At initial testing and assembly - LEDs not lighting up

0bbc7d5f108a5cfd

Hello - I’m at the testing and assembly section of the matrix and only the areas in the red box (LEDs 1 to 21) on the top board are lighting up red. All buttons are unresponsive, if they should do something at this point.

Not sure if the matrix test mode is like the loopa which had the messages pop up quickly, but it’s just the rainbow test screen shown.

Finally, I did goof on some of the micro match cables where the connectors are on the wrong side of the cable, but the polarity tab is on the red side and they are plugged into the boards lined up with the little holes correctly. I can’t see how that would do this but I do not have extra connectors and they are a bit of a pain to remove as you may know. They all work, the wiretester kit you all have is awesome.

Just stumped. Thank you!

EDIT: Revised test results as I read the larger NEO board incorrectly…update below.

0bbc7d5f108a5cfd

All the NEO_68 boards have:
12V measured at 2-pin connector pins on board
12V past diode
5V at the pad closest to the Traco regulator on the other side from the diode(it’s an unconnected C105)

The NEO_85 board has:
12V measured at 2-pin connector pins on board
12V past diode
9.5V at the C105 pad closest to Traco

And the ICs on the NEO_85 naturally have 9.5VCC on all three, which I’m hoping does not fry them although it looks like operating voltage is only 6V 🙁

I’m guessing it’s the TRACO on that board? It is the exact same as the others, oriented correctly. From reading other troubleshooting posts on the Matrix forum I don’t see how that would affect the LEDs.

I even tested just connecting the first NEO_68 board only to the Core PCB and I still see the same behavior, although the switches work, which would make sense since the offending NEO_85 is disconnected.

17703227d9536da1

All looks nicely wired up so far! Looks like all of the micromatch cables are made correctly, it always pays to test them (as you did).

By “switches work”, that means that the serial LED chains are working too? You’re right; measuring 9.5V past the regulator would be bad for the ICs and LEDs. Does the regulator have the correct part number? Also check that the Molex connectors are really making good contact, as a missing rail could allow power to flow in parasitically through the data lines. You did measure 12V though, so it is likely elsewhere.

Generally you can connect any neo board to the core to see if that can pin down the problem.

Do all ICs have the correct polarity?

0bbc7d5f108a5cfd

Thanks for your reply. All ICs have correct polarity. My IC3 (HC595) does look different than the ones in your build guide tho, please see picture below.

Does the really long 8-way cable have enough power to light up LEDs 2->21 by itself? They are quite weak looking.

I do not know what to check to see why power is not getting to the actual LEDs, when testing do I need to have the loopa connected? I have the core connected to one regular neo board and the behavior is:
1) LEDs 2->21 light up
2) switches activate the lighting effect but obviously only the LEDs 2->21 actually participate in lighting
3) I see no indication on the test screen of button presses if there is supposed to be

I may have connected the 47cm cable backwards on the core board at first, what would that have done? Looks like most of the pins go to GND. Since there is a small light effect when a switch is pressed I assume J44 on the core board is working correctly?

And when I hook up the larger NEO board, all switches stop working. Could the non-lighting LEDs and the issue on the larger NEO board be related or different?

Thanks
Tom

17703227d9536da1

Tom9000 Thanks for your reply. All ICs have correct polarity. My IC3 (HC595) does look different than the ones in your build guide tho, please see picture below.

Was the part ordered from the Mouser BOM, or is it a different part?

Does the really long 8-way cable have enough power to light up LEDs 2->21 by itself? They are quite weak looking.

I don’t quite get the question: the cable harness is used to supply 4× 12V to each neo board, and it works. But it requires the J44 cable to supply data to the LEDs.

I do not know what to check to see why power is not getting to the actual LEDs, when testing do I need to have the loopa connected?

It’s not required to have LoopA connected for troubleshooting, as long as you haven’t connected an uSD card it will boot into the test mode (guide 1.27)

I have the core connected to one regular neo board and the behavior is:
1) LEDs 2->21 light up
2) switches activate the lighting effect but obviously only the LEDs 2->21 actually participate in lighting
3) I see no indication on the test screen of button presses if there is supposed to be

It is odd that the first LED doesn’t light up, do you have an issue with the very first LED? Are you sure that the cable has the correct polarity?

Have you tried connecting data to a different neo board? You can debug the LED chain at the breakpoint, there are troubleshooting tips in the build guide (7.1). As far as I am aware, the button press response is not displayed on the screen, but rather the row/column of LEDs illuminates.

I may have connected the 47cm cable backwards on the core board at first, what would that have done? Looks like most of the pins go to GND. Since there is a small light effect when a switch is pressed I assume J44 on the core board is working correctly?

This means one of the micromatch cables? Which one? Power is not present on the input side of the PCB, but you might have shorted out the 5V regulator from the core.

And when I hook up the larger NEO board, all switches stop working. Could the non-lighting LEDs and the issue on the larger NEO board be related or different?

