Analog Memory

[macboy]

Is there any analog component/device/circuit capable of storing analog signal for a specified interval without using any digital stuff?

Other people have provided enough examples. A wire for example can be a (analog) memory device in that any signal showing up on one end of it takes time to reach the other end. If you have a sufficiently long wire and you pick a sufficiently slow transmission mechanism (sound vs. emf), you have a memory device that's fully analog. The delay-lines in old TVs are like that.

I don't think "a wire" can do this, but a transmission line can, that is, a piece of coaxial cable. This is commonly done in analog oscilloscopes to delay the signal to the vertical amplifiers so that the scope can trigger and start the sweep before (only nanoseconds before) the waveform hits the amp and shows on screen. This just barely allows the triggering portion of the waveform to be displayed on the screen. Inside such scopes, you will see a big coil of coaxial cable or sometimes a strange spiralled cable that is designed to give more delay per metre.

So a big long piece of coax can provide a fixed delay at releatively high bandwidth. As already a mentioned, a BBD, also known as a CCD (Charge-coupled device) can provide a short term storage with arbitrary playback speed. They are discrete time (sampling) but do not quantize/decimate the signal (do reduce resolution to N bits). Does that count as digital? You tell me. My old Philips DSO uses CCD devices to sample at up to 250 MS/s, until the device is full, then the stored 'waveform' is played back to the actual ADC at about 100 kS/s to be digitised. For 1980's technology, 250 MS/s at 10 bits was pretty impressive. The limitations of a CCD/BBD is the storage size (usually a few hundred points at most) and the bandwidth limited by Nyquist.

[IntegratedValve]

... for a specified interval ...

That you don't provide a figure for this duration, and don't know what vinyl is, tells me you don't really have a clear idea of what you want, and are very young.
That you're effectively asking for a time machine, but don't realize that, further brackets you.

Anyway, if you only wanted about 64uS delay, you could use a piezo delay line as used for chroma processing in PAL TV receivers. If your signal could be modulated onto a suitable carrier.
Other than that, it's mechanical or nothing. Magnetic tape, rotating drum or platter, etc. Oh hey, maybe the latest innovation, metal wire recorders. Or you could record the signal on a chart recorder, then play it back with some kind of optical sensor tracking the pen trace. I once found a temperature profile control system for a full sized brick firing kiln that used exactly that. So it does work. Just very slowly, as in a week or more to play back the chart.

You didn't mention anything about bandwidth, so maybe that might work for you?

Since you are moaning about it and want to know what I wanted that device for, I can tell:

To see how you can think out of the box. ;-)

[BravoV]

What's vinyl?

I am officially an old man. 

Me too. 

But this is a fun thread to watch, subscribed.

[dannyf]

I don't think "a wire" can do this,

If you had a wire from the sun to the earth, it would have taken several seconds to send a signal to you.

If you had a wire from here to Orion's belt, it would have taken several thousand years to send a signal to you.

If you had a wire from here to the end of the universe, it would have taken 10+ billion years to send a signal to you.

Yes, a wire can do this.

[branadic]

If you had a wire from here to the end of the universe, it would have taken 10+ billion years to send a signal to you.

No, the signal would never reach the end of the universe because of the finite propagation speed of the signal inside the cable/wire and the accelerated expansion of the universe and on the other hand your cable length would have to increase with the expansion speed of the universe D'oh!

[macboy]

I don't think "a wire" can do this,

If you had a wire from the sun to the earth, it would have taken several seconds to send a signal to you.

If you had a wire from here to Orion's belt, it would have taken several thousand years to send a signal to you.

If you had a wire from here to the end of the universe, it would have taken 10+ billion years to send a signal to you.

Yes, a wire can do this.

No, a signal will not propagate along a single wire. It requires a transmission line with a fixed impedance, such as coax or twisted pair.

[dannyf]

No, a signal will not propagate along a single wire.

sound on a steel rod?

laser in a fibre?

electrons in a copper wire?

Of course, if you define "wire" different...

[IanB]

No, a signal will not propagate along a single wire. It requires a transmission line with a fixed impedance, such as coax or twisted pair.

You do realize that if this statement were true then no electric or electronic circuits would ever work?

[gregallenwarner]

No, a signal will not propagate along a single wire. It requires a transmission line with a fixed impedance, such as coax or twisted pair.

A signal will indeed propagate down a single wire at a particular speed, usually a significant fraction of the speed of light. To prove this, take a look at how many modern PC motherboards are designed:



The designers didn't put all those wiggly traces in just for kicks. They did it in order to equalize the lengths of multiple parallel traces, so that all the signals reach the end at the same time. At speeds in the hundreds of megahertz, which is common in today's PC's, the propagation delay down these traces is a significant concern. Those traces can be thought of much like plain old wires; that is, they're not coax transmission lines, yet they still introduce a significant delay in the signal propagation.

[branadic]

that is, they're not coax transmission lines

Have you ever heard about Microstrips?

[gregallenwarner]

that is, they're not coax transmission lines

Have you ever heard about Microstrips?

A microstrip is not a co-ax configuration. The quote I referenced specifically mentioned two types of transmission lines: co-ax and twisted pair. Motherboard bus traces do not conform to either of those definitions.

I suppose if you wanna get technical, then yes, motherboard traces are a form of transmission line, but then, so are free hanging wires floating in the air, because they will also have some coupled capacitance with respect to ground, albeit a very tiny amount. But the point of all this is to refute the claim that free-hanging wires do not suffer from propagation delays, which is false; they do.

[tszaboo]

No, a signal will not propagate along a single wire. It requires a transmission line with a fixed impedance, such as coax or twisted pair.

You do realize that if this statement were true then no electric or electronic circuits would ever work?

From that very little RF I actually understand, I think he is right.
If you use a "wire" you have significant group delay on a transmission line. If you use a coax, matched impedances, and everything is ideal, than there is no group delay.
Now, if there is group delay, the stuff you put in the cable is not necessarily comes out in the same order. Which means it is not FIFO, not a good memory.
Can someone correct me please if I'm wrong?

[TerraHertz]

No, a signal will not propagate along a single wire. It requires a transmission line with a fixed impedance, such as coax or twisted pair.

Incorrect. A single wire in free space _is_ a transmission line with a fixed impedance. The only reason we use coax and twisted pairs is to achieve a lower and controlled impedance, that does not vary depending on the proximity of random nearby objects. Also to reduce losses to free space.

You can easily demonstrate that a single wire works as a transmission line, using a fast pulser and good scope.

Edit to add: I think there is group delay in all kinds of transmission lines, induced by the characteristics of the materials involved. Skin effects in the wires, dielectric parameters of the medium around the wire(s), etc. Trying to minimize that is part of the art of transmission line design, no? I thought the only thing that doesn't exhibit group delay, is radio waves in a vacuum? (But it's not like I know what I'm talking about here.)

[TerraHertz]

If you had a wire from the sun to the earth, it would have taken several seconds to send a signal to you.

Actually, I'm pretty sure if you laid a wire from the Sun to the Earth, several seconds later you'd have a humongous arc of many millions of amps, that would go on for quite a while. As the massive difference in potential charge between the Sun and Earth equalized.
I'd recommend you _don't_ lay a wire between any two celestial bodies. Unless you want to play "Look Ma, I've entirely converted myself to ionized plasma."

Starter reference: google nasa space tether electric universe

[dannyf]

This discussion is evolving into a prime example of people who know so much that they know so little - they don't know what relevant knowledge to bring to a given discussion.