FINALLY figured it out, but need more help!

[jdutcher]

hi everybody,

so i finally found the courage to approach my the sid project again.  after rebuilding the entire thing from scratch using smash's new boards (which are amazing!), i was working on the power module and then it hit me: i forget to solder one of the power lines going to the core!!  i'm truly ashamed to admit it, but that was what the problem was.  figures, i build the entire thing perfectly and simply forget to solder one of the power wires!  anyway, so i've got it running (sends sysex request, etc.) but it seems to be encountering some errors.  here is the midiox input log from start to finish:

00010B5A   1  --     F0  Buffer:     8 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 01 F7

0001129B   1  --     F0  Buffer:     9 Bytes   System Exclusive

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

0001158B   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 34 F7

0001187A   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 3A F7

00011B6A   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 3F F7

00011E59   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 5D F7

00012149   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 20 F7

00012439   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 42 F7

00012728   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 28 F7

00012A18   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 59 F7

00012D0D   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 0C F7

00012FFC   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 7C F7

000132EC   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 10 F7

000135DB   1  --     F0  Buffer:     9 Bytes   System Exclusive

SYSX: F0 00 00 7E 40 00 0F 53 F7

sometimes when i do it, though, i get 10 byte strings instead.

also, i noticed that the regulator on the core is heating up (not too hot to touch, though).  i think this is fairly normal, but i thought i should mention it.

please please please help!

thanks!

james

p.s. when i use the sysexbox program, it give me an 04 error after ever line of code that it sent.

[moebius]

I think after THAT midi-ox dump, you have MIOS loaded.

000132EC   1  --     F0  Buffer:     9 Bytes   System Exclusive  

SYSX: F0 00 00 7E 40 00 0F 10 F7

ect..

After the upload of every 1k block an acknowledge string will be sent by the loader - Action ID "0f" for ok

From: http://www.ucapps.de/mios_bootstrap.html

Bye, Moebius

[jdutcher]

hi moebius,

thanks for the input.  the problem is that i don't see anything on the lcd screen when i power up (except backlight).  perhaps there is something else i need to do?  is there any reason why the sysexbox program would be consistently giving me errors?  i'll try to get some sound out of it and hope that maybe it's just an lcd problem.

thanks

james

[moebius]

Hi,

the problem is that i don't see anything on the lcd screen when i power up (except backlight).

How is the contrast trimmer set? Tweak it! Wiring problem? (You do have your lcds datasheet with pinouts, right?)

is there any reason why the sysexbox program would be consistently giving me errors?

Could this be because of THE Rock Solid, standardized Windows Multimedia, or your midi interface drivers - Maybe it just doesn't work in your configuration:

http://www.midibox.org/cgi-bin/yabb/YaBB.cgi?board=news;action=display;num=1094505393

So, if you have your sid app loaded and working you should be able to play it through midi.

Bye, Moebius

[jdutcher]

i do have the pinout for the lcd, but i don't think it's a wiring problem.  the contrast trimmer has no effect (backlight trimmer works, though).  there is no sound from the box when controlled via midi (and no startup sound with bankstick).  also, the core's regulator IS too hot to touch.

[moebius]

Ok,

Then back to the basic debugging.. at least midi seems to work. ;D

You know how this works: Regulator hot = a possible short. So do a visual inspection first, and then probe around with a multimeter for shorts. ect.

Bye, Moebius

[jdutcher]

ok, so i rebuilt the thing from scratch and i still have the same problem.  the only difference is that the regulator doesn't get too hot anymore, so i guess any shorts were eliminated in round 2.  PLEASE help!  i've been working on this thing for months now!

p.s. haven't tried getting any sound out of it yet though.  it's getting late and i was discouraged by the blank display.  will try some sound tomorrow.

[jdutcher]

in a recent post ("help needed"), sephult wrote:

"How can I test if MIOS is running?

After the startup phase, you can send this message "F0 00 00 7E 40 00 0F F7" to the core, MIOS should reply with the same message immediately. You could also try to upload an application, after every code block a checksum message will be send which acknowledges that the stream has been received correctly."

i've tried this and it works.  the mios seems to be loaded just fine.  the regulator on the core is getting fairly hot after a while, but not as quickly (or as hot) as it did with the old boards.  i'm getting sound out of headphones, but it's just the very faint hum of a SID waiting to play.  i haven't heard back from anyone about this for a while.  does anyone have any ideas?  even just a reply to let me know that people are reading this would be nice.  as i said, i'm using new boards and there are absolutely NO bridges or bad connections.  please help....anybody...

thanks

james

[smashtv]

hmmm.....

