To further Jeff's post, I quickly put together a slide comparing the delta-L
subtractive vs actual de-embeded method. The S-Parameter data was obtained from
Heidi Barnes's video, " How to Use Fixture De-embedding to Match Signal
Integrity Simulations to Measurements" based on the CMP-28 platform from Wild
River Technology. Simulations were done with Keysight ADS. You can get the
slide here:
https://dl.dropboxusercontent.com/u/89802638/De-embed_vs_Delta_IL.pdf
The slide shows the 2" and 8" traces measured. The 2" S-parameter was split
into L-1" and R-1" and were used to de-embed from the 8" trace. The plot shows
the loss per inch at 20GHz for all three cases. Comparing the de-embedded vs
delta-L curves you can see excellent match to 20GHz then ripples start to
appear on the IL_Delta curve afterwards due to the ripples of the 2"
[dB(S(6,5))]curve.
Regards,
Bert Simonovich
Signal Integrity and Backplane Specialist
LAMSIM Enterprises Inc.
“Innovative Signal Integrity and Backplane Solutions”
Email: lsimonovich_at_lamsimenterprises.com
Phone: +1 613 836 7569
Cell: +1 613 795 5895
Web Site: http://lamsimenterprises.com/
Confidentiality Notice: This message and any attachments may contain
information that is privileged and confidential. If you have reason to believe
that you are not the intended recipient or a person responsible for delivering
this message to an intended recipient, you are hereby notified that any review,
dissemination, distribution, or copying of this message is strictly prohibited.
If you have reason to believe that you are not the intended recipient or a
person responsible for delivering this message to an intended recipient, please
contact the sender immediately by reply email and destroy all copies of the
original message.
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of Loyer, Jeff
Sent: April-15-15 12:09 PM
To: hzyltgg@xxxxxxxxx
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Why are there difference when calculating loss per unit
length?
As others have said, it sounds like you are experiencing an issue with your
launch and possibly vias, and you'll need to de-embed these effects. The good
news is that, in my experience (and others'), sophisticated de-embedding (such
as AFR, TRL, etc.) isn't necessary if all you need is SDD21. Using simple
"subtractive de-embedding", AKA "delta-L" (trace_loss_per_inch = [8inch_SDD21 -
4inch_SDD21]/4) might give as good a result as using a more complex algorithm.
When I built some test boards with intentionally flawed launches (noisy vias,
with stubs or which were very long) and compared the various de-embedding
techniques, the delta-L results were as good as other more complex de-embedding
techniques (though it's best to have as much difference in length between the
two DUTs). You'll find more documentation about the delta-L method if you
search for "Intel delta-L". Note that this won't give you any valid SDD11
results, which the other techniques might.
I've heard that, for some cases, the noise in the difference can overwhelm the
simple subtraction, but don't know exactly when that occurs. But, I believe
it will be obvious from the waveforms, it's a pretty basic SNR issue.
Ok - last post from here. Will continue discussion from other e-mail address,
if necessary.
Cheers,
Jeff Loyer
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of ??
Sent: Wednesday, April 15, 2015 12:24 AM
To: Alexander Ippich
Cc: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Why are there difference when calculating loss per unit
length?
Hi Alex,
Got it.
Thanks a lot
Alexander Ippich
<alexander.ippich@xxxxxxxxxxxxxxx>于2015å¹´4月15日星期三写é“:
The influence of your signal launch is fully in your measurement results.By doing a SOLT calibration at the tip of your microprobes, you have moved the
is prominent enough that you see the dB/inch reading of the 4" traces to be
And as you have seen in your results already, the influence of the
launch
----------------------------------------------------------------------------------------------------------------------------------------
The only way to clean up the data is to do the de-embedding.
Best regards,
alex
</mail/u/0/s/?view=att&thcbbf24a80491e6&attid=0.1&disp=emb&realattid™33111bd0e90102_0.1&zw&atsh=1>
Alexander Ippich
Senior Signal Integrity Engineer
OEM Marketing Europe
Tel: +49 7457 / 605 70 79
Cell.: +49 170 / 63 68 571
e-mail: alexander.ippich@xxxxxxxxxxxxxxx
web: www.isola-group.com
length?
Von: 黄涛 [mailto:hzyltgg@xxxxxxxxx]
Gesendet: April 15, 2015 09:03 AM
An: Alexander Ippich
Cc: kevin.voegele@xxxxxxxxxxxxxxxx; si-list@xxxxxxxxxxxxx
Betreff: Re: Why are there difference when calculating loss per unit
influnce of signal injection(pad,contact ,via etc. ) to shorter trace is
Hi Alexander,
Thanks,very clear explanation.
We use micro-probe and SOLT calibration to probe tip. That means the
of
Alexander Ippich
<alexander.ippich@xxxxxxxxxxxxxxx>于2015å¹´4月15日星期三写é“ï¼
š
As Kevin pointed our already, if you calibrate your VNA at the
connectors of your measurement cables and do not remove the influence
of the connectors on your PCB and the via / launch structure, you
will measure the following for the short (4") trace:
S21_4" = loss of connector / launch at port 1 + loss of 4" trace +
loss
ofconnector / launch at port 2
For the long (8" trace), you are going to measure:
S21_8" = loss of connector / launch at port 1 + loss of 8" trace +
loss
constant.connector / launch at port 2
You can see immediately, that S21 for the 8" trace is less than twice
the
S21 for the 4" trace, even though the insertion loss per inch is
--------------------------------------------------------------------------
If you measure much longer traces (so the loss of the trace is
significantly higher than the influence of the launch), you will see
that you get the results that you expect.
Another way would be to use TRL calibration of your VNA, which tries
to remove the launch from the results (but as this method relies on
having identical launches for all structures, this has also it
downsides) or post-process your data to do a mathematical de-embedding.
Best regards,
alex
Im--------------------------------------------------------------
Alexander Ippich
Senior Signal Integrity Engineer
OEM Marketing Europe
Tel: +49 7457 / 605 70 79
Cell.: +49 170 / 63 68 571
e-mail: alexander.ippich@xxxxxxxxxxxxxxx
web: www.isola-group.com
-----Ursprüngliche Nachricht-----
Von: si-list-bounce@xxxxxxxxxxxxx
[mailto:si-list-bounce@xxxxxxxxxxxxx]
unitAuftrag von Kevin Voegele
Gesendet: April 15, 2015 03:51 AM
An: hzyltgg@xxxxxxxxx; si-list@xxxxxxxxxxxxx
Betreff: [SI-LIST] Re: Why are there difference when calculating loss
per unit length?
Connectors are a larger percentage of the trace at short distances.
Have you de-embedded cables too?
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx
[mailto:si-list-bounce@xxxxxxxxxxxxx]
On Behalf Of ??
Sent: Tuesday, April 14, 2015 8:41 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Why are there difference when calculating loss per
calculatdlength?
Hi experts,
Sorry for messy code of this question previously .
We characterize insertion loss by xx.xx dB/inch usually.But after we
measured total loss of trace,the loss per unit length calculated by
different length are difference .For example ,one 4inchs trace
measured and calculated loss A dB/inch .and one 8inchs teace measured
and
littleloss B dB/inch while other parameters of design are same.The A is a
greater than B alwalys.
My question is:why are there difference ?from measurements?design ?or
any others ?
T hanks
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu