When you really want SMDs...

[daqq]

Hi guys,

I've bought a scrap RF amplifier, looked inside and saw that someone probably really hated the people doing his PCB assembly...


By the way, the other photo shows a disconnected cap+resistor that do not look like they died by means of too much power but rather oxidation. Would you agree with this conclusion? Both resistors read around 16Ohms, but the color codes are iffy.


The device is a KALMUS engineering RF amplifier from 1MHz to 230MHz, ~10-15W, probably made around the year 1992.

Was this kind of construction technique (abused trough hole parts with bent legs) common in the era for this stuff?

Thanks,

David

[dave_k]

Yep .. construction style looks fairly typical for a RF power amplifier. Sometimes the only difficult part about working on one of these is finding a soldering iron with enough grunt!

[T3sl4co1l]

Hmmm, can't seem to see any vias.  I wonder if they're plated solid or filled with solder or something.  I would be shocked if there's no bottom side ground.

"Engineering amplifier" sounds like a cost-no-object item, so who cares about cost? 

Should be easy enough to replace those parts -- looks like they arced over, after failing somehow (maybe corrosion, maybe the solder melted?).  The transistors are likely dead too, probably having oscillated and expired in the process.

If you can get schematics, that'll be a great help.  On the other hand, if the resistors were selected values (a possibility!), you may want a sweep + analyzer to make sure the gain is still flat and all, and the output is stable.

Selected parts would be a good reason to design it as "top side load" SMT...

Tim

[SeanB]

Corrosion, plus a capacitor not rated for the RF power that was flowing through it. If you look the capacitor body has lost all the wax conformal coat normal for those age ceramics, and has been well baked.

Looks like SRBP, so the traces will be very fragile, so a big tip on the soldering iron, and gentle handling and lots of solderwick and a dental pick to lift each leg of the transistors are needed, otherwise the trace will come with. The heatsink is the ground plane, with all the mounting screws there providing the connection to it, and is common in older RF designs, as it is all stripline based. the component leads bent like that are standard, but the advantage is that you can do a lot of repairs with large SMD parts in many cases, especially with coupling capacitors, supply capacitors and such, just watch power ratings.

Replace the resistors ( not sure on the component to the right of the cap, might be a ferrite bead or a very low value inductor) and capacitor, but be sure the replacement ones are rated for the RF power they will handle.

[Red Squirrel]

Yep .. construction style looks fairly typical for a RF power amplifier. Sometimes the only difficult part about working on one of these is finding a soldering iron with enough grunt!

That's when the propane torch and solder from the plumbing tool box comes out. 

[tautech]

Looks like SRBP, so the traces will be very fragile, so a big tip on the soldering iron, and gentle handling and lots of solderwick and a dental pick to lift each leg of the transistors are needed, otherwise the trace will come with.

I've not had such experiences with Paxolin/phenolic 1 oz PCB's and made them for SMD with traces down to 15 mil and pads for SOT-23/6 and SOIC, never with any of the fragility you describe. 

[SeanB]

Trust me they get fragile with age, especially with a good number of thermal cycles on them as well.  When they turn almost black you know that you will be doing trace repair, as the copper will already be lifted and only the component lead is holding it in place.

[wraper]

Looks like SRBP, so the traces will be very fragile, so a big tip on the soldering iron, and gentle handling and lots of solderwick and a dental pick to lift each leg of the transistors are needed, otherwise the trace will come with.

It's not. This is FR-4 like material (fiberglass). Fibers going in one direction are seen on the pictures.