**From:** Mary (*[email protected]*)

**Date:** Fri May 12 2000 - 22:14:00 PDT

**Next message:**Brian Seol: "[SI-LIST] : Differential Clock Signal Pair"**Previous message:**ARiazi: "[SI-LIST] : Model Evaluation"**In reply to:**[email protected]: "Re: [SI-LIST] : Trace Impedance Selection"**Next in thread:**Larry Miller: "Re: [SI-LIST] : Trace Impedance Selection"

Mike,

As Neven pointed out and you confirmed, maintaining a constant Zo while

reducing both the trace width and height will result in lower radiation.

I don't understand why you would say that this,

"...seems to conflict with the correlation of radiation to trace (or

structure) impedance, but only if one doesn't understand the whole picture."

I think that Neven described the whole picture very concisely and the

conflict is real. Neven's example is one of many that could be used to

show that there is no general relationship between characteristic impedance

and radiation.

You cited an analogy between a horn antenna and an impedance transformer.

However this analogy is generally applied to describe the transformation

of the wave impedance (ratio of E to H) from a waveguide to free space.

It cannot be applied to ratios of V to I in any intuitive or consistent

manner that I am aware of.

This is my last post on this thread. Interest and information content have

been declining and I get the feeling that there are only a few of us who are

not automatically deleting mail with this subject.

It's been fun,

Mary

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

From: [email protected]

[mailto:[email protected]]On Behalf Of [email protected]

Sent: Thursday, May 11, 2000 8:40 PM

To: [email protected]

Subject: Re: [SI-LIST] : Trace Impedance Selection

To Vinu, [email protected], and others on this thread:

Many of us have (in this thread and related ones) provided only limited

answers to questions because it takes substantial time to generate an

all-encompassing dissertation. However, we may spend MORE time by not being

complete in the first place (I'm often guilty of this).

Vinu asked, "Keeping traces close to the ref. planes", "reducing loop area",

are these not the same as saying lowering the characteristic impedance of

the

line reduces

EMI?"

Someone else seemed to think raising the Zo to lower the trace current would

also reduce emissions. Still someone else expected the higher currents of

lower impedance lines would increase emissions.

The short answer to Vinu is YES. I'll use a prior example I submitted in a

comment to Mary on this same subject to help clarify the relationships of

different parameters, and to address the other speculations.

Surface traces (microstrip) create a loop antenna whose plane is vertical.

The radiation from that loop is linearly proportional to the effective area

and the magnitude of the frequency components of the current flowing through

the trace. The field concentrating effect of any dielectric with Er > 1

makes the trace appear electrically closer to the reference plane, which is

equivalent to making the effective area of the loop smaller (which will

lower

radiation) and will simultaneously lower the trace impedance. The trace

impedance varies in a logarithmic fashion with the height, but the effective

area varies linearly with the height. Now, for a constant magnitude of

signal voltage applied to a line of 69 ohms (which I used in my example), we

will obtain a specific level of radiation. While maintaining the same trace

width, if we increase the trace height (dielectric thickness) by a factor of

four, we get a trace impedance of 119 ohms. We have not quite doubled the

Zo, but we have increased the area of the loop by four times. For the same

applied signal voltage, we reduce the trace current by less than two to one

(which by itself would also reduce the radiation proportionately). However,

the quadrupled loop area (by itself) would increase the radiation by four to

one. The net result is an increased level of radiated emissions for a

higher

trace impedance, even though the trace current is reduced. This result is

easily verified by lab tests.

In some/many cases, a controlled impedance line must be created while still

keeping radiation low. Trace width could be reduced while thinning the

dielectric to maintain the desired Zo. In this case, the area and the

radiated emissions will vary directly with the height (dielectric

thickness).

The lower limit on dielectric thickness would be dictated by the

manufacturing limits for the ever thinning trace width, or vice versa,

whichever hit a practical limit first.

Although not all radiating structures are as simple as the loop

configuration

noted above, they all must either directly conduct current or allow

displacement current to flow, AND (for more efficient radiation... the

undesired result) they must accommodate the transition from relatively low

onboard impedances to the impedance of free space (120*Pi). One gentleman

who had RF experience correctly noted that antennas are more like

transformers (I would quote him, but I have deleted most of the prior

correspondence). Larry Smith also recalled the "near exponential" horn

coaxial cable demonstration which is just another technique for increasing

the impedance of the driving source to better match the intrinsic impedance

of space.

Part of the rationale presented above is why, in many of my earlier

comments,

I keep "simplistically" comparing the impedance match of the transmitting

structure (of whatever physical configuration) to 377 ohms. The radiated

energy can be envisioned to be the energy transfer from a low impedance

transmission line to the "load" impedance of 120*Pi. The higher the line

impedance, the more efficient the transfer (into radiated emissions). This

is a mental aid to help make real-time design decisions only, but it is

helpful to me. The noted exception above of the thinning trace with

thinning

dielectric seems to conflict with the correlation of radiation to trace (or

structure) impedance, but only if one doesn't understand the whole picture.

For the practical, high-density manufacturing case of 5-mil traces with

5-mil

spaces, the dielectric thickness is the only parameter easily manipulated.

Although not too extensive a rationale, I hope the above clears the

confusion

a bit.

Mike

Michael L. Conn

Owner/Principal Consultant

Mikon Consulting

*** Serving Your Needs with Technical Excellence ***

**** To unsubscribe from si-list or si-list-digest: send e-mail to

[email protected] In the BODY of message put: UNSUBSCRIBE

si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.

si-list archives are accessible at http://www.qsl.net/wb6tpu

****

**** To unsubscribe from si-list or si-list-digest: send e-mail to

[email protected] In the BODY of message put: UNSUBSCRIBE

si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.

si-list archives are accessible at http://www.qsl.net/wb6tpu

****

**Next message:**Brian Seol: "[SI-LIST] : Differential Clock Signal Pair"**Previous message:**ARiazi: "[SI-LIST] : Model Evaluation"**In reply to:**[email protected]: "Re: [SI-LIST] : Trace Impedance Selection"**Next in thread:**Larry Miller: "Re: [SI-LIST] : Trace Impedance Selection"

*
This archive was generated by hypermail 2b29
: Wed Nov 22 2000 - 10:50:19 PST
*