latigid on

Hi Laurent, feel free to post a clear photo/part number and I can check for you. The line should go along pins 1-2-3-4 etc. so the text of the part is readable with the line along the bottom. Best; Andy

sr_dout_r1= is for 8x matrices sr_dout_r2= is for 16x matrices I suppose you only use 8x8 matrices, thus keep sr_dout_r1 for the second DOUT. Here's how you address one DOUT PCB: sr_dout_r1=1 sr_dout_r1=2 sr_dout_r1=3 sr_dout_r1=4 and here's how you address the next DOUT: sr_dout_r1=5 sr_dout_r1=6 ...

POIDH!

Hi Adam, Everything works, unless it doesn't . Yes, you should be able to use a controller keyboard and route the USB MIDI to another port for example. Try here; in essence the hardware is the same, only the FET switch that senses current doesn't feature in the wCore/USB. http://midibox.org/forums/topic/18906-usb-host-support-for-mbhp_core_stm32f4/#comment-165136 As noted, there are a few other jumper settings on the 407v breakout. I've read a few things in passing that require a different pull-up or so depending if the device is self-powered or not. Have fun! Andy

Is the Korg self-powered? Did you jumper the +5V on the Core USB PCB? I don't have time to investigate at the moment but you may also look into the jumpers on the Waveshare board. Read up on OTG handshakes; it might be that certain pull-ups or other data routines are required. https://www.waveshare.com/w/upload/5/58/CorexxxV-Schematic.pdf Best, Andy

Hi, Defines the matrix to be addressed. You could have one matrix for 7-seg digits and another as a simple LED dot matrix. Each one needs a unique address, otherwise MIOS wouldn't know what one to turn on when assigned. Simple case: 8 rows is typical. 4 rows will repeat the selection pulses. 16 rows would need two SRs for 16 selection pulses. Use rows-8 unless you need a 16x16 matrix. Determine where the common pin of your LED segments go. Then connect them to this SR in order. If it is the first SR in the chain, then sr_dout_sel1=1. If it is the nth SR in the chain, then sr_dout_sel1=n r1 refers to red1, for the case where you're defining an RGB matrix. Then you have g1 and b1, and also r2, g2 and g2 if you want a 16x16 matrix. Not pull-up resistors, these are series resistors to limit the current and they probably should be bridges to sink more current for the common cathodes. So assign the sr_dout_sel1=n where those bridges are. If you have common cathodes, then inverted_sel=0 is probably correct. At least that is the case for matrices that I have configured (not including those with sink drivers!) Tell me about it! Have fun! Andy

Could be anything from incomplete NGC file, dodgy breadboard or wiring, corrupt SD card (saving often doesn't work well, upload always worked better for me) etc. etc. Draw a functional map of how it should work and see if you connected the dots properly. Adapt the NGC file to suit your hardware and again just playing with it might help.

Define the shift registers connected to the "sel" and "dout" parts, at the moment they all = 0 and so no data are sent.

It's too hard to follow your text sorry. Can you provide the actual wiring that you use and the type of 7-seg displays? How many shift registers are you using and is this accounted for in the .NGC? To light up an LED you need a current source (high side) and a current sink (low side). For common cathode (CC) displays, the segments are sourced from DOUT pins driven high and "sinked" also into a DOUT pin that is driven low. If the sink pin is high, then the potential is the same on both sides of the LED and so no current flows and the segment remains off. The pattern updates with each RC1 pulse for each LED individually, defined by the pin used for CC. I am not 100% familiar with the logic here, but inverted=1 probably means "to turn this digit on, turn the DOUT pin off"- For common anode (CA) displays, the digit is sourced from the DOUT pin and the segments sinked into DOUT pins. Note that the .pdf says Common Anode (inverted_sel=0, inverted_row=0) i.e. not inverted! Try a simpler setup first of all. Just use a simple single LED that you know the polarity of (i.e. test with a multimeter diode mode) and make sure you can get the LED DIGIT pattern working with that LED.

