Re: Joining co-axial cable outdoors

[PoP]

Bugger. Trust me to pick an argument with an MA ;)

??? Tsk.

I spent 15 years as an EE.

PoP

[stefek.zaba]

Agreed. (The wording of the original post did suggest a 3-way "tap" rather than a simple join, but I think that was an oopsie of phrasing). A *neat* solder joint, as taught to radio hammer-chewers, is indeed more likely to perform well at RF than introducing a plug-n-socket; although a crappy soldered joint would be worse than competently-attached plug and socket.

So, to repeat in Stefek-speak the "how to solder coax" instructions: the goal is to create as neat and just-like-the-rest-of-the-cable join as possible, definetely including using the same inner insulation ('dielectric' if you like the fancy words) and resulting cable dimensions, not great gobs of insulating tape, stupidly sharp turns, wierd lumps, odd transition materials, or whatever. So you strip back both ends, carefully: strip outer insulation for say 2 cm, slitting it neatly up to the stripping point and not chucking it away. Push back the outer braiding into a nice bell shape. Cut off the inner insulation/dielectric for say 1cm on both cables; slit one of those slugs down its length. Now lay the inner 1cms side by side, and holding in place using your third and fourth hands solder quickly together (don't melt the insulation higher up) - practice this step on scrap bits. If you don't have a third and fourth hand, you can wrap the two inners firmly together with some thin wire before you solder.

Now put the slit slug of dielectric back over the soldered-together inners; a single turn of thin insulating tape can help hold it in place, but don't overdo it with lots of turns. Now it's time to join the outer braids together. As others have written, if it's copper it's eminently solderable: tease the end of the braid apart a bit, weave the two ends together to create a nice not-noticeably-thicker interlacing of the two braids, spot-solder (don't overdo the soldering - again, the idea is not to melt the inner dielectric). If aliminium or copper foil, soldering won't work: smooth into place, maybe add the bit of ali foil another poster suggested. Cover with the previously slit slug, or a scrap bit of the outer insulation, holding down with insulating tape or the previously- mentioned heatshrink (which you didn't forget to slip on first ;-). For a joint to be subjected to sun and rain, it's worth hunting down the adhesive-lined heatshrink, which will make a much more waterproof joint.

If you think you're too cackhanded to do this neatly, then a plug-n-socket might give better results. BUT how are you going to attach those? If soldering, it's somewhat easier than a cable join, but still susceptible to oopsies; if you get crimp-on fittings, you'd better know someone who will lend you the 100-quid-plus proper crimping tool - you're unlikely to get a quality joint with your second-best pair of slipjoint plumber's pliers ;-) If you really want to louse up the RF properties of the coax, separate out the inner and outer conductors for a good few inches, join inner-to-inner, outer-to-outer in a bit of oversize (30 amp) choccie block, wrap with gobs of self-amalgamating tape, bang a masonry nail through it all to 'secure' it to the wall, and call yourself a registered $ky installer ;-)

HTH, Stefek

[Dave Plowman]

Makes you wonder how the fairly standard:-

aerial - cable - plug - DA - plug - cable - outlet - plug - cable - plug

- inlet ever works at all...

[Dave Plowman]

Better, surely, to use a low loss inductive type?

[geoff]

In message , Dave Plowman writes

The phasors won't take it cap'n

[Chris]

That does indeed help. What a wonderfully clear and helpful description! I am truly grateful. I shall do it today, with a very long extension cable for my soldering iron.

[The Natural Philosopher]

Exactly. As Pop said, it shouldn't work at all.

But it does.

We pissed around in the lab once with bits of cable and a VSWR and a scope and a pulse generator as young graduates in our first job.

Plugs aren't as percfect as cable, and solder is very little different - even if not 'done correctly'. - i.e. bringing the tails out side by side, making the joint and insulating with tape.

Splitting is immediately noticeable tho.

Nearly all connections in the equipment were not done onto circuit board with a coaxial connection: Normally you have a solder tag to the chassis by the socket, and wire to that with a tail from a bit of coax. and the barid then goes to a pin on teh ground plane. Hardly ideal, but in practice very little different from 'ideal'.

The difference between a theoretical and a practical engineer is the theoretician knows what makkes a difference, the practical knows how much difference. And ignores if less than a dB.

