Jump to content

monokinetic

Programmer
  • Posts

    325
  • Joined

  • Last visited

  • Days Won

    5

Posts posted by monokinetic

  1. That behaviour could be explained by PSU being overloaded (i.e. box runs, functions OK, but then stops as PSU can't handle the overload). How are you powering things? Does the regulator get hot (careful when touching, they can get very hot!)? etc

    Also, it might be useful to provide info like:

    - which bootloader you have on the PIC?

    - which operating system are you using on your computer?

    - which version of MIOS studio are you running?

    and so on.

    Good luck!

  2. Darn, I was hoping to do the same. Shame to hear it's not working with the updated firmware. I don't have any Launchpads, yet. But as a matter of interest I had a quick look at the code.

    At a guess, this line here:

    seqregex = re.compile("MIDIbox SEQ V4:MIDIbox SEQ V4 MIDI ([1-4]) [0-9]")

    Needs to be updated for the new v4+ firmware. Have you tried editing it to:

    seqregex = re.compile("MIDIbox SEQ V4+:MIDIbox SEQ V4+ MIDI ([1-4]) [0-9]")

    i.e. just add the + symbol two times to match the updated name you posted in the output of mido.get_input_names().

    Not sure how you tried to modify the regex, but that would be my first thing to try!

    Oh and you might find that the + character confuses the regex, so if that doesn't work, try using a \ before each + which you add. So it looks like this:

    seqregex = re.compile("MIDIbox SEQ V4\+:MIDIbox SEQ V4\+ MIDI ([1-4]) [0-9]")

    Hope you get it up and running....

    • Like 1
  3. @TK. thanks for the tip re: storing the values in the DEFAULT.CV2 file, yes that is much easier. My encoders thank you :)

    FWIW I tried this calibration stuff a bit more:

    something really seems weird with the way values are entered. For example, if I set cali to 5V and try to set the offset, I can scroll up to 255 and then it jumps down to -3840. I see the same with changing the value to 0xffff in the defaults file. Is that expected? If it helps, this is what MIOS_Studio shows during those 3 clicks of the encoder to the right:
    [1164.130] value=f000 (with x0=d800 x1=f000 y0=d800 y1=f000) -> result=ffe0 (offset value shows 254)
    [1164.690] value=f000 (with x0=d800 x1=f000 y0=d800 y1=f000) -> result=fff0 (offset value shows 255)
    [1165.852] value=f000 (with x0=d800 x1=f000 y0=d800 y1=f000) -> result=0000 (offset value shows -3840)

    I have done quite a bit of testing now with Cali out, I can provide more info but I don't want to bombard you all the numbers I measured if this is expected behaviour :)

    In the meantime, I have now found a way to get the output I expected from MBCV. My steps were:

    on Default page, channel type must be set to something other than Note, in my case I use CC. Otherwise it impacts on the output, regardless of any MIDI note ever being actually received. This really got me for a while :)

    on CV out set LFO1 to Sine, reasonably slow rate, with amp 127. Only then can I get reliable level, i.e. 0 to ~10V.

    I have various other notes from my studies of MBCV recently. I will write them up more once I'm sure of some things....

    Cheers, I'm having a blast with this so far. Can't wait to get my scope LCDs hooked up so I can put my multimeter away ;)

  4. Using a fresh compile of master from Github to try out that amazing v4.96 feature set, I tried to set a track to control an AOUT using the LFO as mentioned recently. Indeed assigning the extra out sends the LFO to the AOUT, excellent another source of modulation!

    But on my set up (unipolar original AOUT) the output seemed weird. For example, when I first turn on the LFO effect, and set the output to sine I get a triangle. Then changing through the waveforms gave me just lower volume versions of the wave, it seems to be a saw. Having seen this on the scope, this made me go back and check the output via MIDI CC and it seems the same. If it helps, I have experimented with amp and offset. I expected that with both set to 64 I should see the full waveform, i.e. triangle going from 0 to 127. But what I see seems weird to me!

    FWIW both the AOUT and the CC seemed to have same behaviour. Let me know if you need more info or the session sending :)

  5. Follow up having used the same hardware using MBSeq4 app:

    I'm not totally sure the calibration process will solve this. I tried the MBSeq outputting CC0 to 127 to an AOUT channel. That generates the expected voltage, without me changing the calibration of the AOUT at all. It feels to me more like I need a way to tell MBCV to double the final level of the CV output.

    In the process I also found something new, which I will post in the MBSeq main thread.

  6. Excellent TK! Many thanks for taking this on :) I have pulled from Gihub, compiled and had a bit of a play with it.

    So far I see some things which seem weird to me:

    the "Cali" list of options seems in the wrong order. For me it goes -

    Off, Min, Middle, Max, Wave, -4V, -3V, -1V, 3V, -5V (at this point offset becomes editable), -4V, -3V, -2V, -1V, 0V, 1V, 2V, 3V, 4V, 5V, Max: (with offset, unlike earlier Max value). Is that expected behaviour?