Thanks for the report; the matrix scan would be useful, but it does look somehow like the bottom two rows are connected? If the rest of the SRIO chain is fine probably the issue is with the output side. I would check the cable for shorts once more. Plug a double-row pinheader in and check continuity for adjacent pins starting in the bottom-left corner 1-2 (bottom right if you view from the pin side) but also the top pin to the next in the series 2-3 etc. 2 4 ... 1 3 notch this side The rows are on 8-pin header J5 and the corresponding side of the 24-pin header. Check the male pin headers too. For soldering: Disconnect the IDC24/16/8 cable and power. Are there shorts on adjacent pins of IC18 on lemec_R? Pins 1,2,17 and 18 would be of interest. On IC17, check for shorts on pins 6,7 and 8 especially. On the JA board, the pinout for the LED matrices has row 1 on pin 5 and row 2 on pin 2 (there's a silkscreen indication for pin 1). Check that there are no shorts there especially. With _NG and probably best with no .NGC loaded, you can use the MIOS terminal command set dout x 1 where x is the shift register pin and 1 turns it on. Count the number of SRs in the chain *8 to set the pins of IC17 The idea is to see if you can control the 74HC595 pins independently. You can also use this to test the matrix: turn on row one and then the columns in turn with the subsequent 8 SR pins. Good luck! Andy

https://www.youtube.com/watch?v=dOXIs6Rkhyk&t=7353s contains testing for the JA matrix. Please reload the _NG .hex file and see if you can control the LEDs with CC16. If you can, maybe it has something to do with the track settings not being in sync on the SEQ?

Hi Thomas Let me know if you want a PCB to drive multiple displays: http://wiki.midibox.org/doku.php?id=display_driver current version at least fixed the mirrored J15 connector. The connector for shift registers is J10B for STM32F4 Cores. Best, Andy

Upload your NGC. Are you using SENDER/RECEIVER events? Did you scale/interpolate one CC into another? What about using the native CC scaling with -63/-/64 on the display? Did you look at maps?

Hope it works out! I can't say anything about whether the connections are correct, but here are some thoughts on the layout. Watch the clearance. It should ideally be as large as possible; often I work with ~0.4mm trace width and 0.5mm spacing (1mm trace-to-trace). Take particular care around unmasked copper such as through-hole pads (e.g. on the 2x8) Circled upper-left: the corner can be avoided by routing from the left/bottom left of the pad. Circles right: angles a bit messed. Neatest is to use 45° angles for everything.

@Altitude had a similar idea: http://midibox.org/forums/topic/20525-16pin-sil-to-16pin-dil-adapter/ Check that the fit works in a MB-6582, but I guess so?

Just use MIOS Studio and assign one of the sliders to your CC event, or rather assign your event to one of the sliders. You can probably use J10 or J5 as analogue inputs.

Hi Niels, The power situation is fairly complex and there could be more than one source of noise. Some setups have no issues, some do. With MIDI obviously we have optoisolation so things are different there. The power distribution in the SEQ could be better e.g. by improving the return current path (certain Molex connectors on the lemec boards/USB PCB are meant for this). I think that the LED matrices introduce some digital noise as they switch. I'm not sure if this the source of your noise or whether it is the more commonly observed phenomenon encountered when connecting USB audio devices. Could be that your USB +5V from the computer is not designed for the current that you draw from it or is inherently noisy. Peter tested a laptop and USB power bank, which were both quieter than even a phone charger, but a PC tower had a much noisier +5V line. He could filter out some of that noise but it was still worse than the laptop. His Eurorack PSU is transformer based with linear regulators and uses a copper bussboard. I would consider this system superior to a switching PSU and a flying busboard. I would tend towards the powered USB hub like TK. uses for the SEQ. Another user still had issues with their hub but then one might imagine that that PSU was also noisy. Very hard to guess. There are also solutions to galvanically isolate USB and could be worth a try? Some are quite cheap and some much more expensive. I can't guarantee anything though, so it might be throwing money at the problem without a satisfactory result. For the Euroceiver ground lift, the RS-485 standard recommends to keep the 0V reference connected, even though the power supplies on either end could be mutually floating. Imagine an extreme case where you have a long signal cable and a device either end. You have a massive antenna and a real "ground loop" if the power supplies are connected. In our case, we can have two different power supplies and we keep the signalling within the common-mode range by referencing to a common potential (0V). Without J0 jumpered, the signals may become more susceptible to data corruption as the differential voltages drift apart. You might notice that the setup works in your studio but not at a gig for example. Instead of leaving J0 unjumpered you could try to connect a low-value resistor (say 10R) of suitable power rating across it. This might attenuate the noise that you hear. Best regards, Andy

