Harvester keypad. No lights

[budm]

Re: Harvester keypad. No lights

1) What Voltage do you get on the Anode of the LED's (good board and bad board) with ref to GND?
2) What RESISTANCE do you get when test the LED's on the good board and bad board?

[edugimeno]

Re: Harvester keypad. No lights

I observed something that might be a hint to what''s going on.
When I used the multimeter on diode test mode and applied it into the individual LED pins in the working board, I would get the proper RED-GREEN-BLUE colors, one by one, full lit, in only that LED

When I tried this on the failing board, I would get either a mix of colors, really dim, or no light at all, OR, SOMETIMES, I would get ANOTHER LED dimly lighting up in some color.
LEDs are individually addressed each one on different pins (OUTx) from the LED Driver chip. How would current to one led make ANOTHER LED light up? I can only think of internal shotcuts in the LED driver IC, so when I apply voltage to one LED, it gets back into the IC thru OUTx, internal shortcuts connect it to OUTy, which goes out and drives that other LED.
Does it make sense? Either that or internal shortcut inside the PCB layers...

What do you think?

Thanks!

[stj]

Re: Harvester keypad. No lights

But the leds are mostly plastic, won't hot air melt the plastic before the solder?

i meant the driver chips

[sam_sam_sam]

Re: Harvester keypad. No lights

You heat the board around the device use a small tip with enough air flow to keep the heating element not to burn out

[edugimeno]

Re: Harvester keypad. No lights

hot air needed - those chips have a center pad under them.

But the leds are mostly plastic, won't hot air melt the plastic before the solder?

[edugimeno]

Re: Harvester keypad. No lights

Look into getting a tweeter soldering iron they cost a little bit more than a regular soldering but for you are doing it should work for you

Here is one version of what I am talking about

https://www.ebay.com/itm/938D-220-11...gAAOSwdNRc5EOp

Thanks but this seems suitable for simple SMD components (2 leads), but how can this help for a 6 pin LED or a 28 pin IC?

I've sometimes had access to hot air stations but right now all I have is a normal soldering iron...

[will62]

Re: Harvester keypad. No lights

When I apply low voltage with multimeter in Diode test mode, on good board I get full R / G / B light on each pin, but same test on bad board, I barely get any light (very dim) and sometimes it seems colors get mixed even when Im carefully with leads.

Could be a bad diode, but also the shortcircuit could be in the feeding driver IC (TLC)

If you want to rule out the TLC, check for continuity between the cathodes on an led and the out puts on U1 and U2. Once you find out which output pin(s)
are driving that led, use a soldering iron to lift the pin off the pad. Then recheck led.

[CapLeaker]

Re: Harvester keypad. No lights

Those TLC5943 drivers can be had for less than a buck on aliexpress. I would get a couple of these and put those in first.

[stj]

Re: Harvester keypad. No lights

hot air needed - those chips have a center pad under them.

[sam_sam_sam]

Re: Harvester keypad. No lights

Look into getting a tweeter soldering iron they cost a little bit more than a regular soldering but for you are doing it should work for you

Here is one version of what I am talking about

https://www.ebay.com/itm/938D-220-11...gAAOSwdNRc5EOp

[edugimeno]

Re: Harvester keypad. No lights

I also measured SIN and SOUT on both IC's, both boards, giving same pattern (digital tiny flicker around 3.1v).

I checked the LEDS to determine their internal wiring. Found that 3 corner leads are common anode and 3 remaining are cathodes for R / G / B. When I apply low voltage with multimeter in Diode test mode, on good board I get full R / G / B light on each pin, but same test on bad board, I barely get any light (very dim) and sometimes it seems colors get mixed even when Im carefully with leads.

Could be a bad diode, but also the shortcircuit could be in the feeding driver IC (TLC)

At this point Im stuck between bad diodes or bad driver. I guess I should remove one LED and test it isolated, but Im afraid Im going to fry the board as I don't have any good and thin solder tip neither any special tool to desolder 6 pins at a time

Any other hint on how to diagnose?

Thanks!

[edugimeno]

Re: Harvester keypad. No lights

Ok, I found a proper way to connect both boards without bending the inter-board socket, and some early conclussions, all about the TLC LED driver chip:
Vcc for the chip is 3.3V
BLANK is LOW (mV)
Iref measures 1.2V. Attached resistor measures 3K3. Same as markings on SMD chip
Xerr gives out pulsed Lo-Hi, at the same rate as the leds are blinking in a working board to indicate that CAN bus is not detected

All this, so far, gives same values on both working and not working board. Working board has this Xerr pulses while it flashes so this is not the issue.

