Broken Duplexer

[Melt-O-Tronic]

I was given a Dekolink 800/900 MHz cellular bidirectional amplifier.  One of the two duplexers had its N connector broken off at some point.

I have a new connector, but have not succeeded in getting enough heat up the center pin to remelt the solder on to the element inside.  I'm wondering if removing the top so I can get a soldering iron inside will kill it.  I have a DSA815-TG, but seriously doubt I could get the thing retuned if necessary.

I'd like to repair it if I can.  It's designed for Nextel, but I'm hoping there's a small chance that it will work with GSM since some of the band segments are common.  In any case, it's a nifty piece of gear.  If it's not usable for GSM, maybe the broken duplexer will be a learning tool.   

[Melt-O-Tronic]

And just for fun, here's what the good duplexer looks like on the spectrum analyzer.   

[KJDS]

It won't be too far off tune after removing the lid and putting it back on. You won't be bale to retune it unless you can look at the return loss as well as the insertion loss as you're tuning it. Unless you've been taught, or had a lot of practice, don't even try tuning that, you'll only ever make it worse.

Should you decide that you don't mind weighing it in for scrap aluminium and can get the use of a VNA and want to try tuning, then measure the performance, save in memory, then unlock just one nut, twiddle the screw to see the effect then tighten the nut up again. You'll discover that as the nut tightens it draws the screw up which affects the tuning quite significantly. Once you've got the hang of getting one screw loose and tightened without losing performance, try with two and so on.

[T3sl4co1l]

I bet that thing looks like sex on the inside.

[codeboy2k]

Nice score. To fix that connector, try this:  use 2 irons set extremely high, at around 700 deg F.. then heat both parts and quickly put them together, just be mindful of the heating times.  When I've needed extra time with wires at 700 F it gives me about 2-3 seconds before it cools and solidifies.   Be careful not to melt the center dialectric. You might be able to use a lower temp like 350F at the connector end, and higher temp like 700F inside the box, it can probably take the heat better.

Alternatively, you could try that Chip Quick SMD removal solder, which has  a longer working time at lower temperatures. It melts at 136 F so you can use a iron set to 350F and it gives you some working time.

http://www.chipquik.com/store/product_info.php?products_id=210001

[Melt-O-Tronic]

Thanks guys!  I'll try the two soldering iron method this evening and may try Chipquick eventually.  Great ideas!

So far, my boldest attempt was to put a small dab of liquid flux on the internal element and a small blob of solder on the connector tip, then screw the connector in (I can feel that it makes contact).  Then I inserted a piece of 16 ga copper wire into the center pin, turned the duplexer so the connector pointed to the floor and heated the wire with a fine propane torch flame for a looooooong time.  That didn't get enough heat inside.

I bet that thing looks like sex on the inside.

That made my day!     Totally agree.  I think there's a high probability that even if I get it fixed, the BDA won't work for GSM and I'll end up opening her up anyway.  I'll post pics if it comes to that (secretly hoping it does).

[niflheimer]

From my limited experience with this kind of gear , if you note the mounting screw positions (sharpie or something like that ) you can open it up and reassemble it without losing the tuning. It's the actual tuning screws that do that .

If that were the case the mounting screws would be loctited in as well - I can guarantee that after 5+ years on top of a vibrating tower some of them do loosen up.

[Richard Head]

We used to manufacture similar diplexers at work. It was a real pain when it came to passive intermodulation distortion. Any slightly poor connection on the screening can produce PIMs which are obviously highly undesirable. They used to beat them with a rubber mallet prior to tuning to get the PIMs down!

[Melt-O-Tronic]

With 13 adjustment screws per cavity path and only a few threads of adjustment each, I'm guessing that retuning would be a monstrous undertaking.  I'd turn the rubber mallet on myself.     The compound on the lock nuts looks intact though.  The unit was mounted to the inside masonry wall of a warehouse for years, so no vibration exposure.

Anyway, I couldn't get a proper solder connection, but got the thing temporarily patched and reassembled.  It's working well on the bench -- at least well enough to see if it will work for repurposing.  If so, I think the Chipquik method do the trick.  With a -90 dBm input, I had to use 70 dB of attenuation on the output just to keep it within the limits of my spectrum analyzer.

I need to find a couple of antennas next.  And the site where I'm hoping to use it is a pretty long distance from here.  I probably won't get out there until after Field Day (http://www.arrl.org/field-day).

I predict a teardown in a few weeks.     If it doesn't work, I'll at least harvest a very nice pair of UHF amps, a pair of preamps with step attenuators and a pair of really cool cavity filters to play with!