Re: [SI-LIST] : The Flight-Time/SI Effects of Trace 'Tromboning'

Jon Powell (jonp@pacbell.net)
Fri, 09 Jan 1998 09:41:50 -0800

Orr, Orville wrote:
>
> Has anyone investigated the effects of trace 'tromboning' (i.e.,
> folding the trace to add length) on flight-time and signal integrity
> in general. I'm interested in some rough rules-of-thumb on flight-time
> increases/decreases given a fold with a length of 'x' and a distance
> between the traces of 'y'. (The environment is current
> state-of-the-art 100Mhz+ multi-layer boards). How much separation is
> needed to eliminate any problems (assuming the problem is
> significant).
>
> Also, is this handled correctly in current simulators (such as Quad).
>
> Thanks
> Orville Orr
> Compaq Computer Corp.
> ECG - Enterprise Computing Group
> Phone: (281) 518-8546
> Fax: (281) 518-7445
> Pager: (713) 765-9892
> Email: orville.orr@compaq.com

I like this name, tromboning. I think this is the same thing we call
serpentine nets. A think a good answer for serpentine nets is: Do some
simulations to make sure that the crosstalk across the legs doesn't
effect flight time. Usually something like twice the conductor width for
spacing is OK but you really should just simulate the thing. Yes, the
Quad stuff will work fine for this, in fact, I think there is a demo of
this in the standard release demos. You can also inter-digitate ground
fingers for plenty of crosstalk protection, though it uses more space.

regards,
jon powell
Senior Scientist, Viewlogic Consulting Services.