Q on an RF attenuation problem

Two pieces of advice:

1: Slightly improper installation of any coax connector will substantially degrade performance at the frequencies you are working with. "Heavy soldering" and "heatshrink sheething" that you allude to elsewhere are definitely improper installation. You really need everything to be ship-shape to get the advertised attenuation - using the wrong crimper can make things substantially worse, and soldering is definitely a no-no. Install the connector wrong and the advertised 0.06dB attenuation will be 6 or 12 dB attenuation.

If you have a choice between high-loss skinny cable with properly installed connectors, and low-loss fat coax with improperly installed connectors, I would go against the improperly installed connectors every single time. At 10MHz the hack job won't matter much, but at GHz it will in spades.

2: Anything you can do to get the radio part of the phone nearer the antenna is a Good Thing. Are there models that decouple the handset from the radio?

Tim.

Thank you.

Tim Shoppa wrote

As much as that?

The only bodge I have done so far is soldering the braid joint of an SMA connector, when it is intended to be crimped. The centre pin is normally soldered anyway, and crimping seems to be exactly the same. As I see it, the end result is the same.

The odd thing is that the cable used with GPS receivers (~1.5GHz) is normally a low quality cheap coax, and I would think that with say 2m of that stuff there won't be much signal left.

There is the only one phone I can use.

On a sunny day (Fri, 21 Aug 2009 17:42:27 +0100) it happened Peter wrote in :

I would, perhaps look at cables for satellite dishes. For some the SMA has no middle pin, the copper inner conductor forms the middle pin. And you can buy the stuff per meter. For WF125 foam it is only 23.3 dB / 100 meter, so .23 dB / meter. WF100 has screw on SMA connectors and is .27dB / meter.

And available from many places I would think. But single shield.

Considered hardline? More like 0.1 to 0.2 dB/m at 1.6 GHz.

John

rg400 is a solid teflon dielectric cable. You can always do better with a foam coax, if the temperature isn't too extreme.

Paul Probert

John Larkin wrote

A bit big; 0.4" is the smallest I see.

0.141" and 0.085" are common diameters for which a variety of connectors is available. Take a look at Times Microwave (formerly Times Wire & Cable) and In that last page they probably mean 50141 where they've written 50142 and the scaling on their log plots looks suspect. Whatever, ordinary 0.141 semi-rigid will get you a bit less than 15 dB per 100 ft at 1.6 GHz, and that's a bit less than 0.5 dB/m. Lower-loss versions are available.

Chris

--
http://www.micro-coax.com/pages/products/ProductTypes/CableTypes/SplineDielectricCoaxialDetails.asp?ID=UT-S(3)-250

JF

"christofire" wrote

0.5db/m is about 1/3 better than RG400 is (interpolating the 1GHz and 1.8GHz figures for the latter).

I am seriously looking at redoing the first portion (the 1.5m) in semi rigid as this portion will not get moved. It will be an absolute bastard of a job to preform it though, due to poor access. I can get this stuff in 3m lengths...

Incidentally, is there any easy way to measure attenuation of a cable run? I am thinking of buying some surplus-equipment signal generator and a signal strength meter. Of course, one could use the antenna itself as the signal generator, but the signal would be extremely weak. I have done this kind of thing with multimode fibre and even bought a fancy programmable digital attenuator, but in this case I also had a powerful signal source.

John Fields wrote

The UT-S(3)-250 seems comparable to this stuff

Both are semirigid but the latter is much more bendable and I can get it in 3m lengths.

But it is very interesting to discover what is out there... never knew this stuff existed.

(3)-250

You have to be very careful with spline so as to not deform and degrade the loss characteristics. We use in satellite construction and the techs have special training.

"no_one" wrote

I saw the very large minimum radius of 3"...

The other semi-rigid coax, solid PTFE, has a min bed radius of 0.25" which seems bizzarely tight but at least the stuff can be formed.

The spline cable is about 20-30% lower capacitance (and thus attenuation, presumably - these seem closely related).

You might try an equivalent length of the semirigid (without installing in your aircraft) and see what the phone signal strength meter displays. You shouldn't be losing that much signal in that short a run if your cable is properly built. I use an Antcom at work for GPS receive, but I use it with an LNA. Can you stand the $$ or additional complexity if you can't get the loss down for reliable operation?

Reducing the relative permittivity of the dielectric by introducing air allows the desired impedance to be achieved using a centre conductor with a larger diameter, which reduces the largest component of the loss: the (series) resistance of the surface of the inner conductor. The distributed capacitance may be smaller, but also the distributed inductance because the impedance is the square root of L/C.

Many swear that the reduced loss is because of the (shunt, effective) loss conductance of the dielectric but that may not be true for materials like PTFE and polythene at frequencies for which the cable is intended - the conductor loss is usually dominant.

The inner conductor of semi-rigid usually silver plated, which helps (for the same reason) but it doesn't appear practical to plate the inside of the outer conductor.

Chris

"christofire" wrote

Would this mean that a single centre conductor coax (e.g. RG223) would be better than a stranded centre conductor coax (e.g. RG400) even if the two are of the same dimensions and have the same attenuation at

1GHz?

It's hard to find data for 1.6GHz but I suppose one can extrapolate on a log graph.

I will try to rewire the whole thing with RG402

with soldered SMAs.

-- snip --

I guess that depends on what you mean by 'better'. A stranded inner conductor will withstand flexing more than a solid one and if the loss is the same there might be no reason to choose solid.

The Times catalogue to which I gave a link earlier also includes the two factors that represent the conductor and dielectric losses, with the formula that can be used to find the loss/100 ft for any frequency within the intended range. Times appear to cover the majority of well-known co-ax cable types (e.g RG-nx).

Chris

There are two serious problems with foam dielectric coax: they gradually absorb water, which increases the loss; and when you bend them, the centre conductor tends to squash or cut its way through the foam over time, leading to a nasty reflection at each bend.

The most bendable hardline is basically regular coax whose shield has been soaked in tin, e.g. Belden 1673,

It works almost as well as the copper tube stuff, and is a much better bet than foam coax for most things IMO.

Cheers

Phil Hobbs

-- Dr Philip C D Hobbs Principal ElectroOptical Innovations

55 Orchard Rd Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net

Looks interesting. Unfortunately most domestic broadcast installations in the UK use loose-braid foam-dielectric coaxial cable, or semi-air-spaced (spline-type) polythene dielectric - and it rains a lot here! 'Sky' the major European satellite broadcasting service provider specifies variants of RG6 for L-band down leads (e.g.

Of course, it's cheap (in the short run).

Chris

Perhaps I missed something, but on page 6 of the data sheets you pointed to, the SMA connector is for the GPS.

The TNC connector is for the phone.

Jim