[SI-LIST] TDR and line losses
- From: "Eric Bogatin" <eric@xxxxxxxxxxxx>
- To: "Si-List" <si-list@xxxxxxxxxxxxx>
- Date: Wed, 26 Nov 2003 08:18:02 -0600
While the explanations below for increasing slope of the TDR trace due
to series resistive losses are perfectly correct, there is another,
subtle explanation, which is more commonly the case.
When we look at the front screen of the TDR instrument and interpret
the results, we are assuming that the signal going into the
transmission line under test is an ideal step increase in voltage. In
fact, the skin depth losses in the cables from the front of the TDR to
the device under test will cause the signal going into the device to
have a long rising tail.
When you look at the reflected signal from the T line, you see the
initial voltage reflected back to be slightly lower than expected,
creeping up as time proceeds. This produces an artifact of seemingly
increasing impedance down the trace.
One interpretation is resistive losses in the line. One interpretation
is actual increasing characteristic impedance down the line due to
geometry variation. Another explanation is non ideal rise time
waveform.
The most efficient way of distinguishing these effects is to simulate
the measured response using the initial, measured incident waveform as
the stimulus. TDA Systems IConnect software is a perfect tool that
facilities this analysis. You record the measured step wave with the
cable end disconnected from the source, then record the TDR response.
If you have a simple interconnect you are measuring, you might be able
to model it as a uniform transmission line. If you are able to get
excellent agreement between the measured TDR response and the behavior
expected based on the ideal T line and the real waveform from the TDR,
you know you are seeing an artifact of the TDR.
If you still see the creeping up, try using a lossy line model in
IConnect. This will help identify the problem as a real lossy line
effect.
In my experience, I clearly see lossy line effects in the shape of the
transmitted signal, but I find the creeping up of the reflected signal
is more often due to the non ideal incident step.
--eric
********************************************************************
Recently published by Prentice Hall, www.phptr.com
Signal Integrity-Simplified, by Eric Bogatin
Attend the GTL Signal Integrity University in Sunnyvale, CA - Nov
6-13
GTL 122 - Fundamental Principles of Signal Integrity
GTL 250 - High Speed Board Design
GTL 260 - Interconnect Models from Measurement
----------------------------------------------------------------------
---------------
Dr. Eric Bogatin
CTO, GigaTest Labs
26235 w 110th Terr
Olathe, KS 66061
v: 913-393-1305, f: 913-393-1306
e: eric@xxxxxxxxxxxx
www.GigaTest.com
********************************************************************
Msg: #6 in digest
Date: Mon, 24 Nov 2003 09:34:16 -0800
From: "Kyung Suk (Dan) Oh" <doh@xxxxxxxxxx>
Subject: [SI-LIST] Re: TDR and line losses
Hi, I would like to add one comment to this issue.
The conductor loss definitely contributes to this upward creep but
there is also an additional physics which contributes to this upward
creep and this one is often forgotten and I would like to clarify
them.
The initial impedance level should be corresponding to the lossless
characteristic impedance. After initial impedance level there are
two mechanisms which make the impedance profile to creep upward.
The first one is resistive loss as others pointed out and
the second one is the internal inductance which increases
the characteristic impedance at low frequencies.
It is important to first understand that the upward creep is NOT due
to the reflected wave along the transmission line but it is the
reflected wave of the initial edge at the beginning of the
transmission line.
Mathematically, it is the convolution between the input edge
and the characteristic impedance only and not related with
the propagation constant.
Physically, this reflected wave does not contain any reflection
along the line (assuming it is uniform) until the reflection from
the other end comes back.
At the very beginning, the input edge actually sees the
characteristic impedance at the very high frequency which is
the impedance based on L over C, say Zc_inf.
And the later response sees the characteristic impedance
at lower frequencies. At these lower frequencies the characteristic
impedance is larger than Zc_inf due to the resistance term AND
the internal inductance term.
As you make the line longer, you would see the increasing in the
impedance profile which can be mistakenly thought as due to the
increase in the loss. As this creeping is not due to the "loss"
mechanism along the transmission line, but it is due to the change
in the characteristic impedance due to loss; hence, it is not
depending on the line length.
If you increase the line length to fairly large this creep will
eventually saturate to the characteristic impedance at dc which
would be finite if there is any dc conductance loss. Otherwise it
will continue to grow as the characteristic impedance becomes infinite
at dc without dc conductance.
In reality, the characteristic impedance measurement shows a finite
value at low impedance so the upward creep should be saturate beyond
a certain length.
"The bottom line is that if your characteristic impedance varies
significantly from dc to high frequency, the upward creep will be
there (assuming the impedance changes from high to low as the
frequency increases)"
I have attached the simulated TDR response using Hspice w/
the following three characteristic impedances to demonstrate
the impact of the internal
inductance:
case 1: sqrt(L/C)
case 2: sqrt((Ro+Rs*sqrt(f)+jwL)/(jwC))
case 3: sqrt((Ro+Rs*sqrt(f)(1+j)+jwL/(jwC))
Regards,
-Kyung Suk (Dan) Oh
Dima Smolyansky wrote:
> Suresh,
>
> The upward slope of the TDR trace is indicative of losses. However,
the
> losses will need to be quite substantial for the upward "creep" to
be
> clearly visible. In other words; your transmission trace (TDT) will
show
> even fairly small losses through rise time amplitude degradation;
however,
> when you begin to see the "creep" in the reflection (TDR), that will
show up
> as large rise time and amplitude degradation in TDT.
>
> Also, Howard Johnson did an article once, where he played with skin
effect
> and dielectric loss, and showed how they affect different portion of
the TDT
> waveform. You can do the same in IConnect's lossy line model by
varying the
> skin effect and dielectric loss parameters independently, and
evaluating
> their effect on the TDT (or TDR) waveform.
>
> Thanks,
>
> -Dima
------------------------------------------------------------------
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 technical documents are available at:
http://www.si-list.org
List archives are viewable at:
//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:
