latigid on

Not sure if we're starting new threads for features now, but here goes: The MachineDrum is one hell of an instrument but to make full use of its best features one can't sequence it externally. But it does respond to CC. So in theory, one could fill up a pattern (every step active on the MD), start the SEQ with the MD slaved, then "sequence" it by sending a CC "unmute" event when the step is active, followed by a CC "mute" event when the step is finished, or just before the next one starts. The MD is set to a "base channel" and CCs are assigned using an offset of +0, +1, +2 or +3. MD track Base chan CC Value 1 +0 12 0 = unmute 2 +0 13 >0 = mute 3 +0 14 4 +0 15 5 +1 12 6 +1 13 7 +1 14 8 +1 15 9 +2 12 10 +2 13 11 +2 14 12 +2 15 13 +3 12 14 +3 13 15 +3 14 16 +3 15 Caveats: It would cut a long sample into 1/16th note steps, although you could "tie" a few notes together. In this case only one unmute and mute command should be sent. EDIT no, it would re-trigger on the MD... short sounds only. Swing, delay and other interesting effects would not work as intended. Any thoughts?

Completely off means the other transistors of a row should be checked as well, also all of the pins around those named resistor networks. You mean the extra column is also off? Check around the pins for ICs 10 and 14 too.

Are the rows completely unlit or only half? If it's the latter then it could be related to the transistors. Bases (bottom left) of the transistors connect to the resistor networks: Q29 connects to pin of RN14, Q19 to pin 14 on RN10, Q22 to pin 10 on RN10. Emitters (bottom right) connect to ground, but other pins shouldn't. Check if there's a short to ground.

Hmm, this should already be doable with a drum track assigned. You do lose resolution in 16 voice mode though (maximum step length of 64); I wonder if it could be enhanced with the SEQ V4+?

This one too: I think the sVreg is now EOL; I can't say anything about any replacements.

The original layout from seppoman

Thanks for your input! The next step will be to design and build the PCBs. I'm experimenting with colour as a means of easily recognising parameter groups provisional, some things have changed already

Thanks, I added the PIC info to the OP.

The implementation is unclear and the linked article doesn't exist. What do you hope to achieve? The SEQ already supports 64 (I think) assignable triggers via DOUT pins.

Not enough interest to run sorry, perhaps later.

The connections are listed here but at least for me they are very confusing. It's exactly the same on the Quad IIC. So, I've redrawn the diagram: I guess it's complicated because one box's MIDI out is the other's MIDI in. The connections listed have common power on pins 6/2. SEQ side: MIDI out from the Core is current limited then goes to pin 4. Pin 5 is the pull up to +5V. MIDI in to the Core arrives on pins 1 and 3 and goes to the optocoupler. The optoisolated signal then goes to the appropriate MIDI in on the Core. On the BLM side, if the connections are crimped by colour/order as I've mentioned, just plug the right way into the miniCore and you're done :). (The order is switched relative to the SEQ side to connect SEQ MIDI OUT <-> BLM MIDI IN and BLM MIDI OUT <-> SEQ MIDI IN.) That's exactly how I mounted it; long side of a single 4 pin header to the muck area (raised up by the plastic on the PCB front side), short side to the VERTER PG, PS, G, EN.

Glad to hear it's all glowing now! For the QuadIIC, be careful with the DIN8 connector that you order as there are two different pin configurations. I ordered twice from Mouser thinking it was my mistake, but their description was wrong. Hopefully they've updated it. They can be had on eBay, if I remember the seller was based in the UK.

That's for another time when all the pricing, dimensions etc. are known.

Pretty sweet! For the non-illuminated 5, first check the diodes are soldered correctly. Then do the LEDs have the correct polarity? The main grid and extra column are only common from the cathode side. When viewed from the back, the lower contact on the LEDs should show continuity for all of these 5, in fact the first 9 cathodes should all be connected in a row. The second 8 of a row are driven through a separate transistor. From the topside, is there a slight scratch on the trace just to the inside of the right side of your yellow box? If this trace is somehow cut, please connect cathodes on the 5th and 6th column on the same row.

What an interesting post! If you like I can sell you one of the bugged circuit boards. The potential applications are for floor tiles, doormats, frisbee etc. The board I was talking about is this one: http://midiboxshop.bigcartel.com/product/quad-iic_midi-module-board http://www.midibox-shop.com/quadIIcMIDIR2.html As you can see, it generates 4x MIDI outs from the IIC buss. Notice the extra circuit is dedicated for the BLM and just optocouples MIDI IN (3) and provides a pullup for MIDI OUT (3). It conveniently has a DIN8 connector right on the PCB. So I guess if you wanted to use your PCB, you would ignore the PIC 16F and connect up J4 on the IIC to J5B on the Core. Or the opto circuit could be built on veroboard. I'm not too sure if it should run at 3V3 or 5V. Maybe try the lower voltage and see if the MIDI transfer is stable. It seems sensible to power the optocoupler at 3V3 as this should (?) levelshift the 5V MIDI IN signal from the BLM. Likewise, any 3V3 MIDI IN signal entering the miniCore should be level shifted up to 5V. Please correct me if I'm wrong! Just imagine this connector as 0V, 5V (perhaps not connected if you power the BLM separately), and the standard MIDI connections MI+, MI-, MO+ and MO-. Only two connections are "data" that need to connect to the Core, the other two are a pull up for MIDI OUT and a "virtual ground" for MIDI IN. But I admit I'm not well versed in the lingo of MIDI current loops! I suggest to use a 10 pin ribbon cable and connect the 4*0V and 2*5V to the DIN8 panelmount socket to decrease resistance in the thin wires. These connections are made on the miniCore PCB itself, so no need to add more solder. But if you're like me, sometimes when heating up large ground planes the solder flows a bit thick and bridges a few common pins. That's okay too.

