RE: [SI-LIST] : connector substitution for high speed signals

Steve Weir (weirsp@a.crl.com)
Thu, 10 Jun 1999 19:13:24 -0700

--=====================_87878793==_.ALT
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

That alone will be very risky for high speed. If you really want this to=
work
out right, you should consider one of the handful of expert consultants in=
this
area, such as Dr. Johnson's Signal Integrity Inc. These folks can run
simulations which include the mother and daughter card with your alternate
source connectors to insure that you will still get what you need. There=
are
various techniques which may be employed to cope with the differences which
will occur from vendor to vendor, such as playing with the annular ring
diameters, etc to tweak impedances.

Regards
At 08:34 AM 6/11/99 +0800, you wrote:=20

>
> Hi all SI experts,=20
>
> Thank all of you for giving me good suggestions regarding to=20
> equivalent connectors in high speed environment.=20
>
> What I think is that if I can have same pattern of pin assignment, same
> signal return and signal -ground ratio=20
> as well as compensation of impedance disruption, then I shall be able to=
do
> connector substistuion.=20
>
> Any comment?=20
> =20
> Thanks,=20
> John Lin=20
> CAE Engineer=20
> Quanta Computer Inc.=20
>
> -----Original Message-----=20
> From: Andy Burkhardt [<mailto:andyb@polar.co.uk>mailto:andyb@polar.co.uk]=
=20
> Sent: Thursday, June 10, 1999 10:08 PM=20
> To: John Lin (=AAL=B4=C2=B7=D7)=20
> Cc: Casey; Amit Bhardwaj; Richard S Smith=20
> Subject: Re: [SI-LIST] : connector substitution for high speed signals=20
>
> At 09:01 07/06/99 +0100, you wrote:=20
> >Hi All SI Experts,=20
> >=20
> >One question.=20
> >=20
> >If I want to use one connector to replace another connector because of=
cost=20
> >issue, what criteria are needed to prove they are electrically=
equivalent.=20
> >=20
> >Especially the substitution is happened for high speed / low swing=
signals,=20
> >such as RAMBUS signals, which have 200ps for rising time.=20
> >=20
> >Any comments or suggestions are appreciated.=20
> >=20
> >Thank you for helps in advance.=20
> >=20
> >John Lin=20
> >Quanta Computer Inc.=20
> >CAE Engineer=20
> >Tel: 1+886+03+3979000 ext. 5183=20
> >Fax: 1+886+03+3975000=20
> >Email: John@quantatw.com=20
>
> Dear John,=20
>
> Check physical dimensions carefully as these will certainly have an=20
> effect on the parasitic L & C of the connection.=20
>
> One way to check equivalence is to check the sizeof impedance=20
> discontinuity caused by each connector type.=20
>
> (For Rambs) You can TDR a perfect 28 ohm transmission line and compare=20
> the size of the impedance discontinuity when the connection is placed=20
> in the middle (joining section) of a split idealm 28 ohm line.=20
>
> Minimum disturbance at 200ps rise time will identify the best=20
> electrical connection. Take care though as electrical and mechanical=20
> compromises are often necessary, so reliability, ease of use, etc...=20
> may also be affected.=20
>
> Hope this helps.=20
>
> Best regards=20
>
> Andy Burkhardt=20
> Product Manager=20
> Email: mail@polar.co.uk=20
> Tel: + 44 1481 253081=20
> Fax: + 44 1481 252476=20
> <http://www.polar.co.uk>http://www.polar.co.uk=20
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=20
> World leaders in PCB faultfinding and controlled impedance measurement=20
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=20

--=====================_87878793==_.ALT
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

That alone will be very risky for high speed.  If you really want this to work out right, you should consider one of the handful of expert consultants in this area, such as Dr. Johnson's Signal Integrity Inc.  These folks can run simulations which include the mother and daughter card with your alternate source connectors to insure that you will still get what you need.  There are various techniques which may be employed to cope with the differences which will occur from vendor to vendor, such as playing with the annular ring diameters, etc to tweak impedances.

Regards
At 08:34 AM 6/11/99 +0800, you wrote:

Hi all SI experts,

Thank all of you for  giving me good suggestions  regarding to
equivalent connectors in high speed environment.

What I think is that if I can have same pattern of pin assignment, same signal return  and signal -ground ratio
 as well as compensation of impedance disruption, then I shall be able to do connector substistuion.

Any comment?
 
Thanks,
John Lin
CAE Engineer
Quanta Computer Inc.

-----Original Message-----
From: Andy Burkhardt [mailto:andyb@polar.co.uk]
Sent: Thursday, June 10, 1999 10:08 PM
To: John Lin (=AAL=B4=C2=B7=D7)
Cc: Casey; Amit Bhardwaj; Richard S Smith
Subject: Re: [SI-LIST] : connector substitution for high speed signals

At 09:01 07/06/99 +0100, you wrote:
>Hi All SI Experts,
>
>One question.
>
>If I want to use one connector to replace another connector because of cost
>issue, what criteria are needed to prove they are electrically equivalent.
>
>Especially the substitution is happened for high speed / low swing signals,
>such as RAMBUS signals, which have 200ps for rising time.
>
>Any comments or suggestions are appreciated.
>
>Thank you for helps in advance.
>
>John Lin
>Quanta Computer Inc.
>CAE Engineer
>Tel: 1+886+03+3979000 ext. 5183
>Fax: 1+886+03+3975000
>Email: John@quantatw.com

Dear John,

Check physical dimensions carefully as these will certainly have an
effect on the parasitic L & C of the connection.

One way to check equivalence is to check the sizeof= impedance
discontinuity caused by each connector type.

(For Rambs) You can TDR a perfect 28 ohm transmission line= and compare
the size of the impedance discontinuity when the connection= is placed
in the middle (joining section) of a split idealm 28 ohm= line.

Minimum disturbance at 200ps rise time will identify the= best
electrical connection.  Take care though as electrical= and mechanical
compromises are often necessary, so reliability, ease of use,= etc...
may also be affected.

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
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D
World leaders in PCB faultfinding and controlled impedance= measurement
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D


--=====================_87878793==_.ALT--

**** 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 ****