Core - MIDI checks

tonysnail · November 18, 2009, 6:24am

Hi all,

I’m getting further along the road to building a SID monster, and enjoying the challenge immensely..

I’m trying to troubleshoot the MIDI in socket, and I’m getting some odd results..

I performed the LED check, connecting Anode to PIC Vss(31) and Cathode to PIC Rx(26) - no light on

However, if I connect the Anode to PIC Vdd(32) withe the cathode on Rx, the LED blinks if I play a MIDI file from Midi-Ox, so I can tell that MIDI signals are getting to the core chip.

So I checked the Optocoupler..

with the Anode on Opto pin 6, there’s no light, but the light does blink if I connect it to pin 8..

I performed the other Opto checks and everything’s fine..

I’m doing a thorough check for short circuits and haven’t found anything yet. Does anyone have any ideas what’s going on?

Many thanks,

Carl

tonysnail · November 18, 2009, 2:15pm

Hi all…

I guess the term ‘check soldering thoroughly’ is something I’m slowly getting used to..

However, i’ve got the Sysex ‘Sign of life’ now (F0 00 00 7E 40 00 01 F7) in Midi-Ox (in Midi IN and OUT) - I got my pic from Smash with MIOS preuploaded, and the sysex hex string only appears once at power-up..

I’ve been using MIOS Studio to check the interface between the pic and the PC by performing SRAM read requests of 32 bytes each, peeking around in various memory addresses.. Whichever location I look at, I get the following on the MIDI In monitor:

F0 00 00 7E 40 00 0F 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 F7

I’m guessing I’m poking around the volatile RAM of the PIC - does that explain why the data is empty? Are there any ways of poking at the memory of the PIC itself just so I can make sure it’s alive?

Cheers,

Carl

PS Many thanks to you all for helping someone return back to electronics after 15 years!

tonysnail · November 18, 2009, 3:16pm

Hmm…

I tried to upload the loopback test from MIOS Studio to the pic, and I get the following error, which time-outs after 16 attempts..

Sending block #1 00000000-000000FF

Received error code 05: Write access failed (invalid address range)

Aborting after 16 errors

Does anyone know where I can track this error back to? Is it a problem with the PIC?

I’m also trying to send midi notes to the core in MIDI-OX - i’m only getting note on/off reports in the MIDI out window.

I’ve been reading a big thread from ‘Dubstructor’, who seemed to be having similar problems to me. He was advised to perform this MIDI loopback test - I’m trying to decipher the whole thing, but if anyone has a quick summary of what my problem might be then that’d be really helpful!

Cheers guys!

fussylizard · November 18, 2009, 5:40pm

Carl,

What exactly are you seeing? I’m not sure what to recommend based only on your descriptions of what you’ve tried to diagnose.

Have you looked at the MIDI troubleshooting page: http://www.ucapps.de/howto_debug_midi.html

I would run through the tests there first and report your results here.

Good luck!

Regards,

C

tonysnail · November 19, 2009, 1:34am

Hi There, thanks a lot for your reply..

Here’s the results of my tests…

Test PROG1: I bought the PIC from Smash, so i assume it’s fine

CORE1: There’s a crystal in there, however I may have shorted it whilst probing just now - i swear I saw a spark! Are they particularly prone to exploding?

CORE2: Not for me, I’m using a core V3

CORE3: I think i’ve ironed out all my soldering bluffs, but I’ll keep checking..

CORE4: All Vdds are measuring +5V - they’re a little over 5V, nearer 5.7.. Is this a problem?

CORE5: All ground checks are 5V

CORE6: Midis are checked thoroughly

CORE7: Diode is good

CORE8/9: I checked the resistors with the multimeter and a magnifying glass and all’s good there

CORE10: I’m not getting constant sysex messages, just the initial one at power-on:

F0 00 00 7E 40 00 01 F7

CORE11: The 2200uF seems good, although I’m just running power directly to the board- I haven’t built an optimised supply yet, but power seems stable on the board, if a little over 5V (I was getting about 5.65 on the checks)

TEST PC1/2 - the computer midi ports are fine

At this point, I tried to upload MIOS [see screenshot]

each block was showing up on the In Monitor (on the left), but the Upload Control window was showing ‘Write Failed’ errors, before finishing with acknowledgement responses for the last 8 blocks.