Does the regulator still get hot if you unplug the LCD and run it for a while?  

Even without the CORE sending any text to the LCD you should be able to crank the contrast pot until you see all of the pixels in the display.  If this isn't happening (no contrast) you have VS, VD, or V0 swapped/wired wrong or the display is dead.  Having any of these wrong would keep the display from working and heat up the regulator a bit at least.

Don't give up on it yet!

Best!

Smash

[jdutcher]

yeah, the regulator gets hot even without the lcd connected (at least when i rebuilt it i made sure to use the headers!!!).  adjusting the contrast has a very very small effect on the pixels, but even when it's all the way up, the pixels are barely visible.  the thing is, there's no sound even when the lcd is connected.  also, could it really be a short?  i mean, mios is working, obviously, and i'm getting all of the right values from the ic pins that thorsten lists at the end of the core and sid pages.  therefore, i really have no idea what the problem could be.

[LO]

What happens when you load the SID app?

[jdutcher]

:-[ uhhh......ahem ::).....this is quite embarassing.  :-[

i actually hadn't loaded the sid app.  didn't realize that that was part of the process.  i though mios 1.7 would take care of it.  anyway, i loaded the "main" app from the sid zip file and i can get sound now!!!  this thing is amazing.  i'm also getting full in and out midi response.  the only thing that remains to be fixed is the lcd problem.  the regulator gets burning hot when the lcd is connected and is just hot (not too hot to touch) when the lcd is disconnected.  i'll go ahead and give smash's advice a try and check the lcd connections.  thanks for everyone's help.

oh yeah, which sid app should i load?  i'm running a single 6581 sid with minimal control surface...lcd, one encoder, a few buttons...  i've got the "main" app on there.  actually, i guess what i'm asking is this:  would someone please give me the exact list of things i need to load for my setup?  is the "main" sid app sufficient?  when i hit buttons and turn the knob (can't see anything without the lcd) nothing really changes.

thanks so much

james :-[

UPDATE:

the lcd wiring is 100% correct.  is the only remaining possibility that the lcd itself is messed up somehow?

[smashtv]

:-[ uhhh......ahem ::).....this is quite embarassing.  :-[

hehe I was going to ask, I'll try to remember not to assume next time.  ;)  

<begin teasing>

Welcome to the forgot the last step club!

We meet every other week (next door to the backwards PCB club).  Sometimes we just put both groups together and have a good old time etching bad (but still useable!) PCBs, inserting PIC chips the wrong way (with the power on), and chain smoking while soldering.  

<end teasing>   ;D

oh yeah, which sid app should i load?

I don't have a quick answer for you on this but keep in mind that the minimal surface app is written for a 20x2, so if you run it with a 16x2 you will be missing a crucial menu selection.  I'm told it can be easily recompiled to support 16x2....just a heads up.......

UPDATE:

the lcd wiring is 100% correct.  is the only remaining possibility that the lcd itself is messed up somehow?

Could be a bad display....hard to know.

I have these sweet Data Vision display here that I stocked years ago for spares for my alesis units (I have long since banished all alesis from my outboard racks).

I decided it was time to use one on a box, grabbed the datasheet and started making the cable.  According to the datasheet these have a "HD44780 or equivalent" controller/driver chip.

An hour later (and a second LCD cable built) still no groovy.  Throw it in an alesis unit, still works fine.  Cool, no problem, datasheet must be wrong. (Very common when datasheets are translated to engrish)

Two more hours later and every possible wiring configuration (within reason) tried, still no groovy, still works in the alesis, and that CORE has no problems driving other displays.

moral of the story:  some "or equivalent" displays don't work for the MB, probably a timing issue or some other stupid difference (no fault of TK, there are several thousand companies worldwide fabricating LCDs, it's not unreasonable to assume that while most work fine others will have issues due to "or equivalent" engineering.)

Personally I would set that display -and cable- aside and find anything with a real HD44780 chip on it before spending any more time with the no groovy one.

BTW congrats on getting some sound out of it!  I know its been a long effort but you are definitely on the downhill side now.   :)

Best!

