electrical delay line attempt

[ratstar]

Heres a mouse drawn schematic of a delay line design.

This probably doesnt work! I havent tested it yet but its what im currently up to, I havent got the 60 metres of wire spooling I need for it, and im worried that i need to phase invert the magnetic field or itll induct into itself. I think I can buy a special wire wound resistor with low inductance and high conductivity here->
https://www.ohmic.com.au/products/o...WPDgvRDR02wO4iv-lOh0N5BToX3AwUXsaAmoAEALw_wcB

I want to go for under 10 megahertz cause my multimetres frequency counter only goes up to that, and I think I need about 60 metres to get to that wavelength for the speed of electricity. But thats only a guess.

Has anyone here got any experience with electrical delays? Its basicly just a guitar pedal.

 

[Harald Kapp]

A guitar pedal at 10 MHz?

In copper wire electrical signals propagate at 0.59 to 0.77 time the speed of light (source). Assuming slowest propagation that is ~177 000 000 m/s. It takes the signal 0.34 µs to propagate along a 60 m length of wire. Not very useful for your application, is it?

There's more to delaying electric signals than just adding a length of wire, start reading e.g. here.

 

[ratstar]

Thanks for the reply.

///////////// WRONG ////////////// EDIT//////////////////////////////////////
Your calculation puts it at 29,411 hz. so thats heaps different than my estimate, I think that im the one thats wrong tho. (I'd prefer it to be lower hz in this case, then the wire doesnt have to be as long.)
/////////////////////////////////////////////////////////////////////////////////////////////

Woops I need to think more about this... <edit>

Yes I know theres other types of delays, but a length of wire would have the best signal preservation of all the techniques u think?

Id like to get this delay line memory going a terrahert if I could, because if the pulse width gets to be below 1 micrometre, it counts as if you had transistors that were shrunk that small making the ram bits, but u dont need to miniaturize the same way with a delay line memory.

So instead of shrinking the transistors, you shrink the pulse width of your delay line memory signal and it does effectively the same thing... gives you gigabytes of ram.

 

[ratstar]

Ive got a question, im confused.

I read on the internet the wavelength of a terrahert down a copper wire at 21 degrees celcius is 1 mm - is this right? or is it way off.


<edit>
ok i think ive got it now.

if electricity goes 17,700,000,000 cm/s
i go 3,540 cm every 5 millionths of a second.

so thats the wavelength.

so using that compute a terrahert is 0.0177 (200 micrometres.)
Therefore to beat a nanoscale ram chip you need to get to 200,000 terraherts. (thats the pulse width at a nanometre.)

But then youd probably have 100 kilobits just riding up the leg of your transistor. so the feesable hz would be 200 gigahertz, at 1 mm.

but that isnt bad, id love to get that going.

 

[Harald Kapp]

I'm a bit lost with your calculation.
The relation is: speed = wavelength × frequency or c = w × f
Assuming c = 177 000 000 m/s (depending on the properties of the wire c may be different)
and f = 1 Thz = 1 × 10¹² Hz
you get w = c / f = (177 000 000m/s) / (10¹² Hz) = 0.000177 m = 0,177 mm

But: The wavelength does not determine the propagation delay of the electric signal (wave) along the line. Have a look at the telegrapher's equations for more details.

 

[ratstar]

Yeh thats what I meant.

So how do transistors even deal with super high frequency low wavelength signals - do they end up just acting like low pass filters? =)

Thanks for your patience.

 

[Harald Kapp]

What makes you think so? One wouldn't need a comparatively complex semiconductor to create a low pass filter. If the right transistor is selected and the circuit is correctly designed (incl. high frequency compatible layout) you can use transistors for amplification, for mixing etc.

What are you up to?

 

[ratstar]

What are you up to?

I'm into off grid living, but I haven't made it out of my flat yet!
Why delay line memory? I think maybe it could be as good as other types of memory, and the cool thing its more hacker friendly in a lot of ways. Also I want to make a delay line memory based computer, I'm diving in a little deep because I cant help myself, I always have.

