toneburst Builds An MB-6582

[stryd_one]

LOL That gets my vote for new midibox slogan.

</hijack> sorry!

[toneburst]

Well, progress is being made!

I've installed all 4 Core chips, and managed to upload the SID Testtone application to all 4 (one at a time) via MIDI from MIOS Studio.

All the output buffers seem to work except for SID 7, for some reason, which is encouraging. I don't get any tone when I connect pin 14 of this SID socket to pin 27 of the SID chip socket 8 either, so I'm guessing the signal from Core 4 isn't getting to SID socket 7 at all, for some reason. I will have to check my soldering around that area.

Also, for some reason, the LCD outputs of Cores 2 and 4 don't seem to work (I just get a row of blocks on the top row of the LCD, but no change when I upload a MIOS application etc.). Again, I'll have to check my soldering, I guess. Slightly worryingly, I was getting a clear display from Core 2 yesterday. I guess it could just have been a bad solder joint, and moving the board around today broke the connection.

Also, I found the missing transistor, on a separate little bit of cardboard. Not sure if I cut it off the strip myself and forgot about it, or it was on it's own from the beginning...

Really good to get my first sounds, anyway, even if it was just a 1kHz tone...

a|x

[Max Romantschuk]

Replying so I can follow the thread more easily. ;)

Keep up the good work! :)

[Goblinz]

Also, I found the missing transistor, on a separate little bit of cardboard. Not sure if I cut it off the strip myself and forgot about it, or it was on it's own from the beginning...

a|x

It's on it's own... I misplaced mine for a while!

[toneburst]

It's on it's own... I misplaced mine for a while!

Ah... I didn't think I'd cut it off. Glad for the confirmation I'm not completely losing my memory (if not my marbles).

a|x

[toneburst]

OK, I've been doing some resoldering on the back of the board, and am happy to report that te LCD now works with all 4 cores! When I say 'resoldering' I basically just mean remelting the solder on the dodgy-looking joints, trying to get the solder to adhere to thos stubbornly cold pads. I actually bought some liquid flux the other day, but I'm embarrassed to say I don't know how to use it, so I've been struggling on by just turning the iron to a higher temperature than normal, and trying to hold it to the pad for a little longer.

That 7th SID socket is still not working, though, annoyingly. Anyone any ideas how I should troubleshoot this? I'm now getting a tone from the socket is I connect pin 14 to pic 27 of SID socket #8, so the 1kHZ tone is not getting to socket #7, but nothing is arriving at the audio output. I don't know enough about the signal-path to know where the problem is most likely to lie. There don't seem to be any obvious soldering issues in this are- but then again, there may be issues that a more experienced solderer would notice immediately.

Also, anyone any idea why my MB-6582 should interfere with my FM radio, from several metres away..?

a|x

[toneburst]

I got a bit impatient, and thought I'd try inserting my two old CSG 8580R5 SID chips in SID sockets #1 and #2, and try uploading the SID application from the 'setup_8580.hex' file, again via MIOS Studio.

The application upload went well, though it took longer than I thought it would. After MOIS rebooted, the application successfully formatted all 8 memory chips, which was a relief, as I had trouble soldering the sockets.

I plugged-in my headphones, and was greeted by a loud buzzing noise, and the occasional crackle. I tried triggering different notes from the keyboard in MIOS Studio, but just got slight variations on the same loud buzz. Then I remembered that I should have jumpered across the audio I/O headers, so I added the jumpers.

Unfortunately, that didn't seem to make any difference, and I'm now wondering if I might have trashed my SIDs. This suspicion is strengthened by the fact that both SIDs seemed to get noticeably warm (though not actually hot), almost as soon as I switched the power, on. I'm pretty sure I had them working in my modular SID v.1, so I think they were functional before I put them in my MB-6582.

Is it likely that I have damaged them in some way? I really hope not, though it's not the end of the World, as I have my 8 chip order from Wilba.

a|x

[toneburst]

Ooops!

Very silly mistake! I should obviously have been uploading the MB6582 version of the SID application. I've now done that, and woooo... it works! I've only tried SIDs in slots 1 and 2, but they definitely seem to work now, which is brilliant!! Loads of background noise, mind. Can't remember if I was getting this before, when I had the SIDs in the modular SID board from SmashTV, but I'm definitely getting notes, and it sounds cool!

Have uploaded MB6582 to Cores 1 > 3, and left the Test-tone application on Core 4, until I can work out what the problem is with the audio out from SID socket #7.

Exciting!!!!!

a|x

[Wilba]

You'll notice you only see one line on your display, because the setup_mb6582.hex expects a 4x20 character LCD.

You can change setup_mb6582.asm and recompile (make a new setup_mb6582.hex).

You want to change it to #define DEFAULT_LCD_LINES 2

That 7th SID socket is still not working, though, annoyingly. Anyone any ideas how I should troubleshoot this? I'm now getting a tone from the socket is I connect pin 14 to pic 27 of SID socket #8, so the 1kHZ tone is not getting to socket #7, but nothing is arriving at the audio output. I don't know enough about the signal-path to know where the problem is most likely to lie. There don't seem to be any obvious soldering issues in this are- but then again, there may be issues that a more experienced solderer would notice immediately.

Can you go into more detail about how you troubleshooted this?

Break it down into specific tests... eg. testtone app, SIDs in sockets, #8 pass, #7 fail - testtone app, SIDs not in sockets, etc.

It sounds like you proved the audio buffer for #7 is working and that it's just a SID in socket #7 that doesn't output sound through that audio buffer.

[toneburst]

Hi Wilba.

Thanks very much for getting back to me again!

You'll notice you only see one line on your display, because the setup_mb6582.hex expects a 4x20 character LCD.

Yeah, I noticed that. I'm not too concerned about that, as this is just a temporary screen for debugging. I have a 20x4, but got impatient trying to make up the cable, found this old 20x2, and thought I'd use that, just for testing.

Can you go into more detail about how you troubleshooted this?

Break it down into specific tests... eg. testtone app, SIDs in sockets, #8 pass, #7 fail - testtone app, SIDs not in sockets, etc.

It sounds like you proved the audio buffer for #7 is working and that it's just a SID in socket #7 that doesn't output sound through that audio buffer.

It's definitely not a SID issue, as I've never put a SID in that socket. My testing so far has consisted of

• Insert 4x Core chip

• Upload SID Testtone application to all 4 Cores

• Connect Pins 14 and 27 of each (empty) SID socket in turn

All sockets output audio tone on left/right channels of each pairs audio output, except SID socket #7

• I then tried connecting pin 14 of SID socket #8 to pin 27 of SID socket #7

No audio output

• Then I reversed the procedure, connecting pin 14 of socket #7 to pin 27 of socket #8, and this time got an output from the left channel of output 4.

Soo... I think it must be a problem with the output stage of SID socket #7. Unfortunately, I'm an electronics dunce, so I don't really know how to go about testing the various components in this part of the circuit. I do have a simple multimeter, but I'm ashamed to say, beyond simple voltage tests, I don't really know how to use it. I can't see any obvious bad soldering around this area of the board, but I could well be missing something.

a|x

[Wilba]

BTW good troubleshooting... if only everyone did these kinds of tests before asking for help!

You could start by assuming the capacitors and resistors are "good", it's unlikely you could have damaged them from soldering or from too much voltage or current.

Most likely it's the transistor that's blown, perhaps from overheating while soldering, although you can't rule out bad solder joints, broken tracks, transistor put in backwards, or even forgetting a jumper in J2_SID4. Consider also the capacitor C25_SID4 put in backwards.

However, this brings to mind that you "mislaid" a transistor, and then found it again on a separate piece of cardboard.

Are you sure this was not the BC337 which goes into T1 (near the header J15, for LED backlight brightness) and instead you've put a BC547 in there? I think a BC337 in the audio buffer would probably still work, but just check to be sure.

Take a close-up photo of the audio sections of SID #7 and #8 (i.e. *_SID4)... maybe something will look amiss to me :)

[toneburst]

Thanks for the detailed reply Wilba.

I'm at work at the moment, but I was able to nip home at lunchtime, as I live nearby, and take some quick snaps of the offending area of the board.

Here's the area above SID socket 7, from the top of the board:

..and from the bottom:

Apologies for the ropey image quality. I had to sharpen them in Photoshop, because I didn't have time to setup a tripod decent lighting, unfortunately, so they came out a bit blurry.

I didn't have time to have a close look, but it's entirely possible I used one of the transistors intended for the SID section at T1. Thinking about it, I soldered this one before I did the SID sections, which would explain why I appeared to have too few of the transistors left, and then discovered this lone transistor later. However, the 'left-over one' was soldered into one of the SID channels that seems to be working OK. Is there likely to be a problem with the LCD backlight, if I've used the wrong Transistor at T1? My test display doesn't have a backlight (well, in fact it has an EL backlight, I think, but I don't have the transformer to power it).

a|x

[Wilba]

It could be just the photo, but it looks a bit like your solder joints were done with a soldering iron that was not hot enough, or you're having trouble getting the solder to suck into the hole.

