When I look at the PCB inside the engine control unit for a motobike (for example) I saw the larger components have some kind of resin haphazardly poured around them. It was clear (brownish tinge) and must have been quite viscous when applied because it didn't 'run' much before setting. Not sure how hard it was, I didn't poke it but it was some kind of resin not rubber. I presume larger components (like capacitors) fastened only by their leads can't take prolonged severe vibration and this resin provides mechanical support.
It seems to me rubber compounds would help but unless damped they would mostly just up mechanical resonant frequencies. Hard materials might lead to additional stress when the PCB flexes.
Has anyone experience in this area? What is good to use?
Hot melt glue is pretty good. Leaded parts have horrible resonances, so gluing them down reduces lead flex and raises the resonant frequency. I've seen leaded caps literally fly off boards during vibration testing.
Yup. Vibration can be a real device-killer over time. I believe that the solder junctions (lead to PC-board trace), and the PC-board traces around the junctions are probably at more risk from the continual flexing than the leads themselves.
Even so, this could still help. If the mechanical resonant frequency of the component were moved up upwards, there might be less incoming energy at the higher frequencies available to excite the resonance, and thus less motion. In a motobike engine controller, raising the resonance from (e.g.) hundreds of Hz to thousands of Hz could move it up out of the frequency range most strongly excited by motor vibrations.
By analogy - when building a subwoofer speaker cabinet, you can either try to deal with wall resonances by damping them (which can be difficult to do effectively), or by stiffening the walls and raising the resonances up above the frequency range of the subwoofer where they're no longer a problem.
Differences in thermal coefficient of expansion can also be a problem with rigid encapsulants, for the same reason.
I've seen at least four types of material used for PC-board reinforcement or potting:
- Silicones (sometimes rubbery and flexible, sometimes quite tough and rigid once cured)
- Urethanes
- Epoxies
- Thermoplastics ("hot melt" adhesive) of various sorts.
Hot-melts, and moisture-cured silicones seem to be popular for mechanical reinforcement of individual components. From your description I suspect that a hot-melt thermoplastic may have been used in your bike controller.
For full encapsulations, urethanes, silicones, and two-component epoxies seem to be what's used most of the time. These formulas seem to be selected to remain somewhat flexible when cured, and to have low shrinkage during cure... this helps keep down the mechanical stresses on the potted components.
John and Dave make good points. While pouring resin over components (or the entire board) might work for non critical applications there are two things to keep in mind:
Adding too much mass to the board will make the entire board more susceptible to vibration damage. Better to judiciously glue down larger components than to goop the whole thing.
Pouring goop all over a board will reduce the rate of heat transfer from some components. Unless designed conservatively, things may overheat under the goop.
Also, other threads have dealt with the use of various silicone glues. Best to stay away from these unless you are certain that they do not contain chemically reactive solvents.
--
Paul Hovnanian paul@hovnanian.com
-----------------------------------------------------------------------
Have gnu, will travel.
"nospam" skrev i meddelelsen news: snipped-for-privacy@4ax.com...
Hotmelt glue damps vibration well, it is cheap, non-toxic and therefore very popular.
When I worked in industry we used vibration testing to (literally) see which bits would vibrate, glued them in e.t.c. When "done", the whole prototype would be subjected to a proper vibration stress test. If nothing broke off, the glue blobs would be specified in the production drawings; otherwise more testing & gluing ;-)
Large components are often glued "on principle". Items like heat-sinks, busbars and fixing strips are often "dual-used" to stiffen the board and as anchors for the mechanical fixings.
I am a fan of hot melt glue for fastening, insulating and encapsulating stuff but I am concerned about its performance with varying temperature.
A board I am looking at is effectively enclosed in a heat sink which under abnormal conditions could heat until regulators on the heatsink go into thermal shut down. The whole board might get to 120C where typical hot melt will be soft if not runny. High temperature and severe vibration probably won't coincide but low temperatures (-30) where the hot melt will be harder and vibration is possible.
I prefer RTV for the job. Its characteristics don't change with temperature and it doesn't wear out like hot melt. The joint on a hot melt glue job will often come loose if shaken long enough.
It is also a useful rule to not put the mounting hole in some nice neat pattern. More complex vibration modes tend to be at higher frequencies and thus are easier to damp.
Spot adhesives can be used to produce a 3-point minimum physical attachment, which reduces the repeated bending of lead wires, under vibration in specific planes.
Mixed epoxies or electrical grade silicones work best over temperature.
Use the long cure epoxy. The fast cure stuff tends to be brittle.
Never use "super glue" or TakPak for this sort of job.
The electronics grade RTV takes water out of the air to harden. The skin tends to harden faster than the insides. You need to leave it a lot longer than the point where the skin feels hardened before you retest.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.