Highcooley

After a longer pause, the project is moving on. Find the documentation and schematics of my VCO, waveshaper & VCM design here: https://www.muffwiggler.com/forum/viewtopic.php?p=1856429#1856429 The PCBs of the LP part of my VCF are ordered and the HP portion is in design.

Perfect, you are welcome!

Hi Markus How long is your cable between the LPC17 and the AOUT_NG? I had some trouble in the beginning, when I had a cable which was longer than 10-20cm. I don't see any reason, why you should have damaged the chip with the wrong connection you mentioned. As long as you don't connect these pins to +/-12V, you should be save. cheers Andy

To be honest, I never checked how much distortion an additional AC coupling would cause. I went for the classic 10k, 10uF and noticed that the sharp edges of the sawtooth and traingle get rounded a bit on the scope. That's when I decided not to do any more AC coupling on the VCO. You're right, real morphing would be very nice. However, controlling the upper and lower portion gets more complex and as you say, is also a bit more complex to realise with analog circuits. Ian Fritz has two nice approaches with the "Wavolver" and the "SNICster": http://home.comcast.net/~ijfritz/sy_over.htm Hmm, you also don't have any idea what I could try to solve the DIO_matrix rotary switch problem?

Thanks a lot for your insights in your project! As it seems, I took a very similar approach to your solution. - I actually do exactly the same, voltage controlling a common saw to tri converter to do this kind of transition. And that's where I had the first problem, compensating the voltage level from cutting the saw signal to half and also balancing it around 0V. I solved this with the first LM13700 only being affected during saw-tri transition. - However, I didn't AC couple the signal between tri-saw and saw-rect to keep the distortion as low as possible. As a result, the signal at the output before the AC coupling moves from -10/0 to 0/+10, limiting the different trimming ranges due to the DC component of the signal. - I wanted to use only one CV to blend through all the different wave shapes. Therefore I split the CV, using the lower half (0-5V) for the tri-saw transition and the upper half (5-10V) to transit from saw to rect as well as subsequently modulate the pulse width of the rectangle. The pulse width modulation is only done with the rectangle and not with the triangle signal. - Your solution with the sine converter to get a rectangle sounds pretty neat. I copied the circuit used in the micromoog, which uses a variable threshold to clamp the signal to a positive DC level above the triangle AC signal level. This leads to an instant doubling of the pp voltage. It's not that critical for the signal volume however, since the average volume does not change that much. - The LDR approach is probably very precise, since the average signal volume changes a lot during transition. I managed to get pretty much the same pp values for the three clean mayor signals saw, tri and rect. Of course, the over all volume changes during the tri-rect transition and again during rect PWM. But I'm still happy with the result. By the way, the documentation on the muffwiggler forum is not yet finished. Meanwhile, I have the circuit refined, the PCB edged three times and all three VCO's are working very well together after a lot of tweaking the circuits. I am going to document the final circuit as well as all the findings soon.

Strange DIO_MATRIX rotary switch problem: On the front panel, I use 9 rotary switches to select stuff like octave, modulation routing, etc. The cabling is done like the typical scan matrix application for DIO_matrix modules (http://www.ucapps.de/midio128/midio128_v3_dio_scanmatrix.pdf). I connected the multiple switch positions of my first rotary switch to C0-C7 and the slider via a diode to R0. The second switch is connected to C0-C7 and R1, and so on, and so forth. Unlike momentary switches, one position of each rotary switch is always ON, of course. When I switch on the midibox initially, usually I am able to select each position on each switch as desired. But somehow, after a while/a couple of times switching/operating several different switches/many switches in the same position (not sure yet, what causes the new behavior), I can suddenly not select the same position as a certain other switch has anymore. E.g. switch A is in position 2 and I operate switch B. The LCD display shows each change of switch B reliably. But suddenly, when I try to put switch B to position 2 (again), the display shows, that I have put switch A to position 2 instead. I am then not able to put switch B to position 2 anymore, until I change switch A to position 3 for example. But now, position 3 of switch B is locked until I change the position of switch A again. Does anybody have an idea, what could cause this interlocking situation? As I don't know, how the DIO_matrix's scanning pattern works, it is hard for me to figure out what could cause this behavior. I can only guess that it could be a problem related to: - cable length leading to some kind of interference problem (I have one 20 cm flat ribbon cable from the module to a distribution node and lengths from there to the switches between 10-30 cm) - some kind of saturation problem (because it is occurring over time) - depending on the scan pattern, too many switches in the same position causing unwanted voltage drops I might be able to shorten the cable from the module to the node a little bit, but that won't help on the maximum length that much. Another thing to try would be to completely reverse how the rotary switches are connected and go from C0 via diode to the slider and connect the different positions to R0-R7. Before I start messing up the nice cabling and spend another 5 m of new cables and hours of my time, I would like to hear your opinion on this. Has anybody experienced similar problems?

