FreeLists / si-list / [SI-LIST] Agilent TDR Normalization Questions

["Eric Bogatin" <eric@xxxxxxxxxxxxxxx>]

Chris, and others-
 

I might be able to shed some light on your normalization question in regards
to the Agilent 86100 DCA with TDR plug in. (For those just starting out with
the TDR, I am preparing a series of lectures as part of a Master Class
Workshop on Mastering the 1 port TDR. I've posted one of them already on my
web site at www.BeTheSignal.com. I hope to have three more posted in the
next month.)

 

In the normalization routine used on the Agilent 86100 DCA with TDR plug in,
you, the user, selects the new rise time you want the signal to be
normalized to. It sounds like you have 65 psec set as the default value of
the new rise time. You can access the normalization screen by using the
quick button on the left side of the screen. 

 

After the setting reference plane cal, you will see the "rise time" box near
the middle of the screen on the right hand side, where you can use the
numeric keypad or the large knob to change the rise time to other values. I
find that going to rise times longer than 35 psec, the normalization routine
is very stable and is a great way of emulating how your interconnect will
behave with different rise times. The normalized response with the new rise
time becomes a "response", so to view this trace, you will need to display
the response signal, which can also be accessed with the quick button on the
left of the screen.

 

I find that to really see the normalized response clearly, its more
convenient to turn off channel 1 with the yellow button at the bottom of the
screen, and just display the normalized response.

 

Of course, when discontinuities are electrically short, the peak value of
reflected signals will depend on the rise time of the incident signal and
you will see some discontinuities may not pose a problem with the rise time
of your drivers. Keep in mind, normalization is valuable to get around the
losses and rise time degradation effects present in the cabling and
connectors up to your DUT. It does not get around the masking problems of
multiple discontinuities in your DUT. You can still only interpret the
reflected voltages on the front screen as an impedance profile, only to
first order, past the first discontinuity. To build a more accurate model,
with multiple discontinuities, you need to post process the measurements
with a tool like IConnect.

 

Dima will probably chime in about his software, though he may be very busy
with his new parent company, so I'll comment on the IConnect solution as
well. 

 

You can perform a similar function as normalization with IConnect, and you
can do a few other operations that are equally useful. Their impedance
peeling algorithm will get around both the rise time degradation of the
cables and connectors, and the masking problems of the multiple impedance
discontinuities of the DUT. The impedance peeling method is a great way of
looking at the true profile of the DUT.

 

However, the rise time of the profile will be the rise time of the TDR.
While you can use the normalized signal from the TDR as the reference and
the waveform reflected from the DUT to see the impedance profile with
different rise times, you can also change the apparent rise time with their
built in filtered response. 

 

After you have an impedance profile, or just the reflected voltage signal,
you can apply their filter response, from the right side of the screen, and
see the profile with a new apparent rise time. This provides roughly the
same function as changing the rise time of the TDR with normalization. Just
keep in mind that the filter value is a filter, not a rise time. If the
filter time is long compared to the 35 psec of the TDR, then the resulting
rise time is close to the filter rise time. Otherwise, the resulting rise
time is better approximated by the square root of the sum of the squares of
the 35 psec and the filter time constant.

 

While changing the rise time is a great way of emulating your interconnect
with the TDR for your system rise time, if you already are using IConnect, a
more effective process is to use the software to build a model that works to
the bandwidth of your TDR, and use this model in your system simulation. TDA
will probably also recommend just using a measurement based behavioral model
for your system simulation. While this may work really well to enable you to
take a measurement of your interconnect and evaluate its performance in your
system, if it doesn't work, the behavioral model doesn't give you any
insight into how to optimize your interconnect. This is where a circuit
topology based model will provide more insight.

 

Hope this helps.

 

--eric

 

 

 

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

Dr. Eric Bogatin

www.BeTheSignal.com

Signal Integrity on-demand training

26235 w 110th terr

Olathe, KS 66061

v: 913-393-1305

f: 913-393-0929

c:913-424-4333

e:eric@xxxxxxxxxxxxxxx

 

Signal Integrity-Simplified

Prentice Hall, 2004

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

 

-----Original Message-----

From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On

Behalf Of Chris Padilla (cpad)

Sent: Wednesday, June 29, 2005 2:03 PM

To: si-list@xxxxxxxxxxxxx

Subject: [SI-LIST] Agilent TDR Normalization Questions

 

I've read the Help files included with my 86100B w/ 54754A TDR module and I

follow what happens when one Normalizes to the end of cable and/or test

fixture so I understand that.  There is an IFT and FIR filters being used

and then FFT'ing back.

 

I did a little check, however, and have some seemingly strange results that

I hope someone here can explain.

 

I find that as long as I am using decent high-quality cables, the

Normalization routine will degrade (increase) the ristime of the step from

~30 ps (unnormalized step) to ~65 ps:  a factor of ~2+.  I tried this for

several different cable I had laying around and they are all about the same

as far as Normalization goes.  Expectedly, the normalized step is cleaned up

of some higher-frequency content wiggles at the top of the step.

 

Next, I decided to Normalize to the end of the TDR module or put another

way, I didn't add any cable or test fixtures to the TDR module...I just

screwed my Cal Standards right to the 3.5 mm SMA on the TDR module.

Guess what?  I got the same ~2x increase of risetime as if I had some cable

and/or test fixtures attached!

 

Anyone have an explanation for this?  Seems to me that as long as I use

those "decent high-quality cables", all the Normalization adjustments happen

INSIDE the TDR module!?  What am I missing here...this doesn't seem right.

 

 

My next questions regards even using Normalization in the first place.

I often see a lot of TDR plots (with TDA Systems logo on them! ;o) ) that

don't appear to have been normalized.  It is probably for good reason since

Normalization appears to chop the bandwidth of the scope right in half.  I

have been learning to use I-Connect and I'm confused whether or not I should

be using the Normalization routine on my scope.

 

 

Thanks,

 

Chris Padilla

Cisco Systems

San Jose, CA

 



------------------------------------------------------------------
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