Sign in to follow this  
Followers 0
i45one

MB-6582 LED Matrix issue

30 posts in this topic

I've got an interesting issue with my LED matrix. One of the horizontal rows is dead (third from the top). I have checked continuity on the CS board and I'm pretty sure that all is well to the header at least. I've changed chips to no avail. Any ideas?

Share this post


Link to post
Share on other sites

did you trim the leads on the components under the lcd?  i had an issue there...

also read the voltage at the pin of the sr that feeds the base of the transistor that sinks that row.  if it is working, it will be pulsing, which will read less than 5v dc, but if it reads 0v then you'll know the problem is upstream.  if that's working, test the output of the transistor the same way.

just use a systematic approach and don't try to shortcut the process, trace the signal from one end to the other.  a scope would help of course, but it can be done without it.

Share this post


Link to post
Share on other sites

I removed the LCD just for good measure, same thing, same thing. Forgive me if I am wrong, but since it is an entire horizontal row, wouldn't something around U21/D5 be at fault? I could be entirely backwards but Ive looked at the wiring diagram till i am almost blinded with it. Thats probably the reason I'm not seeing the problem/solution. As far I can tell pulse is present.

Share this post


Link to post
Share on other sites

If it is only one row of the LED matrix at fault, then other horizontal rows are working, and LEDs outside the matrix, so thus the current sinking is working, and the current output of the first 74HC695 is working for the other pins.

So the problem MUST BE somewhere between U21 pin 5 and the start of that connection to the row of LED anodes which is to the right of the matrix.

It could be on the base PCB, or in the cable, or on the CS PCB.

Test between the IC pin itself (from the top) to JD6:D5 on the base.

Then from JD6:D5 on the base to JD6:D5 on the control surface PCB.

Then from JD6:D5 on the control surface PCB to the third row of LEDs... this is visible from the bottom as a horizontal track going through the LED anodes.

Share this post


Link to post
Share on other sites

akk, sorry, this is indeed clearly the source side.  my head still has the seq array in it which sinks a row at a time.  should have looked at the pcb, my bad. 

the troubleshooting is the same, though, it just doesn't go through the transistors.  if continuity along the line checks out, check for a short to gnd on that line.  make sure the resistor isn't the wrong value (this seems unlikely) or that row of diodes isn't backwards (even more unlikely). 

one more thing, how are you testing the functionality?  if you are using the matrix buttons, it could be that button as well.  easy to verify this by switching patches... hope this helps.  feel free to tell me to stop "helping"

Share this post


Link to post
Share on other sites

While I'm sure there are several sources for this kind of problem, I had several of these and they were all hairline splits nearby one the pads of the LED's in that line.  One of my lines had two of these.  One of them seemed to actually have been produced on the board because it was under the red solder mask.  I used a magnifying loupe to find them, and a small amount of solder to bridge the gap.

Kurt

Share this post


Link to post
Share on other sites

If it is only one row of the LED matrix at fault, then other horizontal rows are working, and LEDs outside the matrix, so thus the current sinking is working, and the current output of the first 74HC695 is working for the other pins.

So the problem MUST BE somewhere between U21 pin 5 and the start of that connection to the row of LED anodes which is to the right of the matrix.

It could be on the base PCB, or in the cable, or on the CS PCB.

Test between the IC pin itself (from the top) to JD6:D5 on the base.

Then from JD6:D5 on the base to JD6:D5 on the control surface PCB.

Then from JD6:D5 on the control surface PCB to the third row of LEDs... this is visible from the bottom as a horizontal track going through the LED anodes.

All of these were the 1st thing I checked. I have continuity from pin 5 of U21 all the way to the last LED on that row.  Everything is good there. I even checked continuity from pin 5:J9/J8 to  pin 13? of U21, continuity there too. This one has me stumped. What can I check higher up in the circuit? Is there anything that will make ALL of the lights on this thing wake up? I know all of the encoders are working. The way I checked the lights was to push the matrix buttons. They flash one time when pressed. when pressed on the bottom, the column lights up minus the 3rd row, all rows are functional except for the 3rd.

Thanks for the input so far  ;D

Share this post


Link to post
Share on other sites

ok it's not the switch then.

check for short to gnd on the line.

check for pulses at U21 pin 5.

if you read pulses there, move step by step down the line, testing for the signal at each point, all the way to the row of leds.

if you read 0v there, the only possibilities are a bad chip (several bad chips, since you swapped it, thus very unlikely) or a short to gnd.

Share this post


Link to post
Share on other sites

Again, I've checked for shorts to ground. No shorts. As far as detecting pulses, I have no scope. What controls that line(what is the origin of the signal? I want to trace it from the start instead of the end of the signal. How hard would it be go script something to hold all LED's in a lighted state? I know there is a matrix test, but it doesn't work with this box. Anyone know how to modify it for the MB-6582?

