a 0V-5V signal which is the perfect range for the AIN.

When a 0V-10V signal is received on the linear input it will disconnect the +5V switch and since nothing is on the switch of the bipolar jack the only thing that will happen is the voltage get's devided, so again the signal will be converted to a 0V-5V signal which is the perfect range for the AIN.

You can't use both inputs at the same time though, since the bipolar signal needs to be summed with +5V to end up between 0V and 5V. If your input is more than +/-5V this will be dangerous, as -6V will give a sum of -1V which could damage the AIN (not sure if this is the case but IC's not designed for negative voltages usually have a very low tolerance for it). Also, with an input of +/-10V you will end up with a voltage of 7,5V which could be too hot.

My best advise would be to make sure the output of all your modules falls in the same, safe range (maybe you need to add companion boards on the ouputs of certain modules to get the voltage to specification), and then use companion boards on the inputs of modules that need adapted CV as well. This way you can safely hook anything to anything withous letting out the magic smoke.

[yogi]

=> NorthernLightZ

I'm not sure to which videos you're referring

My bad :o NOT STOCKING, really. Was watching/listening to Videos, must have been Hawkeye's, and reading your build thread at the same time. Just a Senior moment!

If you have a crazy envelope or LFO you want to use, or have other CV equipment that does tricks that the MBSID cannot do, it could be interesting.

Not that I have one NOW, but...

This might be one other downside of the Core8 CV in option; it only responds to 0V-5V

This applies to all uC A/D or D/A converters, and more of a problem with newer 3.3V uCs. In high fidelity applications, external A/D D/A converters, designed for HighZ balanced loads and high sample rates are used. A/Ds on most uCs are included for system control apps, HUI or sensor reading all with in the system's Vcc range.

Going from the analog domain to digital is always going to be an issue at the signal layer, most op amps work best in bipolar mode.

From what I have seen, very few analog synths care to fit in to the needs of the typical A/D converter range 0-5V or 0-3.3V, apples and oranges. There are plenty of good reasons for this, but designing a 'one size fits all' interface is daunting.

I think the best approach involves a header to insert a bridge interface to normalize the desired inter-connection, I.E. for connection of +-5V or 0-10v or whathave you. Very design specific! The lack of a single standard within the Analog synth world is a large hurdle, as you are fully aware from your AOUT project.

Buttons and knobs do introduce recall

YES, that is the beauty of midibox. But how often have you seen the 'humanizer' feature to add imperfections? Not being reproducible is sometimes good.

Well, lots to digest. Will start a new thread when I get my mb6582 together and actually "know from which I speak". Look forward to seeing your progress with your build and AOUT redesign! Did I tell you how cool your front panel is?

Yogi

Edited November 10, 2012 by yogi