Dispersion in Coax

Waveguide near cutoff would get my vote.

In coax the dominant effect is frequency selective loss due to skin effect, which causes a characteristic rounding off of the step response, known as 'drool'. The step response starts off fast, but partway up it slows way down, and it takes forever to settle. The bandwidth goes as

There's no waveguide mode dispersion in coax (due to it being TEM and therefore very small in diameter compared with a wavelength), so it would only be the dielectric's dispersion that you'd see. I doubt that it would amount to much compared with the drool.

Cheers

Phil

n to you.

very long coax cable and look at the output of the cable with an oscillosc ope to see that the high frequencies are traveling faster than the low freq uencies. Anyone in Tapper land do an experiment like this or a similar var iation?

But I wondered if there was some other transmission line

Thanks Phil, it's hard to get that square wave down wave guide. :^)

George H.

If that actually happens, I don't think you'll see it. HF attenuation dominates.

Does that actually happen? I've never seen anything like that.

There can be a little precursor/dispersion on a microstrip, because some of the field is in air and some in FR4.

Older Tek scopes had a coaxial delay line for the input signal so that Tek could implement a delayed trigger feature. If there was any phase distortion or group delay in the coax, the signal delivered to the vertical input would appear distorted on the CRT.

Ummm... It's "droop", not "drool". Methinks "droop" or "sag" is the lack of low frequency response, causing the flat part of the square wave to slope downward. It's not the rounding of the leading edge, which I believe is just part of "rise time".

Ah, almost. If the delay line had no group delay, it wouldn't provide any delay. However it has to be constant (or nearly so) within the bandpass, else the waveform becomes distorted.

My dim recollection from the time I worked at Tek was the delay line's HF attenuation resulted in the waveform "dribbling up". WAG: from lumped-element delay lines, in which there might have been a non-monotonic frequency response.

-F

Helical delay line? They used 'em back in the bad old days of NTSC TV (and often, tube sets at that), to account for the group delay of the chroma circuits (the luma was being delayed so it arrives at the CRT drivers correctly).

Anyway, because any given turn has direct transmission line behavior to its neighboring turn, but also magnetic coupling to more distant neighbors and so on, the dispersion of a helical delay line is quite pronounced.

You also have a sliding scale, from a long, thin helix in free space, to a helix that's so heavily shielded (solid shield inside and outside, and shield interleaved between turns, essentially making a coil of coax transmission line). The ones I've seen in ancient tube TV sets had a foil strip inside, not really providing any shielding but providing an electric ground reference, which keeps the velocity factor and bandwidth low, as is suitable for the application (low MHz).

Speaking of Tek, an example they used was a fairly highly twisted twinax, which was just slightly dispersive, enough that you saw some pre-edge wiggle but not bad enough to be a problem. That's a consequence of having the turns just close enough together (and not shielded between), to get some useful advantage in velocity factor.

Tim

Huh.. some C coupling between loops that was faster than going around the loop? I have no idea what a dispersed voltage step looks like.

George H.

We call it drool. The HF components of the step are attenuated such that you get a fast rise up to, say, 80% of the ultimate amplitude, but it creeps up the remaining 20% in a most annoying way.

Here's 12 feet of RG58, step response to a clean 30 ps step.

Prop delay is 19 ns. There is no suggestion of a fast precursor.

Longer cables look much worse.

Tek had elaborate equailzation networks to make up for the drool. Most of their delay lines were spiral-wound coax, to keep them short, so they may have had a bit of precursor.

Tight PCB meander lines can have a little pre-shoot, same idea.

I've always defined group delay to be the variation (shortest to longest) phase delay versus the frequency. "...group delay is the time delay of the amplitude envelopes of the various sinusoidal components of a signal through a device under test, and is a function of frequency for each component. Phase delay, in contrast, is the time delay of the phase as opposed to the time delay of the amplitude envelope." As I understand it, the problem is variations in the delay at different RF frequencies in the coax, which in my humble opinion is the group delay.

My guess(tm) is that I'll find a termination resistor and network that matches the characteristic impedance except at the highest frequency that the vertical channel is expected to pass, where it forms a small mismatch in the form of a "peaking network". That's to get a voltage boost at a frequency where the vertical amp that follows the coax delay line is lacking in gain. I've used the same trick in broadband RF amplifiers.

Well, here's the schematic of a Tek2445 scope delay line: The C404 tunes some kind of inductor to give the frequency response a little help at the high end. R403 adjusts how much help C404 provides. The "Gain Flattener Network" for U600, which incidentally is a chronic failure part with no easy replacement[1], probably also provides some manner of high frequency boost.

Ok, I yield. "Drool" is good enough to distinguish the effect from "rise time". However, Google can't find any use of the word "drool" along with square wave, waveform, step response, or similar buzzwords. The quotes around the word "drool" insures that "drool" is in all the search hits. Perhaps you should consider registering the term with one of the tech jargon dictionaries before someone uses it for something else?

Is that pre-shoot on the signal line, or direct dispersion across the ground plane underneath it?

The hybrid used in the 24xx scopes that fails is U800, the horizontal output driver.

U600 is the vertical output and is not a failure prone part.

Only us real insider experts use the scientific term "drool."

Some people in the picosecond business like to specify risetimes as

Straight microstrips don't usually have a visible pre-shoot. I think the signal is making short-cuts between the loops, instead of going all the way around. A looser meander won't show this.

Not to worry, in a few weeks google will find this thread as the number one hit to "coax drool" which was my zero hit search term. I have never seen coax drool, let alone had a word for it.

George H.

That was my picture, C-coupling is faster than going around all that length of copper.

George H.

Straight microstrips have straight ground paths the same length, not shorter.

But it's notable that looser meanders don't show it.