I2C clock speed 40 kHz valid?

[DaQue]

I'm working on reverse engineering talking to a very old chip (2000) and when I measure valid command sent to it in a working circuit the clock pulses are 25 us leading edge o leading edge. The 1st clock is about twice as wide as the rest. Does this seem odd to anyone else. I guess it's still valid, right?

 

[tonton81]

you can use a lower frequency to talk to it, even if its 20khz bus speed of mcu

 

[DaQue]

It just seems awefully slow. Data sheet says it supports 400k. I guess the old micro in pager could just be bit banging it.

 

[tonton81]

doesnt hurt to start low, once you see it working you can go higher
lower rates tend to be more stable speciallly over longer distances