Please explain scope delay lines?

[Axtman]

So a Tektronix 365 scope has a delay line that is probably 20 feet (6 meters) long. Electricity travels at nearly the speed of light. I thought I read that light would circle the earth seven times at the Equator a second.

How can such a miniscule piece of wire have any sort of delay?

[KungFuJosh]

"In its ideal state, electricity travels at the speed of light or one foot every nanosecond...Of course, this assumes that there is no resistance, no inductance, and no capacitance in the circuit."

[johansen]

It takes time to trigger the scope and start the sweep, in that time, information is lost.

If the delay line is long enough you can put the information on the screen after the sweep has started, preserving linearity.

https://www.everythingrf.com/community/what-is-propagation-velocity-in-a-cable

[bdunham7]

How can such a miniscule piece of wire have any sort of delay?

Well, certainly you agree that it has SOME delay, right?  So if you signal goes 25,000 miles in one second, then it goes one foot in 1/(25,000 x 5280 x 7) seconds.  That comes out to about 1 nanosecond per foot, thus 20 feet would result in a delay of about 20 nanoseconds.  The signal moves through the typical wires a bit slower than light in a vacuum, about 1.5ns per foot.  Delay lines are specially constructed and things move even slower.  So a typical delay line provides 40ns or more of delay, enough for the trigger and sweep circuitry to start up.

[Circlotron]

How can such a miniscule piece of wire have any sort of delay?

Even relatively short pcb tracks have a tiny but measurable delay. This often necessitates having equal length tracks so that one signal doesn't arrive before another related signal. One way of achieving this is to put wiggles in the tracks to extend their length to match their otherwise longer partners.

[Someone]

The explanations are easily found:
https://w140.com/tekwiki/wiki/Delay_line
tens of nanoseconds, meters of line, yep thats as expected.

[jpanhalt]

How can such a miniscule piece of wire have any sort of delay?

Who said it only "goes around" once?  Early calculators with memory used sound and a metallic loop of only a few feet.  Relaying a signal can be quit quick compared to analyzing and displaying it.

[ebastler]

So a Tektronix 365 scope has a delay line that is probably 20 feet (6 meters) long.

For starters, it would be good to know which particular type of delay line we are talking about. I could not find any "Tektronix 365" scope online. Is that the correct model number?

Assuming we are talking about a delay line made from rigid coax cable, which was used in various Tek scopes and seems in line with what the OP describes. The "velocity factor" by which signal propagation speed v is reduced vs. the speed of light in vacuum c depends on the dielectric constant epsilon of the cable's dielectric material: v = c / sqrt(epsilon).

For polyethylene, epsilon = 2.3 and hence v = 0.66 c. That's 5 ns delay per 1 m cable length. Is that what the scope can do -- i.e. with a 6 m delay line, show you what has happened up to 30 ns before the trigger event? If not, there's something else going on which we would still need to figure out.

(Dielectric materials with a higher dielectric constant could be used to obtain slower signal propagation. But I am not sure whether that was done, since you also want a material which limits signal losses. Also, to the OP: How sure are you about the 6 m length estimate? I looked at the service manual of the Tek 465 scope, which states a 120 ns delay. It does not specify the length of the delay line, but just looking at the drawing it appears to be significantly longer than 6 m. Might well be the 24 m which would be required for 120 ns delay according to the above estimate.)

Who said it only "goes around" once?  Early calculators with memory used sound and a metallic loop of only a few feet.  Relaying a signal can be quit quick compared to analyzing and displaying it.

That's interesting technology, but certainly not what would be used here. These torsion-wire memories had acoustic waves going through the wire; not quite supporting the signal bandwidth you want in a scope.

And the various ultrasonic delay lines (mercury-filled tubes, torsion wires, glass plates) were either used in a single-pass, open-loop configuration when used as a delay line, or in a closed-loop, going-round-in-circles configuration when used as memory. I can't see how a multi-pass delay line would work reliably, without massive crosstalk problems.

[Axtman]

Oops. I mean Tektronix 465, like the one in my avatar.

Thanks for the explainations.

[TimFox]

In the Tektronix 465, the built-in delay line is approximately 120 ns.
I believe it is built from L's and C's, rather than a long coax.

[ebastler]

Going by the drawing in the Tek 465 service manual, I think the delay line is a spool of coax cable.

Just skimming through the manual quickly, I couldn't find more details. But if the drawing, including the indicated wire thickness, is roughly to scale, I'd say 24 m of cable is not an implausible length. (Say 60 cm circumference, 40 turns?)

[TimFox]

Yes, that appears to be a spool of cable in the 465.
Solid polyethylene coax has a delay of about 5 ns/m = 1.5 ns/foot.
Higher-impedance and slower cables were made using a helical center conductor to increase the inductance per unit length.

I remember the delay lines in vacuum-tube Tektronix scopes, where the impedance level was higher than practicable for coax.

Anyway, so long as the time required from the trigger input to the start of the horizontal sweep is less than 120 ns, the CRT is able to display the signal shortly before the trigger event actually happened.

[kloetpatra]

yes the length is about right
here are some pictures of the 466 delay line which should be nearly the same

[TimFox]

Yes, that is a balanced transmission line, apparently with solid polyethylene insulation.

[CatalinaWOW]

The how it works part has been well described here.  It is worth thinking about why you want to see signals before the trigger level has been achieved.  The simple answer, implied above is to see the beginning of the pulse you are triggering on.  Does it rise uniformly or are there steps and bumps in it.  But you may also want to see other signals.  Triggering on the output of a logic gate you would never see the signals that caused the state change without a delay line long enough to account for the gate delays in the logic gate.

Modern DSOs with deep memory can make this ability to see the precursors to a triggering event seem nearly infinite.  Before the DSO became practical the delay line gave a window into this world, with the length of the delay line being a compromise between many factors including signal attenuation, signal fidelity loss due to dispersion in the delay line, the physical volume occupied by the delay line, cost and other factors.

[ebastler]

Modern DSOs with deep memory can make this ability to see the precursors to a triggering event seem nearly infinite.  Before the DSO became practical the delay line gave a window into this world,

Travelling back in time, kind of. Cool, eh?   

Much to my chagrin I couldn't afford such a fancy CRO as a student. My modest 20 MHz Hitachi could (and still can) only show what came after the trigger...

[CatalinaWOW]

Modern DSOs with deep memory can make this ability to see the precursors to a triggering event seem nearly infinite.  Before the DSO became practical the delay line gave a window into this world,

Travelling back in time, kind of. Cool, eh?   

Much to my chagrin I couldn't afford such a fancy CRO as a student. My modest 20 MHz Hitachi could (and still can) only show what came after the trigger...

Same here.  And same at home for quite a while after joining the workforce.  But at work we had the good stuff.  Storage scopes and Polaroid cameras.