PCF8575- the chip died

[sh@sha]

Hi,
I'm using PCF8575 from Texas Instruments for I2C function. After several time of testing, the chip died. It still happened even after I replaced the chip. Anyone having the same experience? Any help would be greatly appreciated.
Thanks in advance.

 

[betwixt]

You will have to give more information, "the chip died" doesn't tell us anything about possible failure mechanisms. Please at least post a schematic.

Brian.

 

[sh@sha]

This is my connection diagram.
VCC= 5V.
R=1k ohm.
C=1uF

 

[betwixt]

I can't see anything wrong with the schematic. Is it possible something you have connected to the I/O pins is causing the damage?

Brian.

 

[FvM]

Loading each output with 1K pull-up causes up to 80 mA total output current, which is already near to the maximum rating of 100 mA. Still safe, though.

 

[IchoTolot]

If R17 and R18 is 1k too, then the current into SDA / SCL is to high.

Kind reagrds
Holger

 

[betwixt]

??? SCL is an input, it only draws CMOS leakage current. SDA has to draw enough current to pull the line down to low level, 1K should be fine, fast systems often use 470 Ohms or less.

Brian.

 

[IchoTolot]

Don't know if the chip is capable of mastermode. If not you're right with scl if yes SCl is OC too.
SDA is OC and regarding to I2C 3mA is max. The datasheet mentions this too.

 

[FvM]

SDA is OC and regarding to I2C 3mA is max. The datasheet mentions this too.

No. Allowed current is 20 mA. 3 mA is maximum value to guarantee 0.4 V Vol under worst case conditions.

As said, 1k I2C bus pull-up is pretty standard and can't cause damage of I2C devices.

 

[sankar.m8]

How do you think that the chip is died?..check your power VCC impedance / check for any burned mark..
What are all the IOs connected in S1 to S16 ?

 

[IchoTolot]

No. Allowed current is 20 mA. 3 mA is maximum value to guarantee 0.4 V Vol under worst case conditions.

You're right I've overlooked the notes.

As said, 1k I2C bus pull-up is pretty standard and can't cause damage of I2C devices.

Maybe it wouldn't cause damage on this chip. The standard permits only 3 mA and as you mentioned 0.4V wouldn't be guranteed at 1mA. So for a reliable function of the the I2C bus it is not wise to use 1k at 5V.
But the chip seems to die due to some other reasons.