It's a bit hard to describe really - if the iron isn't hot enough, then the solder tends to be non-sticky, like it's got a crusty coating... when it is hot enough, it's more liquid and flows more. Similarly, if a solder blob has lost it's rosin (flux) eg. if you are carrying a blob on the iron tip to the joint, then it won't stick. **edit** that was a really bad description of not-hot-enough-solder... think of it more like soft-serve icecream or whisked egg whites... it will be malleable and hold a shape and with the iron you can pull out "peaks" and "horns" from the joint, instead of it being liquid and pulling itself into a blob.

Ideally you should be placing the tip so it touches both the pad and the lead, then applying solder into the joint (or even onto the tip) so it melts and gets sucked into the hole. You can see it happen suddenly, the blob on the joint gets progressively bigger as you feed more solder, then it gets sucked into the hole and forms like a tent around the lead.

The joints I can see on the SID IC's socket look good - I'm only commenting on the ones in the audio buffer section (last pic) where there isn't a neat "tent" or "dome" rather it's more of a spiky blob.

I would advise getting a solder sucker (solder pump) - the cheap one-shot-suck kind, not something with an actual motor-driven pump - and sucking out the solder from these joints and trying again. You can't really fix these joints by adding more solder until you've taken away some. You don't need to remove the parts, or get all the solder out, just make the joints "flat" again.

(Find attached an atrocious Paint-annotated criticism of your solder joints. Sorry.)

SID7_bottom_02.JPG

SID7_bottom_02.JPG

[toneburst]

Thanks a lot for the soldering advice. You're absolutely right. Funnily enough, most of the other joints look much better, but I did have trouble with this particular section, and a few of the other pads.

I've got a mechanical solder-sucker, and I've actually been trying to do what you describe with some of the other dodgy-looking joints- remove the old solder and add more.

I've been nervous of turning my iron up all the way though, for fear of melting stuff other than the solder. I think I might have to try and find a pointier tip for the iron, too, as the one I have is quite long and thin, but the actual point is quite blunt, making it hard to heat the pad without running the risk of melting the material of the board itself, next to the pad. I have another tip somewhere that is fatter, but has a sharper point.

I'll go over the joints in that section again and resolder. Are there any particular components I should be particularly careful to avoid heating too much? I guess the transistors are the most temperature-sensitive parts in this section of the board...

Thanks very much for your detailed advice once again, Wilba.

a|x

[chimchim]

You shouldn't really need more than a 20w iron as that is what I did most of the board with (excluding audio jacks etc.).  The pads are definitely a little smaller on this board compared to others I've worked with so I did notice sometimes the leads would get hot enough but not the pad.  It can be hard to make a solid connection against he pad with a blunt tip (Clean it often!)

My advice it to hold the iron almost vertical and try to almost wedge the point of your tip into the pcb hole with the lead; I found this was able to sufficiently heat both on the trickier joints.

[toneburst]

Thanks chimchim, that sounds like good advice.

a|x

[Wilba]

Things with semiconductors (diodes, transistors, voltage regulators, ICs) are most heat sensitive I think.

Capacitors are fairly tolerant.

Resistors are practically impossible to destroy with a soldering iron (most often they are just chunks of carbon).

It's OK to have high temperatures for a short time... i.e. it's better to solder a joint quickly at a high temperature for just 3 seconds, than use a lower temperature for 10 or 20 seconds. With ptractice, you can solder a joint in 1-2 seconds, and if it needs a bit more solder, do another 1-2 seconds. For temperature sensitive things like transistors, it's a good idea to let the part cool after each solder joint.

Don't worry about melting stuff other than the solder! You can't melt the PCB, the pads, or even the solder mask (the red coating on the PCB), and as I said before, it's better to have a good, high temperature and solder the joint quickly, than stuff around for several seconds trying to warm up the pad and lead enough to make the solder stick. Crank it up to 350° C and try a few resistor joints. Try also adding a tiny blob of solder to the end of the tip, then use this blob to get heat into the pad and lead. I mean TINY, not enough to make the whole joint, just enough to improve the heating of the pad and lead. Also (and sort of contradictory advice) try holding the iron at a shallow angle, like 20° from the PCB and use more of the side of the tip rather than just the end of the tip. Your aim in all this is to heat up the pad and lead at the same time to a hot enough temperature that solder you feed in just melts instantly, turns into a blob on the pad and then ultimately gets sucked into the hole.

