TK. Posted April 8, 2007 Report Share Posted April 8, 2007 MIDIbox V1.74685 has been released as an interim solution for everybody, who is planning to upgrade to V2 later, which will require a PIC18F4685. A first alpha (incomplete) release of MBSID V2 can be expected in one or two months. Until then, just use this version and enjoy the new possibilities! :)This is the best "V1" ever released, because the MBNet approach speeds up the transfer of patches to the slaves significantly. WT patches are now directly edited in RAM, which is also a big advantage.Details are described in the ChangeLog: http://www.ucapps.de/midibox_sid_changelog.htmlAnd the release package can be downloaded from http://www.ucapps.de/mios_download.htmlNote: the PIC18F4685 is a very new microcontroller, it's hard to find it in local or mailorder shops.But since it provides so important features for MBSID V2 (CAN and bugfree EUSART), this shouldn't prevent me from using it for the upcoming projectSmashTV will do batch orders directly from Microchip in future and sell the chips in his webshop for an extremely fair price: $9.95 per pieceAnd the other good news: doc is doing the distribution for Europe, he offers the chip for EUR 7.90 per piece (+ shipping costs) - just contact him via PM, 30 pieces are already available! :)The chips are preprogrammed with bootloader and MIOSBest Regards, Thorsten. Link to comment Share on other sites More sharing options...
QBAS Posted April 8, 2007 Report Share Posted April 8, 2007 WOW- thanks x10000000000000000000! Link to comment Share on other sites More sharing options...
Wilba Posted April 9, 2007 Report Share Posted April 9, 2007 My PLED display from OSD doesn't work properly in 4-bit mode, and MIOS detects this and disables the display (it looks like the display freezes but the app still runs normally). If anyone has similar problems, I can send my workaround... a custom LCD driver that uses two pins from port E as substitutes for the two pins being used for the CAN bus. So a little tip: only temporarily convert your Core->LCD cable connections to 4-bit. In 4-bit mode, you can (should?) leave D0-D4 pins on the LCD unconnected, so just desoldering those wires on the LCD is enough. If you have problems you can reconnect it as 8-bit... in other words, break the tracks to D2 and D3 on J15, connecting those pins to port E pins, keeping D0 and D1 connected to port B.Also the link to the diagram in the changelog is to a low-res gif, the better version is here:http://www.ucapps.de/midibox_sid/mbsid_v2_communication.pdfOne final thing TK hasn't mentioned... this version supports all switches and LEDs in a common matrix... should you want to do the same hardware changes in your current hardware (i.e. only 5xDIN and 3xDOUT for the entire "step C" control surface!) then contact me and I'll give some advice. Link to comment Share on other sites More sharing options...
Wilba Posted April 9, 2007 Report Share Posted April 9, 2007 BTW, you can still use your JDM burner for the PIC18F4685!http://www.midibox.org/dokuwiki/doku.php?id=jdm_with_picpgm Link to comment Share on other sites More sharing options...
Altitude Posted April 9, 2007 Report Share Posted April 9, 2007 Does everything also hold true for the 4682?Wilba: care to post your fix? I have the same problem with my LCD which was bought from Smash TV so this may be a common problem for a lot of ppl Link to comment Share on other sites More sharing options...
Wilba Posted April 10, 2007 Report Share Posted April 10, 2007 Does everything also hold true for the 4682?If you mean does JDM work for 4682, probably yes, try it with PICPgm.Wilba: care to post your fix? I have the same problem with my LCD which was bought from Smash TV so this may be a common problem for a lot of pplUntil I find a better place for it, (or TK puts it in the downloads section):http://www.midibox.org/dokuwiki/doku.php?id=pic18f4685_8bit_lcd_driverThere's a complete lack of documentation ;D basically RE1 replaces RB2, (connect PIC pin 9 to D2 of J15), RE2 replaces RB3 ( connect PIC pin 8 to D3 of J15). This means you sacrifice two of eight analog input pins, but they're the 7th and 8th ones. Link to comment Share on other sites More sharing options...
TK. Posted April 10, 2007 Author Report Share Posted April 10, 2007 I will integrate the driver into the MBSID V2 application if really so many LCDs are affectedBest Regards, Thorsten. Link to comment Share on other sites More sharing options...
Altitude Posted April 10, 2007 Report Share Posted April 10, 2007 If you mean does JDM work for 4682, probably yes, try it with PICPgm.I was wondering more along the lines of do all the features of the 4685 hold true for the 4682. I have not done an in depth comparison but superficially, the 4685 looked like it only varied in the size.As far as burners go, I use the DIY kit 150 burner and that works for everything, I added an entry to its chipinfo file identifying the 4685 and was all I needed to program the bootloader Link to comment Share on other sites More sharing options...
stryd_one Posted April 10, 2007 Report Share Posted April 10, 2007 The 4682 has a smaller code memory. I suspect that TK will use every bit available in the larger 4685 for V2. You might be able to alter this V1 code for that PIC but it's not worth the effort. Link to comment Share on other sites More sharing options...
