Stuck at Rebooting MIOS after midibox_sid_v2_044 upload

MIDIbox SID
1

Greetings!

I am stuck at the Rebooting MIOS LCD output after uploading the MidiBox SID v2 hex.

Setup: MBHP_CORE _V3 with PIC18F4685, 2x MBHP_SID_V3 for stereo SID, 2x20 CLCD, “PSU” with +5 and +14V like in this schematic, powered by a C64 PSU.

What I have done so far:

  1. Burned the bootloader 1.2b with my own pickit on a fresh PIC18F4685. The LCD showed black bars - great.

  2. MIOS Studio displayed “Bootloader up & running”, MIOS8 as OS, MBHP_CORE or similar as board and PIC18F as Core Family.

  3. Uploaded MIOS 1.9h for the PIC18F4685 using a MIDISPORT Uno and Windows 10 (64Bit) which worked like a charm. “READY.” on the LCD and MIOS Studio tells me “Application is up & running!”

  4. Uploaded midibox_sid_v2_044/setup_8580.hex, LCD tells me “Rebooting MIOS”, MIOS Studio goes back to “Bootloader is up & running!”

  5. Uploaded MIOS 1.9h again, tried different troubleshooting apps like LCD connecion test, SID connection test which all worked fine and came up after “Rebooting MIOS”.

  6. Uploaded different midibox_sid_v2_044 like setup_6581.hex or setup_tk.hex - all failed like in 4.

  7. Tried mbsid_testtone_v3, which failed like in 4. too.

  8. As someone who does not like to bother others with his problems, I googled and searched the forum and double checked the manual and guides. I found out, that CAN bus was now a thing on D0-D3, which

was at that time connected to my display. I disconnected D0-D3 from the LCD, connected D3 with a 1k resistor to +5V, and put a 1N4148 diode between D3 and D2 like shown in this schematic.

  1. Tried all the steps from 2. to 7. again with the CAN Bus set up - still same issues like in 4.

  2. Found note in the changelog of MIOS v1.9d, that D0-D3 should be clamped to ground. I have not done that, as it conflicts with the schematic from 8.

So, I am fresh out of ideas now, besides from testing to clamp D0-D3 to ground now, as the schematic using a pull up resistor and a diode looks pretty much like a wired and to me.

Suggestions are very welcome =D

Edit: I have not connected any kind of frontend yet, so no DINX or DOUTX attached yet. Might this be a problem?

Best regards

mrzottel

2

Hola,

Maybe not very helpful, but I have an old Core8 lying around here, with a running version of MBSID. There is only a display attached, no other modules. Just to confirm, that you don’t need DIN/DOUT modules and also the SID module needs not to be attached - the application will start up fine and should show the MBSID start screen (an empty patch).

It is also not necessary to modify any CAN bus/display lines for this test. You could just leave D0-D3 unconnected, the display should be initialized in 4-bit mode by MBSID (only using the upper 4 data lines).

Can you disconnect the MBSID module and see what happens? Maybe there is a connection/power problem there, just try to run only the core and the display for now.

Which display do you have? Is it 100% HD44780 compatible? If not, this might be a problem (maybe the other apps used 8bit mode for the display, which worked fine).

So, here are some ideas:

* disconnect the MBSID module, check for success

* make sure, the display does work in 4bit mode

* in MIOS studio, after the MBSID app should have started, you can send some notes via the virtual keyboard (make sure, the PC MIDI out is connected to PIC MIDI in). If the SID module works, but the display does not work, you should hear some basic SID sounds :).

* measure voltages, the PIC should have at least around 4.75-4.8V, otherwise it might be unstable or the app could reboot/never start.

Many greets,

Peter

 

3

Greetings and thanks for the quick response!

Quote

Which display do you have? Is it 100% HD44780 compatible? If not, this might be a problem (maybe the other apps used 8bit mode for the display, which worked fine).

I got this LCD which is supposed to be HD44780 compliant (Datasheet says KS0066 or equivalent controller) and 4/8-Bit compatible. It also does work, showing me either

a black bar (only bootloader) or “READY.” after uploading MIOS or finaly “Rebooting MIOS” after uploading midibox_sid firmware. When I reboot by turning it off and on again,

it is back to black bars.

Quote

* disconnect the MBSID module, check for success

Without the MBSIDs no success as well. Uploading midibox_sid still “destroys” MIOS on the PIC and goes to “bootloader only” mode.

Quote

* make sure, the display does work in 4bit mode

It does :wink:

Quote

* in MIOS studio, after the MBSID app should have started, you can send some notes via the virtual keyboard (make sure, the PC MIDI out is connected to PIC MIDI in). If the SID module works, but the display does not work, you should hear some basic SID sounds :).

The MBSID app never starts =(

Quote

* measure voltages, the PIC should have at least around 4.75-4.8V, otherwise it might be unstable or the app could reboot/never start.

Voltages seem to be fine:

MCLR(1) - VSS(12) = 5.16V

VDD(11) - VSS(12) = 5.16V

VDD(32) - VSS(31) = 5.16V
 

I also tried disconnecting the LCD now, and check if MIOS Studio tells me wether an app or only the bootloader is up and running after uploading the MBSID app, but even without a LCD, it is only the bootloader comming up.

edit: added link to the LCD :wink:

Best regards

mrzottel

4

Hi mrzottel,

I have seen a few people over the years have similar issues with the 4685s.  Statistically very rare but worth mentioning if it helps you get to goal…

Port latch-up (and a crashed app) due to a very noisy PSU during MIOS or app upload - was fixed with a few erase cycles on a programmer setup for HVP mode.

Weak VPP on a PICkit2 clone during programming - the write looked good but the chip had issues after MIOS was uploaded.

