Extreme decoupling

[Bud]

Made this pic a while ago, I think it was an old laptop board. look at the decoupling caps soldered right on the leads.

[T3sl4co1l]

Someone got fired for that 

[Fred27]

Well, they're supposed to be as close as possible. That man's a genius.

[richard.cs]

They forgot to put pads for them on the board. It happens and sometimes a hand mod is cheaper than a re-spin. A good PCB house could have machine placed them across the legs and got them to reflow ok but it doesn't like like that was done here.

[PE1RKI]

they putted capacitors on the clock output to ground.
maybe they failed emi tests and this solved the problem ?

[Dave Turner]

If they were pick and place then someone has kept an excellent machine well hidden otherwise someone screwed up if we're talking about a commercial run.

[FrankBuss]

No decoupling, it is the clock outputs, see https://www.engineering.uiowa.edu/sites/default/files/ees/files/NI/pdfs/01/19/DS011920.pdf

And look at pin 7 on the left side But I guess it doesn't matter if a low wage Chinese worker needs a minute to solder it and to cut the pin, would increase the board cost by a few cents, and perhaps the next board revision was already planned.

[richard.cs]

If they were pick and place then someone has kept an excellent machine well hidden otherwise someone screwed up if we're talking about a commercial run.

We do a lot of small production runs and prototyping where I work, occasionally we get our assembly contractor to place something like this and they can usually manage. They would be placed at the bottom of the leg naturally although they might move in reflow. We have certainly had production boards made with resistors floated between SMT chip legs with excellent results.

[Dave Turner]

richard.cs - are the between pin components placed by hand or machine, is any extra paste required? Given that you pointed out yourself 'misfloats' I was wondering what the failure rate was.

[richard.cs]

Machine placed. So far as I know it was without extra paste, we certainly didn't remake the stencils but maybe some was machine-dispensed.

When we asked for the extra part we expected hand rework but they came back and said they thought they could machine place and reflow so long as we would accept that manual rework might still be needed. They seem to have a pretty good idea what is and isn't possible with their process and the occasional part-on-the-wrong-footprint doesn't raise eyebrows. Things like "put a 0R in whatever package you think best across those diode pads".

[Bassman59]

Made this pic a while ago, I think it was an old laptop board. look at the decoupling caps soldered right on the leads.

I remember seeing some rework where some pins on a QFP were lifted and series termination resistors were installed between each pin and its pad. Impressive.

[Bryan]

Hmm, wonder if it was by design and they had to be as close to the pins as possible, can't be a coincidence that the legs with a cap are right next to each other.

[jeremy]

I once removed the plating from a via and soldered a capacitor inside as a dc block. Worked great!

[richard.cs]

I once removed the plating from a via and soldered a capacitor inside as a dc block. Worked great!

I've run wire through vias but I hadn't thought of that one. The length-diameter ratio seems like it would make it fairly difficult but I suppose the hole could be bored bigger if you were confident about internal track locations.

[rexxar]

I once removed the plating from a via and soldered a capacitor inside as a dc block. Worked great!

That's crazy!! I'm adding that to by box of tricks, of course

[janoc]

I once removed the plating from a via and soldered a capacitor inside as a dc block. Worked great!

Pfff that's a bit hardcore!  I wouldn't want to be the one troubleshooting such board when there are components inside vias!

[SeanB]

Ever try fixing cordwood construction units, especially those lovely ones that were potted to keep the arcing down.

There the ultrasonic cleaner with the tricholrethylene and the vapour phase cleaning section worked a treat. It would get the epoxy soft after a 30 second immersion so you could pick it off, then as a final step a 3 second dip and a 30 second dwell in the cold zone to get the vapour off and you could then fix the unit, then pot it again. You could buy new, but as they were made from Unobtanium you had a lead time issue. Regular 30kV diodes, high voltage resistors and capacitors were easier to get hold of.