I2C communication

In article <MPWdnRMAEcTw7U3WnZ2dnUVZ_jWdnZ2d@giganews.com>, 
nospam@nowhere.com says...
> 
> 
> Andrew Smallshaw wrote:
> 
> > On 2010-04-21, Vladimir Vassilevsky <nospam@nowhere.com> wrote:
> > 
> >>Done quite complicated systems with I2C shared in multitask OS. Never 
> >>had a need to analyse the bus. There is no point to do that: if you have 
> >>control of the master, you know what is going on.
> > 
> > 
> > Which is fine until you have a multi master bus,
> 
> That great attention paid to multi master I2C thing always surprised me.
> Had anyone ever seen a system with multiple masters on I2C in reality?
> 
> > or you are not
> > entirely sure your master is working as intended - e.g. it is being
> > bit-banged.
> 
> If one can't implement such a trivial thing like I2C master bit banging 
> so to be entirely confident in it, he is not fit for embedded.
Yeah, you never had something NOT work as expected.
YOU never had to learn how to do anything, it just comes to you.
Blah, blah, blah.
We get it, you are the Master and the rest of us are not worthy, now 
shut the fuck up.  Didn't your mother ever tell you that if you don't 
have anything nice to say, don't say anything at all.

> 
> > Debugging and analysis tools are completely pointless
> > if you know exactly what is going on - you have no need even for
> > a voltmeter if you are able to compute accurately what the voltage
> > at a given point will be in a given situation.  However it often
> > helps to be able to measure things since your assumptions about a
> > system are not always correct.  Sure, I don't have an I2C analyser
> > here and I've always managed to get by without one, but that doesn't
> > mean I seen times when one would have been useful.
> 
> I verify the code with LEDs, scope and control printouts; never had any 
> need for fancy debug tools. Best debug is not making techical mistakes 
> at first time.
See above, oh great one....

I love to see you puffed up peacocks strutting for the sake of doing it.  
What are you 12 years old?

On 4/22/2010 11:51 AM, Vladimir Vassilevsky wrote:
>
>
> Andrew Smallshaw wrote:
>
>> On 2010-04-21, Vladimir Vassilevsky <nospam@nowhere.com> wrote:
>>
>>> Done quite complicated systems with I2C shared in multitask OS. Never
>>> had a need to analyse the bus. There is no point to do that: if you
>>> have control of the master, you know what is going on.
>>
>>
>> Which is fine until you have a multi master bus,
>
> That great attention paid to multi master I2C thing always surprised me.
> Had anyone ever seen a system with multiple masters on I2C in reality?
Yes SMBus Batteries are MultiMaster.
>
>> or you are not
>> entirely sure your master is working as intended - e.g. it is being
>> bit-banged.
>
> If one can't implement such a trivial thing like I2C master bit banging
> so to be entirely confident in it, he is not fit for embedded.
Try Coding the master and slaves at the same time.
There is only so much you can see on a DSO.
>
>> Debugging and analysis tools are completely pointless
>> if you know exactly what is going on - you have no need even for
>> a voltmeter if you are able to compute accurately what the voltage
>> at a given point will be in a given situation. However it often
>> helps to be able to measure things since your assumptions about a
>> system are not always correct. Sure, I don't have an I2C analyser
>> here and I've always managed to get by without one, but that doesn't
>> mean I seen times when one would have been useful.
>
> I verify the code with LEDs, scope and control printouts; never had any
> need for fancy debug tools. Best debug is not making techical mistakes
> at first time.
Normal I agree. But I had the monitor, ( from a former job where there 
was finger pointing) I used it, it saved me time.
>
> Vladimir Vassilevsky
> DSP and Mixed Signal Design Consultant
> http://www.abvolt.com