Then I measured LED outputs from the drivers (OUT0-OUT15) on both working and failing boards. This differed... Values were aproximately the same, if OUTx was HI on one board, it was HI on other board, if it was pulsing on one board, it was pulsing on other board. BUT, some of the outputs had the same voltage on working/non working board, while some other outputs had like 0.2 - 0.5v less in non working board.

Could it be that the LEDS individually failed? I don't see any black spot inside them, I see visually exactly like on the working board

What else could I test? Thanks again!

[R_J]

Re: Harvester keypad. No lights

What is the source for the led voltage? and is it there? If the source voltage is steady, then the fault will be in the sink side which is the ic. All functions like brightness etc. are data controlled, It could be that the ic's are failing but it seems odd that all drive would be shut down and not just one or two of the outputs.

[edugimeno]

Re: Harvester keypad. No lights

I just talked to my contact (who called me to try to repair these). He said that some customers reported that sometimes SOME lights fail at the beginning, and then it ends up failing all at the same time. I tried to clarify if suddenly HALF of the lights fail, and then as a second stage ALL the lights fail, which would indicate U1 fails, then later U2 fails...or maybe the lights start failing one by one, in which case this could be due to an overcurrent on the leds so they just blow, could this be?
He didn't have a precise answer, he would try to ask a customer tomorrow, as he hasn't seen them in the process of failing by himself.
Also he told me that when they are about to fail he can recognize the situation by a really dim flicker in some of the lights, really really dim, that you can only see at dark. does this help at all?

Thanks!!!

[edugimeno]

Re: Harvester keypad. No lights

As RJ mentioned above, it's unlikely that both drivers (U1,U2) failed at the same time. However, if they are daisy-chained, a problem on U1 could take U2 out. Check for continuity between U1-pin24 (sout) and U2-pin5(sin) to be sure.

This is something I will be checking either tonight or tomorrow, good point

Also, you mentioned that the led backlights can be turned on for night use. I assume that the board gets a signal when you turn the headlights on. Need to see if one of the pins into B board is changing when you turn on the headlights. You'll need to replicate that for out of circuit testing.

No sorry, I was misunderstood here. What I mean is that each key has 2 LEDs behind, one usually light in white, and is used to easily locate the function of the specific key. Some harversters have up to three of these keypad blocks, so 27 functions to turn on/off, so it's hard to remember what does each button do. So this is why each button is illuminated so the function icon can be seen. This light is permanent, no way to turn it on or off. Then, the other LED is used to show that the function is engaged or disengaged

Lastly, I doubt that you get any info, but have you tried contacting the manufacturer and asking for either a schematic or test point data? (EiA board 105225 pcb 307452s). Might be worth a try:

Grammer EiA Electronics: https://www.grammer.com/en/contact-us/

Not a bad point at all, it's worth to try, Im sending it in a minute, thanks!

[will62]

Re: Harvester keypad. No lights

As RJ mentioned above, it's unlikely that both drivers (U1,U2) failed at the same time. However, if they are daisy-chained, a problem on U1 could take U2 out. Check for continuity between U1-pin24 (sout) and U2-pin5(sin) to be sure.

Also, you mentioned that the led backlights can be turned on for night use. I assume that the board gets a signal when you turn the headlights on. Need to see if one of the pins into B board is changing when you turn on the headlights. You'll need to replicate that for out of circuit testing.

Lastly, I doubt that you get any info, but have you tried contacting the manufacturer and asking for either a schematic or test point data? (EiA board 105225 pcb 307452s). Might be worth a try:

Grammer EiA Electronics: https://www.grammer.com/en/contact-us/

[TechGeek]

Re: Harvester keypad. No lights

Sorry, yes, it was a type. Your guess was right, thanks!
(I can't find a button to edit original message!!)

You can edit your message for up to 1 hour from the time it's posted.

[edugimeno]

Re: Harvester keypad. No lights

I think that was a type-o. Here's a TLC5943.

Sorry, yes, it was a type. Your guess was right, thanks!
(I can't find a button to edit original message!!)

[edugimeno]

Re: Harvester keypad. No lights

Most of the test points are barely accesible when both boards are in place, connected. I need to connect them with jumper wires to be able to fully access them all, the TP's I checked yesterday gave all the same values in both working and failing sets
I will fully check more points today as soon as I get time to connect boards with jumper cables
Thanks everyone

[R_J]

Re: Harvester keypad. No lights

I missed that about the led's in post #1, I guess I would check if there is supply voltage for the led's, It seems the tlc5943 sinks the led's so there should be full voltage on the output pins when the led is not on. If that voltage is there, the ic is not sinking and I doubt both ic's would be bad so it could be a data problem. (corrupt eeprom?)
I see there are different 6pin 5050 led configurations, some have 3 pins common and each other lead controls a color. Others have VCC, Clock and Data pins like the WS2812