Altitude Posted April 10, 2007 Report Share Posted April 10, 2007 what about the EUSART bug? Link to comment Share on other sites More sharing options...
stryd_one Posted April 10, 2007 Report Share Posted April 10, 2007 Fixed, see below Link to comment Share on other sites More sharing options...
dj3nk Posted April 10, 2007 Report Share Posted April 10, 2007 But since it provides so important features for MBSID V2 (CAN and bugfree EUSART), this shouldn't prevent me from using it for the upcoming projectDId I misunderstand bugfree eusart ?greetz Link to comment Share on other sites More sharing options...
stryd_one Posted April 10, 2007 Report Share Posted April 10, 2007 DId I misunderstand bugfree eusart ?No, I think I just missed it! Fixed my post so as to avoid confusion, thanks dj3nk! Link to comment Share on other sites More sharing options...
visibleman Posted April 11, 2007 Report Share Posted April 11, 2007 MIDIbox V1.74685 has been released as an interim solution for everybody, who is planning to upgrade to V2 later, which will require a PIC18F4685. A first alpha (incomplete) release of MBSID V2 can be expected in one or two months. Until then, just use this version and enjoy the new possibilities! :)wow! great news!!!I remember reading that with the PIC18F4685 it is now possible to connect two Sid modules to each Core, for 'kind of stereo' effect. The Sids cant be controlled separately in this case, they are always linked.I think Wilba used this in his wonderful MB-6582.Can anybody post wiring schematics for such connection? TK? Wilba? Thanks in advance! Link to comment Share on other sites More sharing options...
Wilba Posted April 11, 2007 Report Share Posted April 11, 2007 I haven't got around to a wiring diagram yet, sorry... you can look at the PCB layout and work it out ;DIn the v1.74685 firmware, it will drive two SIDs at the same time, but the same data goes to both so there's no stereo effects (other than mismatched filter caps causing differences in cutoff freq.). So essentially it's no different to connecting two SID modules with the same cable.But if you're planning ahead for V2, then do the same thing, connect two SID modules with the same cable, except for the wire that connects to the CS pin on the SID (SID:J2:SO), for the right channel SID, connect Core:J14 instead.I've optimized it on my PCB as both SIDs are connected to the same pair of shift registers... if you use two SID modules, the shift registers on both are outputting the same data since they'll be connected to the same serial input ;D Link to comment Share on other sites More sharing options...
NorthernLightX Posted June 17, 2007 Report Share Posted June 17, 2007 I'm probably posting something really stupid, so forgive me ;DIf wilba is borrowing pins from another connector to be able to use an 8-bit LCD driver, wouldn't it be easier to use those borrowed pins for the CAN bus instead, and leave the LCD connector intact? Link to comment Share on other sites More sharing options...
stryd_one Posted June 17, 2007 Report Share Posted June 17, 2007 The CAN bus is partially implemented in hardware and so the pins for CAN are restricted to the ones being used. </short version> ;)The display driver works best on the normal LCD pins, but using other pins creates a minimal performance loss (so small that you don't notice it) This is because the LCD connector uses a whole 8-pin port, which means you can write a whole byte to the screen in one go. If you use other pins, you have to drive those special pins individually...but we're talking about a few instructions, like 1/1000000ths of a second, it's no big deal.The CAN bus uses hardware built into the PIC that takes care of lots of stuff that would take up a great deal of uP cycles, in fact I think it would be too fast for software to drive it. That's page 275 of the data sheet, it's a 70 page chunk of geekyness that I couldn't explain much better :) TK has a significant amount of work being done in software to make it easier to use for us mortals, so not all of the datasheet applies... Link to comment Share on other sites More sharing options...
NorthernLightX Posted June 17, 2007 Report Share Posted June 17, 2007 Thanks Stryd, that was a very elaborate explanation.So, I presume that in a future revision the LCD port pins get rewired in some way? And the CAN bus gets it's own pins? Or is that not likely to happen, because most midiboxes do not use slave cores and therefore have no need for CAN? Link to comment Share on other sites More sharing options...
TK. Posted June 17, 2007 Author Report Share Posted June 17, 2007 This is only an issue for devices which get use of the ECAN peripheral, and for MIDIboxes which are using excotic displays which don't support the 4bit mode properly (currently I'm only aware of issues with PLED (they will be discontinuied anyhow) and VFD displays, common LCDs are working ok - I tested this with all LCDs I own, and I own a lot of different types! ;-)The drawback of the 8bit mode workaround: two IO pins get lost - so this is no general solution.To avoid misunderstandings I repeat this in other words: for MBSID V2 users this issue is not the end of the world, as for the case a LCD doesn't support the 4bit mode there is a workaround. But this workaround is only sufficient for MBSID V2 - it wouldn't work with a MIDIbox which allocates all IO pins.Therefore don't be afraid, the layouts don't need to be changed. :)Best Regards, Thorsten. Link to comment Share on other sites More sharing options...
Recommended Posts