Dear all,
for us the situation is currently the following. We have defined the tests and limits in our group standard, which all suppliers need to fulfill (GS95002,
2013). As this is not a totally public document, I do not feel comfortable to share the document. However, the following graphic might give some insights. The BCI test with substitution is for us a very important test on ECU/board level.
Note that other car manufacturers might have other limits and that limits (also ours) can change. These are from 2013 and are currently used for all our Ethernet implementations.
Kind regards,
Kirsten
Von: Galin Ivanov [mailto:Galin.Ivanov@MICROCHIP.COM]
Gesendet: Montag, 14. Mai 2018 16:21
An: STDS-802-3-10SPE@xxxxxxxxxxxxxxxxx
Betreff: Re: [802.3_10SPE] Follow up on presentation from 11.Apr.2018 - DM Noise Level
Hello Mr. Cordaro,
I can follow your explanation – essentially it all comes to the right expectations/assumptions!
In that respect I must admit I’m unaware of an OEM requesting the mentioned 355 mA test level for operational mode (right test category!?) immunity testing and will consequently forward that question to all the OEMs participating:
What is the target level for EMI stress to design the 10SPE system against (considering testing categories/criteria)?
In order to get realistic estimation on noise levels we need to clearly define our starting point,
so feedback from OEM with specific levels is required at that stage!
In that respect we need answers to the following specific questions:
-
Is BCI method considered the preferred/worst-case EMI stress type? (e.g. vs. Stripline?)
-
Applicable BCI method: substitution(open-loop) or closed-loop?
-
Target test limits for in-band (<25 MHz) and out-of-band (>25 MHz) frequency:
o
Target test limit for operational mode / Test Category (i.e. full functionality maintained during EMI stress)
o
(optional) Target test limit for non-destructive mode / Test Category (i.e. functionality can be impaired during EMI stress, must resume after stress stops, no damage to hardware)
Also, related to your presentations, allow me to notice, that the assumption of CW type of stress may not necessarily be representative for the worst case!
Most OEMs require also AM modulated and some even Pulse modulated stress – autocorrelation function can yield worse results in specific cases in those scenarios!
Best regards,
Galin Ivanov
From: Jay Cordaro <jay.cordaro@xxxxxxxxxxxx>
Sent: Mittwoch, 9. Mai 2018 19:52
To: Galin Ivanov - M21104 <Galin.Ivanov@xxxxxxxxxxxxx>
Cc: STDS-802-3-10SPE@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_10SPE] Follow up on presentation from 11.Apr.2018 - DM Noise Level
Hello Mr. Ivanov,
To follow-up to your question on the call, With nodes floating 5cm above ground plane NodeA measured 105mV
at 16.67MHz. Mode conversion for harness was measured at -51dB at 16.67MHz, limit is -43dB. See page 9 of presentation. Scale measured CW noise from 105mVpk-pk by 8dB. That's where I got 263mVpk-pk. Assuming linearity, for 355mA
test level, scale 263*(355/200) = 467mVpk-pk. In band. Let's call it 500mVpk-pk or 178mVrms. Hope this helps.
regards,
Jay Cordaro
On Wed, May 9, 2018 at 3:51 AM, Galin Ivanov <Galin.Ivanov@xxxxxxxxxxxxx> wrote:
Dear Mr. Cordaro,
in your presentation, named “Follow-up to 10BASE-T1S Immunity Measurements” from April 11, 2018 you refer to differential noise levels measured – e.g. on pages 7 and 8.
The frequency ranges for those measurements and the evaluation of the results from those span to frequencies much greater than what is defined by the PSD masks found in the current draft of the specification.
Most notably - your worst case is derived at 33 MHz, which according to the above mentioned PSD distribution shall not carry any significant signal power and can be effectively suppressed when evaluating the RX signal quality!
If you re-evaluate your diagram with the PSD distribution of the target signal in mind, then you will see that, according to your measurement results, the significant maximum peak is found at 16 MHz and has not more than 110 mV, according to the diagram found on page 8.
In this relation, the table on page 10 list a value of 126 mV (!?!) as a base for the calculation. Can you please explain the discrepancy?
Even if we accept the table value to be correct, and take into account other possible sources of mode conversion and Xtalk, based on our current data, we don’t see a justified reason to anticipate more than 300 mVp-p of DM noise in the relevant signal band!
Best regards,
Galin Ivanov
Microchip Technology Germany II GmbH & Co. KG
Galin Ivanov
Principal Engineer, AIS Applications Engineering
Emmy-Noether-Straße 14, 76131 Karlsruhe| Germany
Tel: +49 (0)721 625 37 335 Fax: +49 (0)721 625 37 119
galin.ivanov@xxxxxxxxxxxxx | www.microchip.com
Firmensitz: Ismaning
Registergericht: Amtsgericht München HRA 98692
Haftender Gesellschafter: Microchip Technology Germany GmbH mit Sitz in IsmaningRegistergericht: München, HRB 197 653
Geschäftsführer: Mohammed Nawaz Sharif, James Eric Bjornholt, Ulrich Hallerberg
To unsubscribe from the STDS-802-3-10SPE list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-10SPE&A=1
To unsubscribe from the STDS-802-3-10SPE list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-10SPE&A=1
To unsubscribe from the STDS-802-3-10SPE list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-10SPE&A=1