MB-6582 Base PCB Errors

[Wilba 3]

I hate to create a topic like this, but it's the truth!

People found errors on the PCB before I did because I have NO TIME to pick up the iron and solder the base PCB myself and find them first. Damn.

1. I failed to connect the B+ pin of the LCD port (J15) to +5v. Since it wasn't connected in the schematic, it didn't show up as a missing track in the layout.  >:(  This should just mean that the backlight doesn't light up, but I will have to validate this won't kill your LCD. You will have to connect the B+ pin (pin 9) to +5v (pin 12).

2. The C64 power socket works with my C64 PSU but not with some other peoples. See here:

Bridging across those pins should solve the problem.

[Davo 0]

Will these fixes appear in the next batch of boards if/when that happens?

I don't think there would be a problem to bridge pins 4 and 5 if you don't need it.  It should be okay to put that fix in the PCB design files.

[Jaicen 0]

Jeez, this is a catastrophe! I can't believe an error of this magnitude slipped past on such a simple board! ;)

Seriously though, that's a really minor thing to have to correct, if that's all you did wrong on this layou you can feel pretty satisfied with yourself in my view.

[bugfight 0]

Jeez, this is a catastrophe! ...

hehe, sarcasm is the tongue of the devil...

seriously wilba, don't be down about any of this.

it ain't no thing.

regardless of these few little errors, i'm very grateful

for you hooking me (and everybody) up like you have.

[Wilba 3]

I said "people found errors" but it was jimp.

I know it's no big deal, I was just hoping a repeat performance of the prototype, which was faultless, right up until the point TK changed to using PIC18F4685 and it needed a rewired MIDI Out and a CAN bus ;D

[DavidBanner 0]

I was wondering if anything like this would show up, doing the PCBs is pretty much like doing a print run and anytime  I've done a print run, there is usually at least one typo - it's sod's law.... It's even harder not to let a tiny glitch slip though when you don't have at least one extra set of eyes looking over the whole job to proof it.

given the complexity of the project I'd say this little glitch is definately "within tolerance", and definately a big win for Wilba and everyone else who's benifiting from the time and effort W has put in.

I can't wait to get my iron fired up.....

[Wilba 3]

Oops again!

I put a via under the crystal, which means it might conduct from the via pad to the metal can of the crystal if it's flat against the PCB. Probably won't matter, but to avoid any potential problems, mount the crystal a little above the PCB, use a strip of card under it while you solder and remove after.

[Altitude 52]

what are the smd pads for by the cores on the back

[Wilba 3]

what are the smd pads for by the cores on the back

They're for bridging with solder blobs should you need to use 8-bit mode for an LCD or PLED display.

[julienvoirin 4]

there is maybe a new error :

for the PSU option B :

Solder a bridge at J71 between pin 1 and 2. (This connects V1 common pin to 5v, so it outputs 14v relative to ground).

doing this connect the 5V supply line reserved for the Cores to the ground, so no more supply

Solder a bridge at J72 between pin 1 and 2. (This connects V1 output pin to 14v supply lines, to go into V2 and V3).

9V entering a 12v regulator can't give 12v

Am I right ?

i discovered this by soldering the base PCB and testing voltages.

The solution i found is not very elegant. don't bridge as mentioned, i connect the input of V1 to the input of V3. But it bypass C12 and C11.

{edit} : stupid me, i was using the same heatsink for V1 and V3 ,connecting ground lines together and causing this prob.  do not use heatsink

[Wilba 3]

There is nothing wrong with the PCB layout. PSU Option B works.

You are getting confused by the 9V DC being added to the 5V DC to produce 14V DC relative to ground.

doing this connect the 5V supply line reserved for the Cores to the ground, so no more supply

No, V1 common pin is not connected to ground.

In this wiring, V1 will regulate the input to 9V DC, 9V above the common pin which is 5V above ground.

Thus V1 output will be 14V above ground, as stated.

9V entering a 12v regulator can't give 12v

As above, V1 is outputting 14V above ground, which is then connected to the V2 and V3 regulators, which have their common pin on ground.

i discovered this by soldering the base PCB and testing voltages.

The solution i found is not very elegant. don't bridge as mentioned, i connect the input of V1 to the input of V3. But it bypass C12 and C11.

Better solution appreciated

You probably did not test the voltages correctly.

If you test the output of V1 relative to common pin (middle pin) it will be 9V. If you test the output of V1 relative to ground, it will be 14V (for PSU Option B).

To reiterate, in PSU Option B, the 9V AC from the PSU is rectified and regulated into 9V DC (by V1) but this 9V DC supply does not have the same ground as the rest of the PCB, V1 common pin is connected to 5V so that V1 outputs 14V relative to ground.

As mentioned in the wiki, PSU Option B is like the "C64 PSU Optimized" wiring diagram:

http://www.ucapps.de/mbhp/mbhp_8xsid_c64_psu_optimized.pdf

the only difference is a common 7809 and 7812 regulators supplying multiple SIDs instead of one regulator per SID module.

