Hello AAC forum,

Working on an LED display that uses LEDs in groups of six. In parts of the display
two groups of six are needed. In other groups three groups of six are needed.

The single module has six LEDs. Specs:
5mm LED
Forward voltage: typical 3.2 v max 4.0 @ 20mA
Peak forward current 100 mA
Steady current 30mA
Data sheet on 5mm pink LED

Figure A is the single LED module.


When voltage is applied to the 12v+ terminal and ground at DMOS the modules
lights the six LEDs as designed.
(Perhaps I should be showing the circuits as schematics instead of PCBs. Although
some assumptions are required, like the connections of the resistor pads with a 100 ohm
resistor, the PCBs are thought to be simple enough to be understood. If schematics
are required to communicate the circuit, please indicate so in a post and schematics will be
drawn and posted.)

So two modules are connected as in Figure B

Both modules work independently but when they are connected as in Figure B, Module Two comes on but Module One does not light.
When module two is disconnected from module one, module one lights as it did in the unit test.

So it is conjectured that while both modules have the 12v+ connected correctly, the ground is not as it should be. That is the 12v source at module one traces directly to the 12v output to module two. But the ground to module two is in series with R1 which causes the ground to malfunction
when the modules are daisy chained.

So to test this supposition a third configuration was made.


In this arrangement both modules operate as intended. The trace from R1 to the DMOS output was drilled out and a jumper wire ran from the terminal at module two to module one.

So the question is to confirm that the PCBs should be refactored so that the ground (marked DMOS) should be connected in a manner similar to the pattern used to connect the 12v+ from input directly to output. That is, there should be a trace directly from the ground input to the ground output. Right?

It seemed that if the R1 resistor pad closest to the ground terminal is connected to the ground terminal, then that would provide sufficient connection to the ground on terminal to provide a ground connection at module two. Apparently the contact to R1 interferes with the ground connection to terminal two.

Thanks.

Allen in Dallas

 

My thinking is that there is a gap in the conductor somewhere along the path that is replaced by the wire. Probably that gap is at the point of connecting to a solder pad. An ohm meter check between the two solder pads will probably reveal an open circuit that is hidden by the solder mask material.
Was the artwork for these boards done with tape on mylar, or with a PCB layout program, or was it a cad program used for doing the layout??
And why is the trace above the lower row of LEDs so very close to the solder pads when there is enough room for it to be farther away.

 

I see no problem with Fig. A except for trace size. I would connect my meter to see possible problem.

 

I see no problem with Fig. A except for trace size. I would connect my meter to see possible problem.

I DO see an excessively close spacing for a layout that may be hand soldered. And how many boards have you designed that would be hand soldered, without solder mask? When it takes no more effort to provide greater distance than one trace width it makes no sense at all to place the traces so close.

 

So the question is to confirm that the PCBs should be refactored so that the ground (marked DMOS) should be connected in a manner similar to the pattern used to connect the 12v+ from input directly to output. That is, there should be a trace directly from the ground input to the ground output. Right?

Electrically, there is no difference between the two versions of wiring you've shown. They both electrically connect the same way.

Regarding the PCB, I would have placed and routed the PCB a differently, but yours should work.

Below is a simple diagram of what your trying to accomplish. The wiring configuation is really not a "daisy chain". The modules are connected in parallel so it is really a "bus" connection configuration.



Perhaps you have used a non-plated thru hole somewhere when should have used a plated thru hole?
Or, are you testing without certain components installed?
(The resistor value I've shown is an arbitrary value)

 

My guess is that the problem, resuming it is a single sided board, is an unintended gap in the trace somewhere between those 2 solder pads.

 

Hello MisterBill2, Bernard, eetech00 and AAC forum,

Thanks for the replies.

Must have stated the supposition unclearly.

But it has been figured out.

Figure A works by itself.
Figure B does not work.
Figure C works.

And eetech00's reply shows that characterizing
the connection of the two modules has a
a daisy chain is misleading. It is a simple parallel connection.

Thanks again.

Allen in Dallas

 

Here we have an issue as to the meaning of terms. Circuits are either series or parallel, daisy-chained means boards linked in sequence, and does not imply either series or parallel, but rather simply board to board instead of all fed from one set of terminals. And I stand by my assertion that the failure was caused by a gap in the trace at some point.Some times the gaps are invisible. I have located conductor gaps by application of enough voltage to make the gap arc over, The arc usually is easy to spot, Some times it does require decreased lighting to see the arc.

 

There may also be some additional confusion, as there are two sets of power terminals, but it's not specified which was used to connect power during configuration A testing. Were you applying power to left or right set of terminals? Did you test both? If your circuit works when power is applied to the left set but not the right set, then I would agree with MisterBIll2 and suspect the top gnd trace (bypassed by the wire in configuration C) to be at fault.

 

Here we have an issue as to the meaning of terms. Circuits are either series or parallel, daisy-chained means boards linked in sequence, and does not imply either series or parallel, but rather simply board to board instead of all fed from one set of

I disagree, but it doesn't matter. You can call it what you wish.

 

Hello MisterBill2, Bernard, eetech00 and AAC forum,

Thanks for the replies.

Must have stated the supposition unclearly.

But it has been figured out.

Figure A works by itself.
Figure B does not work.
Figure C works.

And eetech00's reply shows that characterizing
the connection of the two modules has a
a daisy chain is misleading. It is a simple parallel connection.

Thanks again.

Allen in Dallas

Your statements imply that there is an electrical difference between the different wiring schemes and the problem was solved with version "C". Perhaps there is a difference, but it only means something is wrong with the board layout, wiring, or, testing.
The only thing that seems certain is that if the schematic (post #5 which should be the same as yours) matches the board layout, and the PCB was fabricated correctly, then any of the wiring configurations should of worked. That said, I can't really say whether the PCB should be "refactored" or not.

 

The fact that some words have different meanings to some folks at some times can have very large effects on occasions. Sometimes those effects can be very important and sometimes very costly as well. Portions of my responsibilities required writing that could not be interpreted with more than one meaning. That requirement has affected my writing style and even my speech style, in that I tend to think carefully about conveying only the meaning that I intend. My intention was to point out that confusion of meaning can lead to a lack of understanding the information that is intended to be conveyed. I did not and do not intend to be critical, but only to avoid confusion.

 

Back-lighting a board and examining it with a magnifier may reveal very fine gaps in traces; I've encountered those on homebrewed and professional circuit boards.

 

Back-lighting a board and examining it with a magnifier may reveal very fine gaps in traces; I've encountered those on homebrewed and professional circuit boards.

I’ve had the opposite problem. A design that looked fine in software... A design that passed a visual inspection when received from the board house...

But there was a short from a via to a trace that didn’t belong. I had to grind the space between the two on 25 boards.