Smash

[jdutcher]

thanks for the support, smash.  so i sent the sid 6581 with cs app to the box and i can now change sounds, etc.  i think it still might be getting a little too hot even without the lcd though.  also, there's a pretty heavy background hum (maybe a grounding issue?) and it reset a couple of times on my for no reason (i mean, i could literally hear it "pop" off and then "pop" back on and start playing the default sound).  this concerns me.  as for the display, i'm using the 40x2 from bg micro.  as far as i know, other people have used this one with no problems.  one thing i'm thinking: i'm actually using part of the circuit board from the original c64 for my power circuit.  i mean, i literally cut a small square from around where the power input jack and switch are on the c64 and wired directly from there to the cap, diode, and, ultimately, the core/sid circuit.  i've checked continuity, etc. and i think it checks out, but i'm wondering if this is fishy?  the thing that makes absolutely no sense to me is this: whatever the problem is, why am i hearing anything at all?  if it's working well enough to function (seemingly) normally except for the lcd, that probably limits the possible trouble spots significantly, right?

i've got a show tonight, so i'll have to make due with an lcd-less box.  i would really really love to get this thing working 100% though, cause it's annoying to have to push random buttons and turn the knob blindly to change the sounds.

thanks

james

p.s. after fiddling with it for a bit, i think the ground hum that i was hearing was maybe just because the volume out of the sid was low.  when i blindly adjusted the volume to pump a little louder, the ground hum was much less noticeable.

[jdutcher]

ok, so a few updates...

i was actually able to use the sid box in my show last night and it was a hit!  after using it for a whole day and then last night, i now have a more informed description of its functionality (or lack there of).  so here goes:

1.  the regulator does get hot after a while, but it seems pretty stable.  i chalked that up to normal functioning and threw a heatsink on it...problem solved (i hope... :-/).

2. when i first turn on the box and the sound kicks in, the actual sid sound is very quiet and there's a fairly prominent humming sound.  usually, when i fiddle around with the buttons and the knob for a few moments, the sid sound jumps to its normal level and (i think) the hum goes away.  i'd estimate that the "normal" volume i've just mentioned is aboout 5 times louder than the initial volume.  i don't think this is a wiring issue, because it's clearly altered by the control surface.  is this normal?

3. the box no longer resets itself after a few mintues as long as i leave it alone.  in other words, i concluded that it only is susceptible to resetting when i really go crazy with the buttons and the encoder (i.e. esentially holding a lot of buttons down at once, etc.).

4. it's really important that i confirm that the 40x2 displays from b.g. micro are compatible.  i read on the boards a few months ago that other people were using them (and actually, i think thorsten posted a link on one of the midibox pages).

i really think i'm in the final stretch here.  if i could just resolve the screen issue, i think i'd have a fully functioning sid synth.  if anybody has any insight into what the problem may be or if anybody can tell me if the above obeservations are "normal" or not, please let me know.  i know i seem a bit frantic and overwhelming on the boards right now and i apologize for that, but i've spent a LOT of time with this thing and i really want to get it working.

thanks!

james

[jdutcher]

no one has any ideas?

is it safe to say that it's just the lcd at this point?  if so, should i try a different brand or something?  or should i order a new one of the same model and assume my current one is busted?

my wiring is 100% correct according to the lcd data sheet.  is there anything i can do to make sure the pinout is correct?

please, i'm begging for help on this.  i think i'm almost there!

[johnh]

no one has any ideas?

is it safe to say that it's just the lcd at this point?  if so, should i try a different brand or something?  or should i order a new one of the same model and assume my current one is busted?

my wiring is 100% correct according to the lcd data sheet.  is there anything i can do to make sure the pinout is correct?

please, i'm begging for help on this.  i think i'm almost there!

Are you in the US?  All Electronics has a  24x2 that should work for $4.

I'd give it a shot...

http://www.allelectronics.com/cgi-bin/category.cgi?category=365&item=LCD-107&type=store

[mr_chombee]

I have exactly the same problem here. I do the midibox seq and use the same displays which thorsten uses (from reichelt). I already saw some data on screen when I uploaded other applications (i.e. midibox64 main.syx) to the PIC, but these were totally messed up and the core sends out strange sysex strings. also connected a 2x16 display to the core - same problem. I think it's definitely not the LCD.

is there probably a problem with a wrong burned bootstrap loader? (I got a pre-burned PIC).

[smashtv]

I have exactly the same problem here. I do the midibox seq and use the same displays which thorsten uses (from reichelt). I already saw some data on screen when I uploaded other applications (i.e. midibox64 main.syx) to the PIC, but these were totally messed up and the core sends out strange sysex strings. also connected a 2x16 display to the core - same problem. I think it's definitely not the LCD.

First off, please keep in mind that it's very hard to help you when you spread your questions about the same problem(s) over multiple posts....

Did you check the 2x16 using the same cable as the other one?

The row of black blocks on all character displays comes when the display is initialized correctly.  The black block display at init is built into all HD44780 LCD controller chips, it's not a funtion of MIOS.  

That said it is possible in some situations for the display to correctly initialize -with the data lines wired wrong-.  Wrong wiring on the data lines will cause it to try to display characters, but all the wrong ones.  Also a bad or questionable connection on any of the lines could cause this.

Is the "strange sysex" merely the MB64 app sending the random data from inputs that are not clamped properly?

is there probably a problem with a wrong burned bootstrap loader? (I got a pre-burned PIC).

Seems people are starting to be quick to point fingers lately before checking their own work.  

I buy 18F452's by the tube, and I pull one from every tube for my own certification.  This chip gets a burn test (4x erase, program, verify) and a custom I/O test before I burn any other chips from the lot.

That sticker I put on your chip only gets put there after it is verified.  I take this effort very seriously because if I don't people could be chasing unexpected issues related to the loader.

Trust me, when you cut open that anti-static sealed baby chip tube to get the PIC out, the bootstrap was on there and ready to run.

Out of a dozen or so PIC's returned to me as not working/bad bootstrap, NONE have been bad.  Not one.

That said, I invite you to send it all back to me for a full refund if you are not satisfied.  But know that I will troubleshoot it and post what I find here, only to let prospective customers know what was actually wrong so they can order from me without fear.  My intent is not to embarass you or question your electronics skills with this, but you called me on it in public instead of privately, so now a private response about this from me would make it seem that I don't care if anyone has success with the stuff I sell.  That is obviously not the case.;)

Sorry if I am grouchy over this, but I grow weary of people first thinking that my boards/kits are at fault for issues during construction.  For every issue posted here, every problem with LCD's, with Banksticks, with Bootloaders, there are at least ten times more boards/kits out there that worked fine first try.

Best!

SmashTV

[smashtv]

Wow

Note to self--don't post before morning coffee.   ;D

Sorry about grouchysmash that made that last post, you guys caught a rant that should have been posted separately (and probably worded a bit nicer)  :-[

I don't mean to imply that your question about a bad burn on the PIC could not be the source of your problems, It is just very unlikely based on the info we have so far about your box.  Since you have a meter give me more detailed info about what's going on and I'll talk you through it, without the stupid american 'tude in the last post, and know that I have a shelf full of MIDIbox parts here to insure that we get your box up and running.  ;D

(or if you are sick of messing with it my offer stands to send it back, we can -all- learn from figuring out the cause of your specific failure modes and we would help countless future builders by linking the symptoms to the cause of the problem!)

I would really look the LCD cable over, an intermittent connection on a data line or a short between data lines could definitely cause this.

Let me know everything about your LCD connection please and we'll move on from there.....

Best

SmashTV

[mr_chombee]

Wow

Note to self--don't post before morning coffee.   ;D

Sorry about grouchysmash that made that last post, you guys caught a rant that should have been posted separately (and probably worded a bit nicer)  :-[

hehe...I was about to reply, but then saw your 2nd post. cheers for helping me out! hope jdutcher will not be upset if I hijack this thread, but we seem to have the same problem, don't we?

btw, the PIC is not from you. I ordered it here in germany from mike. I think he does a great job as well, but maybe there's a slight possibility that the bootstrap is defective somehow(?). is it possible to upload mios with a defective bootstrap?

however, let's start from the beginning:

the other (2x16) LCD I tried, has been connected with a different cable. I'm pretty sure that the wiring is ok on both cables.

interesting:

I again uploaded the mbox64 code, got the random data stream, but then, after I switched back to  the mbox seq application, the display shew something like "rebooting Mios" or "updating Mios". this makes me think that the display connection should be fine. right?

other ideas?

cheers,

nico

[smashtv]

hehe...I was about to reply, but then saw your 2nd post. cheers for helping me out! hope jdutcher will not be upset if I hijack this thread, but we seem to have the same problem, don't we?

I don't think so.....His LCD requires a negative contrast voltage to work......

btw, the PIC is not from you. I ordered it here in germany from mike. I think he does a great job as well, but maybe there's a slight possibility that the bootstrap is defective somehow(?). is it possible to upload mios with a defective bootstrap?

Whoop foot in mouth!  I can't/don't relate the alias' people use here with my customers.....