[julienvoirin 4]

OK PRBLEM SOLVED

Thanks wilba for the explanations it gave me the idea to look ... the heatsink shared by  V1 and V3. It was linking the 2 ground lines.

So for midiboxers : DON'T USE HEATSINK

[NorthernLightX 2]

So for midiboxers : DON'T USE HEATSINK

It's perfectly OK to use heatsinks, just don't share a single one among the different regs  ;)

[julienvoirin 4]

or add mica plate between heatsink and regulators; i told that for beginners

[Davo 0]

or add insulated copper plate between heatsink and regulators; i told that for beginners

Insulated copper plate?  Don't you mean "mica sheet"?

[Wilba 3]

OK PRBLEM SOLVED

Thanks wilba for the explanations it gave me the idea to look ... the heatsink shared by  V1 and V3. It was linking the 2 ground lines.

So for midiboxers : DON'T USE HEATSINK

Ahhh! That would explain it...

Yeah, it's OK to use a single heatsink between V2 and V3, these share a common ground. For PSU Option A, you won't need a heatsink on V1.

[julienvoirin 4]

sorry wilba , i well feel that you were upset by my dummy remark.

i think that a link to your flicker would be a good idea, as it explain a lot :

http://www.flickr.com/photos/wilba/408737117/in/photostream/

and gives an image to the components used.

about that i tried pcb jacks but it seems they do not fit. Which one did you used ?

and the led used upper the power switch, how do you do to connect to j2. Do you bend legs ?

in the wiki, i think that v1 and v3 not sharing the same heatsink would be a good line to add, i was confused and without thinking a lot i let them share the same ground.

[Wilba 3]

sorry wilba , i well feel that you were upset by my dummy remark.

No, I'm not upset! While I have tried to explain as much as possible in the wiki, there will always be things I didn't explain, or expect people to do (like connecting a heatsink between V1 and V2/V3!)  I am actually happy that you didn't destroy your C64 PSU by connecting 5V to ground!

I've added a note to the PSU Option B construction notes.

[/tilted/ 2]

It's perfectly OK to use heatsinks, just don't share a single one among the different regs  ;)

Shock of shock! The Heatsink fetishist strikes again...  ;D ;D

[LaRssA 0]

hey all! :)

Just finished my 6582 today (yay! :D)

allthough, I found another bug on the base PCB.

as mentioned, you need to solder a bridge between pin 9 and 12 at J15.

This connects +5V to the backlight, but I also had to solder a bridge between pin 10 and 11 (GND) to make it work.

Wilba, can you check if this is just something I had to do, or if it is an error?

Anyway, I love how it has turned out, and I will post pictures if anyone's interested(?) :)

Lars

[MTE 0]

and I will post pictures if anyone's interested(?)

No ;D Im satisfied by seeing all the time the same SIDs  ;D  ;D  ;D  ;D

[Wilba 3]

Pin 10 goes to the BC337 which is sinking the current for the backlight.

Grounding this means you're connecting +5V/ground to the backlight without any current limiting.

Perhaps your transistor or other joints were bad or the brightness pot wasn't turned up, or a break in the other tracks.

Try removing the solder bridge at pin 10 and checking continuity for pads of T1, R3, R4, P1.

[nebula 3]

1. I failed to connect the B+ pin of the LCD port (J15) to +5v. Since it wasn't connected in the schematic, it didn't show up as a missing track in the layout.  Angry  This should just mean that the backlight doesn't light up, but I will have to validate this won't kill your LCD. You will have to connect the B+ pin (pin 9) to +5v (pin 12).

2. The C64 power socket works with my C64 PSU but not with some other peoples. See here:

http://www.midibox.org/forum/index.php?topic=9775.msg71430#msg71430

Bridging across those pins should solve the problem.

Wilba (or anybody else who can answer this):

I would like to confirm that these are the only two known errors on the base PCB.  I thought I might have seen something else in another thread... ?

A short summary of known MB-6582 PCB errata would probably make a good Wiki page, as an nameitative "single source" of this important info :)

[Wilba 3]

OK nag me if I do not wikify this within a week  ;)

FYI there are only those two "missing features" on the base PCB (backlight and power socket) ;D

There's one less-than-perfectly-placed diode on the control surface PCB (left of the Shift button).

Also, the "SID R" LED hole on the panel is partially overlapped by the PT-10 mount, requiring some trimming, but that's technically not a PCB error because the LEDs are exactly where I want them (the "SID L/R" button is aligned with the menu encoder).

[Jack 3]

Hi,

I have measured on the 5V line, 8-12V DV, on the naked Base PCB.

I am using the PSU Option A and the 9V line is OK.

I think this the overvoltage is caused from loaded caps, because no load is connected.

Can someone confirm this?

Best Regards

Jack