After over a year of development, I would like to start documenting my Midibox_NG project in this thread to collect as much information as possible for similar future projects. The idea is to keep all findings, questions and struggles concerning this projects in one thread to keep myself organised and to help future digital/analog synth developers. As the title predicts, the goal of this project is to design and develop an all in one box digitised analog synth, similar to the MiniMoog Voyager. The synth will be controlled solely over midi. An LPC17 with Midibox_NG running acts as the main brain to read digital pots and switches of the control panel, save and recall presets as well as to drive the synth through multiple analog and digital outputs. Analog synth hardware: 1x LFO 2x Modulation Bus 3x VCO incl. waveshapers as well as additional loops for other effects than the waveshapers (documented on: http://www.muffwiggler.com/forum/viewtopic.php?t=108794) 1x Noise generator 1x Ext_In 1x Mixer stage 1x 5Pulser (http://home.comcast.net/~ijfritz/sy_cir8.htm) 1x Ring modulator 1x Phaser effect 2x Moog ladder LP filter (to create a stereo effect through filter cutoff spacing) 1x Moog T904B HP filter (to create a band pass together with one of the LP filters (http://www.freeinfosociety.com/electronics/schemview.php?id=944) 2x ADSR (for filter and volume) 1x Stereo VCA 1x Headphone preamp 1x +/-15V & +5V switching power supply (I learned, that it is much easier to develop circuits with the additional "headroom" of +/-3V in comparison to a +/-12V supply) 1x Power board to generate +/-12V for the AOUT_NG modules as well as +10V clean reference voltage. Resources All the analog circuits are redesigned and relayouted to be uC controllable. They are mainly based on the following modules: - Ninni Bergfors' Moog based bergfotron modules (http://hem.bredband.net/bersyn/) - Original Moog D schematics (http://www.fantasyjackpalance.com/fjp/sound/synth/synthdata/16-moog-minimoog.html) - Original Micromoog schematics (http://www.fantasyjackpalance.com/fjp/sound/synth/synthdata/07-moog-micromoog.html) - Yves Usson's Yusynth (http://yusynth.net/Modular/index_en.html) - René Schmitz's modules (http://www.schmitzbits.de/) - Carsten Toensmann's Moog Modular Clone (http://www.analog-monster.de/mmschemos.html) - Vinnui's Modbus concept (http://vinnui.blogspot.ch/2011/09/concept-of-modulation-bus.html) The original idea was to use the Moog D schematics solely and as many original parts as possible. But after already acquiring many old parts for a reasonable amount of money, I dumped the idea when the last Chinese seller of UA726 remakes went off market. However, I don't regret the decision, as I am very happy with my results so far with more modern parts, which are not nearly as energy hungry as the old stuff. Midibox Hardware: 1x LPC17 core module 1x KS0108 based graphic LCD 1x SD card reader 1x AINSER64 module 1x DIO_MATRIX module 1x DOUT module (own design with partly high power shift registers for switching relays) 4x AOUT_NG modules Development Roadmap (fully implemented, partly done, not touched yet): - HARDWARE Power Board - HARDWARE Midibox modules (excl. DOUT) - HARDWARE VCO & waveshaper board (incl. trimming & tuning) - HARDWARE LP filter - HARDWARE HP filter & filter switching circuit - HARDWARE 5Pulser - HARDWARE Ring Modulator (on the breadboard and PCB layouted) - HARDWARE ADSR (PCB layouted) - HARDWARE DOUT module - HARDWARE control panel pots & switches - HARDWARE Mixer including overdrive LED circuit - HARDWARE VCA & headphone preamp - HARDWARE LFO, MOD BUS, Noise generator circuits and PCB - HARDWARE Casing - HARDWARE Phaser (planned as a seperate module, since not enough AOUTS available) - SOFTWARE Utilising AINSER64 and AOUT_NG at one port () - SOFTWARE Controlling multiple AOUT_NG modules at once () - SOFTWARE VCO octave switch option (thanks TK) - SOFTWARE Multiple rotary switches locking each other out (softwarebug ironed out, thanks again, TK) - SOFTWARE Controlling Digital switching ICs (binary input) by DOUTs (signal switching related topic) - SOFTWARE synth fine tuningrange (+/- 3 semitones, steps as fine as possible) -> solved in the analog section, without midibox assistance - SOFTWARE VCO detuning -> semitone detuning works, free detuning not implemented in MBNG) - SOFTWARE Midibox_NG menu structure -> initial work done - SOFTWARE Preset save and recall - SOFTWARE Midi control of parameters (excl. tone & pitch bend)