Bad flash on the PIC itself - extremely rare due to good testing/quality control at Microchip.  I have seen a couple where the only thing changed was the PIC/exact same process and it worked as expected.

I have attached a known good/well tested hex file to load directly into the pickit 2 software.  Use the G_0000 file and it will set the correct options and write both the bootloader and MIOS onto the PIC in one operation.  If the symptoms change that will eliminate quite a few possible problems.

Let us know how it goes!

 

Best regards

 

Tim

4685_G_00-03.zip

5

Thanks a lot smashtv!

I will try as soon as I get back to the lab and let you know how it worked out.

 

best regards

mrzottel

6

Hi,

I’ve done the following now: 

* 25 Erase cycles using the picKit3-clone, flashing the regular hex (only bootloader), uploaded MIOS (came up good, READY. prompt), uploaded SID App - still not working.

* Flashing your .hex with the picKit3-clone, came up with READY. prompt, uploaded SID App - again not working.

(All of the above using the supplied picKit3-clone voltage set to 5V)

* Using external power (5V) and switching off internal power from the picKit3-cone, flashed your hex, uploaded SID App - again not working.

Symptoms remained the same - “Rebooting MIOS” shows up, and upload requests from the bootloader appear in MIDI in. 

I will try some more stuff with the external power…

 

Best regards

mrzottel

7

Which parts of the MBHP_CORE module are not mounted?

E.g. the two pull-ups for the IIC EEPROM are essential.

And the pull-up + diode for the CAN interface as well (as you mentioned in 8. - please ensure that these components are still connected).
Topic 9. would work as well - it’s important that the CAN (and EEPROM) interface are at a defined voltage level and not floating.

Best Regards, Thorsten.

8

R2 and R12 are mounted, so IIC should be fine. The pull-up resistor between +5V and D3 as well as the diode between D3 and D2 is mounted as well. Both D3 and D2 measure at +5.15V while powered. (My C64 seems not to be too exact with the 5V rail).

Edit: You asked for the not mounted parts, I forgot: The 7805 is not mounted, as I use the C64 PSU, but in/out are shorted, so that the caps are all connected. I did not ommit anything but the 7805, even the bridge rectifier is in place :wink:

9

Might be a hint: D0-D2 are spot on 5V, D3 is 5.15V. +5V power rail is at 5.18V.

10

So after double and tripple checking every path, every solder point, voltages, resistances, diodes (at first I thought the diode for ecan is shorted, as it opens at 0.5V in the right direction, but also 0.8V in the other direction, but this was in circuit. Out of circuit, it works fine. So I guess there is another path through the PIC.) I gave up on finding an electrical issue, even though the board looks quite bad by now. :wink:

I tried the software side with merging hex files into a single one, removing overlapping segments by letting the “newer” hex file win, but flashing this did not help either.

Now I ordered another PIC, which should arrive some time after christmas to test out the last option: Bad PIC. 

I will let you know how it worked out :wink:

Edit: After my vendor send me the wrong PIC today (PIC16F877A) it will take a few days to get the right one… grrr…

Best regards

mrzottel

11

After waiting almost a month to get the right PIC delivered, I tested with a brand new IC. Still no luck, so it wasn’t the PIC.

12

As my problem seems very similar, i’m writing in this thread.

my hardware consists of 1 core and 2 sid modules, one 6581 and one 8580. Control Surface is planned, but not finished yet (so DIN/DOUT are not connected yet). PSU wiring is exactly as in this link, but only one core/sid/sid instead of four. A 4x20 LCD is connected and working. I omitted the pull-up resistor and diode for CAN-Bus, as i only have one core (correct..?).

Uploading MIOS worked well, but the SID Firmware does not seem to work. When turning power on, at first a screen with MIOS shows, then the following message: 

Quote

MIDIboxSID V2.044
Launching CS

  • after then, it crashes, and again the mios screen shows. 

As there is no control surface yet, it makes sense, but i thought i could do a test without any CS..?

Any ideas whats wrong with that?

 

best regards

 

 

13

The core should start without a cs installed, at least it does for me. You should not omit the resistor and the diode, even if you have no other cores on your bus, as this would leave the pin floating, which is never a good thing. Simplified it would mean, that the pic does not really know which state the pin is in. Can be high, can be low, can be somewhere in between - and that can cause troubles. The pull-up resistor should pull the pin to about 5v, so it is in a predefined state. The diode does the same to the other pin at a slightly lower voltage. (Instead of a diode, it might work with a simple jumper, if you do not intend to use the CAN-Bus at all, but not 100% sure about that.) 

In any way, your boot process gets further than mine did, so the missing diode/resistor is almost certainly something you want to check.

Edit: You might also want to connect your display in 4-bit mode, as this frees up the pins for the CAN-Bus, so do not(!) attach the resistor and diode with the display still using those pins. This page has the necessary information for 4-bit display mode and CAN-Bus modification. Both seem to be required for MidiBoxSID to work.

Btw: In my case, it was a short between two pins of the PIC. It seemed, that the socket of the PIC was damaged somehow. After replacing the socket, everything worked great. Sorry for the late follow up, but the forum seems to dislike my password. I have to do “forgot password” and set a new one each time to log in. :wink:

14

ok, thanks. i added the pull-up and diode and it boots fine now… 

after i did not get correct sounds, i did the interconnection test and got weird results 

so i checked the wiring and soldering once again (no errors..), attached only one sid module and repeated interconnection test; perfect this time. So i stuffed the SID again and voila- works fine  (but one sid only).

next step will be to check the other sid module individually…

edit: both work now. No idea why, but it works :smiley: -basically, i unplugged the connection wires, checked both SID modules individually and plugged them together again, now everything is fine