Share this post


Link to post
Share on other sites

Again, I've checked for shorts to ground. No shorts.

again??

As far as detecting pulses, I have no scope.

a scope is easier, but not an absolute requirement for this

What controls that line(what is the origin of the signal? I want to trace it from the start instead of the end of the signal. How hard would it be go script something to hold all LED's in a lighted state? I know there is a matrix test, but it doesn't work with this box. Anyone know how to modify it for the MB-6582?

you don't need that, the matrix buttons do this for you, just press a row button and a column button simultaneously for each led you want lit.  you can light all of them, if you want, in about 10 seconds...

Share this post


Link to post
Share on other sites

Now that my ISP is back up...

Here is what I've found after digging a little deeper:

Leds that don't work:

LFO/3

Control/Knob

Waveform/Triangle

Waveform/Random

Mode/Sync

Mode/Edit

As the plot thickens, I've not had a chance to look at diagrams to see if all of this is related to one thing. If this lights a light bulb, please feel free to share.

Share this post


Link to post
Share on other sites

Mode/Sync LED isn't being used yet.

Osc Control/Knob is actually Osc Asn in the diagram (I changed the artwork label but not original pin assignment table).

Find attachment which is an updated LED matrix test, which should work the same as the original LED matrix test but with the MB-6582 configuration and control surface. I hadn't realised until I ran it that it doesn't just show one LED at a time but sort of a trail of the last 8 LEDs turned on. It's easier to just look at it than explain it. It only shows LEDs in the 8x8 modulation matrix, not the other LEDs which are also in a LED matrix (all LEDs are being multiplexed by the last DOUT with the transistors, half are in the modulation matrix, the other half are in the rest of the CS). I'll work on another test app for showing all the other LEDs too.

mb6582_led_matrix_test.zip

mb6582_led_matrix_test.zip

Share this post


Link to post
Share on other sites

OK, here's a LED matrix test that lights up all the LEDs.

It does this the same way that the application does it, i.e. each "column" of 16 LEDs is lit in turn, one vertical column of LEDs in the mod matrix plus (up to) 8 other LEDs in the control surface. Just thought I'd mention that, because you should not sink all the columns at once, because this means up to 8 LEDs will be driven in parallel by one DOUT pin through one resistor... might not work, certainly not a good idea.

I still think your problem is a short or broken track somewhere... and that there is more than one problem... the 3rd row of the mod matrix is one problem and the other missing LEDs something else... perhaps even dead LEDs or other breaks in the track. I'll have a look at the PCB wiring and see some likely places that a break in a track would cause those other ones not to light.... but it's probably best you run this test app and report all the LEDs that don't work first.

mb6582_led_matrix_test2.zip

Share this post


Link to post
Share on other sites

Also you should check that there are no shorts between any of the pins of JD6, JD7 and JD8.

Obvious culprits would be two adjacent pins, but maybe a short between two leads of a LED or somewhere else on the PCB might be causing trouble.

Share this post


Link to post
Share on other sites

BTW I just noticed my CS wiring PDF is wrong... "LFO Sin" and "LFO Tri" are swapped.

"Sync", "CC", "Edit", "Play" are in the wrong order (should be "Play", "Edit", "CC", "Sync")

Thus the following LEDs are all on the same DOUT output tracks:

--->D6---+--->Osc Control Knob (Assign)
        |
        +--->Env Control Assign
        |
        +--->LFO 3--->Mode Edit---> SID 2
        |
        +--->LFO Random--->LFO Triangle
[/code]

I've tried to show also how it branches... not that it really is important...

This would then explain most of the following missing LEDs:

LFO/3

Control/Knob

Waveform/Triangle

Waveform/Random

Mode/Sync

Mode/Edit

My guess is the "SID 2" LED doesn't work either, which I predict you'll tell me as soon as you run that test app that lights all the LEDs.

IF there's a break in the track leading from JD7:D6, it will probably be before the Osc Control Knob LED not after...

Share this post


Link to post
Share on other sites

Thank you SOOO much Wilba! I'm going to give the app a shot. LED issues aside, this thing is a freaking beast of a synth. I can't stop playing with it.

Ran the app, and unless I am missing something, it only goes through the matrix led's. Cool little app though. And you are correct in your assumption that sid 2 doesn't light. I think your diagram will be VERY helpful in what to look at to try to solve the non matrix issue.

Share this post


Link to post
Share on other sites

I hear you are still stumped by this problem... and I missed the update to your last post  :-[  I would have replied sooner.

The second app I uploaded, the one in this post: http://www.midibox.org/forum/index.php/topic,10390.msg78577.html#msg78577

i.e. mb6582_led_matrix_test2.zip should light up all the LEDs, in the mod matrix and everywhere else.

