Here's a picture of the mercury delay lines for CSIRAC in Melbourne,
Australia.s museum
,http://www.pcauthority.com.au/Gallery/299284,in-pictures-csirac-australias-first-digital-computer.aspx/14>.
It's the world's oldest original computer still in working condition.
Although it's normally on static display, it was saved complete and a
few years ago restored to working condition again by volunteers, many
who originally worked on it. It was first built in 1949 and improved in
following years. Retired in 1966 (not a bad working life for a computer
even by today's standards) it performed computations for over 1000
projects, the last for the futuristic ICI building opposite the
Melbourne hospital. It's still a very impressive thing to see in the
flesh. From memory there were about 30 mercury column delay line
memories, see if you can count them.

Don Black.

On 19-May-13 3:30 AM, Robert wrote:

Sorry for the empty message, hit the wrong button.

Someone told me they worked on a 1950s computer that used a mercury
pool as a delay line. One end would generate a wave and the other end
would detect the wave with the transit time being the delay.

Try and get that around ROHS!
Bob

--- In TekScopes@yahoogroups.com <mailto:TekScopes%40yahoogroups.com>,
"Robert" <go_boating_fast@> wrote:

--- In TekScopes@yahoogroups.com

<mailto:TekScopes%40yahoogroups.com>, Don Black <donald_black@> wrote:

What I call lumped (artificial) delay lines are delays made of a

number

of discrete inductors in series with capacitors to ground (or the
complimentary side) rather than a cable using its natural

inductance and

capacitance. In the case of the deflection plates, they are made of a
number of short segments coupled with small coils and trimmer

capacitors

from each one (segment) to trim them). Of course you need a special
vacuum screwdriver to adjust them through the glass without losing

the

vacuum (joke). Tektronix usually used a coiled coaxial delay line for
the signal delay to allow viewing the leading edge of a signal.

However

artificial delays can also be used. We used to have a 3M dropout
compensator for video recorders which had to delay one horizontal

line

of signal. I was switchable for the different standards, up to 100 µs
for 405 lines and it had row upon row of coils making up the delay

lines.

Others more familiar with Tektronix please jump in and add or

correct me

as you see fit please.

Don Black.

On 17-May-13 7:19 PM, cheater00 . wrote:

Regarding delayed deflection plates.

What is the basic operation of the lumped delay line circuits? Is it
just a long delay cable with a network which compensates the
impedance?

Regarding the digital vs analogue cancer that spread into this
conversation:

On Fri, May 17, 2013 at 6:10 AM, Steve <ditter2@
<mailto:ditter2%40yahoo.com>> wrote:

Modern real time scope run up to 65 GHz. No analog scope could

get

within an order of magnitude of that. And few applications

simply look

at a Y-T waveform. I use fast RT scopes every week for tasks

such as

decomposing jitter components in serial data streams running at 28
gigabits per second. I can easily measure these down to 300
femtoseconds, with an accurate resolution of 50 femtoseconds or so.
Users buy digital scopes for their analysis capability, seldom as a
replacement just to look at waveforms that an analog scope could do.

I love old analog scopes for their place in history. They are

fun to

restore. It is not different then people who restore antique autos.
May of these had design features that were decades ahead of their
time. But no one claims that an antique auto has the utility or
reliability of a modern car that gets you to work every day.

The tasks that design engineers need to perform today simply

can not

be performed by an analog scope.

<rant>
Because everyone is just like you and has exactly your job,

right? Let

me tell you, in 100% of my use of an oscilloscope I don't need *any*
of the doohickeys you mentioned. And even if I took the technical
level of my work to the very frontier of what's being done today, I
still wouldn't need any of it. Who cares about 300 picosecond jitter
when designing analogue audio circuits? You've got to be out of your
mind to think everyone needs 65 GHz bandwidth (so why bring it up at
all). My area, audio circuit design, will forever be fine with a 200
MHz bandwidth analog scope, maybe with minimal digital or analogue
storage capabilities. There are loads, loads, loads other areas in
technology which don't need *anything* provided by digital scopes.
Electrical and electronic engineers of all kinds and types have

needs

that are simply not provided for by a digital scope, while the

digital

scope provides a lot of stuff they expressly do *not* need. The same
goes for hobbyists: very few end up needing high bandwidth; if

an eval

board or kit is provided you don't need to care about clock lines
being bad. There's a huge amount of those people, and I dare say

many

more than those who need 65 GHz scopes.

It seems like you think the moment NASA started building the space
shuttle everyone stopped building, using, and servicing

bicycles. What

a load of junk. Look at your wrist watch or your mobile phone. Think
that needs a 65GHz scope to build?
</rant>

Please let's not derail this topic further. If you or anyone else
wants to continue the DSO vs CRO conversation, open up your own

thread

(which I likely won't take part in).

Kill it with fire.