Microcontroller vcc and gnd short

[Harish Achar]

Hi,

I am using a NXP family microcontroller LPC2148 from past 3 years for my product.
But in recent batch we got too many microcontroller rejections.

Reason for that is vcc and gnd pins of the microcontroller shorted after several days of testing.
I thought this might be because of 3.3V supply regulator LM1117. But there is no problem with the regulator.

I have a doubt on microcontrollers. How do I test each and every microcontroller before assembly and without loading the program.

Please suggest.

 

[KlausST]

Hi,

A short beween VCC and GND may be caused by ESD.
An ESD error may cause immedaite fail, but it may cause a delayed fail - even after months.

How can you test it?
I´d say it is very difficult to test - especially because there may be delayed fail.
ESD usually causes a break in the isolation barrier.
Thus there may be increased leakage currents.
--> Test every pin of the IC for increased current against VCC and GND.

But be aware, that ESD problems may be caused by the testing procedure.

Consider to ask the microcontroller manufacturer or any specialized test house to analyze the die to find out the reason for the short circuit.

Review all the way of the IC for ESD problems:
* delivery
* goods receipt inspection
* IC handling
* electrical test of IC
* baking
* assembling
* PCB handling
* programming of microcontroller
* test of PCB
* ...

check
* people
* soldering stations
* tables
* ESD mats
* storage / stock
* packaging
* test equippment
* floor
* wrist bands
* humidity
....

Klaus

 

[dick_freebird]

Test methods can ruin parts, for sure. Especially beware
"force current, measure voltage" as even with compliance
limits you can break down an "off" transistor with ease.
Seen it done to one of my parts, one of the tristate leakage
tests would blow the DUT up every so often.

You should be able to determine where the "first wave"
of failures tends to be discovered, and work backward
along the chain of custody from there.

 

[Harish Achar]

Hi KlausST,

Thank you for the suggestion. I agree with you. There are some other factors I wants to discuss with you.

I did check the LM1117 output which gives constant voltage 3.3V to microcontroller. This 3.3v is only connected to microcontroller section not to any other ICs.

But whenever I plug the output load then I got 0.5V spike on regulator output terminal.
So total voltage is 3.8v for 90us and after it comes to normal 3.3V. So I plug the load several times and got this pulse sometime. This pulse is not regular.

Does above said spike cause the microcontroller to malfunction?

And

Load I am using is an inductive coil and coil section is isolated from the microcontroller. But coil uses same 12V supply which I connected to LM1117 input terminal.
And 12v drops to 11v sometime when I connect the load. Same time I get 0.5v spike.

What do you suggest?


- - - Updated - - -


Hi Freebird,

So I tested micrcontroller vcc pin resistor w.r.t. gnd and the resistance values vary from 8.91M to 9.6M and all of the microcontroller doesn't have the same value.

Does it tell something? Please suggest.

 

[KlausST]

Hi,

Does above said spike cause the microcontroller to malfunction?

Read the datasheet.
In "Absolute maximum rating" it says: VDD must not be higher than 3.6V.
This means: every (short time) Vdd voltage higher than 3.6V may cause immediate damage to the IC.

(My personal opinion: It is likely that the micrcontroller survives 3.8V for 90us, in detail it also depends on current and energy)

What do you suggest?

we need more detailed information:
* schematic (especially LM1117 and supply of microcontroller)
* but also possible that (at dowerup/powerdown) capacitive or inductive circuitry causes damage at other pins than VCC and GND. Thus the complete schematic could be helpful.
* PCB layout
* points on the PCB where you measured the spikes
* waveform of the spike (scope picture)
* test conditions
...

Klaus

Added:

Load I am using is an inductive coil and coil section is isolated from the microcontroller. But coil uses same 12V supply which I connected to LM1117 input terminal.
And 12v drops to 11v sometime when I connect the load. Same time I get 0.5v spike.

Inductive and capacitive loads, wrong signal routing, ringing....
...all needs to be reviewed.

 

[Easyrider83]

1117 has very low maximum voltage. Much lower than 7805