jrp Posted March 12, 2014 Report Share Posted March 12, 2014 (edited) Hello, I rechecked everything so many times... i must see i am stuck :-( I build a core with 18LF4685 (LF version is compatible, right?) , and a sid module together on veroboard. The Core is working and recognized, the test-tone application gives me a triangle just as it should. Now the strange things: Right now the interconnection test fails with all connections coming from the shift registers. Only CS# (directly driven by the pic) is working correctly. I changed the shiftregisters to new ones with no effect. Strangely, when i first did the test everything was fine (although the other problems still where present). It just stopped working. As soon as i upload any of the MB_Sid applications (setup_6851, setup_6852, etc) I get the message "bootloader is up and running" after reboot. With every query i get random messages : No connection, bootloader up, application up. Playing on the keyboard gives me nothing, usually: Sometimes for a brief moment i can see the square wave on my oszi. Every other application i upload to the core works seems to work, although connectiong a din(1) and uploading MBIO128 gives me CC messages and Note ons. Should be Note On only i think? But a query will always say : "application up and running", it seems like only the sid applications are affected. I checked the connections so many times (and i am willing to check again), also all voltages are where they should be. I included all pullup resistors from the core schematic. My system is MB Pro with snow leopard and MIOS Studio 2.4.6 I tried a generic 4*4 Midi interface, a Midisport 4*4 and the midi ports from a Motu 8pre. All seem to do the same. Edited March 19, 2014 by jrp Quote Link to comment Share on other sites More sharing options...
jrp Posted March 12, 2014 Author Report Share Posted March 12, 2014 I am powering my 7805 regulator from 15v during testing, parallel to my 7812. But i am only running core and sid module (no lcd, no leds) and i put on a heatsink. 5V looks nice and solid on the scope so i see no problem there. Quote Link to comment Share on other sites More sharing options...
jrp Posted March 12, 2014 Author Report Share Posted March 12, 2014 Now i had the chance to test on a win7 pc. Same results. I also cut all connections to the sid-module leaving only the pic-core with all pullups and midi i/o. No LCD, no DIns or DOuts. No bankstick, no wires, soldering on any other pin. I even erased the pic and burned the bootloader for the 18f4685 again for a fresh start. It is all the same: Uploading MIOS 1.9gfor pic 18f4685 works flawless. Application is up and running. Last Message from core: f0 00 00 7e 40 00 0f 00 00 f7 Uploading Setup_6581.hex returns lots of f0 00 00 7e 40 00 0f xx f7 messages Then i get Bootloader is up and running and a f0 00 00 7e 40 00 01 f7 every couple of seconds. Then, if i hit query i get the same, OR: No response alert, and then f0 00 00 7e 40 00 0e 03 32 f7 f0 00 00 7e 40 00 0e 03 32 f7 f0 00 00 7e 40 00 0e 06 70 f7 after that it returns to the ever repeating f0 00 00 7e 40 00 01 f7 Sometimes query returns a application is up and running and the message f0 00 00 7e 40 00 0f 00 00 f7 f0 00 00 7e 40 00 0e 03 32 f7 f0 00 00 7e 40 00 0e 06 01 f7 only to return to the good old upload request.... All other applications i tried (midibox16e, midibox_seq, Midibox_cv and some more) load without any trouble showing a happy "application is up and running". What could be the cause of this? Quote Link to comment Share on other sites More sharing options...
Marxon Posted March 13, 2014 Report Share Posted March 13, 2014 Hi jrp! All other applications i tried (midibox16e, midibox_seq, Midibox_cv and some more) load without any trouble showing a happy "application is up and running". What me wonders is that afaik there is no 18f4685 version of MidiboxSeq and MidiboxCV. So how could you get them working? Best regards Marxon Quote Link to comment Share on other sites More sharing options...
jrp Posted March 13, 2014 Author Report Share Posted March 13, 2014 Well, i have no idea but that the 18f4685 is a newer chip and is it downward compatible. Also i did not test if these applications are running as expected. I have no lcd and no other interface connected. All i could see was that MIOS Studio was saying everything is ok. So i can see that the core seems to be fine and MIDI transmission seems to be working in both directions. Quote Link to comment Share on other sites More sharing options...
jrp Posted March 13, 2014 Author Report Share Posted March 13, 2014 i just tried this with a different core module (and i know it´s working, it´s part of the controller i use every day). Same result with the files i want to use Mios_v1.9g_pic18f4685 : Application is up and running setup_6581: Bootloader is up and running. But what is even stranger: When i try to upload apps meant for the other pic chips like MB_SEQ i get an error: Detected Mios8 and Mios32 response - selection not supported yet! Quote Link to comment Share on other sites More sharing options...
