latigid on

Hmm, I still think there's some merit in this, for example with a single PCB and 4 displays (e.g. MBCV) you could get away with 4 standoffs and simpler wiring/less POR circuitry etc. Did you consider FPC/FFC sockets? A quick search revealed that all of these are SMT but the metal pins are more amenable to solder than the flat cable. 0.5mm pitch correct? Tricky but doable.

The correct side is the rear, otherwise you will also need to mirror the cable or its connector. You may find the easiest way to remove it is to cut or slide off the plastic part of the header, then remove the pins one by one. (Technically you could also melt the solder and push the pins through, but the result will probably be quite messy.)

Keep in mind there are two main types of LEDs: older yellow, green and yellowy-green which each have a forward voltage of about 2V and newer diodes (especially green at 525nm, blue, white etc.) with Vf= to 3-4V. Check the data sheet and compare both the forward voltages and brightnesses in mcd. Brightness is somewhat non-linear with respect to the current passing through the junction but it can be tuned somewhat.

Okay, please check the build guide thread for drawings and graphics files of the side.

Here are a few images of the side in case you want to make wooden end cheeks. Dimensions are in inches, please check them against your case to be sure. temporary base.DWG temporary base.DXF temporary base.EDRW EDIT: suppose you could make some metal flange rack ears too! Oh, these screw threads are 6-32 and all screws are stainless steel so the aluminium parts should last for a very long time.

I don't have my unit here to test, it's at a friends awaiting a few modifications (my case had fewer standoffs so needs mechanical strengthening). TK. should be able to give you an idea of current draw with all LEDs lit, probably he has a nice bench supply running :). Perhaps a small caution to check the board for errors before starting. They have been electrically tested but there's a remote possibility of unmasked ground plane which could give a short if covered with a lot of solder. I have a few spares here which I will go over carefully when the light is better, will let you know if I find anything.

Glad you're enjoying the cases! I was equally impressed with the quality and maybe it's a road to go down for later MIDIbox enclosures. They had to do a few minor modifications during production, so I'll just get an update and provide the side dimensions for chunky wooden add-ons :).

For the older 32 bit Cores, yes you can chain an ETH and SD off the same cable because they are pinned to accept different RC lines if I recall properly. For the F4 Core, you have SD on board, so why the need for another?

Cases have shipped, please let me know if there are any problems (e.g. defects etc.) as soon as possible.

Okay, well I'm on holiday at the moment, so I won't be able to remote debug very easily. Even with a short on this resistor it would only kill that particular section and not the whole 5V line. 5V short could be anywhere on the board, not necessarily at that IC. Check the vias and mounting holes too. Bonne chance!

Interesting approach, what is the intermediate board between the Core and IDC6 connectors?

Thanks for the tip, originally I had 5 extra RNs in. For the transistor/resistor pairs, what short do you mean? Short to ground? You might have to upload photos. Note that the bottom right transistor pin should be connected to ground. All 74HC165 inputs have a pullup to 5V. Please check carefully around this pin, is the solder mask scratched? If you try to reflow the solder it can help.

Good news is that these WS2812B 5050 LEDs will probably fit (just!) underneath the clear shaft encoders I have, everything on the top PCB layer. No cutouts will also make the PCB routing easier, this was a challenge with the BLM. One caveat that I read about these is that the sample clock must be very accurate (to within 150 ns). Seems like the F4 should handle this though? Now to look for a transparent tact switch :)

Can happen. Start with the R1 bridge, you may have a small short if you scratched some soldermask off. Check every cap/ pins for 5V and pin 16 of the 10k resistor networks.

Interesting solution! But we have to consider how the illumination will work with the hardware too. There's one option using a clear shaft encoder but it only has a 4mm hole. Maybe through hole addressable LEDs bent over a PCB hole could do.

Thanks for the tests! Can I ask the specs/datasheet of the LEDs you are using i.e. luminous intensity in mcd? And when you say "directly to DOUT pins" were you using current limiting resistors? I think these will be needed to smooth out differences in brightness one often sees with RGB LEDs, not to mention the lower forward voltage for the Red die. Without matrix connections the PCB will also be much simpler to route :).

I see we are coming from different points of view: menu-based vs. knob-per-function. I know that I prefer the latter approach, for instance the sammichFM is a great compact synth but editing parameters on a 2x20 LCD and one encoder is quite unergonomic for me. I can see how having only depth controls makes panel labelling easier, but I prefer full access using the encoders (albeit requiring ninja skills to learn where each parameter is positioned). Selecting each channel one at a time and the configuration of its LFOs, ENVs, SEQ (also amounts from MIDI data such as velocity, AT, CC) etc. is also my preferred workflow. For example, I would see that Channel 5 is active from its illuminated button, so I would know the encoder banks adjust the modulators' depths for this Channel. To change parameters for ENV5B (using my nomenclature) I press the left hand buttons "B+2" which shifts the encoder bank to ENV control. LFO5A = "B+3." These could be indicated with multicolour LEDs e.g. Channel Selected = green, "Bank" Selected (in my concept one of CONFIG, ENV, LFO etc.) = blue, and if the Channel and Bank numbers happen to match you will get cyan. So you do have some multipurpose buttons, but thinking on the idea for a while this is the best "full control" solution IMO. I agree, there could be one PCB/PCB set with different faceplates depending on your preference :). Buttons/encoders could also be left out for instance if you didn't want a dedicated tuning control.

