philetaylor

If the BC337 was dead then you wouldn't see 5v between the A and K of the backlight (which I think you said you have). Are you SURE that you have it the right way round? pin 15 on the LCD is cathode (-) and pin 16 is anode (+) so they should go to B- and B+ on the CORE. Another possibility is that the backlight has solder bridges that may not be linked, can you check continuity between pins 15/16 on the LCD and the actual connections to the backlight? Looking at the datasheet, it is possible to switch the polarity with different bridges. Assuming you have the DEM 40271 SYH-LY, this might be your problem, take a look at: http://www.soselectronic.hu/a_info/resource/d/dem/dem40271sy-ly.pdf Phil

That's great news. Your project looks really good! Have you tried adjusting P1 as that controls the backlight brightness? I would be surprised if both backlights are faulty! Thanks Phil

As they say "a picture tells a thousand words!" :) So MB-SEQ is up and running... That is a good sign! Can you describe how you have connected the LED's? As they are all on, can you check that you have wired them as per this: http://www.ucapps.de/midibox_seq/mbseq_v3_dout.pdf On the DOUT module, you should have a 220R resistor between each output and the LED with the other LED leg connected to Vs. Can you check that you have done this? If you have connected the LEDs to Vd by mistake then you would get the effect that you are seeing. Also have you connected the buttons/encoders like this: http://www.ucapps.de/midibox_seq/mbseq_v3_din.pdf You don't say if the buttons/encoders are working or not? Thanks Phil

IMHO the CORE32 is much easier to program than the CORE8, everything (including MIOS) is written in well documented C code. The down side is that not all of the MIDIbox apps are currently available for the CORE32 which may be a problem as it is often easier to modify an existing program than create a new one "from scratch". The compiler that the CORE32 uses is also the industry standard GCC, whereas to program in C on the CORE8 you must use SDCC which is a much less "feature rich" compiler, or use PIC assembler.... The CORE32 is much faster and has much more available memory than the CORE8 so would certainly be my choice for any new projects! Thanks Phil

When you say there is no light, do you mean no backlight? What model display are you using? Are you sure you have A (Anode) and K (Cathode) connected the right way round? It is an LED backlight and not EL? Have you loaded the MB-SEQ software? What is on the display? It may be easier to ask on the French board http://midibox.org/f...um/24-francais/ although your English isn't that bad :) Giving as much information as possible is always a good idea! Thanks Phil

Hi. That display uses the UC1701 controller, I have never tried that one but it does look to be very similar to the EA DOGM128-6 the initialization looks slightly different but it should still work and if not would be VERY easy to fix! Just for information, it is the EA DOGXL160-7 that uses the UC1610 and I have written a MIOS32 driver for this but due to the way it works (2 bits per pixel) it is a bit more complicated to write a MIOS8 driver and I am unlikely to try. As far as Midibox LC is concerned, if you plan to use 8 displays then it will need some modification to work with all 8 displays but as LC is written in C it is much easier and it will mainly be correcting positions of things. As the current driver assumes that the display is 128x64 then you can either modify the driver (pretty easy as you just change every occurrance of the number 128 to 102 in app_lcd.c or just make sure you don't display anything between 103 and 128 (as they will not appear!) The driver just appears to MIOS as a single LCD which (for 8 LCD's) is 256x256 the layout should be fairly obvious if you look at http://svnmios.midib...g%2Fapp_lcd.inc If you do decide to go down this route I will happilly help you get it working. (it is always good to stay on the right side of somebody with CNC access :) ) Cheers Phil

Yes pretty much, again the length of the cable really doesn't matter with LEDs either. The worst thing that can happen is that the LED is slightly dimmer if you use a 'really' long cable! Cheers Phil

As long as you are only talking about (digital) buttons then 22" should be fine as the PIC only has to tell whether the input is 5V (off) or 0V (on) so cable length (within reason) isn't a problem. Where you would see problems is with analogue inputs (AIN) for faders and pots etc as even 7 bit values rely on a voltage difference of 0.04V per value (even less on MIOS32 as it works at 3.3v) and there is much more chance of noise getting in and causing "jitter" when using longer cables (although this can be mitigated by using screened cables. Cheers Phil

Hi. MIOS only supports 2 character LCD modules although it would be possible to increase this is as long as you have enough spare i/o pins available on your PIC (and a knowledge of PIC assembly to modify the code :rolleyes: If you want more than 2 LCD's, the easiest method is to move to a serial driven graphical LCD, my personal favourites are the EA DOG series ( http://www.lcd-modul...oducts/dog.html ) These are fully supported by both MIOS8 and MIOS32 for up to 8 displays (I know I wrote the driver!) Another option is the PCD8544 based displays (from mobile phones) these are small 84x48 GLCD's and up to 8 are supported. The reason that up to 8 displays can be supported on these GLCD's is that they use serial rather than parallel data so the 8 data lines can be used as enable/select lines instead. EDIT: If you go for the EA DOG displays, make sure you get the graphical not character ones as the MIOS driver only supports 2 of the character type (as it uses the parallel not serial interface) Thanks Phil