    When I select 5V from above list and try to change offset, as I scroll up it seems to jump down to -3840. Which is not within the range you mentioned (between -2048 and 2047 right?)

    Anyway, seems like it's close. If it helps, my testing process is to set a CV channel to output the triangle from LFO 1, set the rate reasonably slow and then the -CV output of LFO1 to 127. Also, how do you recommend to set the offset values if I want say the triangle wave LFO sent to CV with level 127 to range between 0 and max voltage (i.e. ~10V)? I had planned to set 5V to maximum offset, but wasn't sure how to proceed after that! My guess is that I set -5V with offset to output 0V, 0V to output ~5V and 5V to output 10V. Is that the idea?

    Oh and final suggestion: would it be possible to support the new caliaout at the Mios terminal i.e. in MIOS Studio? I ask because the way the offset jumped, now to reset I have to scroll a long way. From memory, MBSEQ stores these calibration values on the SD card. Does MBCV do the same i.e. could I just tweak the values on the SD?

    Thanks muchly in advance :)

    David

  7. Hi there Thorsten,

    aha yes I remember the new calibration stuff, it'd would be great to have in MBCV as well.....

    If you would be able to do it that would be an amazing new year's present, I had a look last year and it seemed a bit complicated for my skills :)

    But I'd gladly do some testing on both apps!

    Cheers

    David

  8. Hi folks,

    I've been using MBCV2 for a while now and I always felt like it wasn't working quite right with my synth. I went back to it this weekend and think I understand the problem now, indeed MBCV is not outputting full range CVs for me.

    I'm running an STM34F4 based standard core with an AOUT. I'm not using the "Balanced output extension" as mentioned on ucapps.de because my synth runs unipolar CVs in a range of 0 to 10V. When I use this setup with MBSEQ4 to run CVs I can output CC 0-127 and get 0-10V, so I know my hardware is good.

    Now, when I run MBCV2 and say for example assign an LFO to CV out 1, what I get is a 0 to approx 5V output. I think the reason is that MBCV2
    assumes that the AOUT is set up with the "Balanced output extension". I base this theory on the fact that I can tweak the mod matrix to force
    the LFO to a full range CV by setting offset high. I realise I could probably set up the opamp output stage of the AOUT to handle this
    (double gain I guess?), but I'd prefer a s/w solution :) 

    So I was wondering if that is correct, and how I could tweak the software to support unipolar AOUTs such as mine? I think a compile time option would be enough, or maybe the AOUT driver could be extended to have an option to represent unipolar vs bipolar set ups?

    Thanks for reading!

    David

  9. FWIW I use Lemur on an old iPad v1 and use it for the BLM of the sequencer, and with the MBSID. That communicates over OSC, is rock solid and make me think that going the SPI i2C route would be unnecessary. Only thing is that requires you using a computer to to go USB MIDI -> OSC -> Lemur. I might go computer free at some point using Midibox + an OSC interface, but so far things work so well and I given that I always end up having a computer on to capture MIDI and audio from my jams that it seemed like if it's working I'll get on with making some music :)

    Maybe before you complicate things trying to communicate using an internal bus, give that a try? Just a suggestion!

  10. 18 hours ago, TK. said:

    The MIOS repositories are now hosted at Github:

    aha excellent, I agree that this should make collaboration easier....

    10 hours ago, Antichambre said:

    I hope git is easier

    Heheh, define easier ;)

    I'm no Git expert, but if it helps, here's how I'd imagine doing things:

    Make a directory for this repo i.e.

    cd ~
    mkdir mios_stuff
    cd mios_stuff

    Clone the repo to your local dir:

    git clone https://github.com/midibox/mios32.git

    Let's check to see which branches are available in the new repo:

    cd mios32/
    ~/mios_stuff/mios32$ git branch -av
    * master                                      79292db corrected comments in source_me_* files
      remotes/origin/HEAD                         -> origin/master
      remotes/origin/feature/juce_5_4_3_migration 5e8148f MIOS Studio: updated release script
      remotes/origin/feature/mcan                 b7b78c2 corrected mios32_* namespace
      remotes/origin/master                       79292db corrected comments in source_me_* files

    I used the -av option to show all of the branches on the remote repo. Note the * by master, that shows which branch you have checked out at the moment. It looks like TK has set you up the feature/mcan branch. Let's assume you want to do some work on that, so switch to the MCAN branch like this:

    ~/mios_stuff/mios32$ git checkout feature/mcan
    Branch feature/mcan set up to track remote branch feature/mcan from origin.
    Switched to a new branch 'feature/mcan'

    Here is where things are a bit different to SVN. I'd try and explain it but you'd be better reading this:

    https://git-scm.com/book/en/v2/Git-Branching-Branches-in-a-Nutshell

    Then basically you can do your work, doing regular commits. Those go in to the branch you have checked out currently. Then when you are ready to share them with others you push to the repo, ensuring that your push goes to the feature/mcan branch. Eventually, once we know that everything you have done in feature/mcan is OK to add to the full mios32 then someone has to merge the changes in feature/mcan branch into the master branch. Then finally everyone can use MCAN and blow their minds with so many MIDI ports to play with ;)

    I also recommend looking up how Git stash works (I find that invaluable) :)

    Sorry for the slightly long post, and apologies in advance for things that a Git expert would tell you are wrong. But that's how I plan to use it so far!

