wc32 output problem

Hi guys,
              I use a wc32 board to turn on 12vdc relay #rtb14012f.
No i don't use wc32 direct output, i use a little IC named uln2803a (NPN darlington transistor). So everything work well about 15 minutes, then nothing.
 
 
I discovered that my outputs do not give me 5vdc amymore. Some output like op9, only have 2.5vdc and op10 to op13 only have 600mvdc. op14 to op16 have nothing, .... 0vdc. 
           
 
Can you tell me what ive done wrong? 
 
 
My design looks like this : http://www.panu.it/ardupower/schema.png
 
I use the most recent FW.
I only use op9 to op16.
 
 
regards,

okay, I think the Octal Bus Transceiver (LJ245A) is burned. From side B which is 3.3v, when I turn on the Output to 1, i get 3.3v, but at  the port A which is 5vdc i read 400mvdc.
 
VCCA Is 4.97vdc as well.
 
 
So the question is, why my setup burns lj245a circuit? Need i place an opto isolator between uln2803 and wc32? Or at least a diode to make sure there is no voltage peak can reach the wc32 board .?
 
 
 
thank you for your support.

If the output is short to ground, also OPx is turned on, that is like forcing 5V supply short to ground through the buffer IC, it will burn the buffer IC out.  That buffer IC is there to protect the CPU pin not being burn out by shorting to ground.

not sure, lj245a maybe not broken?

You probably needs to put a current limitation resistor in front of ULN2803, since that input can sink much larger current.

okay what value do you suggest me?

ULN2803, 2003 etc already have an internal 2K7 (typically) base resistor to the leading transistor.
Specifications indicate a maximum input current of 1.35mA @ 3.85V in (and typically only 0.93mA)
 
No way it will be drawing more current than the buffer can safely deliver all day.
 
Is there any chance that you've had a short-circuit or wiring problem - even if it is no longer present - that could have done the damage?
 
I have used (and have had in constant operation, 24/7) relay boards using the ULN2803 and ULN2003 drivers for several years without problem.
The only benefit adding a resistor to the WC output would have, is if it's right at the WC terminals - to minimise damage if the output were pulled to ground or +5V, by limiting the current. A 1K resistor should limit your output to 5mA (sink or source) - if it gets shorted to ground (while the output is high) or shorted to +5V (while the output is low).
Again though - the resistor is only to save the board incase there's some OTHER event that would have blown it up. It isn't required for the normal, safe operation of a UNL2003/2803 device.

[sharedmedia=gallery:images:1005]
 
 
Maybe here is the problem:
 
 
i used an external 5v bush button to turn on relays but there is no diodes to protect wc32 from incoming 5vdc. Am I right?

If you have the WC32IO board (which I suspect you may, since outputs 9-16 don't have screw terminals on the bare WC32), then outputs 13, 14, 15 and 16 will have been ok because there are open-drain FETs (but you'd need pull-up resistors on the ULN2803 inputs).
 
However outputs 9, 10, 11 and 12 are straight through - so if the WC32 output was low (off) and you operated the switch/pushbutton to turn the relay on, yes you will have destroyed the output driver.
 
There won't be any inductive spike back through the ULN2803, but the switches - as drawn - will certainly destroy the output chip.
Sorry to be the bearer of bad news

Haha no problem rossw ,I just removed destoyed parts from the w32, i will order new one soon and replace it.
 
No I don't use the wc32IO board. I use a board that i designed for special application. Anyway, do you think adding diode could be a good choice?
 
:blush: Switch button could turn on relay directly instead of turning on uln2803..... Maybe this is the best way to do, less expesive, less routes and more secure. I can turn on relays even if wc32 is dead, even if uln2803 is burned...
 
 
thank you CAI and rossw for quick support.

BTCAD said:

do you think adding diode could be a good choice?

Click to expand...

I assume you want the switch to force the output on, regardless of the output state of the WC.
That is an "OR" function, rather then messing with diodes why not just use a two input OR gate
or depending on logic levels a NOR? A 14 pin package gives you 4 gates, enough for 4 outputs.

For several of my projects wanted a little more complex switch override: Force ON, Force OFF, Automatic WC control.
Doing that was just a matter of adding a few more gates.

If you have 4 spare WC inputs you can accomplish what you want with firmware.

/tom

Not a bad idea Tom, but if i connect the switch to the ground and second leg to the coil, i'll get the same result. No additional device etc.