Personally I think white (or blue) led's are a bad idea for this sort of project however to answer your question, these LEDs (in bulk from China) are unlikely to be subject to particularly stringent testing or they may even have failed quality control testing hence the variance in brightness. As white LED's are effectively blue ones with chemicals added I think that Wilba's comments on blue LEDs apply: Phil

As I guess I would say it is the size of the opening of the PT-10 case.... From: http://www.pactecenclosures.com/Plastic-Enclosures/PT-10.html 232.41 x 163.27 (mm) Cheers Phil

Hi. IIRC, the MIDIbox64 (or 128) can do this in software so no need to design a flip-flop circuit, buttons can be programmed to toggle values and DOUT's can be configured to show the current 'state' of the button. Thanks Phil

Personally I use Electrolube Fluxclene. It come with a useful brush head attachment for scrubbing any difficult to clear flux.

Certainly interesting if you put that word into Google. Without reading German, I think I get the general idea from the first result :shocked: Phil

The opto's can be a bit difficult. I have blown a few in my time but luckilly they are usually quite cheap. You will have to have done something pretty bad to kill the PIC. There is another thread around at the moment that is listing the worse things people have done to PICs and they survived. The best so far is one from arumblack who put it in the wrong way round and it got so hot that the paper label started to burn.... After all that it is still working :) Phil

I have never killed a crystal by soldering it but I am always careful to not overheat them. Have you soldered your crystal in place or "just inserted it"? It must be soldered to make a good contact. Cheers Phil

Hi. The Core has plated through holes (well the ones from SmashTV do anyway) so no top soldering required. I can tell from your posts so far that you are careful and methodical so we can assume that you have all of the correct components in the correct places! If you aren't getting the initial sysex message then this is often caused by a faulty opto but you have tried swapping that. I have seen posts where the problem was poor connection to the crystal, this would also cause this problem. Have you run through the troubleshooting http://www.ucapps.de/howto_debug_midi.html as this will usually give a good indication of the problem. Cheers Phil

I ordered a front panel for my Sequencer (using Wilba's design) from Julian at www.thebeast.co.uk. The panel looks great but as anybody following the discussions will know, it isn't easy for Julian to make threaded blinds so we decided to go for blinds that are large enough to fit a hex spacer in. They are 5.72mm diameter x 2mm deep. As JB Weld isn't that common in the UK, I thought I would try Loctite "Bond Loc". This is available from Maplin http://www.maplin.co...?ModuleNo=46006 and claims to bond metal very well (it contains stainless steel!). I thought I would try the spacers for the LCD, there are 8mm deep (including the 2mm blind) so the LCD stands off 6mm from the back of the panel. Once I have some 2mm perspex/acrylic for the window, this should be fine (hopefully). While waiting for the Bond Loc to cure I built the Core32 module that I had earmarked for this. That went without a hitch and seems to work fine! The LCD's fit fine, bit of a cock-up though, I bought them cheap off t'ebay and they are Electroluminescent backlight ones. Oh well. One 5v inverter order later (only £12 so not too bad), That works quite well the picture shows it is a blue/green backlight so not unpleasant. I am still undecided whether I like the displays but they will do for now. I am now stopped until my PCB and other kit arrives, hopefully in January. Phil

Hi Wilba, I did notice that (after I asked Julian to manufacture the panel). I thought that I would try it with the fixings as is and if the board does flex too much in use then I would just epoxy some 10mm standoffs direct to the panel. Thanks Phil

Oh Julian, I wanted to show off the panel but you beat me to it :rolleyes: :rolleyes: Phil

Oh well , that's a shame. I'm sure that vinylboy would have been a real asset to the community! Too bad, all the more for us jumped up geeks I suppose :) Phil

I think Lylehaze has hit the nail on the head, you can do some really complex merging/routing with MIDIbox. Have you looked at the Midi Router project? http://www.ucapps.de/midi_router.html It supports up to 4 IIC modules for a total of 5 MIDI IN/OUT ports and can be easily modified (with C knowledge) to merge/route between any combination of devices. An off the shelf MIDI merger is usually pretty dumb..... Phil

I would say not very fragile at all. I have done some "really" stupid things to PICs in the past. Installed them the wrong way round, shorted out pins and generally abused them and everytime they just work! Are you getting the upload request MIDI message? If so your PIC is still alive. Cheers Phil

You qre quite right.... From the core_stm32 page: ADCs cannot be supplied at 5V, 3.3V is the limit. This increases the signal/noise ratioCheers Phil

I have had quite a few jitter problems using USB power and unshielded cables so yes I would definately recommend using shielded cable and external power. If you are using an AINX4 then I believe that the range is 0-5V but don't quote me on that! The latest updates that TK has done have helped to reduce jitter by setting an idle threshold. Cheers Phil