    • Like 1
  11. Hi folks,

    nice to know this info is of use to some!

    @Rowan I'm afraid both of my attempts at building an AOUT_NG failed so I gave up on that path. But yes to what Zam wrote, basically all I did was recalculate the opamp resistors. Don't know why it didn't occur to me years ago, but that's life :)

    Good luck with your builds!

  12. As you probably know, the AOUT module is set up to support 1v/octave scaling, and that has been fine for me when I was using it with a Serge modular system. But I have been wondering how I could use my AOUTs with my Buchla-ish modular synth, which uses 1.2v/octave scaling. I contemplated building an opamp based solution, boosting the voltage a bit. But this seemed a bit of a hack to me, so I spent a bit of time understanding the opamp based outputs from the MAX525 DACs. I realised that by tweaking a resistor value I could get the AOUT to put out 1.2v/octave. I've tested this for a bit with two different oscillators (258r and 259r) and it seems fine for my uses. I'm no electrical engineer though, so if this mod blows up your rather expensive equipment I'm afraid you're on your own...

    That said, for any other Buchla users who want to use an AOUT to output 1.2v/octave, just change RY to 1.8K on any channels you want. That should give you the extra bit of gain required, and then you can trim it to 1.2v/octave. FWIW I usually use the MBSeqV4 app, then use the nice new calibration stuff to get things super accurate. I'm able to get submillivolt accuracy over the whole voltage range. Top voltage is obviously a bit higher in this case, almost 13v from my measurements. That seems to have been fine in my system, but YMMV.

    Also, I have set my AOUT so that V1-V4 have this tweak (1.8K for RY), and I use them for controlling Buchla oscillators. V5-V8 use the 2.2K resistor for RY still, and I use those channels for modulation CVs and to control the Serge oscillators which run at 1v/octave. All in all I'm most happy with this system and am glad I don't have to mess around with adding more circuitry.

    Hope that helps someone in the future!  
     

  13. @TK: oh my, did Christmas just come early :)

    Excellent to see some updates, thank you! Will test over the coming days.

    re: the AOUT extended calibration, is there a chance to merge this into MBCV2.0 at some point? I would find it really useful in that app as well.....

    Just a thought!

    Warm and festive regards

    David

  14. On 6/22/2018 at 11:53 PM, TK. said:

     I added an experimental AOUT calibration feature:

    
       o V4+ AOUT: support for AOUT channel calibration.
         In the CV configuration menu, turn GP ENC #7 until 0V/1V/2V/.../Max will be visible.
         Calibrate the offset of the target value with ENC #8.
         Each V has a dedicated calibration value which can be configured this way, the output will be interpolated accordingly.
         Note that with exiting the CV configuration menu the calibration values are stored on SD Card in the MBSEQ_GC.V4 file
         (-> CV_Cali <cv-counting-from-0>).
         You could backup/set/restore the values from there if required.
         To reset all calibration values: delete all CV_Cali items in MBSEQ_GC.V4, store the file and enter "reset" in MIOS terminal.
    

    Please try this version: http://www.ucapps.de/mios32/midibox_seq_v4_095_pre13.zip

    @TK wow this works really quite nicely. FYI I tried v4_095_pre14 on my STM32F4 based MBSeq. I observed the following:

    - when scrolling through the options using ENC #7 it seems to go min/middle/max/wave. When set to wave there's a full voltage range triangle type wave coming out of CV, is that intended?

    - once I scroll through to the end of possible values using 7, the last value is Max:. That seems a bit weird over here. When I initially scroll to Max I get almost 0 volts out of the CV channel, then if I touch ENC #8 the offset value on screen jumps to -3700 ish! At that point I get the full voltage out of CV. I guess this might weird out the interpolation a bit? And I started to scroll ENC #8 to get it back to around 0 but it was taking ages and I gave up :)

    Otherwise, this change seems great. I have used it to play a couple of basic patches from MBSeq to my modular and it seems really accurate, excellent work, as usual!

    Tiny feature idea: could presses of GP8 cycle through min/0/max offset? for quick changes?

    Cheers

    David

  15. Quote

    Thank you for the response. The reason I need an additional NG core is that I need more CV outs controlling other functions and also need to control Leds and scan pots and buttons.

    Aha OK makes sense. Yeah the cores are cheap enough to make splitting tasks between them worth it IMO.

    Quote

    It is good to know that it should be capable to make EG, LFO and mod matrix. Is the control via MIDi implemented or it is still in wish list?

    Not sure which type of MIDI control you mean. The Lemur interface is using NRPNs from memory, so any other interface could use that I guess. I'm not sure of the status of CC control, I've not got things hooked up at the moment, otherwise I'd check for you. Prob best to check the code if you want to be 100% sure...

×
×
  • Create New...