Hi all,
I need to cut three 2ft lengths of coax, but these lengths need to be all of the same length to within 0.002". The 2ft spec is not critical; the fact that they should all be of identical length with respect to each other, however, is. How best to achieve this? Thanks.
I am afraid you are
(a) on a hiding to nothing as a simple tug will stretch 2ft of cable by more than 2 thou, as will a hot day in June
(b) likely to have extremely similar problems if you are attaching any termination whatsoever to them in terms of pluggs sockets or resistors.
Back in the day this was a problem often encountered in radar horns, and in general there ware 'variable length' wave-guides or bits of solid coax made up to tune the responses of two/three receivers to get phase differences accurate.
I see you had sufficient self-control to refrain from asking 'why?', and 'what sort of coax?'!!
Presumably the hot day in June will affect all three the same amount. OTOH, it wouldn't surprise me in the least to discover that a sharp tug on one of them will completely destroy the phase-matching.
So how do you make sure that the plugs have the same phase response?
I suspect the answer to your question is that you have to have some sort of phase measuring system. You cut one cable to length, and then you cut the other cables to have the same phase output. I don't know how you would do this though. Others here might.
Also, *why*?
to 2 thou in 2 ft? I think not.
maybe 125 thou...
Well the way I would do it to get 'as good as it gets' would be to make up one thousand, and sort them into actual time delays on a rig designed to do that, and then select triples that matched and throw the others away, and charge a lot of money for them.
Of course as I said, they would stretch if abused.
Depends on the wavelength. But assuming you can get by with lambda/10 as a working tolerance your implied working wavelength is 0.02" ~ 0.5mm
Put another way 3x10^8/0.0005 = 600GHz - that's in the terahertz band!
Q band ~7mm was about as high as the VLA ever operates.
Liberally coated in snake oil and cosmetic gold leaf?
What are you actually trying to do?
its easy enough with a TDR to measure signal delay, or use an unterminated cable as an oscillator tuning element so selecting near identical cables is not hard..
My worry is that the specification is impossibly tight so that you would need to handle te cables with kid gloves to avoid stretch, and arrange them to be temperature synchronised as well.
In my life, it was to allow a triple radar horn to be absolutely aligned with a target when the signals were identical in phase.
Useful for radar seeking missiles and sometimes beam riders. He is probably rebui8lding a bloodhound or summat..
No that's 20 thou: not 2 thou
Only if they are kept at the same temperature within, I think, about 0.2 degrees.
(Mind you, that's just the copper core. I'm bugggered if I know how to address differential thermal effects between core, insulator, shield and jacket.)
Indeed. The reason I made the original comment is that when I worked on what became Sea Wolf, in 1969, IIRC that was what I worked on, the three way IF strip that had to be calibrated for differential phase drift every few milliseconds.
We essentially shoved the same signal into the receivers, adjusted them with varicap diodes (IIRC), and then 'went live' for a millisecond
I'm trying to make up a set of cables for my vector network analyser. I pulled the 2 thou figure out of my backside but for all practical purposes (IOW coming up with a workable solution to accurate cutting) it's neither here nor there. But for the benefit of NP, here's my revised thinking aloud:
My VNA goes up to 1.3Ghz, so that's the highest frequency of interest. The wavelength at this frequency is 230mm. Measurement resolution needs to be ideally 1/3600th of this (or better) so the difference in the cable lengths should be < 0.64mm. So, nowhere near as bad as 2 thou, but still a bit of a PITA to get right. Any suggestions?
The only vaguest hope of getting anywhere near this accuracy of delay balance would be to use semi rigid coax and to maintain the length matching including the connector pins too.
The proprietory sets (cost a bomb) I've seen use fully flexible coax; it's very-low capacitance but other than that has exactly the same feel as ordinary TV coax. Loss is of no consequence in this application, but it *has* to be 50 ohm coax.
Won't you need, in any case, to recalibrate every time you use it?
Bill
That's further up the line at the point where the device under test (DUT) is measured. I can't calibrate out variations in these cable lengths which patch the VNA to the bridge where DUTs are checked.
I had this problem with my VNA (which goes up to 3GHZ). I needed to link the front panel connectors to the T/R test set. The solution was to cut a set of semi-rigid cables (with solid copper outer sheath) all from the same length of cable to within a few tenths of 1mm aided by a binocular microscope. This was difficult. Stripping them identically was even more difficult. Then I soldered them into the appropriate N connectors (which needed a very big soldering iron).
After all that I discovered that inside the VNA the signals were being sent from the front panel connectors to the sampler board on "ordinary" flexible coax.
Also, nearly every time I use it, I just set the delay adjuster to make the Smith chart look right with a short circuit. This compensates for any minor error in cable length.
John
I cant see any hope of maintaining anywhere near that spec when soft coax b ends in normal use. Rigid would give a big improvement. In fact I cant even think of any mechanical way to cut cables that doesnt result in a good bit of distortion and thus grossly violate your spec. Laser cutting would leav e carbon etc behind.
I've little clue what tolerance you could get on some type of sanding setup with 3 rigid cables glued together.
NT
This is sort of what Martin suggested in an earlier post: cut one to approx 2' and then trim the others to match it for length by phase. The best suggestion yet *in theory* but the devil is in the practical challenge, as you say, of somehow fine-honing the other two cables to the length of the first. 'Finely sanding coax' doesn't sound feasible and then of course adding male plugs to each end could easily trash the earlier efforts. :(
fine sanding metal tube sounds doable to me, but as you rightly say, attaching plugs will make a mockery of the fine tolerances.
NT
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.