Electrical setup help

GR9

New elf
Joined
Dec 11, 2023
Messages
43
I’m just going to preface this by saying that I thought I knew what I was doing, but I quite apparently don’t. Since I don’t feel like starting a fire (I think I almost have at least twice now), I really want to know what I am doing wrong, and what I’ve already destroyed and need to replace.

Ok, so I’ve got a 12v 30A Alitovo power supply from Amazon, a bunch of 12v WS2811 pixels, a raspberry pi 4, and some custom circuit boards and fuses.

The setup works perfectly fine when all 500 pixels are connected to the pi and PSU via jumper cables, with absolutely no issues.

The problems started happening after I tried to add in fuses, pigtales, and bump the 3.3v output from the pi to 5v (Which seems to be recommended by a lot of people).

Below, I’ve attached a (mediocre) drawing of my setup. The second image is the same circuit diagram that it embedded in the big drawing, but in readable format.

IMG_1371.jpeg
IMG_1370.jpeg
I’ve blown about 10 fuses at this point, and after the first fuse blow, only had one string connected so I did not mess the pixels up or anything. lights do one of the following every time I start it up and run a sequence over FPP.

1) Work properly. This is rare, and every time I turn the system off to add one more string, after turning it back on again it never works right. If I add all the strings (With a 4A fuse, and connection to the other pigtale halfway though the line), it ago does not work.

2) Blow a fuse somewhere. The 8A fuse on the pigtale that does from the PSU to the lights is the most common one, but the 4A fuse at the lights themselves did blow once.

3) Do absolutely nothing at all

- - -

So I pulled out a multimeter to try and figure out if a wire was not connected right, if something was backwards, not throwing enough power, etc.

I tried a bunch of things, but here are the ones I remember:

1) Set it to “20DCV” and went from Pin 18 on the pi to the data in on the pixels, while the pixels were powered as shown in the diagram. No fuses exploded, the lights did not work, and the reading was well under 1V. I bypassed by custom circuit board for this. Since I thought the reading was supposed to be around 3.3, I tried a different pi I had lying around connected to the circuit as shown above. Don’t recall what happened, but the lights did not work right. Testing it the same was as described above, it also read well under 3.3. Connecting it to the custom circuit board and testing the output did not change the reading. At this point, I re-soldered the custom circuit board with a new board and all new parts, did not resolve it.

2) After this I quit, but i thought the PSU might be an issue, so I unplugged the pi and the pixels (But left the cables and fuses plugged into the PSU), set the meter to “20VDC”, and touched the black lead to “V-“, and the red to “V+” on the PSU. I swear I did this before with no consequences, but this time something popped and a bunch of smoke came out of the PSU. It shut off, then restarted itself and appeared to run fine (But I did not plug anything into it to actually test that theory). Someone I talked to said I probably blew a capacitor in the unit, so assuming they are right, is it still safe and reliable to use? For the record, the 2 of the 3 8A fuses connected to the unit blew out as well.


summary:

I’m an idiot and know significantly less then I thought, and really need help figuring out what the heck went wrong and how to fix it as this project is taking both too much time and too much space (Let alone all the frustration).

If you have any ideas / thoughts as to what went wrong or how to fix it, they would be greatly appreciated!
 

merryoncherry

Senior elf
Joined
Apr 2, 2022
Messages
650
Location
Cherry St., Hudson MA USA
The data signal from to the lights is digital pulses, so it will not read all that well on the multimeter, at best it will read a fraction of its peak voltage.

I can't really comment on the rest of it, as I am not sure how you wired the pigtails, etc.
 

GR9

New elf
Joined
Dec 11, 2023
Messages
43
as I am not sure how you wired the pigtails, etc.
What do you mean? They are connected to the fuses via CAT5 cable (Pulled it apart to get one singular wire). On the pixels side, the positive is connected to the fuse (which is connected to the positive line of the lights), and the negative is soldered right to the lights.

at the lights:

image.jpg

In the box

image.jpg
 

Wayne1

Apprentice elf
Joined
Dec 6, 2021
Messages
65
On the 3rd drawing you have the Ground going to the +5 rail (my guess is that you have the wires crossed at some point therefore the fuse issues)

The smoke that came out of the power supply sort of would indicate a blown cap but you cant be sure without checking.
 

MikeKrebs

Full time elf
Joined
Dec 8, 2014
Messages
170
I am concerned about the safety of using 2 24AWG wires to run 8 amps. That seems like a lot for those little wires. And the V- (ground) side should be equivalently size.


I would go back the very basics. Here is how I would trouble shoot.
Check all pigtails for continuity. Make sure nothing crosses over unexpectedly.
Disconnect all outputs on the PSU.
Test all the V+/V- (ground) terminals against each other to make sure you are getting 12v out. You should as they are bussed together.
Connect your fuse box using 18 awg spt1. Measure voltage coming out of fuses. (Your fuse picture is concerning. Stray wires and bare connections too close together. At least put some tape on them if you don't have shrink tube.) Test for 12v. Adjust PSU if needed.
Connect your 5v DC2DC circuit board. Check for 5V.
Connect the addon board. Check for 5V.
Connect the pi. Check for 5V on the input pins. Check for 3.3v on the 3.3v output pins. As Merry pointed out, you can't really get any good number off the multi-meter on the data pins.

Go back to the pixel outputs and see if you have 12v on the appropriate wires. Measure voltage at end of line.