I've uploaded two images of the PCB with the traces at fault. Hopefully these can assist you track down a break in the track or something.

I would follow the following troubleshooting procedure:

Take out shift registers in U21, U22, U23.

Connect two wires to J2 on the base PCB, which is between the power switch and the power socket. This is normally used to light the power on LED so it already has a 220 ohm resistor in series.

With these wires which we'll call red=5V and black=ground:

Directly test the LEDs that don't work in the 3rd matrix row. Put black wire on the lower of the two pins. Make sure the wires are good as well by testing ones that DO work. It's unlikely you have a whole row of blown LEDs but you should validate they all work individually.

Assuming the LEDs are good, validate the 8th column of the mod matrix. Put black wire in JD8:D7, red wire in JD6:D0 to JD6:D7.

Tell me what you see ;) and I'll think some more

MB-6582_CS_PCB_JD6_D5_thumb.png

MB-6582_CS_PCB_JD7_D6_thumb.png

2108_MB-6582_CS_PCB_JD6_D5_png867ddcb2be

2110_MB-6582_CS_PCB_JD7_D6_pnga32418bae1

Share this post


Link to post
Share on other sites

Thanks Wilba! I will try this when I get home and let you know the result.

From previous testing I know the LEDS are functioning. My meter in continuity mode presents a led friendly current. I am able to light each LED individually from the main PCB points, and CS PCB points. I can also rule out bad registers as I have rebuilt the whole DOUT section with the exception of the transistors. I don't have any spares lying around. If you think that might be the cause I will replace those too.

I don't know if this tells anything, but I noticed that when I provide current to that row, those LEDS are a little brighter than the other rows. This led me to believe that I might have a bad resistor pack. I replaced all of that with different resistor packs and got the same result.

Share this post


Link to post
Share on other sites

gah! i fixed everything except for the 3rd row which is still dead as a doornail. that app is a GREAT thing to have. It turns out that there was a cold solder joint on one of the resistors, I dont know which one because I just reheated them all and everything else lit up.

Still cant figure out whats up with that 3rd row.

Share this post


Link to post
Share on other sites

Problem solved....

Thanks again for everything. Just in case some other person has a similar problem, check BOTH sides of the board with a magnifying glass before ruling out bad joints/shorts.

Share this post


Link to post
Share on other sites
I just reheated them all and everything else lit up.

<solder police>

Remember kids: you should always apply a drop of flux, or add a little flux-cored solder to any solder joint you re-heat.

</solder police>

Share this post


Link to post
Share on other sites

....before ruling out bad joints/shorts.

More advice: DON'T rule out bad joints.

Share this post


Link to post
Share on other sites

I want to know exactly what the problem was, exactly which joints were bad, where the shorts were, etc.

This will help me diagnose any other problems people have in future.

For example, you said all the LEDs would light if tested via the JD6, JD7, JD8 pads on the PCB.

So I need to know why testing this way (simulating DOUTs) would be different to when it was powered on.

Oh and congratulations on fixing it!!!

Share this post


Link to post
Share on other sites

If you noticed I talked about both sides of the board. On those connections the traces are on the top of the board. I soldered the resistors from the bottom originally. Best I can tell the "through hole" plating was off because the joints were fine on the bottom surface. Soldering the joints from the top solved the problem. I found it using your app. I found that by pressing on the area around JD6 lights would come on, go off etc.

The exact joints that were off were JD6:D3 and D4. I soldered all of those resistors from the top and the thing is rock solid. If it's not on the wiki, you should put that app on there as a general testing tool. It saved my butt, and quickly at that.

Now if I can get those two crippled 6581's out of it will be pretty much perfect. Last two 6582's should be here Monday.

Any recommendations on pots? The ones I got were way too big to fit the back panel.

Thanks again!

#78 :-)

Share this post


Link to post
Share on other sites

If it is only one row of the LED matrix at fault, then other horizontal rows are working, and LEDs outside the matrix, so thus the current sinking is working, and the current output of the first 74HC695 is working for the other pins.

So the problem MUST BE somewhere between U21 pin 5 and the start of that connection to the row of LED anodes which is to the right of the matrix.

It could be on the base PCB, or in the cable, or on the CS PCB.

Test between the IC pin itself (from the top) to JD6:D5 on the base.

Then from JD6:D5 on the base to JD6:D5 on the control surface PCB.

Then from JD6:D5 on the control surface PCB to the third row of LEDs... this is visible from the bottom as a horizontal track going through the LED anodes.

hehehe

I thought you did this already ;)

I used 16mm dual gang 500k audio/log taper pots. Linear is fine though, it doesn't really matter.

Look like this one but shorter shaft:

http://www.altronics.com.au/index.asp?area=item&id=R2296

If someone finds this on Mouser or Reichelt, please post a link.

#1, #2 ;)

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0