Jump to content


  • Content Count

  • Joined

  • Last visited

  • Days Won


Posts posted by FantomXR

  1. I use this method a year or so on different cores with different configurations. My last build was a faderbox with 9 100mm faders which can send 10bit resolution with USB-power. If you only need 7bit, there shouldn't be any problem.

    Yes, it has 4V Output. Of course 5V would be better. But as @Zam already said: You can not really clean up 5V USB and end up with 5V. You will loose something.

    I'm okay with it, because it works perfectly. Even on breadboard. No need for a high-precision PCB-design to get this working. Just try it. It costs just a few cents. 

    I have a few capacitors connected to the USB-5V-rail and an inductor in front. But even that is not necessary. My earlier designs didn't have that filter-caps and worked very well too. 

  2. The solution is easy:

    Go and get a MCP1541. You connect USB 5V to the input and it's output you connect to the VREF input on the AINSER8 (of course you need to cut the trace between VREF and +5V on the AINSER) and to the all analog controls that are connected to the AINSER8. 

    This little IC solved all my problems I had with jitter. Especially when running just at 7bit, this should do the trick.

  3. Hey people,

    I have a question:
    I use OLEDs with a resolution of 128 x 64. I use them in a row. I set up a sysex-stream which shows text on the displays. This stream is set to lcd(1:1:1). 

    If the text takes more space than 128px, the rest gets printed on the next screen(s). I know that this could be useful. But not in my application. Can I somehow disable that? I took a look into the code but I didn't find it yet. 

    Each OLED should get it's own stream. If I do not push the streams in the correct order (f.e. first stream 2 => OLED 2 and then stream 1 => OLED 1) it will overwrite stream 2 of the 2nd OLED.


  4. Hey people,

    I try to understand how the DIN is even working in the STM32F4-Core. 
    PB14 is not going through the HCT541 buffer. It goes directly into the IC. If I understood it correctly the purpose of the HCT541 is to do a level-amplification. It takes the level from the STMs outputs (which is 3.3V - 0.4V for HIGH for CMOS or 2.4V or TTL) and amplifies them. As the HC165 and HC595 are powered from +5V they need 3.15V to see a HIGH-signal. As the HCT541 delivers 4.5V for HIGH the HC595 does of perfectly work.

    But why do the HC165 even work? As I said it doesn't get an amplification. So in theory the HC165 sees 2.9V for HIGH but that's not enough so it should stay LOW and doesn't work. But it does...

    Can someone explain this behavior?


  5. My mainboard uses the 407.

    Hm... there is still something I do not get.

    The discovery-boards use the micro-USB-connector for USB-Host-Mode. That means, when connecting a device (mouse or MIDI controller) to the discovery-board via this board it's not possible to connect to the core itself through USB (because the USB-port (PA12 & PA13) is blocked through the device). But in your video I see that you still have your core connected through USB to the computer. How did you do that?? :-) 

  6. Hey people,

    here is the problem:
    FATAR has changed the diode-matrix of their keybeds recently. So the DIO-matrix is not usable for those keybeds anymore. I want to get it running and of course I'm happy to design a new DIO-matrix and of course post it here so everybody can use it.

    So... I'm not 100% sure about the diode-matrix yet. In the old diode-matrix both keyboard sides (left and right) had separate rows. So the left side had 8 rows, the right side had 8 rows:

    Look at this pictures. With "rows" I mean T0-T7. 

    In the new diode-matrix it looks like both sides use the same rows. Because if I plug in a new keybed (via an adapter) into a DIO-matrix (which I did many times before with old keybeds though old diode-matrix) it outputs correct values from a-1 (lowest key) till middle-C. With the next key C#3 the left side starts. But C#3 outputs a-1. Further tests confirmed my assumption, that both sides share the same set of T0-T7. 

    So that means: I need a shiftregister more than is available on the DIO-matrix. At the moment I have T0-T7 connected to the 8 inputs of the HC165, MK&BK are connected to the 16 outputs of the two HC595s. So in total I can connect 64 keys to the DIO-matrix. I now want to add another 595 to connect the last keys. Of course connecting the 595 itself is easy and done quickly. But I need to adapt the code and what I've tried so far was not working. I need to add a dout_sr3 to the code...

    Any help?? Maybe @TK. could point me in the right direction!

    Thanks people!


  7. Hey people,

    for those of you who are looking for a nice online gerber view, I can recommend AISLER:

    You can either upload the gerber files, but they do accept also right away eagle-brd.-files which makes it very easy to visualize it.

    Prices seem a bit expensive. But anyway: for visualizing the PCB this is the best tool I've ever seen.


  8. Do you send single note off events?

    This is not necessary. I assume that Thorsten took care of the classical MIDI implementation. 
    If you send CC #123 from NG to KB, KB should interpret this as "All notes off". This should speed up the process significantly. 

  • Create New...