[LAD] [OT] 3ghz coax and soldering...

Robin Gareus robin at gareus.org
Wed Jan 19 11:56:23 UTC 2011


On 01/19/2011 06:35 AM, gene heskett wrote:
> On Wednesday, January 19, 2011 12:12:05 am Robin Gareus did opine:
> 
>> On 01/19/2011 03:39 AM, gene heskett wrote:
>>> On Tuesday, January 18, 2011 09:16:00 pm Robin Gareus did opine:
>>>> Hi Joern,
>>>>
>>>> If it is an option: use Leerrohre (DE for "empty tubes" ?) to make it
>>>> future-proof, rather than to rely on cable-standards. In a few years
>>>> you may want to replace coax with optical or whatever.
>>>
>>> I think that would translate to wave-guides in English
>>
>> nope. I mean tubes like pipes in the wall or floor that allow one to
>> easily replace cables that run inside those tubes. It'd still be a major
>> re-wiring task but at least one can change the wires easily compared to
>> in-wall mounted cables.
>>
> 
> Ahh, so, here that is called "conduit" and can be metallic or more likely 
> in recent years, plastic, glued up just like water pipe and smoother to 
> pull wires through than the metallic stuff.  Both need plenty of "snot" on 
> the wires for a temp lubricant if going very far, and about 3 bends is 
> still the limit for one pull run.

I guess Joern's Studio is only 3 or 4 rooms. It's mostly because he
mentioned wall sockets that I suggested this.
It can also be the wall (or ceiling-mounted) type. I think they're
called raceways or just cable-tray.

[..]

>> It is currently running inside the ALICE detector @LHC.
> 
> Yikes.  If that ever collapses, it will emp into smoke, and and all 
> electrical stuff for many meters around it. 

nah It will pull itself into a black hole of course :)

> That, it can be said is NOT a 
> friendly environment.  Even your $15 Casio wrist watch is in danger if you 
> move too fast in that.

The magnetic field you mean? The tricky part is get close enough. You'll
need a few weeks to undo all the screws on the "enclosure" and not get
kicked out in the meantime. Once you get in there you don't need to
move, the electrons inside the watch are already moving fast enough. The
only workaround is to have only short interconnects and lots of bulk in
between them.  Luckily this is an OT post:
http://blogs.discovermagazine.com/badastronomy/2010/09/21/dont-cross-the-lhc-stream/

>>> I will allow the comment that when fabricating wave-guide parts and
>>> filters for 7Ghz work, which I have done a few of decades ago, those
>>> were usually silver soldered because the regular tin/lead solders
>>> surfaces oxidized with time much worse, screwing with the skin effect
>>> losses.  Silver oxide may look fugly, but is still a pretty fair
>>> conductor when frequencies are in the realm where skin effect reigns
>>> supreme.  Just as true inside the wave- guide as it is on the skin of
>>> a coax conductor.
>>
>> interesting. I do remember a drawer with at least 15 different kind of
>> solder. Now I know what that silver stuff in there was in there for :)
> 
> Silver solder, or just silver bearing solder?  

likely both. It is a very well equipped lab. There was a small
hand-torch in there as well. Though I think it was mostly used as a
spare lighter :)

> The latter seems to top out 
> at about 3% siver in an otherwise eutectic allow, but is a noticeably 
> stronger solder mechanically that doesn't oxidize near as fast.  I don't 
> use anything else myself but it does raise the iron temps needed by 
> 50-100F, which leads to needing to clean the tips more often.
> 
> Real silver solder needs a propane or better torch or a tig for heat and 
> you use a borax based flux that melts and forms an airtight puddle of 
> liquid glass over the joint so the oxygen is sealed away from the hot 
> metal, facilitating a nice clean joint when you feed the silver wire in to 
> make a 'sweat' joint, like the plumbers do for copper piping.  But the 
> borax should be chipped away and removed once it has cooled as its a bit 
> lossy at those frequencies.  Properly done, you get the 'sweat' started, 
> and pull it along by moving the heat (it crawls to the heat) until you have 
> the joint level full of silver but no extra beads sticking out.  Don't 
> apply any more silver than what it takes to just seal the joint.  Dry fit 
> accuracies of less than a 5 thou gap work very well, wider gaps are 
> correspondingly harder to control.  You really need a milling machine to do 
> wave-guide parts correctly.
>  
> Thanks robin

Thank you for the lengthy explanation.

ciao,
robin



More information about the Linux-audio-dev mailing list