If we can agree on the hardware there's no reason why two different front panels couldn't be created. My approach places editable parameters for each modulator on its own encoder bank, on a per channel basis. It also needs a paging key for a second set of controls on the 4x4 encoders. Note your proposed encoder assignments do not include rate controls for the LFOs, A/D/S/R for the envelopes etc. -- how would you adjust these parameters?

Don't know what happened, so I'll put the pictures back in here:

I think it will be simpler to identify modulators this way. I.e. Channel 1 LFOA = LFO1A, Channel 5 MOD C = MOD5C. Then the MOD is easy to assign across Channels. Can you explain what you mean by this? I'm thinking about the following structure hierarchy, sorry if the pictures weren't clear: Choose a channel. Here on one page you can adjust all 8 modulators, activate the sequencer and apply velocity/aftertouch. Click "Config" and you have a second page of options. Bank+channel number will select each ENV, LFO, MOD and SEQ, some have two pages switched by Config such as the 16 steps of ENV2/B and SEQ. To me this is the most intuitive way to access all parameters of a modulator. I think a more flexible approach would allow 2, 4 or 8 voices. Especially in a modular setup the concept of 8 CVs is very powerful, going beyond just a MIDI to CV converter. ENC1-3: not required as there is a dedicated octave/fine encoder + a chromatic keyboard for semitone transpose ENC5-16: for the most flexibility I suggest all modulators of a channel are accessible on one "bank" Is it not better to have all 16 steps on their own encoders?

I'm aiming for 3U, although I don't have definite PCB dimensions yet. Not quite sure if it will go between a pair of rails. The next question is: should it? The older idea with lots of LED rings consumes about 700mA if I remember correctly. And it would be around a whole block of 86 HP. But the IO is flexible. Either you can use the line drivers and connect up a breakout panel in the rack or else put your jacks/bananas/whatever on the box itself. As I've mentioned before my plan is to enclose a SEQ and CV in some sort of 19" case with a breakout panel and a decent PSU.

Okay, I've thought about a new MBCV v2 implementation after a chat with TK. and some time on my train journeys. I've followed the Lemur templates and here is what I've come up with. We don't have a panel or PCB yet so all assignments are flexible! But as I imagine multiple pages I've tried to group similar functions. Here with channel 1 selected, the encoders adjust the amount of each modulator applied to it. Bipolar control could be quite interesting! The right hand column of encoders is kind of a "meta mod." Instead of these fixed assignments (as currently in the Lemur concept) two modulators could be applied for each of of LFOA/LFOB/ENVA/ENVB and sent to any destination on the MBCV. This is a bit redundant considering the actual MOD section, but could free MOD up for interesting combinations, also for AIN control. Easier to label this way! + destination? The aim would be to build a new control surface for MBCV without LED rings but using illuminated encoders as supported by TK. This way the PCB routing is simpler with less DOUT pins and less space/holes are required on the panel. The layout has: Left hand buttons for selecting CV channel, and choosing the section (CONFIG, LFO, ENV, MOD, SEQ). I discussed with Thorsten about different "banks" and "scenes" although I'm not yet sure how these are implemented. Would a scene be like a preset of CV assignments?4x OLEDs for scopesSCS (standard control surface of LCD, encoder and 6 buttons)Dedicated octave/fine tuningChromatic keyboard layout for transpose and SEQ programming, with some dual purpose buttonsBank of 4x4 encoders which change parameters depending on which modulator is selected As far as I can tell there are two possible uses of MBCV, either as a MIDI to CV converter or as a crazy LFO/ENV/SEQ source for a modular synth. Here are some possible architectures: 8 channel mono: 8(VCO; Gate), so 8 monophonic synths. Probably the best option as a modulation box. Could also operate as 8 voices on one channel, but I think few synths would use this4 channel: 4(VCO, VCA; Gate, Trig). This could be controlled over 4 separate channels or 1 channel of 4 voices. (Trig is a short pulse sent at note on.)2 channel: 2(VCO, VCA, VCF, MOD; Gate, Trig, Retrig, /Trig). Again two separate monophonic channels or one duophonic. (Retrig is a second pulse sent after a fixed delay time or using another threshold such as aftertouch or a modulator. /Trig sends a pulse when a note is released.) Anyway, those are my thoughts without too much consideration into what's possible on the programming side :). Please feel free to add ideas or critique.

Just wait for the 1N4148s ;) I thought long and hard about what LEDs to use. Yes, the package size is similar, but look at the brightness: 45/35 mcd. Recommended is more like 300 mcd. And the green is 571 nm, more of the yellowy-green. (Look also at the forward voltage, true green is always around 3.3V.)

People interested in MBCV v2 should get onto this with transparent clear knobs. It's very likely some sort of illuminated encoder will be implemented.

On a sort-of side note, I've tested both the Bourns and Sparkfun illuminated encoders for brightness. I'd guess both are actually made by Topup corp judging by the very similar packages. The illumination is sufficient when I used some analogue PWM, although I don't have clear "Waldorf" knobs. It's now a question of this easy solution or a cheaper option with hollow shaft ones. It would be possible to bend a square/flat 5mm RGB LED across a PCB cutout, or maybe there's an SMD version that would fit into the 3 mm hole on the same side of the board.