[toneburst]

Great news!! I've been hearing sounds from my MB-6582!!!

I've installed 2x 8580 SIDs in slots #1 and 2, and I've managed to get the V2 editor working, so I can at least hear the default LEAD, BASS and DRUM sounds.

The Bassline sounds great, though it would be nice if the resonance went a little higher. I may end up having to build one of Seppoman's external CV-controlled filters, I suspect.

Drums seem very quiet, for some reason

Multi mode doesn't seem to work, but I guess this is because I only have one working Core.

Unfortunately, I haven't yet been able to load the v2_vintage sound bank, so I can't hear what the SIDs can really do, but I'm excited to have got this far.

On the downside, still not a peep (or a bleep or a bip) from SID socket #7. I have a horrible feeling that I've done something really nasty to some of the pads in this area. A little while ago, I bought some liquid flux from eBay, and attempted, without really knowing what I was doing, to use it to help solder some of the stubborn pads I've been having so much difficulty with throughout this project. It looks like it might have done something nasty to my board, because around some of the pads I used it on, the board has gone black. It looks like one pad in particular has been completely destroyed. I can't see any sign of the metal of the pad, and the hole through the board seem bigger than it should be, which is rather worrying. I tried to remove the black stuff with the tip of a scalpel, which I think was probably a Very Bad Plan. Although I tried to be very careful about it, it looks like I've managed to scrape off some of the red coating of the board, around the pad. I'll take a couple of snaps of the offending area tomorrow. I'm not sure if this is the cause of Socket #7 not working. That particular solder joint seems firm, at least, so I guess there is some solder inside the hole, at least, so despite it looking horrible, the component may in fact be adequately connected. I'm beginning to think that the transistor may be trashed. Replacing it will involve more desoldering though, a prospect I relish hardly at all... :(

a|x

[toneburst]

Just managed to get the v2_vintage bank uploaded, after re-downloading the sysex file (thanks for the suggesting I try that, TK), and am having fun flipping through the presets. One thing I do notice; there's an awful lot of 'oscillator leakage' going on. I was under the impression this was much less of an issue with the 8580/6582 SIDs. I was (probably naively) expecting a lower noise-floor. Think I'm going to be reaching for the gate effects when recording this...

a|x

[Futureman]

With regards to the dead SID slot, did you try buzzing it out with a multimeter etc? no IC's in the IC holders, just using the schematics and seeing if you have a short or open circuit somehere?

Regards

Mike

[toneburst]

Wow.... Just connected a keyboard for the first time, and.. this thing is addictive!! It sounds really GOOD!!!

I'm worried I'll never get the CS done now- I'll be too busy playing it :D

Thanks for the advice re the non-working SID slot. I'm not quite sure how to do what you recommend Futureman- sadly I'm a bit of an electronics newbie, unfortunately. Sounds like a good idea though. Any tips on which components I should test?

a|x

[TK.]

It also makes sense to ground the Audio INs of each SID, e.g. with jumpers connected to all J3_SIDx headers

This should lower the noise floor dramatically!

Best Regards, Thorsten.

[toneburst]

Hi Thorsten,

good call. I did do that, actually, which cleared some of the noise. I've found applying a gate to the output in Live clears things up nicely, anyway, and the noise isn't noticable while a note sounds, so it's really not a problem. I might have to invest in some kind of hardware multigate if I were use it live, though, as the cumulative noise of 8 ungated SIDs might be more of an issue.

I'm loving the sound though! Enough to make any slight concerns over noise levels insignificant.

a|x

[toneburst]

I've killed another LCD :(

I must be one of the most inept MIDIBox builders out there, I think...

I put the connector on the wrong way around. Just like I did the last time (when I trashed my really expensive red 20x2 I bought via a friend in the States....

Ordered another identical unit on eBay. Next time I'll try and make a neater job of the soldering, too.

a|x

[chimchim]

There's an old carpenter's adage:

"measure twice, cut once"

You will save yourself much time and frustration by being overly anal retentive when it comes to this stuff.  For instance I just got done soldering all the diodes for the CS pcb and I must have verified each was oriented in the correct direction 5 times before soldering; not to mention one more time after I had one leg soldered on each.

The time spent triple/quadruple/quintuple checking your work is minuscule compared to the lost time de-soldering and replacing components (add in frustrating lead time if you must mail order replacements) which WILL happen if you are too hasty.