Reply by sankar murmu July 10, 20042004-07-10

I have some suggestion, before you call that there is problem in dsp chip.
Check the following points
1. ADC clock is not more than 5MHz.If it is more, you can get some wrong data from ADC.
2. if you are using SDK, you don't get the correct 12bit value, the 12bit value is scaled to 16bit value by SDK.
3. Check ADC reference supply, it must be free from any kind of switching noise or ripple.
4. if it is your own target board, check the ADC input pin interface respects the design guide lines given by Motorola.

Thanks,
Sankar Murmu.

On Tue, 06 Jul 2004 Bende Georg wrote :
>Yes, I have the same effect. The voltage is about 0,15V and I get values around 32000 in the ADC register. Seems this isn't the bevhaviour mentioned in the errata.
>Did anyone find a workaround for this issue?
>What annoys me that it seeps to happen only to some of the units.
>
>Georg Bende
>Softwareentwickler
>Abteilung Elektronik
>Dr. Fritz Faulhaber GmbH & Co KG
>Daimlerstr. 23
>71101 Schaich
>Tel: +49 7031 638294
>
>      -----Ursprgliche Nachricht-----
>      Von: Douglas Holub [mailto:d...@comcast.net]
>      Gesendet: Dienstag, 6. Juli 2004 16:38
>      An: Bende Georg
>      Betreff: Re: [motoroladsp] 56F807 DSP failures >      Herr Georg,
>
>      It is a mystery to me. Errata item 4.6 says that measurements < 24 mV may return a value of 0.  Errata 5.6 reccomends which clocks to use for better accuracy.  I had an A/D input on my board which was > Vdd.  It caused some of the A/D measurements to return 32760 all of the time.
>
>      Douglas Holub
>
>           ----- Original Message -----
>           From: Bende Georg <mailto:g...@faulhaber.de>
>           To: m...@yahoogroups.com
>           Sent: Tuesday, July 06, 2004 5:05 AM
>           Subject: [motoroladsp] 56F807 DSP failures
>
>           Hello all,
>
>           Recently we have started some test production, and on about 1% of the units claimed to be defective by our test software. I took a closer look on these units and I have noticed a huge offset on 1 or 2 of the ADC channels. At 0V level it showed the same value as the other -good- ones at 1V, while the whole measuring range is calibrated to 3V. Did anyone meet such a symptom before? Is it a typical DSP fault or some kind of design flaw? I have checked all voltages corresponding to the analog inputs such as Vdda and Vssa and their values are OK.
>
>           Any suggestions and ideas are welcome,
>           Georg Bende
>           Softwareentwickler
>           Abteilung Elektronik
>           Dr. Fritz Faulhaber GmbH & Co KG
>           Daimlerstr. 23
>           71101 Schaich
>           Tel: +49 7031 638294 >           _____________________________________
>           Note: If you do a simple "reply" with your email client, only the author of this message will receive your answer.  You need to do a "reply all" if you want your answer to be distributed to the entire group.
>
>           _____________________________________
>           About this discussion group:
>
>           To Join:  m...@yahoogroups.com
>
>           To Post:  m...@yahoogroups.com
>
>           To Leave: m...@yahoogroups.com
>
>           Archives: http://www.yahoogroups.com/group/motoroladsp
>
>           More Groups: http://www.dsprelated.com/groups.php3 >click here <http://companion.yahoo.com>" target="_blank" rel="nofollow">http://us.ard.yahoo.com/SIG9rpa8ku/M)5196.4901138.6071305.3001176/D=groups/S05771855:HM/EXP89204457/A!28215/R=0/SIGse96mf6/*http://companion.yahoo.com>
>  <http://us.adserver.yahoo.com/l?M)5196.4901138.6071305.3001176/D=groups/S=:HM/A!28215/rand4001248 >  _____
>
>           Yahoo! Groups Links >           *     To <http://docs.yahoo.com/info/terms/> .



Reply by Bende Georg July 6, 20042004-07-06
Nachricht
Hello Howard,
 
We use the same supply voltage for the tests; there is normally some variation in the uncalibrated ADC readouts, up to 2-3 LSB I can handle it. But this seems like a "stuck" bit - I can't explain it any other way. Did you ever have such a problem?
Another question if you have high-volume product experiences: what were the most frequent errors regarding to the DSP? There was a discussion whether we should make a flash/RAM R/W test, does it make any sense? Our product is has to be tested as thoroughly as possible, but I would rather abandon the senseless tests....
 
Georg Bende
Softwareentwickler
Abteilung Elektronik
Dr. Fritz Faulhaber GmbH & Co KG
Daimlerstr. 23
71101 Schaich
Tel: +49 7031 638294
-----Ursprgliche Nachricht-----
Von: Ebersman, Howard [mailto:H...@microwavedata.com]
Gesendet: Dienstag, 6. Juli 2004 16:50
An: Bende Georg; m...@yahoogroups.com
Betreff: RE: [motoroladsp] 56F807 DSP failures

We have used the 807 in several high volume products and have not noticed any A/D problems relating to the processor.
 
We did have some minor issue on our part due to boards being calibrated with a different power supply than used in the field (and off by a tenth of a volt or so) , resulting in slight calibration error until we realized the problem.
 
Howard
-----Original Message-----
From: Bende Georg [mailto:g...@faulhaber.de]
Sent: Tuesday, July 06, 2004 6:05 AM
To: m...@yahoogroups.com
Subject: [motoroladsp] 56F807 DSP failures

Hello all,

Recently we have started some test production, and on about 1% of the units claimed to be defective by our test software. I took a closer look on these units and I have noticed a huge offset on 1 or 2 of the ADC channels. At 0V level it showed the same value as the other -good- ones at 1V, while the whole measuring range is calibrated to 3V. Did anyone meet such a symptom before? Is it a typical DSP fault or some kind of design flaw? I have checked all voltages corresponding to the analog inputs such as Vdda and Vssa and their values are OK.

Any suggestions and ideas are welcome,
Georg Bende
Softwareentwickler
Abteilung Elektronik
Dr. Fritz Faulhaber GmbH & Co KG
Daimlerstr. 23
71101 Schaich
Tel: +49 7031 638294


_____________________________________
Note: If you do a simple "reply" with your email client, only the author of this message will receive your answer.  You need to do a "reply all" if you want your answer to be distributed to the entire group.

_____________________________________
About this discussion group:

To Join:  m...@yahoogroups.com

To Post:  m...@yahoogroups.com

To Leave: m...@yahoogroups.com

Archives: http://www.yahoogroups.com/group/motoroladsp

More Groups: http://www.dsprelated.com/groups.php3


Reply by Bende Georg July 6, 20042004-07-06
Nachricht
Yes, I have the same effect. The voltage is about 0,15V and I get values around 32000 in the ADC register. Seems this isn't the bevhaviour mentioned in the errata.
Did anyone find a workaround for this issue?
What annoys me that it seeps to happen only to some of the units.
 
Georg Bende
Softwareentwickler
Abteilung Elektronik
Dr. Fritz Faulhaber GmbH & Co KG
Daimlerstr. 23
71101 Schaich
Tel: +49 7031 638294
-----Ursprgliche Nachricht-----
Von: Douglas Holub [mailto:d...@comcast.net]
Gesendet: Dienstag, 6. Juli 2004 16:38
An: Bende Georg
Betreff: Re: [motoroladsp] 56F807 DSP failures

Herr Georg,
 
It is a mystery to me. Errata item 4.6 says that measurements < 24 mV may return a value of 0.  Errata 5.6 reccomends which clocks to use for better accuracy.  I had an A/D input on my board which was > Vdd.  It caused some of the A/D measurements to return 32760 all of the time.
 
Douglas Holub
----- Original Message -----
From: Bende Georg
To: m...@yahoogroups.com
Sent: Tuesday, July 06, 2004 5:05 AM
Subject: [motoroladsp] 56F807 DSP failures

Hello all,

Recently we have started some test production, and on about 1% of the units claimed to be defective by our test software. I took a closer look on these units and I have noticed a huge offset on 1 or 2 of the ADC channels. At 0V level it showed the same value as the other -good- ones at 1V, while the whole measuring range is calibrated to 3V. Did anyone meet such a symptom before? Is it a typical DSP fault or some kind of design flaw? I have checked all voltages corresponding to the analog inputs such as Vdda and Vssa and their values are OK.

Any suggestions and ideas are welcome,
Georg Bende
Softwareentwickler
Abteilung Elektronik
Dr. Fritz Faulhaber GmbH & Co KG
Daimlerstr. 23
71101 Schaich
Tel: +49 7031 638294


_____________________________________
Note: If you do a simple "reply" with your email client, only the author of this message will receive your answer.  You need to do a "reply all" if you want your answer to be distributed to the entire group.

_____________________________________
About this discussion group:

To Join:  m...@yahoogroups.com

To Post:  m...@yahoogroups.com

To Leave: m...@yahoogroups.com

Archives: http://www.yahoogroups.com/group/motoroladsp

More Groups: http://www.dsprelated.com/groups.php3