@TK. cool! When removing the offset resistors, consider that the LEDs on the octal PCB will be driven by a higher voltage (current). So consider at a minimum doubling the LED resistor values or even swap the LEDs on unipolar channels to different colour (with appropriate resistors).

I think you're doing great! :) MB_NG is very flexible and comprehensive and you just need to know where to look. I have used it mostly to prototype/test designs for other systems and have never built a controller based on it. I would also say that I (still) don't fully understand the syntax and just muddle my way through the examples. That said, the manual chapter that I linked should serve as a complete documentation source for you. I think your blog-posting style is fine if it serves as a waypoint, no problem! I hope that you don't expect me to answer all of your questions though, because I think that it is your job to learn as you go, and probably the easiest way is just to experiment with the hardware. This is my experience anyway; because I did not write and do not understand the code structures, I need to apply them in hardware and learn by doing. So if you change this parameter you will get that result, which translates something abstract (unintelligible code structures) into something tangible (LCD messages with the desired padding and string format). Still, if I can point you in the right direction I will (and anyone else should also feel free to do so!). Chin up! All good! Andy

This is >128 values though, so you will need to use NRPN. Again uCapps.de has all of the syntax explained:

This is no longer Liquid Crystal but segment displays :)

Consider an LED matrix with 7-segment displays.

Great work! The universal solution will be much appreciated as we have spare UARTs on the upcoming BLM :)

Good news, nice! This also seemed to eliminate the issue for the other user. I think it still remains to connect up the modular and the computer at the same time... Have fun, that looks like a really cool rig to use . Happy jams in 2020! Best; Andy

Thanks for the diagram! Just to preface, Eurorack is the Wild West of electronics. There are many ways you can put things together and some will give better results than others. The "fundamental flaw" in Eurorack is that the 0V line serves as both a reference and a current return path (i.e. an unbalanced system). If there are potential difference mismatches on the 0V lines, noise can result. So the idea is to have the 0V resistance as low as possible and to run power rails as much in a "star" configuration as possible. Searching a bit on MuffWiggler, the Row Power seems to have some history of noise problems. It might be that this is not the ideal way to power your Eurorack though there are a few things to try/clarify. Is the mains connector to the DC brick 2 or 3 pins on the brick end? With the power off and any voltages discharged (take care!), what is the resistance from the earth pin to the DC barrel? Your MG rack has two Row Powers, correct? Are things actually connected like this on separate rows or more like the hand-drawn diagram? What happens when you disconnect the ES-9 USB? Is it a desktop or laptop computer? If a laptop, what happens when it's running on battery? What is the power distribution in the case? Flying buss or bussboard? Is the Euroceiver powered with the case +5V or the onboard Vreg? (I suppose the Vreg) What happens if you power the Euroceiver with USB (check J5 jumper!) derived from the Mac while also powering the SEQ with USB? Better might be to use a powered hub to increase the available current. What happens if you power the SEQ from e.g. a phone USB charger rather than the Mac? What happens if you remove the Euroceiver J0 jumper? Generally you do want to leave this connected, but disconnecting it will cut the 0V connection between the SEQ and Eurorack. Cheers, Andy