Dear TK, god of the Midi Box :queen: Finally I managed to finish all my three VCO boards, hooked them up to the uP and tuned the oscillators. The result with your transpose options are great and the whole note pitch as well as octave and semitone transpose only need 4 AOUTs for 3 VCOs. However, is there a possibility to have a finer resolution than semitones to transpose, (like for example pitch-bend) which I can use likewise mutliple times to detune OSC2 and 3 individually as well as finetune the whole synth to other instruments? I tried to use semitone transpose but with NRPNs instead of CCs. But the result is still a semitone transpose. Thank you very much for these awesome features!

All right...I'll give you some feedback as soon as 4 AOUT_NGs are working :yes: ... btw. octave transpose is working perfectly! Semitone transpose works with CCs. Next will be a test with NRPN to see, if I get a smooth semitone transpose over a 15 semitone range without audible steps. @lukas412: THX!

Sorry, everything working smoothly now. Haven't got a clue yet what the problem was, but consider it as resolved. @Admin: Please delete this topic, thank you!

To daysichain multiple AOUT_NG modules, I connected J2 of the first module to J1 of the second AOUT_NG. And as much as I understand from the Midibox_NG manual, the config for two AOUT_NGs should simply be: AOUT type=AOUT_NG cs=1 num_channels=16 (cs=1 as I use an AINSER64 at cs=0) After that, I should be able to address CVs 1-16 However, the second module doesn't respond. Am I missing something? If I hook my osciloscope to J2, I do see the clock signal at pin 5 but no DOUT at pin 4. Thanks for your help! Regards, Andy

Wow, fabulous work! Thank you very much for this great effort! I am looking forward to test it, as soon as I get my VCO PCBs.

Since nobody seems to have an opinion on this topic, I quickly sum up what I plan to go for: Signals lower than -5V --> relays --> no suitable digital switch found in a normal price range Audio path --> relays --> no distortion Critical CV signals (e.g. note CV) --> relays --> no distortion Uncritical CV signals in the positive voltage range (0-10V) --> CD4051 and CD4053 --> cheap, compact, easy to drive, energy efficient Uncritical CV signals in the range of -5 to +5V --> CD4066 --> cheap, compact, easy to drive, energy efficient Single CV signals on a board --> relay --> no suitable solution found yet

Hey guys, I am wondering what your preffered way for switching CV and audio signals with DOUTs is. So far, relays (in combination with power shift registers), FETs or digital switches come to mind. Relays aren't as compact and power saving as the other two options. On the other hand, I fear distortion and signal voltage level problems with digital options. If you go the digital way, can you recommend certain parts which you use and do the job best? In my case, I have two problems to solve. One is switching certain VCO modes like AUDIO/LF as well as feeding back outputs of other VCOs for syncing and FM. The other problem is a modulation matrix, where I have 8 modulation sources, 8 destinations as well as 4 shaping sources. Here I could think of an audio matrix switch. But which one to chose as there are virtually hundreds with different features and in different price ranges? Sorry, if this topic was discussed before. Since the forum's search option is still down, I tried to search manually without success. Cheers Andy

:-) wow, sounds very good! Especially for the tuning function. However, this way it wouldn't be possible to use the base pitch for multiple VCOs and adjust pitch and octave for each voice individually but still follow the base pitch. My goal is to implement a VCO section like it is used in classic synths like the MiniMOOG. Thanks and regards, Andy

Hi Thorsten I would like to ask, if you already found the time to implement this option to do some basic math with several AINs to one AOUT or dedicate a finetune option to the Tone out as well as a combined octave + semitone output? Regards, Andy

Nice filter, Magus! Yep, definitely add a small ceramics cap...it helps a second ton :turned:

Hey guys Time run fast during the last couple of days and I didn't have any time to work on this problem. So, I would like to give you guys a quick followup on my quest to silence my 5V fan on the power rail. First, I wanted to try to solve it as easy as possible, by applying a soft ferrite bead to the fan's leads. However, as the leads where already cut to length, I would have had to cut and extend them again. So in the end, building a propper LC filter didn't take much more time to solder together. I followed Janis' tip with the computer motherboards (thank you for pointing me in the right direction!) and knocked together a quick and dirty two-stage LC filter out of computer junk and four cm of veroboard: As I don't have an LC meter in my lab, there was no reason to try to do the math for a proper filter anyway. But from right to left, a large coil (with a little amount of windings) around a ferrite ring and a 3300uF cap from the cpu's power mains (which both where obviously put there to solve a similar problem) should suppress low frequencies in the first filter stage. The second stage consists of a small coil with a lot more windings from the reciever stage of a cordless mouse (again, with no selfspeaking information printed on it, but hey, it's from a HF stage) and a 100nF ceramics cap for higher frequencies (spikes). Of course, this filter is not capable of fully cutting the noise as it is not specifically tuned to a certain frequency. Additionally, people from R/C forums reported, that it is pretty hard to fully cut noise caused by brushless motors. Thy have to deal with similar problems in their quadcopters and their solution higher up from LC filters is to implement $40.- DC/DC converters to get rid of such problems. However, my filter now does the job to suppress the fan noise to a level where it neatly blends in to the basic noise floor of the synth, caused by leaking ADSR gates of the SIDs and the HF noise probably caused by the PICs. I hope this solution might help others with similar noise problems. Please feel free to comment and share your experience with other solutions (e.g. soft ferrite beeds). Thank you!

Does anybody know a cheap source for the SEQ style button caps (e.g. on ebay)?

Hi Jerome Wow, that was super fast! Thanks a lot. I didn't expect you to instantly redesign the board. However, it seems to be an easy solution for most applications (toggle switch, state, application feedback, etc.) even as you had to break your premise to keep it as a DIN only board. Regards, Andy

Wow, amazing work! How definitive is the layout of the MB-E4x2CS board? Some feedback LED option in case somebody wants to use it as a tact switch board would be awesome (single color or dual color). However, this is just an idea for a minor improvement of the already awesome board. I'm in for a couple of these anyways.

Sorry, I think this is a misunderstanding. I didn't want to say that a switching supply won't be a good choice at all. What I mean is, that a switching wall wart will likely cause troubles with the MFOS design, as the caps as well as the backward diodes are specifically designed for old school transformers with a 50Hz cycle. I just mentioned this, as nowadays it is not always clear what you get if you have a wall wart. However, it is very unlikely that there are many switching wall warts on the market which produce AC. But I think this input about using a switching PSU instead of a conventional one has definitely to be considered in the first hand.

Thanks guys for all your suggestions! Yea, LC would be my first choice anyways... if I find some ferrite core or another inductivity laying around, which I think is currently not the case in my lab :-) Maybe I can scrap soome filter parts out of an old radio or something.

It definitely depends on what you are planning to drive with this PSU. MFOS' design is not a bad choice, especially if you want to go for a small footprint and don't need a lot of power. However, it's a bit like sucking every last mV out of the wall wart to produce +/-12V, hence the huge cap bank and the feedback diods. Also make sure you have an old school transformer in your wall wart and not a switched PSU, or directly go for a transformer. This design is suitable for basic functionalities without critical VCOs or circuits which need an absolute rock solid voltage source or a lot of power. If you can afford the cost, space and weight, I would go for a +/-15V center tapped transformer, as this will leave you with enough head room for mains noise, drops and ripple. You can also go more for the 2200uF-3300uF range than the 9900uF tank bank, which cuts cost on the other side. However, that's the opinion of an analog synth builder who needs a stable +/-12V source with about 1.2A. You're talking about 167mA transformers. One thing to consider is, that the AOUT_NG alone sucks about 40mA without the core, any other module, display or subsequent circuits attached. So keep power consumption in mind.

Hmm, I haven't measured the heatsinks yet (can't find my temperature probe at the moment :rolleyes:). However, the front panel gets noticeably warm, which worries me a bit. So, the fan would definitely not be a bad thing. I'm wondering what effort would be necessary to get rid of the voltage ripple. It's probably a combination of either multiple fast ceramic condensers or a clever RC. However, I didn't try to measure how much energy the fan draws in between two ripples. Maybe it helps if I correctly measure the ripple frequency and calculate the LP filter. So far I directly measured 5V to GND. The 5V supply is a big mess with four Pics attached to it anyways, which is clearly audible as a quiet hissing sound as well. My fan is a 3300rpm model. So the ripple should actually not be around 0.5-2.5kHz but rather in the 110 or 220 Hz area, depending on the amount of coils which drive the motor or the amount of steps for one revolution respectively. It should be four steps (i.e. four switching cycles) with a standard DC fan which makes 220 steps per second for 55 rps. If I remember correctly from my electronics class, the LP filter would be calculated as fc=1/(2*pi*R*C). This leads to about 15 uF with a 50 Ohm series resistor. However, the resistor would slow down the fan significantly to only around 2000rpm, which would in the end lead to a 23 uF condenser. Maybe I am a bit too academic and a simple 100 Ohm resistor in series to operate the fan at the 12V source would do. However, I don't know if the a ripple directly in the 12V source might affect the audio out even more, as it is the direct source power of the SIDs.