2G Gauge Cluster PCB Blown

[spyderdrifter]

Anyone have thoughts on what would cause the PCB on the gauge cluster to blow out? I built my LED board and hooked it up the way I always have, and the flexible PCB smoked and burnt up in a couple spots. My LEDs still work, but the PCB is shot. No idea why, except maybe too many LEDs, but they use less power, so I'm a bit lost.

 

[TSiAWD666]

Did you include in-line resistors with your LEDs to provide current limiting?

 

[llxkevinxll]

Is it possible some traces were damaged during the Led install and caused a couple shorts?

 

[spyderdrifter]

Did you include in-line resistors with your LEDs to provide current limiting?

Every LED has its own resistor.

Is it possible some traces were damaged during the Led install and caused a couple shorts?

I think if there was a short in my board, it wouldn't work right.


My guess is that maybe I have too much of a power draw from the cluster. I have the board powered by soldering the wires to 2 spots on the back of the PCB shown here along with the burned out sections.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

 

[spyderdrifter]

Did you include in-line resistors with your LEDs to provide current limiting?

You can see what I built here to see the resistors and the rest of the wiring.

You must be logged in to view this image or video.

You must be logged in to view this image or video.

 

[DogWhistle]

I’ve seen damage like that from over-voltage, but It appears to be over-current this time. LEDs vary in their characteristics, even out of the same batch, so some of those LEDs could be out of spec, per your load calculations.

Also, the resistors, although limiting the current to each LED, still “draw“ current too.

You could take the time to use an amp meter to measure the actual draw, but in order to reproduce consistent success, maybe leave the LED pigtails out of the back of the cluster housing, and use a scotchlock tap to bite into the harness wires, leading to the cluster.

 

[spyderdrifter]

....maybe leave the LED pigtails out of the back of the cluster housing, and use a scotchlock tap to bite into the harness wires, leading to the cluster.

This is the only part that would suck. I've been wanting these to be a simple plug&play unit so when other people order them, they don't have to send me their cluster first, or do any hacking of the harness. I'm gonna have to research more things to make this right without blowing the PCB in the process. Tapping into the harness is good as a temp solution, I think, but to get it where it needs to be, I need to do more.

 

[steve]

How hard are you driving each LED?
Assuming each is 20ma and discounting the power lost in each resistor, you have 23 LEDs on the main board and that's almost half an Amp being drawn. Get your meter out and measure it in your test setup. Plus your drawing it from almost the end of that the flex circuit.

 

[spyderdrifter]

How hard are you driving each LED?
Assuming each is 20ma and discounting the power lost in each resistor, you have 23 LEDs on the main board and that's almost half an Amp being drawn. Get your meter out and measure it in your test setup. Plus your drawing it from almost the end of that the flex circuit.

What do you mean by "how hard am I driving the LEDs"?
I've never really done testing of my LED builds, they all just sorta worked. I know that's not the right way to go about things, and I'm trying to change that.
What would drawing power from the end do to things? The end of which circuit? (assuming you mean the oem flew circuit)

 

[steve]

What do you mean by "how hard am I driving the LEDs"?

Simple, how much current are you running through each LED? I don't see any voltage regulation in your design so with a simple dropping resistor the current through the LED is going to vary and the cars electrical systems voltage changes from 12.6V down to 9V while cranking and back up to 14.4V while the alternator is charging the battery.

Each LED type has a nominal forward voltage drop and a nominal operating current in it's specs and you use there values to compute the correct dropping resistor for a given system voltage in this basic circuit. What are your design parameters?

As an example; A basic 5mm red LED has a Forward Voltage of 1.8V to 2.2V and a recommended forward current of 20mA. Assuming worst case voltage of 15V and a 2V forward voltage you would use a 650ohm resistor. Under those conditions each LED Resistor combo would dissipate 0.3W and consume 20mA of current. (0.04W for the LED and 0.26W for the resistor)

If the voltage dropped to 12.6V the same resistor would only allow the LED to get 16mA of current and at 9V the LED would only get 10mA. This changes the brightness and helps with the current and power.

In a parallel circuit like this current and power are additive. 23 LED's and resistors are 23x the 20mA current and 23x the 0.3W of power. 0.46A and 6.9W with most of that power being dissipated by the resistors dropping the voltage. These are the expected values and before connecting the circuit to your cluster you should measure them with your multimeter to make sure you don't have anything out of spec.

What would drawing power from the end do to things? The end of which circuit? (assuming you mean the oem flew circuit)

Why does it matter where you connect? Each inch of the flex circuit has some resistance and the more inches you have (there and back) the more power is going to be turned into heat along it. In this case you exceeded the power handling of the flex circuit traces and they cooked.

I'm not sure that drawing 500ma should have smoked the traces but I don't see an indication of a short in your display but the damage sure looks like there was.

You can reduce the power usage by letting the LED forward voltage drop work for you. Combining 3 or 4 LEDs in series with a resister means it only has to drop the difference in voltage (15V - (3 x 2V)) / 0.02mA or (15V - (4 x 2V)) / 0.02mA (Assuming 24 LEDs, 8 or 6 strings of 3 or 4 LEDs) Each string only draws 20mA and the resistors are dropping much less voltage so their power dissipation goes down (8 x (0.12W + (9V x 0.020A)) = 2.4W or 6 x ((0.16W + (7V x 0.020A)) = 1.8W) # strings x ( LED Power + Resistor Power)

If none of this makes sense I'll blame the lack of coffee this morning and old age.

 

[spyderdrifter]

Simple

From this, it just means I need to go back and relearn all of this. Been years since I dealt with all that so it's all been forgotten. You might as well be speaking Russian at this point. Thanks for the info.