I carried on: The MIDI In tests seemed to pass..

TEST IN1A: resistors are fine

TEST IN1B: optocoupler is in correctly

TEST IN2: MIDI plugs are fine

TEST IN3: The LED/resistor between RX and Vss lights up when I try to send a data stream (by uploading MIOS)

Further MIDI In tests - the opto’s fine, and the LED blinks when I try to upload MIOS

I then tried uploading the SW Loopback test.. [see screenshot]

Again, the upload displayed an error message, followed by an acknowledgement response..

Then the MIDI monitor started throwing up loads of gibberish before stopping, and not responding to my turning the power off/on (whereas before it would display a power-up/power-down notification.

I also then the feedback loop by playing notes - nothing is being returned on the Input monitor though. Not too suprising though, given the little tantrum the PIC just had..

So that’s where I am..Basically to summarise my question:

Is the Hex file I’m trying to upload to the PIC getting there? Something definately happened after the loopback test was uploaded, although ‘a load of MIDI In gibberish followed by nothing on the In Monitor’ is not what I expected!

Many thanks,

Carl

lylehaze · November 19, 2009, 6:33am

I can give you a few tips..

The upload request (sign of life from PIC)

You get these repeatedly if you only have the bootloader in the PIC..

You only get one of these if MIOS is already loaded.

I’m pretty sure Smash loads MIOS for you.. so getting just one is correct.

I notice you have selected “manual mode” in MIOS Studio.. This is only advisable

if you are SURE you have everything just right.. You’re not there yet.

You need good communication in BOTH directions for the suggested “Smart Mode” to work.

Figure out how to get it working in Smart mode.. It’s worth your time.

Do a loopback test from PC MIDI OUT to PC MIDI IN, without having the core connected at all.

Figure out how to receive the messages you are sending, through a MIDI cable..

confirm that thoroughly.. then remove the cable and make sure the messages do NOT come back in..

(this will make sure it’s not a software loopback you’re really seeing)

Then test it again with a second MIDI cable, to make damn sure you have two good MIDI cables.

After all that is proved.. THEN connect core OUT to PC in, and PC Out to Core In..

Switch MIOS Studio over to smart mode..

And load a MIOS APPLICATION (not MIOS itself, that’s already in there)

Have Fun!

LyleHaze

fussylizard · November 19, 2009, 6:38pm

Like lylehaze suggested, try Smart mode.

One other thing that occurred to me when looking at your second screenshot. I was pulling my hair out a few months ago with upload problems seeing similar messages about overwriting MIOS, etc. like you are getting. Turned out I had ordered the wrong PIC from SmashTV. Double-check that the .hex file you are uploading will work with the PIC you have. I’m not sure if this can be a problem with all apps or just with the bootloader or what, but double check that the PIC you have will work with what you are trying to do. MB-SID projects require a different PIC than other MIDIbox projects (due to the CAN bus) so you might want to double check PIC versions, or see if there are multiple .hex files for some of these test apps and be sure you are using the one for the PIC you have.

tonysnail · November 19, 2009, 8:47pm

Hey thanks guys.. I’ll try these out..

My first port of call is to make a proper power board, so I can be sure that the power getting to the core is stable.

After that I’ll do the loopback test.

By the way, I’m using the newer PIC18F4685 - could that be why the little loopback hex prog doesn’t work?

Thanks a lot!

Carl

fussylizard · November 20, 2009, 4:41pm

Hey thanks guys.. I’ll try these out..

My first port of call is to make a proper power board, so I can be sure that the power getting to the core is stable.

After that I’ll do the loopback test.

By the way, I’m using the newer PIC18F4685 - could that be why the little loopback hex prog doesn’t work?

Thanks a lot!

Carl

If you’re running the IIC loopback test off the MIOS8 download page, then that won’t work which might be part of the issue here.

Are you building a standalone core board or the MB-6582? I’m guessing the former.

Have you tried using Smart mode for uploads?

Do you have an LCD attached? If so, what does it say?