Why these pcb lines curvy?

Greetings,

I was tinkering in youtube then I saw EEVblogs latest video. There is something that I didn't understand about the circuit design of that product. Why are these lines are curly instead of straight?

Thanks.

It looks like they are trying to avoid skew by having the traces be the same length.

Maybe they are curvy to get identical length of the different lines.

Thanks for the answer.

Why do the lines have to be in the same lenght?

To avoid (timing) skew

Like I said, skew. So all the signals on the various pins arrive at the same time.

Very much necessary for high speed parallel data.

When logic signals were low (in frequency, or long in time) the traces were not critical. However now that logic signals are very fast, the delay caused by copper traces (simplification) can be significant.
By making the traces all the same length the signals are delayed the same so when they reach the receiving IC, the signals are correctly aligned with each other (wrt time).

Thanks for the answer.

OTOH, all those bends, at high frequency, become inductors, and the parallel lines become capacitors...
Nightmare.

Hi @Burakimportant.

This "construction" is known as "delay lines" and causes signal delays.

RV mineirin

Delay Lines Draw Upon Stable Dielectric Constant | 2014-10-25 | Microwave Journal.

Thanks for the answer.

I will look at them.

If you want to read a good explanation on what is going on with these signals google:

Rich Hartley, THE EXTREME IMPORTANCE
OF PC BOARD STACK-UP

This is mostly about impedance matching to reduce signal reflections.

Reflections can reduce and even cancel out a signals level.


Arriving at the same time, not so much since clocking is usually involved.


For our sssllooww stuff here, non to worry.

No, since they are transmission lines over a groundplane, they are actually pretty close to purely resistive. The thickness of the FR4 prepreg laye between top traces and the plane underneath is something like 100µm, so electrically the traces are almost unaware of each other, each acts as separate microstrip of around 50 or 60 ohms. The majority of electrical and magnetic field exists directly under the traces. There also are few 100 ohm differential pairs also on this board.

Microstrip signal propagation velocity for FR4 is of the order of 15cm/ns, so the skew being removed is of the order of a few 100ps or so.

The zig-zag traces are basically a sort of printed delay line or inductance. For some situations or values this provides an easy and cheap way to get the components without actually using any physical components.

There are no inductors in the posted image. Those are entirely delay lines.

Yes, but any length of wire has inductance, albeit small right? Increasing its length by any means will only increase the inductance. So in this particular application this inductance is probably within acceptable limits.

Hi,
If you go to 8:00 in the YouTube video, you will see Dave refer to them as;

"Controlled Impedance Wiggle Wiggle Wiggle Yeah Traces"

Tom... :grinning: :+1: :coffee: :australia:

Sorry, but that is not how a transmission line works. The wire has a distributed capacitance between itself and the backplane, in addition to the inductance. When the inductance and capacitance are balanced (by careful choice of geometry), the impedance of the line becomes non-reactive because the L and C cancel each other, so long as it matches the terminating resistance (load resistance). I have glossed over the minutae to convey the general picture.

That is why you can have a 1,000 km coaxial cable and not worry about the inductance increasing so much that the bandwidth becomes almost a few Hz.

When the wavelength of the signal is of the order of the length of the connection, such connections become components with relevant electrical properties (capacitance, inductance, characteristic impedance, and delay characteristics). The line length and the terminating resistor become important. The properties also depend on the mechanical structure and the material. In this range, it is even possible to build filters with only lines. A piece of wire can thus become a resonant circuit - which sometimes disturbs, or is intended to be so . Antennas, for example, also work in this way.
Designing circuit boards for such frequency ranges is a very challenging task.