I seriously doubt you could load MIOS with any defect in the bootloader, but the chip could have an unrelated issue that's corrupting things.  I have run across a few pics with bad I/O lines, so know that a partial failure is possible.

There are a few other things that can cause these symptoms:

Timebase/crystal issues- sometimes if the crystal is wrong, or the capacitors for the crystal are wrong, or you live under a radio broadcasting tower the clock on the CORE can get way funky.  Usually when this is happening the CORE will only start up one out of many tries...(and only if you hold your mouth right/planets in alignment/etc. ;) )

Power supply-are you using the power supply section on the CORE or an external supply? If you feed the CORE with a power supply with very much "ripple" (a small amount of AC voltage leaking into the DC power) it will act very strange, with some extra gibberish on the screen.  If it is an external supply we'll jump off on the topic of how to measure ripple and figure out if it's within acceptable limits.

interesting:

I again uploaded the mbox64 code, got the random data stream, but then, after I switched back to  the mbox seq application, the display shew something like "rebooting Mios" or "updating Mios". this makes me think that the display connection should be fine. right?

Yes.  As long as you are sure there is no chance of an intermittent connection this rules out the display and cable as cause.

I have noticed posts from a few people that had troubles until they uploaded their app several times.  Try as I might (with weird settings, etc.) I can't duplicate that issue here.

You might try uploading MIOS and the app a few more times, just to see if it decides to play nice.

Best!

Smash

[mr_chombee]

uuuuhhh oohhh uhhhh....all blame goes to me! I found out that the 6th bridge at the bottom side of the core has been connected to a wrong pin. how embarrassing.

displays are working fine now. yeehaw!

thanks for your time...

[smashtv]

uuuuhhh oohhh uhhhh....all blame goes to me! I found out that the 6th bridge at the bottom side of the core has been connected to a wrong pin. how embarrassing.

displays are working fine now. yeehaw!

thanks for your time...

Great!

Congratulations!  Don't worry, I have been soldering for money since the age of 14, now 20 years later I still miss a few (only when the planets are out of alignment, and I don't hold my mouth right!).  I have a reputation in the coin-op business for people not being able to tell where I solder on a board when I'm done, even had clients accuse me of not changing the parts on the invoice! (that was easily proven by the bag of dead parts and the fresh date coded chips on a board from 1980)

The method that works best for me is to keep a meter at hand while soldering.  It takes less time to test any connection for a short or open than to look it over closely to decide if it needs tested or reworked..... ;)

Best!

Smash

[moebius]

Well,

as this thread is dead ;) (or let's say problems solved) I use it as an excuse to blaa blaa..

Checking your own work is a one thing - but doing it to somebody other.. ;) Once you get used to a soldering iron (and that is MUCH easier with a proper one), you know when you have done a good job, and where problem might be..

I once went to meet this service technician, focusing to audio stuff "only", and he had a gig, so I went along.. The problem was described as "one channel on studio monitoring system being damped" - yeah, that it was, left monitor channel was down at least -6dB compared to right one.

With physically switching speaker cables problem was isolated to this World Famous active nearfield monitor (Finnish brand ;))..

After voltage measurement and opamp replacement procedure (HIS standard: not an advice nor recommendation. He's got over 20 years experience on the field - knowing the failure rate of opamps (and the rate he's charging), It's just "cheaper" to check, if there isn't power problems and replace all opamps and see if it helps, than do a time consuming fault hunting, just to see it was a blown opamp), problem was still there.

He was about to take it to his lab and do the proper checking with scope ect... but then I had a hunch and suggested, that he would resolder leads of a power resistor (10w or so) on the power supply. It actually worked ;D There wasn't any signs of a cold solder joint or anything. PS voltages looked good with a DVM.

Probably years of use and heat generated by that resistor had revealed a cold solder joint left from manufacture, that injected a nasty ripple voltage (which you couldn't see with a DVM) to the power lines causing improper functionality..

Problem solved: Boy, Did I feel good ;)

----

Another thing: I have been repairing a Korg Polysix...

Anyone of you having a 20 years (or so) old synthesizer with patch storage - Watch out!

There are a good changes that it uses a Ni-Cad battery as a memory backup.. As those get old, they stop working, but loosing your custom sounds isn't the worst thing: They will eventually spill their guts out to your precious synth PCB.. This can corrode copper traces so badly that it's very hard to repair.. SO check and change Ni-Cad batteries, NOW! If there is already corrosion, it must be scraped off. I secured bare copper traces from oxidation with solder plating.

Bye, Moebius