making a SMD 500G shock proof

[ConKbot]

hm I didn't know you could get leaded surface mount

SOIC, TSSOP, are all leaded the leads allow for some flex,  vs something like a DFN or QFN where its solder pad on solder pad.

[Simon]

oh i see, never seen any 1206 leaded parts

[KJDS]

unfortunately military vehicle manufacturers don't tell you all about the vehicle , All we can guess is a gun but then why 3 axises.

MIL-STD-810 is the universal source for military vehicle vibrations.

[wraper]

Be very careful if you use any inductors. They will easily crack before any other parts are affected.

[Frank]

As others already said, you will need to know exactly what the "500g" means for the customer. If this figures came from some mil spec, probably you will have to test for shock and vibration.

Do you have any connector on this design? With this kind of shocks, the contacts will chatter, and maybe even fail and break.

[Simon]

yea we have connectors, honestly if it's really this sort of shock it's pretty gloomy. I was never told of shock requirements when i originally designed it so I'm sort of pissed off and starting to loose interest

[KJDS]

yea we have connectors, honestly if it's really this sort of shock it's pretty gloomy. I was never told of shock requirements when i originally designed it so I'm sort of pissed off and starting to loose interest

I'd be surprised if you have any real grief unless you've got anything that is very vulnerable on the board. Connectors should be of a locking type but from memory 500g shock isn't very high for a lot of applications.

[Simon]

I have got military grade pcb connectors, basically rugedized and version of molex that go down to -55C. But they are tall

[mikeselectricstuff]

My instinct would be to go as small and light as possible - less mass = less force under acceleration.

[Zero999]

oh i see, never seen any 1206 leaded parts

Some tantalum capacitors are often in leaded 12069 packages.

How about replacing some of those larger capacitors with several smaller leaded SMT tantalums? Of course ceramic capacitors and resistors can be replaced with through hole, if space will allow.

[georges80]

I designed a high power LED driver years back that was going into housings (massive) that were used to light a couple of areas of the shuttle during liftoff. i.e. these poor lights were essentially at ground zero in terms of copping some serious badness.

The capacitors (electros) sheared off their solder pads when the driver was run on a shake bed at the contractor testing facility. We ended up potting the entire board to have it survive some VERY serious vibrations, in the 10,000g range.

Got to see a video INSIDE a bunker near the shuttle launch pad and all the equipment was on air suspension racks and even then everything was bouncing/swaying around. Mind boggling vibrations...

cheers,
george.

[Niklas]

Are you going to thermo cycle the PCB? If so, be very careful and check the thermal expansion coefficient of the potting or conformal coating. I recently had an issue with SMD connectors, potting and thermo cycling and spoke to an expert at the Fraunhofer Institute about it. His comments were "If you can avoid potting, avoid it at all costs" and "SMD or through hole, both will crack, but the through hole mounted component will in general hold for a few more cycles together with potting". If the potting compound has a larger thermal expansion coefficient than the PCB and components and it can find its way under the components, then it will act as a small balloon and rip the components of the board.

See if you can find some info about space electronics. I think they are using both glue and some kind of conformal coating.

For the wiring, use multi-stranded wire, not the one with only 7 strands.

[Simon]

oh i see, never seen any 1206 leaded parts

Some tantalum capacitors are often in leaded 12069 packages.

How about replacing some of those larger capacitors with several smaller leaded SMT tantalums? Of course ceramic capacitors and resistors can be replaced with through hole, if space will allow.

Well i told my boss I'd want to take the current electrolytics for ceramics that are much smaller and he asked if the current boards "support that" duh, well i could replace the electrolytic with one but I think i need more but they don't want to redo the boards.

[mzzj]

I'm dealing with the customer and my company seem to have a policy of just saying it can be done and not asking for details rather than try and get as much information as possible.

How is the previous hand-dryer project with 20m3/second supersonic airflow and 50$ blower motors going? 

[Simon]

I'm dealing with the customer and my company seem to have a policy of just saying it can be done and not asking for details rather than try and get as much information as possible.

How is the previous hand-dryer project with 20m3/second supersonic airflow and 50$ blower motors going? 

technically it worked out very well but we got lost in paperwork and the customer came up with new requirements at the last minute because they did not realize they'd have to tell us about something requiring a redesign. But for the hand drier their comment was "can it be turned down" so I think that bit of it worked very well, trick is to use a 1.1KW ametek vacuum cleaner motor which is the same design used by comercial hand driers unless your looking at the electric turbo that the dysons have.

[TerraHertz]

apparently it's all 3 axises. We think the customer is just trying to cover themselves and being anal about it not realizing the hassle it causes. We are going to pot the whole thing. As usual we don't want to spend money as I'd just do the thing again with 90deg through hole connectors and 1206 or 1812 parts only

Problem is, most potting compounds are at least a little flexible. Which means that at 500G the potting will be wanting to deform quite a bit, and will apply that force to the components that are mounted to the more rigid fiberglass PCB. How far do the components need to be pushed, to break PCB adhesion, tracks, through-holes, etc?

Tldr; potting may make the problem worse, not better.

I'd guess you'd have to somehow find a potting material with very similar modulus of elasticity to the PCB.

[Simon]

presumably the amount of deflection across the area of something like a 1206 part won't do much though will it ? the board is only 40x20mm

[ConnorGames]

Probably pretty obvious,but most xtals are not rated for 500G. You'll want to look for ones that are, or use MEMS oscillators, which need more power but are generally rated for ~10000G (IIRC). I am using the CTS 403 series for a personal project, it is rated for 1500G shock but says nothing about vibration in the datasheet.

[Simon]

well I'm running off the watchdog timer in the MCU, it's just a timer circuit, infact I could have probably done it with a comparator and RC input, i think i used MCU because my boss wants to be able to change the delay in commissioning, he can reprogram it but not re-solder a resistor.

[Richard Crowley]

Those of us who regularly dissect commercial products (some of us have been doing this since before Dave was born!) regularly see hot-glue "goop" squirted around larger components to hold them in place.

Even more secure would be to simply "pot" the circuit, but I would strongly consider thru-hole vs. SMD.

[Simon]

Well potting is what we came up with. why through hole ? the parts are much heavier and surely the flexing will break component leads.

[Richard Crowley]

Well potting is what we came up with. why through hole ? the parts are much heavier and surely the flexing will break component leads.

But there is MUCH MORE MECHANICAL support for a thru-hole component. SMD components are mechanically attached only by a thin layer of solder.  OTOH, it is true that they have lower mass, so that offsets the attachment problem.

[Simon]

yes, in fact i am pleased that I am using through hole connectors

[Alphatronique]

Hi

my PCB assembler experience  ,go whit tiny SMD for part(0402)  + potting + really good board support

and for connector trough hole was mandatory 

if you not believe me ,do a simple test  solder 2 wire to a 0402 part then make a pull test
you will find that before break connection it need couple of pond  and since part have very low mass it
may handle lot of G and not to forgot that it potted to .. 
but secret was board support it must not have any flex on the PCB  ,SMD not like flex  ,but smaller was part less it affected

as for connector normally pin may accept more abuse and flex  and you cannot reduce mass of connector

for active part avoid BGA   was preferable like SOIC or TQFP Chip  maby QFN if it small part < 8mm

that may help you a bit ..

http://en.wikipedia.org/wiki/File:M734_Amplifier_and_Oscillator_.jpg

think that motar fuse handel lot of G ;-)

[luky315]

Can someone recommend a coating material which improves mechanical reliability and is applicable without special equipment?