[SI-LIST] Q on TDR coupon measurements

Hi folks-
 

I'll add a #5 item to Scott's excellent list below on possible artifacts
that affect the measured impedance you see on the front screen of a TDR:

 

5. The long, rising tail of the incident edge of the signal going into the
device under test. We implicitly imagine that the incident signal going into
the TDR is really an ideal step edge, maybe with a finite rise time, but
after the rise time, we think to ourselves, the top of the waveform is flat.
Unfortunately, due to the losses in the cabling and some artifacts in the
TDR sources, the waveform entering the DUT will most of the time, have a
long sloping up tail, of as much as 10% in some cases.

 

Never forget that what we are really looking at on the front screen of the
TDR is a scaled version of the reflected incident signal. If the incident
signal has some variation in it, a perfect, uniform, lossless line will have
a TDR response with the variation of the incident signal, scaled by the
reflection coefficient. 

 

If we are always looking into a nearly 50 Ohm line, the reflection
coefficient is a very small value, and we will only see a small variation in
the TDR response. If it is a low impedance, we will see the TDR response
typically, looking like it is dropping ( the TDR response is the negative,
scaled version of the incident signal). If the impedance of the line is
high, a narrow line, for example, we see the TDR response sloping up, since
this is the scaled version of the incident signal.

 

I think this is often the artifact present when some folks say they are
seeing the series resistance of their narrow line, or the conductive loss in
the dielectric of their low impedance line. While, it is possible to see
these effects under the right cases, I think the non-ideal shape of the
incident signal often masks these intrinsic effects.

 

As Scott points out, there are only two ways of knowing what is real and
what might be an artifact: using Iconnect or another modeling tool, measure
the incident signal with a good open or short (good shorts are often easier
to get than good opens) and using this as the source, use inverse scattering
to build a model for the DUT to see if it behaves like a losses line or has
other effects, or use the reference source to peel the impedance, or,
secondly, use the normalization feature in some TDRs to normalize out the
effect of the cables from the TDR to the DUT. This effectively measures the
distortion in the incident signal, and corrects for it to make it look like
an ideal Gaussian. 

 

For those interested, these, and other TDR techniques are illustrated in the
online lectures associated with the Master Class Workshop MCW620 that are
available on our web site, www.BeTheSignal.com <http://www.bethesignal.com/>
. 

 

--eric

 

**************************************

Dr. Eric Bogatin, President

Bogatin Enterprises, LLC

Signal Integrity Online Training

26235 w 110th terr

Olathe, KS 66061

v: 913-393-1305

f: 913-393-0929

c:913-424-4333

e:eric@xxxxxxxxxxxxxxx

 <http://www.bethesignal.com/> www.BeTheSignal.com

****************************************

From: Scott McMorrow [mailto:scott@xxxxxxxxxxxxx]=20
<mailto:scott@xxxxxxxxxxxxx%5d=20> 

Sent: Wednesday, March 22, 2006 12:50 PM

To: Grasso, Charles

Cc: si-list@xxxxxxxxxxxxx

Subject: Re: [SI-LIST] Q on TDR coupon measurements

 

Charles,

 

I'll follow on to what Ray and Lee have already said:

 

There are several "things" that color a TDR measurement of a test

coupon:

 

a.) The launch ringing that occurs at a poorly designed test contact=20
area, or RF connector transition.  This can be inductive, capacitive,=20 and
in many cases is actually a resonant structure that rings at high=20
frequency.  This can be eliminated or reduced by better test=20
fixture/connector design, lower edge rate on the TDR, and, if the=20 ringing
is symmetric, can be averaged out.  Unfortunately, launch=20 ringing is
generally never symmetric, and as such will color the TDR=20 measurement of
a trace by shifting the reflected impedance following the

 

launch structure.  This is where software like Tektronix iConnect earn=20
their money.  Using reverse scattering methods, the impedance=20
discontinuity of the launch can be theoretically removed.

 

2) The impedance "wiggle" seen in a TDR measurement, which is due to=20
effects such as laminate weave Er variations, laminate press thickness=20
variations, and trace etch variations.  This can generally be averaged out.

 

3) DC resistive losses cause a slope in the TDR measurement.  This can=20 be
partially removed by computing the slope and using linear=20 extrapolation
back to the launch point.  However, even in a perfectly=20 matched coupon
launch with no ringing, and with no variation in=20 impedance due to
materials or manufacturing, this extrapolation will=20 always indicate a
higher than actual impedance.  Why?  Because there is=20 also a concurrent
actual change in impedance due to internal inductance,

 

which is related to skin depth.

 

4) AC resistive and dielectric losses cause an additional coloration of=20
the rise time of the pulse itself, which together determine the ultimate

 

resolution of impedance in any TDR measurement, without some sort of=20
normalization and reverse scattering process.

 

Bottom line is that the closest measurement of the actual impedance of a

 

trace is made with an exceptional launch, that has no ringing, near to=20
the launch point.  If you compute the rise time degradation of the=20
reflected pulse, you can use that to compute the upper frequency=20
resolution of your measurement.

 

As for "average mean", who knows?

 

 

Scott

 

 

 

Scott McMorrow

Teraspeed Consulting Group LLC

121 North River Drive

Narragansett, RI 02882

(401) 284-1827 Business

(401) 284-1840 Fax

 

http://www.teraspeed.com <http://www.teraspeed.com/> 

 

Teraspeed(r) is the registered service mark of Teraspeed Consulting Group
LLC

 

 

 

Grasso, Charles wrote:

> Hi all,

> =20

> 

> In researching the "state-of-the-art" for TDR measurement I ran across 

> an interesting measurement that I would welcome comments on from the 

> group.

> 

> After the coupon has been probed, the impedance is only measured over

a

> small portion (<25%) of the resultant waveform and then using an 

> "average mean* technique for reporting the impedance values.

> 

> =20

> 

> In the opinion of the group - is this a valid technique? What are the 

> reasons for only using a small portion of the waveform ? Does anyone 

> know what an *average mean* value is??

> 

> =20

> 

> Thanks in advance!!

> 

> Best Regards

> Charles Grasso

> Senior Compliance Engineer

> Echostar Communications Corp.

> Tel: 303-706-5467

> Fax: 303-799-6222

> Cell: 303-204-2974

> Pager/Short Message: 3032042974@xxxxxxxx

> Email: charles.grasso@xxxxxxxxxxxx;

> Email Alternate: chasgrasso@xxxxxxxx

 



------------------------------------------------------------------
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:
http://www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List FAQ wiki page is located at:
                http://si-list.org/wiki/wiki.pl?Si-List_FAQ

List technical documents are available at:
                http://www.si-list.org

List archives are viewable at:     
                http://www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu

Other related posts: