Wilba

OK that makes it easier for me to follow, thanks! :) I agree, it appears like PIC 0 is not behaving like the other PICs... you've proved it with the interconnection tests. I wish everyone's troubleshooting reports were like this, you've basically done every logical combination to identify what does and does not work and where the fault lies, and the fault definitely lies with PIC 0. Perhaps if you dare, peel back the label on PIC 0 and PIC 1 and compare the ICs... in the highly unlikely event that SmashTV has sold you something that is not a PIC18F4685 :) It is just too strange for a PIC to be working but have a problem with only a few output pins, unless perhaps those outputs were fried by high voltage or static electricity... very strange. I think you could use the changeid application (carefully!) to change PIC 4 to have ID 0, then use this as your PIC 0 in Core 1 in the short term... for the control surface to work, you need a PIC with ID 0 in Core 1, and for master->slave comms you need a PIC 0 to be the "master".

To start with, stop calling it "CORE1 PIC"... refer to each PIC by the PIC ID that you (or SmashTV) burned into it, in this case they should be 00, 01, 02, 03, which get installed into Core 1, Core 2, Core 3, Core 4. Try uploading testtone to each PIC in Core 1. Note J11 needs to be set to 1, and MIOS Studio upload dialog needs Device ID set to match PIC ID (0-3). Always remember, MIOS Studio directs the upload to the PIC, it doesn't matter which Core it is in, only J11 needs to match the Core you are uploading to not the PIC You need to fix the upload problem before moving on to the SID/SID module testing. Then, you can test each PIC in each Core to determine which is a "known good" PIC and which are "known good" SID modules. Try uploading something else (like the interconnection test) so you know the upload really worked. If you fix your uploading problems, then start testing SID output sections again and report results, try to report what works like "testtone app in PIC 0 in Core 1, pin bridging both SID sockets, both audio sockets have tone" etc. The key to good troubleshooting is only changing one thing at a time and noting the differences (what starts/stops working).

That would explain intermittent behaviour :) one encoder position would affect the state of pins of other encoders :) Yes. If this is not happening, perhaps the .syx file is not being received, or the bankstick is not working, although you can test the bankstick separately by storing patches in each bank (A-G) and storing the ensemble parameters.

Yes. They should be fine, they have been used on other MIDIbox projects for a long time (including my original MB-6582 "prototype").

I would like you to check the resistor networks are the right ones, and are inserted the correct way (dot on resistor network should be in square pad). If you say "I use SmashTV's base PCB parts kit" then they are the right ones, but still check they are in the correct way. This kind of intermittent problem suggests either shorts between pads (or breaks) on those tracks/wires between the shift registers and the encoders, or maybe even some kind of problem with the ground connections to those encoders... i.e. the encoders are connected by ground tracks and are the only things using ground on the control surface, switches and LEDs do not (they are all in the switch/LED matrix) therefore an intermittent ground on some encoders will make them stop working. This may be irrelevant, but... which encoders are you using? Make/part number/datasheet etc. Is there any chance the encoder "case" is shorting pads on the PCB?

You don't need the wires for R2 boards from SmashTV, only R1 boards bought from my bulk order. R2 boards have tracks instead of wires. Ignore J21B-J24B, these are just alternatives to using stereo sockets on the PCB. J70 has L, R and GND pins. This is left channel, right channel and ground. Your stereo jack has six pins probably because it is switched. If there is no plug in the socket, the three on the left will connect with the three on the right. When there is a plug in the socket, three pins will connect to the plug, the other three pins will be open (not connect to anything). You will have to work out which three connect to the plug. The easiest way is to take a plug which is connected to an amplifier, plug it into your socket, and touch pins of the socket, listening for "hum" :) It should then be obvious which three pins connects to the plug, then connect these to J70, note "tip" is left channel, the next one ("ring") is right channel, the next one ("sleeve") is ground. When you are finished, you can update the wiki :)

also bottom of this page: http://www.midibox.org/dokuwiki/wilba_mb_6582

Very cool.

We don't lend, we rent him out at one TimTam per square inch of PCB.

Big is good, the bigger "throat" means you can clamp further in from the edge. I use a mix of big and small.

I have touched over 3000 SIDs without any anti-static protection and they all worked. :) But when I took the motherboard out of my PC to replace the CPU fan, I did it in the nude standing on wooden floorboards with an earthed cable in my ground socket. The moral: do what stryd_one says. Don't take chances when it really matters ;)

Absolutely not. :) In the first gluing phase (the corner screws), you need PCB and panel attached via clamps with just a little bit of slack so you can align the PCB perfectly to the panel holes using the shafts of some (unsoldered!) switches as a guide. I can't even imagine how that would be possible without clamps. In the second gluing phase (the threaded spacers), it's important to clamp at points where the PCB is warped and is holding the spacer away from the panel. That might be at a few points in the middle or on the edge. Again, I can't see how that's possible without a clamp. I don't know about other countries, but here you can buy a pack of 6 or 8 various sized small C-clamps (from China of course) for $10. They don't need to be that good, you're not applying lots of lateral force, so if it's equivalent to a good firm grip with your hands, that's enough for this job. Four clamps is a bare minimum I think, but I suppose you could do it with two big ones and some other kinds of clamps, ones like clothes pegs or bulldog clips, etc.

OT ALERT! :) Conveniently (by design) the 8x8 mod matrix LEDs are arranged so each horizontal row of LEDs is supplied by one pin of a 74HC595 so it's quite easy to put different LEDs in each row and make the brightness match. If you're going to do this, you might as well invest in some machine pin strips or chop up an IC socket, so you can swap resistors after everything is soldered and the control surface attached. However, different colour LEDs within the mod matrix might look a bit wrong in the other modes, but if you don't mind that then go for it, you should have no trouble matching brightness.

I don't have any fancy gear to measure it... nor any sound engineer learnin'... I just hooked it up raw to my Sony MDR-7506 and can hear the SID's internal "noise" and not any noise from the PCB itself, if that makes any sense.

Do those capacitor footprints allow caps with 2.5mm or 5mm lead spacing? Brilliant idea! *cough*

It's clicky, it's homeless, and it comes with a 100% money-back guarantee until the universe implodes, all for the incredibly low, low price of ¡¡¡¡¡ $0 / €0 / £0 / ¥0 / ?0 / ?0 / ?0 !!!!!

:rolleyes: when I pack up your 20x4 negative red Orient LCD (which just arrived btw), the Soundwell encoders (which arrive next week) and Re'an P401 knobs (you wanted some of them too right?) and whatever other stuff I'm using to bribe favours out of you, I forget...

Yay, moderator super powers are cool...

Yay, moderator super powers are cool...

hehehe I only offer because it's so easy to put a dozen LEDs in an envelope, $2 and you get exactly what you need in 4-10 working days :)

The LEDs are (effectively) in one big 16x8 matrix, with 8 "columns" sinking current by the 3rd 74HC595 shift register (the one with the transistors), and 16 "rows" sourcing the current by the 1st and 2nd 74HC595 shift registers. 8 "rows" are used by the 8x8 mod matrix LEDs, the other 8 "rows" are used by all the other LEDs. It's very easy to use LEDs in the mod matrix that are different to the other LEDs, you just use different resistors connected to the 1st and 2nd 74HC595... and in theory you could identify all LEDs in one or more "rows" and use a different resistor for those rows... but this solution isn't very good, as unfortunately those LEDs would be in different groups on the display, rather than being in one group (like LEDs in the SID 1/2/3/4 group, or the LFO waveform group). The best solution is just get some more red LEDs from SmashTV... if buying another pack of 100 is too much, I'll just send you some spares that I have here. You can solder the LEDs in two stages, it won't make a difference if you are soldering them while the panel and PCB are attached (as I recommend for perfect alignment).

Noise is a factor of fan speed, not width. Just run the 40x40x20mm one with 5V and it should be OK. You could put a 7805 into V4 and use this as a separate 5V regulator to drop the unregulated 9V - this may reduce possible noise on the supply for the rest of the PCB, and thus reduce possible noise in the audio outputs. Connect J25:9V-11V to J74:1, then you can connect the fan to J74:2 and some other ground pin (preferably J25:GND). Maybe you can use a 7809 in V4 instead... probably better, as a 7805 might get pretty hot. Try your fan on the existing 5V and 9V supply pins, if it's not too noisy and doesn't introduce noise in the audio, then don't bother with this separate supply circuit. Remember though, the fan is just for looks. If you cut slots in the right side of the case, the warm air can escape and there's no heat buildup.

You mean this one? http://www.retrojunk.com/details_articles/130/

You guys will laugh at anything... 352 SID Stack

PIC18F4685 using 4-bit mode, Core32 using 8-bit mode?