Marxon Posted March 13, 2014 Report Share Posted March 13, 2014 Are sid and core circuits connected all the time on your veroboard? I would try to disconnect the sid part from the core part and check it again. Quote Link to comment Share on other sites More sharing options...
jrp Posted March 14, 2014 Author Report Share Posted March 14, 2014 (edited) As mentioned in my posts, i disconnected the sid circuitry (no difference, and the testtone app was giving me the 1khz triangle) i used a different core (no difference) I tried three midi interfaces on OSX and Windows (no difference) Only thing i haven´t changed is the PIC as it´s the only 18f4685 i have. Maybe i got a fake chip? Any tests i could do? Edited March 14, 2014 by jrp Quote Link to comment Share on other sites More sharing options...
jrp Posted March 14, 2014 Author Report Share Posted March 14, 2014 (edited) Very strange: I used a third core module, this time not on veroboard like the other two but on a pcb from mike. It´s running! So for one thing i am happy, but on the other hand i would like to know what´s wrong with my core build. Has anyone experienced problems like me before? A core running only certain apps? The only difference on my veroboard layouts that i can see is that i tied pin 1 directly to 5v, ommiting the 100r resistor. In my understanding it´s not necessary if the core is not to be resetted under power, right? I´ll try to trace everything, but i really see no other difference. Strange, isn´t it? Edit: Added the 100r on pin 1, no difference. The Pic that was running the Sid app on the other core is saying bootloader is running again. scratching my head... Edited March 14, 2014 by jrp Quote Link to comment Share on other sites More sharing options...
Marxon Posted March 14, 2014 Report Share Posted March 14, 2014 (edited) I have experienced your story in similar form many, many times. My Midibox Sid, Seq, FM and NG are built to 95% on veroboard... Has anyone experienced problems like me before? A core running only certain apps? I guess the error should be found in the parts that are only used from the sid app. Please recheck the soldering/wiring of PIC pin 29, 28, 24, 23, 22 and 17. Edited March 14, 2014 by Marxon Quote Link to comment Share on other sites More sharing options...
jrp Posted March 17, 2014 Author Report Share Posted March 17, 2014 (edited) i unsoldered all the pins related to the sid module. It didn´t change a thing. Now with everything connected again i still have: Testtone working fine Sid interconection test all good, but one thing i noticed is that when i play the key to set CS# low, the last digital pin i checked remains high. According to Mios Studio all pins should be low. Don´t know if this indicates a problem or not. MB_Sid app is saying bootloader up most of the time, sometimes no connection, sometimes application is up. When that happens i still get no output from the sid, only a steady voltage around 6v. But sometimes when i play a key that voltages steps down mybe 0.5V. Also i see the clock signal (pin 17 of pic, pin 6 of sid) going on and off every few seconds. With the testtone app it´s steady. Supply voltages is 5v and stable. Edited March 17, 2014 by jrp Quote Link to comment Share on other sites More sharing options...
TK. Posted March 17, 2014 Report Share Posted March 17, 2014 Also i see the clock signal (pin 17 of pic, pin 6 of sid) going on and off every few seconds. This could be a watchdog reset caused by corrupted app upload. Do you still get error messages when you are trying to upload the application? (the upload window will notify about the number of errors) Best Regards, Thorsten. Quote Link to comment Share on other sites More sharing options...
jrp Posted March 18, 2014 Author Report Share Posted March 18, 2014 (edited) Hello, I was hoping that some of the behavior might ring a bell. I can see no error message in mios studio. Trying to contact the core setup_6581.hex contains 80960 bytes (319 blocks). Range 0x0003000-0x0015aff (76544 bytes) - PIC Flash Range 0x0017000-0x0017fff (4096 bytes) - PIC Flash Range 0x00f00000-0x00f003ff (1024 bytes) - PIC EEPROM Upload of 81664 bytes completed after 36.24s (2.20 kb/s) If i insert this pic into my other core module it is working fine! i hooked my veroboard sid to this core and i can play the sid. So upload is working fine on my trouble-core, but something else must be wrong with the hardware. I am really confused by now. I checked a lot of times and my build is actually quite neat. Also i have nothing connected to my core, only the pullups are in place and the pins for di/o chain, sid and bankstick have been soldered. (+ all the vital components, psu, midi i/o of course). So it´s really not much on there and i would really like to know what hardware mistake can cause this certain application to crash while upload and all other applications are working... Edited March 18, 2014 by jrp Quote Link to comment Share on other sites More sharing options...
jrp Posted March 19, 2014 Author Report Share Posted March 19, 2014 I just solved this strange issue by connecting a lcd (in 4 bit mode) and a pullup resistor to pin 36. That was it. Everything seems to work now. Still i find it strange that this only troubled the pic when running any of the sid applications. There were no problems with interconection test, testtone or anything else i uploaded. So if this is a new bug nobody ever saw since everyone oses a lcd from the start - then now we know :) Being confident about my build and soldering I came to the conclusion that the only difference between my cores is in the lcd connected. The core i had the thing running on already had the lcd in 8 bit mode, but since i´m happy to solder less pins i looked at the 4 bit configuration and saw the pullup on pin 36. Quote Link to comment Share on other sites More sharing options...
TK. Posted March 20, 2014 Report Share Posted March 20, 2014 Probably the missing pull-up resistor was the reason, because it can stall the CAN handler due to random events, so that only the watchdog will bring back the PIC into a controlled state by triggering a reset. That's also documented in the FAQ, see: http://www.ucapps.de/midibox_sid_manual_ki.html Best Regards, Thorsten. Quote Link to comment Share on other sites More sharing options...
jrp Posted March 21, 2014 Author Report Share Posted March 21, 2014 Ah, i see! Most have overlooked the info. And i thought i have discovered something new ;) Happy it´s running now, really great project! Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.