This view of the board has J5B on the rear. So four pairs of AIN should be jumpered to ground towards the middle of the PCB, the bottom-most is +5V and shouldn't be shorted. Do you have the Quad IIC installed? If not, you need to provide another MIDI connection (the input of which should be optoisolated). You can supply power via the DIN or else by using the extra case holes for a separate power connection/regulator.

Smart move! The little notches in the top left corner were meant to be for a diode jig but unfortunately I didn't quite get the distance right. Best of luck for the remaining parts.

Something else to check is the orientation of resistor networks, they have a dot which must align with the PCB legend. To test if the chip survived, you could try to swap some around and see if you can get a bit of functionality. Depending on the buttons you ordered, it is still possible to misalign these by 90 degrees.

I found a few (old, like 25 y/o VSOP!) "Eurorack" PSUs. Not in the modular synth style but built for DIN rail electronics. They offer 5V 3A and +/-12V 0.3A. It's a linear design with a large transformer and heat sink, but uses LM723 as a precision rectifier and driver transistors to boost the current. I should really replace the old caps but at the moment it's running the SEQ and BLM just fine. Model is TRACOPOWER TCP531

If you need, I suggest ordering from Sparkfun et al. Instead I'm working on a different concept using hollow shaft encoders and WS2812 LEDs arranged on a PCB. This should be more flexible in terms of multicolour possibilities, also the PCB routing is much easier!

I don't see a schematic for the SEQ v4 lite. But one trick I found for the BLM is that Schottky diodes from the base to collector help to turn off transistors faster and can reduce ghosting on the current sink side. Also in my case there are 10k resistors from the base to the emitter to reduce power draw, and to provide a current path if the base is ever in a high impedance state.

Wrong topic?!

A problem... and a solution. Testing the BLM the last couple of weeks I was getting intermittent behaviour. Notably the SEQ was having trouble communicating at power on and the miniCore would reset if too many blue LEDs were lit. I think what happened is large spikes were causing the supply to drop out, which might appear as a reset pulse on pin 1. The problem remained if I used different 5V PSUs (switching and linear) and adding extra caps to smooth the rail didn't help. TK. managed to get all of his LEDs lit without any comment, so I can't rule out that it is particular to my build, or is caused by faster switching due to the Schottky diodes that I added. I'm interested to see how others go with their BLMs. The power situation is not ideal. Running everything connected to the DIN cable means there's quite a lot of cable impedance/inductance and the regulation is far away. Using 5V means there is no headroom for a normal regulator, and as the Quad IIC board uses the same rail to drive the MIDI IO circuit, using a higher voltage here isn't really feasible. One option is to use the extra case holes to connect a higher (e.g. 9V) DC input, then regulate down from there. A simple regulator could be built in the "muck" area present at the top (but beware that the acrylic spacer sits over the other side of the PCB). Or, continue to use the 5V line from the Quad IIC board but with a regulator injected in. 5V input means you need a "Buck Boost" converter. I bought an adafruit "Verter" board, which are around $10. I thought the simplest place to put it would be in place of the R1 bridge wire, but the problem is really on the miniCore side. So I cut the 5V and ground lines coming in after the DIN socket. For mounting I used an SIL header in the blank muck area; there are some unused functions on the Verter PCB like enable etc. I only had one terminal socket handy so I soldered directly to the output. The result is much better: the BLM initialises as soon as the SEQ is ready and there is no problem illuminating all LEDs. The Verter specifies that its normal 5V output is actually 5.2V, which is still fine for the PIC and its analogue inputs. This also helps with the inevitable voltage drop observed when many LEDs are lit. Happy BLMing! Someone asked about power draw (also added to the first post): Using the Verter, standby voltage and current is 5.2V and about 30-60mA.With all LEDs lit (both colours) the draw goes up to 647mA, although this would never be done in normal operation.In track mode with all 16x16 blue LEDs lit the draw is 467mA and the voltage drops to 4.72V. I don't really perceive a difference in brightness, but it could only be a good thing! As long as the ADC inputs are scanned relative to the supply voltage you shouldn't lose control at the top end.

Other way is just with four of these: Make sure not to jumper the +5V line!

The performance of Arduino-based controllers is generally worse than many of the alternatives e.g. Raspberry Pi etc. I haven't tried running the STM32 F4 Core in Host mode with a computer connected or not. The STM32 F1 Core works just fine standalone. For your wiring question: you have all the info you need in the schematics. As you choose to use a non-standard method (would be very simple with the carrier PCBs) you will have to join the dots yourself.