Re: [SI-LIST] : FPC impedance control

Andy Burkhardt (mail@polar.co.uk)
Wed, 21 Apr 1999 18:22:11 +0100

At 04:34 PM 16/03/99 +0000, you wrote:
>From: johnlin@ccmail.arima.com.tw
>Received: from earth.sun.com (earth.EBay.Sun.COM [129.150.69.3])
> by engmail1.Eng.Sun.COM (8.8.8+Sun/8.8.8) with ESMTP id RAA12626
> for <si-list@silab.eng.sun.com>; Mon, 15 Mar 1999 17:24:07 -0800 (PST)
>Received: from ms17.hinet.net (ms17.hinet.net [168.95.4.17])
> by earth.sun.com (8.9.1/8.9.1) with ESMTP id RAA19022
> for <si-list@silab.eng.sun.com>; Mon, 15 Mar 1999 17:24:08 -0800 (PST)
>Received: from ccmail.arima.com.tw (ccmail.arima.com.tw [203.74.188.253])
> by ms17.hinet.net (8.8.8/8.8.8) with SMTP id JAA21973
> for <si-list@silab.eng.sun.com>; Tue, 16 Mar 1999 09:24:07 +0800 (CST)
>Received: from ccMail by ccmail.arima.com.tw (ccMail Link to SMTP R8.20.00.25)
> id AA921605105; Tue, 16 Mar 1999 09:25:08 +0800
>Message-Id: <9903169216.AA921605105@ccmail.arima.com.tw>
>Date: Tue, 16 Mar 1999 09:23:14 +0800
>To: <si-list@silab.eng.sun.com>
>Subject: [SI-LIST] : FPC impedance control
>MIME-Version: 1.0
>Content-Type: text/plain; charset=US-ASCII
>Content-Transfer-Encoding: 7bit
>Content-Description: "cc:Mail Note Part"
>Sender: owner-si-list@silab.eng.sun.com
>Precedence: bulk
>Status:
>
>Dear all SI gurus,
>
>One question.
>
>Recently, I design a stackup structure for a FPC, flexible printed circuit
>board, to get right controlled impedance.
>The FPC is an embedded microstrip structure with a thin silver epoxy layer as
>the ground layer and 20cm trace length.
>
>Then I measure the trace impedance of the prototype of the cable from one
>end of the trace with TDR.
>
>I find that its impedance smoothly rises up from 50 to 70 ohms.
>However, measuring from the other end of the same trace, I find that the
>impedance curve looks flat ,around 60 ohms.
>(The FPC cable has a U turn at its tail).
>
>Why I got two different results by measuring the two ends of the same trace?
>What causes the impedance ramp up?
>
>Any comments on this phenomenon?
>
>Thank you for your helps in advance.
>
>John Lin
>CAE Engineer @ Arima
>
>**** To unsubscribe from si-list: send e-mail to
>majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE
>si-list, for more help, put HELP. si-list archives are accessible at
>http://www.qsl.net/wb6tpu/si-list ****

Dear John,

Sorry for the late response. (I had a great vacation!)

The rise in impedance can come from two sources:
(1) Skin effect losses due to very thin traces (or in your case
perhaps a lossey GND return path).
(2) A true change in impedance of the structure along it's length.
eg a tapered trace (thick to thin) in your case.
(other progressive changes in structure geometry will cause similar effects.)

Resistive losses are linear, so you should see the same rise when
testing from either end of the test trace.

A tapered trace will might cause an impedance change of 10 ohms
over its length, but add to that another 10 ohms of resistive skin effect
loss and this gives you your 50 to 70 ohm rise.

When testing from the other end you might expect to see the 10 ohm
drop due to taper, but you must add 10 ohms of resistive skin effect
loss in a linear manner over length, so this gives a flat 60 ohm.

I have seen similar effects on PCBs, so a close inspection by
microsection at various points may be in order. Non-tapered
traces will still exhibit skin effect loss, so other sources of
geometry variation can also cause such results.
The GND plane provides the return path for current flow, so any
form of cross-hatching will increase the inductance of the GND
plane and reduce capacitance leading to an increase in Zo.

Hope this helps.

Best regards

Andy Burkhardt
Product Manager
Email: mail@polar.co.uk
Tel: + 44 1481 253081
Fax: + 44 1481 252476
http://www.polar.co.uk
=====================================================
World leaders in PCB faultfinding and controlled impedance measurement
=====================================================

**** To unsubscribe from si-list: send e-mail to majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****