4ms delay line at 1GHz

时间：04-08 整理：3721RD 点击：

Hi,
Does anyone know how to realize 1-4 ms delay at freq around 1GHz?
Are there any off the shelf components or maybe some app notes how to do it?

Hi,
do you need real 1-4ns or milisec!? Must be it "analog" or for logic signals?
K.

Added after 1 hours 22 minutes:

If you only at 4 nsec needs, some vendors are; Sprague, PCA, Siemens...

I remember yet, Elmec was a very good address for some nsec Delay-lines: www.endrich.com/en/site.php/48271

If you possibly mean milliseconds though, you are aware, that you are talking of 300 to 1200 km electrical length?

Existing implementations of electrical delay lines with GHz bandwidth have been limited below 1 μs, even with superconductiong
cables. Fibre-optical delay lines can achieve a multiple μs range, however.

If you really wanted to do it, you could. I would make a really high Q cavity filter, either whispering gallery mode or superconducting, and make a 1 to 3 pole banpass filter. 1 KHz bandwidth would give me ~ 1 mS for a 1 pole.. To bump that up to 2 mS I would reflect the signal back thru the filter and use a circulator to form an input and output port.

It would, of course, be a little jumpy on centerfrequency--might need some sort of active center frequency tuning to keep it right on 1 GHz.

It might be possible to string together a bunch of saw or baw bandpass filters centered at 1 GHz to get a big time delay too.

Hi,
Thank you for replies.

It wasn't a typo, I need 4ms of the delay, analogue RF signal ~1GHz carrier, FSK modulated.

I heve done in the past a system that used 500km of optical fibre and Mach-Zender electrooptical modulators to get such a delay. this time I just dont have the budget for such a system.

I besically need to simulate a large distance between radios and it has to be done at the carrier frequency.

Hi,
Interesting job...
Than is Brian`s solution with 2-4 (tunable)cavity filters the best (& relative cheap) solution for you.
Good progress!
K.

Well, last time around you did NOT do it at the carrier frequency, you upconverted to light frequencies, and then downconverted at the other end. So explain a little more exactly what you are trying to do, including required bandwidth, etc.

In radar systems, we used to make "blitz boxes" that were placed on a hill top some distance away, and had the function of repeating the radar pulse with a time delay related to the path distance. In your case, you might be able to mount a repeater on some distant hill top, and use high gain antennas at both locations. You would probably need to shift the return signal slightly in frequency to keep interference down, or possibly use cross-polarized antennas for the two paths.

Other than that, filters are probably the best way to get the delay you need...unless your modulation bandwidth is too large that it would not pass thru narrowband filters. Just string a bunch of filters in series with 3 dB pads between each filter.

Fiberoptics are still possible. They have pretty low cost nowadays. You can get two way transmission by using two different laser wavelengths, like 1310 and 1550 nm, so you could share the same fiber. A couple km of fiber is pretty cheap. You could probably throw it together for S3000 in hardware costs. But you would be using digital lasers, not analog ones. ie you would convert your rf signal into 1 gbps digital data, and at the other end turn the digital signal back into rf with a lowpass filter. should be ok if the modulation is only fsk! I recently did a 656 MHz clock signal distribution network for a VLA telecsope array, and sending the clock around was pretty easy using off the shelf fiberoptic parts.

There might be some way using locked oscillators and analog delay lines (like L-C audio delay lines) to simulate the path. You would have a tracking VCO either frequency or phase locked to the input signal, and the time delay would happen in the control loop. Not sure if that would violate your system requirements.

Rich
www.MaguffinMicrowave.com

my radio network should be closed in a room.
the problem with all filters is that they will be very narrowband and my system uses frequency hopping. look here
http://en.wikipedia.org/wiki/Link_16
it is the type of radio network that i need to close in a room and simulate the distance (delay).

Hi,
I found on the pp 345 a filter what, I think, you can use (Center frequency 985 MHz, Fractional bandwidth FBW 10.359%, 40dB Rejection bandwidth 125.5 MHz. On a 120x50x1,27mm substrat, Epsilon 10.8...

Best thanx for "alpuslu", www.edaboard.com/viewtopic.php?t=384786#1247286
K.