Unexpected problem moving from ST10F167 to ST10F168

[caprioec]

Posted on February 28, 2005 at 19:07

Unexpected problem moving from ST10F167 to ST10F168

[caprioec]

Posted on February 24, 2005 at 18:39

Hi all, I am new to this forum, and I need urgent help!!

I have a couple of products originally developed using ST10F167.

Since it it no longer available, I have purchased the F168 after carefull reading of the ST document about the differences, since apparently no issue was relevant for my applications.

On one board I just plugged the F168, downloaded with the new Flasher the same executable code, and everything worked fine.

But the second application is a nightmare: I use all the 16 AD inputs on Port 5 to check signals to be within preset limits, and I get continuosly random errors, stating the signals are out of the limits, while an high-bandwidth scope is confirming that everything is OK!

The same board, with F167, is OK!

The only issue about AD in the compatibility document from ST is about a concurrent injection and normal conversion, and this is not my case.

I have used the same executable code as F167: may be needed to recompile it? there is a need to change the RAM memory allocation? I am using the CAN module, but not the XRAM.

What gets me mad is also that I cannot find a USER manual for F168, Ok I have the datasheet and the programming guide, but the really needed document is the user manual: where can I get it?

Anybody can give me a valauble suggestion?

Thanks a lot in advance!

[caprioec]

Posted on February 25, 2005 at 10:28

Thanks McGyver,

I understand that I have to refer to the ST10R167 as if it was an F168, OK, not obvious but the important thing is to know it.

About my AD problem, of course I either do not understand the problem!

Apart the joke, my problem is that it looks like that with the F168 the analog values read seem grossly wrong. The conversion is completed, the program is running correctly, but the values are not correct.

It is not an accuracy issue, the error is too big, in some cases 50% of the full range!

As I told, placing an F167 in the same board gives proper results.

Is there a need to modify the settings of the ADC?

I have used the same Hex code as for the F167: should I recompile it?

Thanks again

[caprioec]

Posted on February 25, 2005 at 18:44

After many trials, may be I have found a solution, but the reason is still unexplained.

I have discovered that increasing the series resistors to the ADC inputs from 1K to 10K greatly alleviates (and probably solves) the problem.

However I cannot reference this to any change in the device specifications, since the input capacitance maximum value is the same.

Unless the actual input capacitance is much lower, so making everything more sensitive to noise.

Can this be an explanation? anybody else has experienced similar problems?

[najoua]

Posted on February 28, 2005 at 05:31

Hi,

Refer to the attached application note. It will help you to find an explanation to your problem. It exaplains all sources of ADC errors and how minimyzing them (by putting suitable analog source impedance, etc...)

I hope this helps you.

Najoua

________________

Attachments :

8597.pdf : https://st--c.eu10.content.force.com/sfc/dist/version/download/?oid=00Db0000000YtG6&ids=0680X000006I0JL&d=%2Fa%2F0X0000000bXS%2FySdM6x6NRbSkU9fJM629ZwL1zGput.h6nwQCF6OGEEg&asPdf=false

[caprioec]

Posted on February 28, 2005 at 19:07

Thanks Najoua,

that application note is useful indeed.

However there is no explanation about the big difference between an ST10F167 and an ST10F168.

The only reason could be a much higher sensitivity to noise and spikes, but there is no warning on this aspect in the datasheet.