[The Natural Philosopher]

Yes. but you still lose the 3db :) Sctually it must be more with resistors mustn't it?

[stefek.zaba]

Oh, I forgot one more crucial tip for the budding install monkey: wrap up the joint in self-am tape, *but* make sure you leave the braid exposed futher away from the joint itself, so rain can make its way up the cable by capillary action ;-)

Thank you for your words of appreciation. Do have a practise in a more comfortable place on some scraps of coax first - and be warned that soldering with the wind blowing can be a pain in the bum, as you lose heat quite a lot faster than indoors. Fine if you have a 40W temp-controlled iron, as it'll simply keep up with the added loss; less fine if you are waving that 15W little yellow Antex about ;-)

Cheers, Stefek

[Chris]

In article , [e-mail] writes

It's done. And it works. Radio 3 reception is wonderful. But I made one stupid mistake which made me burst out laughing. I joined the cable neatly - and tested it - and then started fixing it to the wall - and found I hadn't taken the cable behind drain pipes etc.! So maybe I'll do it again - but not today!

[PoP]

But you didn't ask us for advice about putting it in front of or behind the drainpipes, so how can it be a mistake? ;)

PoP

[norm]

Did you actually pass your HNC because you seem to be talking cobblers ?

As a former BBC transmitter engineer with 25 years experience of RF issues I offer the following advice.....

The aim of the exercise is not to introduce discontinuities of impedance along the cable. You can get pretty close to this using the method someone suggested earlier with exposing the inner, soldering it, covering the joint with a slit piece of the insulator (same dielectric constant you see), then covering the now insulated middle bit with the two bits of braid that should be dangling etc etc.

It's a load of arsing about. Get two connectors and a back to back socket and wrap in in self amalgamating tape.

Couple of points...its unlikely the terminations an either end of the cable are very close to the characterisitc impedance of the cable anyway cos TVs are built on the cheap and so are aerials. Secondly, most TV wall sockets are appalling from a design point of view but work OK.

Finally, if your TV signal is weak enough that using connectors and a piece of tape makes a difference then your picture is probable borderline to s**te anyway.

n

[PoP]

I certainly did. However, admittedly I went the digital electronics route rather than the analogue :)

Which is exactly what I have been suggesting all along. The suggestions relating to soldering etc weren't mine - are you sure you aren't confusing me with someone else?

As I commented earlier, the acid test is to make the modification and feed the signal into the back of the TV (or radio). If it works then great, if it doesn't then maybe the signal was not helped with the modifications to the aerial feed.

PoP

[Edwin Spector]

That'd be a technical term, then?

Edwin.

[Edwin Spector]

I have just done some measurements, and it hardly makes a gnat's c*ck of difference!

Aim Compare signal losses due to various types of join in coax cable.

Apparatus Coax cable (unknown type, about 2-ft length, 50 Ohm, terminated in SMA connectors) Network analyser (HP 8753E)

Method

1. Measure insertion loss and return loss (or VSWR) of unbroken cable. Use these results as reference.
2. Cut cable near the middle.
3. Re-connect the ends as crudely as possible (inner soldered, braid just twisted) and re-measure IL and RL at 650MHz and 750MHz.
4. Re-connect ends as well as possible and re-measure.

Results Original return loss: >27dB (Very good) Original insertion loss 0dB (relative, not absolute) Crude join: IL = 0.3dB at 650MHz, 0.31dB at 750MHz. Better join: IL =

[The Natural Philosopher]

What was teh VSWR?

[stefek.zaba]

[ ... ]

Ah yes, but think of the pride in workmanship ;-) But thanks for injecting real data into the speculation. RF folklore doesn't travel all that well from its original context; so for radio-hams and wannabes pushing 100W of ERP up coax, and needing to be sure it's going only where it's supposed to, a degree of paranoia about cable/joint integrity creeps in which isn't necessarily relevant to broadcast signal reception in non-marginal areas.

Cheers, Stefek

[Edwin Spector]

Return loss of better than 20dB over 650 to 750 MHz, which is equivalent to a VSWR of 1.22:1. Theoretical loss due to this mismatch is 0.043dB (I looked it up).

Edwin.

[Chris]

In article , Edwin Spector writes

I am very impressed by someone who actually finds out something! Thankyou.

[Edwin Spector]