So transistors can switch on and off at a terrahert? That would be cool if its true.
(looking here)
https://electronics.stackexchange.com/questions/209921/how-fast-can-a-transistor-switch
https://www.reddit.com/r/AskElectronics/comments/3p9wlo/how_fast_do_transistors_switch_on_and_off/
https://ieeexplore.ieee.org/document/1392539

 

[WHONOES]

I think 200,000 terahertz is somewhere in the range of X-rays to gamma rays. Whilst 1 terahertz is in the range of far-infra red. i.e. from jsut under visible light through to stuff you don't want to mess with.
I'm sure someone will correct me if I have got it wrong.
For a transistor to switch at those frequencies, if at all, it would be pretty exotic.

 

[davenn]

I'm into off grid living, but I haven't made it out of my flat yet!
Why delay line memory? I think maybe it could be as good as other types of memory, and the cool thing its more hacker friendly in a lot of ways. Also I want to make a delay line memory based computer, I'm diving in a little deep because I cant help myself, I always have.

So transistors can switch on and off at a terrahert? That would be cool if its true.
(looking here)
https://electronics.stackexchange.com/questions/209921/how-fast-can-a-transistor-switch
https://www.reddit.com/r/AskElectronics/comments/3p9wlo/how_fast_do_transistors_switch_on_and_off/
https://ieeexplore.ieee.org/document/1392539

As far as I'm aware , there are no terahertz switching transistors and none of those links you provided state that there are

Your thread is wandering all over the place like a drunk walking home from the pub

you never stated WHY you wanted a delay line
Your math and thoughts in your first few posts are really oddball and that is being kind

You were asked what you were trying to achieve, but you didnt answer that either

I'm very close to moving this thread to to WooWoo section unless it improves substantially

Dave

 

[AnalogKid]

Transistors that can switch at terahertz frequencies are very rare and very expensive. Nowhere in this thread do you say what it is you want to achieve. What is the source of the signal you are delaying? How much delay time do you need? What is the destination for the delayed signal; what kind of circuit and application?

In the 60's, RCA used a 25-50' coil of coax cable to form a monostable for the FM demodulator in a videotape machine. Other than that, most delay line applications use an electrical delay line (a long string of L-C sections or a long tapped inductor with caps at each tap), a mechanical delay line (coiled springs), or an acoustic delay line (an FM carrier through a glass bar). Here is an example of that last one:

https://www.ebay.com/i/274399577897...97&pmt=1&noa=1&pg=2386202&algv=DefaultOrganic

ak

 

[ratstar]

Thanks for the input.
I want to make a memory line with it for a home made computer.

Heres a picture of a THZ oscillator. I wonder whats inside it.
http://terasense.com/products/terah...8swAznAMe_axYQK_4o_SG0l_KHJvo-wAaAgreEALw_wcB

That glass delay unit you showed would store 30 bits in it, if you used the centre frequency and it was a 1 millionth of a second delay.

 

[Harald Kapp]

I'm very close to moving this thread to to WooWoo section unless it improves substantially

No improvement recognizable. Move?

 

[ratstar]

I don't know what you guys problem is... cant even be yourself these days.

 

[WHONOES]

Purely out of interest I did a search for any terahertz product and came up with the following link.
https://vadiodes.com/en/products/custom-transmitters

The next link is for a document that discusses the construction and possible uses of terahertz devices.
https://www.mdpi.com/1424-8220/19/11/2454/pdf-vor
There are others and are all fairly exotic.

 

[Harald Kapp]

I want to make a memory line with it for a home made computer.

How are you going to read and write data at THz speed?

The classic electric delay line was made of LC components, not a single strand of wire.

 

[bertus]

Hello,

In old PAL color tv's there where also delay lines.
One long one of 450 nS ans a little block for 64 μS.
The wiki shows them:
https://en.wikipedia.org/wiki/Analog_delay_line

Bertus

 

[AnalogKid]

Ampex pioneered videotape timebase error correction (AMTEC) and color phase error correction (COLORTEC). Both used voltage-variable video delay lines - a long tapped inductor with varicap diodes at each tap. In resurrecting an old meatball TR-22, I made them by hand-winding the inductor on a ferrite rod. Later, we moved the design to use binary-incremented fixed delay segments ala the AVR-1 for the timebase, but kept the variable line for the color phase.

ak