Make sure you are running the ground from the SN74 to the pixels (along with the ground from the 12v). After you verified all those voltages, you should be good that way.
 

GR9

New elf
Joined
Dec 11, 2023
Messages
43
On the 3rd drawing you have the Ground going to the +5 rail
Can you point out exactly where you see this? I’ve got a few images next to each other so I’m not quite sure which one you are calling #3


Test all the V+/V- (ground) terminals against each other to make sure you are getting 12v out. You should as they are bussed together.
How would I do this safely with the multi meter? Last time I touched it the thing kind of exploded, so I don’t exactly want to start a fire this time. Can I just set to 20VDC and touch black to V- and red to V+?


Stray wires and bare connections too close together.
I had them all taped up, but took it off for the picture so you could actually see it. Glad I’ve got at least one thing right :)


Check for 5V on the input pins. Check for 3.3v on the 3.3v output pins.

Go back to the pixel outputs and see if you have 12v on the appropriate wires. Measure voltage at end of line.
Same question as above, how would I do this without starting a fire or blowing a fuse?


but you cant be sure without checking.
Took the cover off, and everything looks OK to me. Is there a specific component that should check?

image.jpg
 

MikeKrebs

Full time elf
Joined
Dec 8, 2014
Messages
170
Yes 20V DC should be where you set your multi-meter and either direction for your wiring. If you do it "backwards", you will see negative voltage. Do you have the wires connected to the right terminals on you multi-meter?

If you had "a bunch of smoke" come out of the PSU, to my untrained eyes, I don't really see the problem from your picture. I assume only traces are on the bottom of the board?

Test your multi-meter on some other devices like a 9v battery, a 1.5 volt battery, a 12v battery.

Edit...now that you have the cover off, test with the cover off so you can see the smoke should you cause it. I would be gun-shy but once I was pretty sure my MM was okay, I would test again. You are looking for something wrong. If something is wrong with the PSU, you want to find out while you don't have anything connected.
 

Skymaster

Crazy elf
Global moderator
Generous elf
Joined
Dec 19, 2021
Messages
1,093
Location
Western Sydney
I'll just point out that that the majority of the board in the bottom half of that photo is at lethal voltages. Be EXTREMELY careful.
I'm somewhat concerned that things have gone bang when multimeter probes are attached and honestly do not think a forum is the best place to get advice. I would be seeking a professional or someone who has strong experience in dealing with this stuff who can assist in person.
 

Grozzy

Oh great, the idiot's back
Joined
Aug 16, 2018
Messages
107
With it working before you added the pigtails, suggests your Power Supply *was* fine (maybe not now).

It intermittently works would indicate that you don't have pigtails that were accidentally wired wrong from the factory. But might be caused by a broken/bare wire in the pigtail cable or connector that when bumped or moved shorts.

As someone mentioned above check for both continuity and shorts on your pigtails. Do this with them removed from power. Your multimeter should have a continuity setting that beeps or shows all zeros.

It should beep when measuring one end to the other same pin to same pin but not across pins.
 

GR9

New elf
Joined
Dec 11, 2023
Messages
43
I am 100% going to agree with Skymaster here, I am not running that thing with the cover off. Those giant capacitors don't look fun.

For testing continuity, would I just use the resistance mode? I don’t see a continuity mode on my multimeter, but I suppose I could have missed it. See image below.

image.jpgThanks for everyone input thus far!
 

Skymaster

Crazy elf
Global moderator
Generous elf
Joined
Dec 19, 2021
Messages
1,093
Location
Western Sydney
For testing continuity, would I just use the resistance mode? I don’t see a continuity mode on my multimeter, but I suppose I could have missed it. See image below.
Yes, use the 200 ohm mode; you should see close to zero ohms.
You might be able to use the diode mode too - some will show voltage drop and a beeper on conduction (the one between 200R and the hFe mode) - but that behaviour is not consistent across meters.
 

GR9

New elf
Joined
Dec 11, 2023
Messages
43
Ok, I did some testing with the multimeter, and now that I know what I'm looking for :)

I think the problem is that I am an idiot. I haven't touched the setup since I blew whatever it was that I blew, and I was getting number around 120 when testing with the multi-meter on my custom circuit board. I went to type the readings into this post, and forgot one, so I went back to re-test it. Just so happened to notice the pigtale that went from that custom circuit board to the car transformer in the box did not have those little arrows on the side lining up. (Yea, the pigtales I bought apparently go in the wrong way without much extra force needed).

Assuming that was the actual cause, since re-testing always returned the "1" (Which is the same number that appears on the screen when the leads are not attached to the meter, so I assume that's the default state).

In any case, I still don't want to turn on the PSU until I am almost certain it won't cause another explosion on its own, so I'll leave the question I typed before re-formatting this post below:


One more note. When I tested the PSU (V-) to the PSU (V+) I got a reading of 129.6, which I am also not sure if that is normal or not.

Thanks for everyone's help thus far!
 

Grozzy

Oh great, the idiot's back
Joined
Aug 16, 2018
Messages
107
Testing across the power supply with the meter set to Ohms/Resistance will always give off weird values as you have a whole bunch of components on the other side. I would just disconnect everything off it except the AC connection, power it on and then with the meter set to DC check across the DC 12v output to see if you are indeed getting 12v. And avoid touching the AC 110/220 side.
 
Top