Ulrich,

If it stops after a while you have it made. Set the trigger to normal and the 
scope will record the last transaction...

Well, I am sure it is more complicated than that, but my toughest I2C problem 
was a serial EEPROM that I was polling (for read) continuously several times 
per second. That device would have ONE memory location (always the same) 
corrupted every 3 months or thereabout of continuous operation. 

It turned out to be a non-initialized variable in the i2c driver in my code 
that once in a great while would recycle to a particular value that caused the 
read to become a write...

I have an I2C analyzer that was useful to confirm the problem, but it was not 
helpful in identifying the problem. It is nicely made and not very expensive. 
It has a USB port so you can dump the data to a PC for off-line analysis.

http://www.mcc-us.com/SMB-SW-DS.htm

Shortly after finding and fixing the problem, we decided to run Lint over the 
code. Lint would have spotted the uninitialized variable problem immediately, 
it did find a couple of other places where problems were just lurking...

Didier KO4BB

---- Ulrich Bangert <[EMAIL PROTECTED]> wrote: 
> Gents,
> 
> I know some of you are hardware hackers like me too. In the 25 years of
> electronic development I have always refused to work with logic
> analysers and always have claimed "Give me a fast scope and some hours
> and I will do the job". However, these days I had VERY strange problems
> with an I2C bus based device. 
> 
> Communication on the bus would work flawlessly over hours and then stop
> due to... yes, due to what? That clearly is an situation where the "fast
> scope" mentality not applies. On the search for something affordable I
> came over this:
> 
> http://www.pctestinstruments.com/
> 
> I still have the device not here and still do not know the reason of my
> problems but the technicians at Intronix said to me "That is exactly
> what we have built it for" when I asked them very specific if their
> device could be helpful in this situation. It it were true this thing
> were worth its price in gold.
> 
> Best regards 
> 
> Ulrich Bangert
> www.ulrich-bangert.de
> Ortholzer Weg 1
> 27243 Gross Ippener 
> 


_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Reply via email to