Hard to say, but you can debug the LED chain first and then looking at the switches will be easier. If everything stops working when neo85 is connected, it tends to suggest an issue with that board or cables, short circuit on the power line or bad regulator, shorts around the shift registers, etc.

0bbc7d5f108a5cfd

latigid on
Thank you. I’m not sure how to reply inline with your comments but here are my answers in order:
1) The part was ordered from the Mouser BOM, but if a different part would be better I can try those.
2) I was grasping at straws, thinking maybe the data cable provides a small amount of power but that does not seem to be the case
3) No SD card, no loopa connected sounds like the right method for testing.
4) I believe the first LED is wired correctly and does work, I posted an unlisted video to YouTube, link below. I am like 90% sure of the cable polarity but it’s not totally clear to me from the images in the build guide. I’ve tried the three smaller Neo boards they all exhibit exactly the same behavior, weirdly enough.
5) It was the long micro-match led cable J44. I did check the TRACO regulator on the Core board and looking from the bottom of the board oriented with the midi ports at the top, fro the three pins of regulator, the left pin is 5V and the right pin is 12V, so I assume it’s working correctly?

https://youtube.com/shorts/zKNCIvwspXs?feature=share
Please excuse the shaky camera I was holding the phone with one hand and testing with the other.

17703227d9536da1

Thanks for the video, it’s often worth a thousand pictures!
Looks like pretty normal behaviour to me: as you don’t have the axes board connected, the colour would be set at a default (I think red is correct), but you are able to get the white colour when pressing a button row/column, so it says that the serial LED transfer is working. If you connect the axes board, then the sliders start to change the base colour. You can see here

that it looks like you have a break in the LED chain. So where the LED stops lighting up, check the output pins on the previous LED, or the input pins on the LED that doesn’t light up. It could also be an LED mounted with 180° rotation.
Like that you should be able to work through the “snake” (year of the snake will be here soon) and see how far down the LEDs go.

0bbc7d5f108a5cfd

That is something to try. Are the LEDs in series so if one is flubbed then it breaks the chain for the rest of the board? If so that makes sense. I must have flubbed ones in the same general area on the board. They look correctly oriented but perhaps too much heat was applied. Is there a method to test an LED in place vs removing and replacing as I only have so many spares? My meter’s Diode test doesn’t seem to trigger any of them and I’m wary to try a coin battery test like the other LEDs unless that is recommended. Is there a way to bypass one perhaps for testing purposes? Maybe shunt something?

17703227d9536da1

Tom9000 That is something to try. Are the LEDs in series so if one is flubbed then it breaks the chain for the rest of the board?

Yes, the LEDs are driven with a serial protocol, so any breaks in the chain will prevent data flowing downstream.

If so that makes sense. I must have flubbed ones in the same general area on the board.

If they stop at exactly the same position on every neo68 board, I’d be very surprised, but it is not impossible I suppose. The data coming out of the LED is supposedly buffered/re-driven, so if a signal makes it to a few LEDs, it should reach all of them.

They look correctly oriented

Did you happen to check the LEDs before putting the keys on (build guide 3.12)? You probably won’t be able to see the notched pin/pad any longer after the solder has been applied.

but perhaps too much heat was applied. Is there a method to test an LED in place vs removing and replacing as I only have so many spares?

Could be, especially if some LEDs look a bit toasty brown. To copy (again…) from the guide, the following could have befallen your LED(s):

Dead LED (remove and replace)
Inverted polarity (remove and replace; the reversed polarity probably will have killed the LED)
Input pins shorted (unbridge input side)
Output pins of previous LED shorted (unbridge previous output side)
Dry or missing solder joint (rework)

At least one of those has to be true: if it is shorted pins (you can test for a short circuit), then rework the joints, if it is dead or rotated, then replace the part. You’ll only see how many are dead once the chain gets going, and that is the easiest way to test.

My meter’s Diode test doesn’t seem to trigger any of them and I’m wary to try a coin battery test like the other LEDs unless that is recommended.

As they are data-driven RGB LEDs, a simple diode test won’t work in this situation.

Is there a way to bypass one perhaps for testing purposes? Maybe shunt something?

Theoretically you could bridge the data over, but I would tend not to suggest that at the risk of shorting outputs together or so.

Take my word for it 🙂, please try to touch up the pins and then swap out the LED that doesn’t light up.

0bbc7d5f108a5cfd

Ok thanks so much. I printed out the instructions so I may have missed the “test leds before switches “ although you have designed these to be easily replaced. I’ll give it a go.

Like the loopa these pcbs are like the highest quality ones I’ve received in kits.

0bbc7d5f108a5cfd

Quick update - these ePixel LEDs must be pretty sensitive or touchy. I have had to rework a few of them and then go to plug it in, a new row lights up! Some I think I was too conservative with the solder and time to begin with.

Removing them is a huge pain but doable. So far I have two boards lit up and working on the third!

Will post back with results.

17703227d9536da1

Nice! Keep at it!

0bbc7d5f108a5cfd

Quick followup that all LEDs now work! There were maybe ten LEDs that were causing a break in the chain. About half of those needed reflow and the other half needed total replacment, so I appreciate all the extras you include in the essentials kit!
Now to get started on switch testing….

0bbc7d5f108a5cfd

Testing the switches without the axis board connected, had a few diodes that were not making contact so fixed those. Most work and so far this is a pretty cool light effect!

I have run into 4 switches in a vertical row on a neo board (SW8, SW25, SW42, SW59) that are not working, and there are no obvious issues with the diodes or solder joints like the singles that weren’t working.

The fact they are all in a row/column on an entire neo board leads me to wonder if it’s some type of controller or IC issue? The cable is fine and I do not see any issues with the ICs on that board. Any thoughts?
Thank you

17703227d9536da1

Great that you could get the LED chain going. Now we’re onto guide 7.2, but as you gave the switch column, that can be narrowed down to

All diodes soldered with reverse polarity; you mention that these are all fine, but just to double check
R17 (vertically in line with the diodes too) is missing or soldered with a dry joint
IC1 pin 5 is soldered with a dry joint

Hope that does it!

0bbc7d5f108a5cfd

It was IC1 pin 5! I have all LEDs and all switches working. Of course I hook up the Axes board and no such luck lol. I believe I am cursed at this point.

Here is the behavior: When both the MOW10 cable is plugged in and the AXES board is hooked into the core board, there is clearly a short since I saw some micro-sparks at the power inlet when I plug it in. Obviously I did not leave it plugged in and luckily the LEDs and Switches still work (whew). When I unplug the MOW10 OR the Axes board from the core board, no short behavior. I’m not sure where to look or what MOW10 and Core J1 header should be connected by design vs accidentally.

17703227d9536da1

Check for a general +5V power short (with the board disconnected) by measuring continuity/resistance/diode across C4 or C5. Maybe C4 0V is bridged to the adjacent +5V via?
Check for a general +3v3 short across C2, though the 3v3 is limited to that multiplexer chip.
If you plug the axes board into the core, can you control the LED matrix with the sliders?
Did you accidentally install or connect to the micromatch FOB10 socket J89A instead of J89? If you did, then when you plug in the MOW10 cable, you will bridge the +5V supplied by the neo85 board and the +5V supplied from the core, and generally regulators don’t like that.
Again the MOW10 cable should not typically connect power or 0V between boards unless you have installed axes D1 or bridged pins there. Did this MOW10 cable get tested too?
All ICs have the correct orientation? A clear photo of the board could help to spot errors.

0bbc7d5f108a5cfd

Ok tried all of your suggestions, no luck. The only thing that may be different than your standard build is I had to order the analog to digital convertor chip since I could not find it in my kit, again probably since it was from quite awhile ago. I ordered the MCP3208-CI/SL from Digikey, seems to be the same.

What should happen when I plug the axes CON1 into the core board but leave the MOW10 off? Because nothing is happening.

17703227d9536da1

Please post one or more clear photos of each side of the board.
Could you measure voltage anywhere? Might be hard connected directly.
CON1 connects the power and ADC to the core, so all of the analogue elements should work.
MOW10 connects the digital lines, controlling the switches and LEDs.

0bbc7d5f108a5cfd

Sooooo…probably going to laugh at this one but it turns out I used the 14mm standoffs instead of the 10mm and CON1 was not making contact after all. I found the correct ones, switched them out and all the sliders and joystick work great. I hooked it up to the switches/led boards and it’s very cool how you have the sliders and joystick visualized in the Matias switches!

The six buttons do not trigger anything though, except the switch BLM ports which changes from green to red and there is a fast popup message in the OLED. The other 5 buttons do not light up or register. Attached are pics of the board but I feel like I’m pretty close here!

17703227d9536da1

Tom9000 Sooooo…probably going to laugh at this one but it turns out I used the 14mm standoffs instead of the 10mm and CON1 was not making contact after all. I found the correct ones, switched them out and all the sliders and joystick work great. I hooked it up to the switches/led boards and it’s very cool how you have the sliders and joystick visualized in the Matias switches!

That’s a new one I think! Glad that it works, but I really wonder about the sparks from before…!

The six buttons do not trigger anything though, except the switch BLM ports which changes from green to red and there is a fast popup message in the OLED. The other 5 buttons do not light up or register. Attached are pics of the board but I feel like I’m pretty close here!

That is a good sign, if that LED changes colour when pressed it suggests that the digital signals are being read back through the board, and that the 74HC595 chip is the right way around. I’m not sure what the other MEC switches should do; maybe this is the time to connect to the LoopA and give it a spin?

If the LEDs on the MEC switches don’t light up when pressed, then it could be that the LEDs are inserted around the wrong way. On bottom side of the PCB, the encircled pin is the anode and the flat marking indicates the cathode. Don’t PANIC (Positive Anode, Negative Is Cathode). With your multimeter in diode mode, which way around do the probes have to be to light up your LED?