making a SMD 500G shock proof

[Simon]

So boss asks me to design a circuit, I do that, Nice little SMD board just a micro and power supply to it. Now he wants it to withstand 500G shocks. Hm I think soic and 1206 packages are ok, what worries me are some though hole connectors and the 4mm diameter electrolytic capacitors that are the only things with any real mass versus their footprint.

Any idea's. He is thinking of potting it which will be messy. As I can do wthout the capacitors I am in favor of starting again and using 1206 ceramics maybe a few in parallel.

[kizzap]

Use through-hole components? Might give a little more support. At a guess though potting will be the most stable.

[Simon]

yea it's just that we are going front a neat little light board to a monstrosity. next I'll need loads of fixing holes so that the pcb does not rip apart or whatever mounting strategy I use will need loads of mount points.

[kizzap]

another idea I just had:

Some capacitors come with screw down supports. maybe you could make something similar to that?

[Richard Head]

Lucky you don't build proximity fuses. They have to withstand 10 000G!
Can you imagine a 1g smd component suddenly weighing 10kg?

[Simon]

ah so is it like for every G/g the weight increases by the amouint of G ? like my 1g part at 500G will weigh 500g ? useful for some basic testing

[Richard Head]

Simon

Yes, that's how it works. Can you imagine 500g for each and every smd resistor?? (Assuming they weigh 1g, but I'm not certain about that)
So, if your populated board weighs say 200g it will become 100kg for a brief instant! Can you imagine 10 000g?

[Kjelt]

First thought in my mind, this might require to use special glue on each and every component before soldering. I wonder if SMD components can handle 500G in the first place without potting to distribute those forces.

[miguelvp]

Is the 500G on a predetermined direction?

For example in a car for front side collision, you will know the acceleration vector so how the board is mounted in the car will make a difference. I mean it's not like you have to make it 500G proof when someone drops a car flat on it's wheels.

Also, will conformal coating help?

[Simon]

Yes glue around the part is one thing that was suggested. See I'm worried that once posted we go from a 20g board to a 100g lump, so now we have 5 times the strain on our fixing holes that started out as just 2 holes in diagonal corners that now have to take 50Kg so we make the board tougher and heavier. So really designing with 1206 caps will be best and using 90degree connectors instead of connectors sticking up in the air. No 1206 parts will weigh much less than 1g probably more like 1/100 of a gram so I'm quite happy to use those.

[Simon]

Is the 500G on a predetermined direction?

For example in a car for front side collision, you will know the acceleration vector so how the board is mounted in the car will make a difference. I mean it's not like you have to make it 500G proof when someone drops a car flat on it's wheels.

Also, will conformal coating help?

I'm not sure but now that I understand what the effect of the forces are I'll look into finding out as if the board can be appropriately mounted it will make life simpler. But chances are I'll not get much info as I don't think we know much about the vehicle it goes in.

[Simon]

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

[DmitryL]

So boss asks me to design a circuit, I do that, Nice little SMD board just a micro and power supply to it. Now he wants it to withstand 500G shocks. Hm I think soic and 1206 packages are ok, what worries me are some though hole connectors and the 4mm diameter electrolytic capacitors that are the only things with any real mass versus their footprint.
Any idea's. He is thinking of potting it which will be messy. As I can do wthout the capacitors I am in favor of starting again and using 1206 ceramics maybe a few in parallel.

Nice Do you have any facilities to test how your stuff deals with the shock ? Otherwise it might be a bit pity to find out that the homebrewed shock-proof design fails miserably in reality

[hans]

It's in essence all about forces. Gravity is 1G force pulling us down to the surface, as in F=m*a. When free falling all objects will accelerate with approximately 9.8m/s^2 (varies a bit across the earth), until terminal velocity (which is just air resistance building up to an equilibrium with gravity).

If that object suddenly stops, it experiences a force to take away the momentum. Depending in how short a distance it stops/stopping time will lead to an acceleration applied, which in turn is a particular force in Newton. For example, 100G would be 980m/s^2 of acceleration, on a 1kg object is 980Newton. That is more or less the same as 100kg at gravity (1G). For imagination purposes it is often expressed as "as large as weighing X kg for a brief moment".

You could fake the Newtons by using gravity and have something pulling on it, but I think you need to have a very clever mounting system to apply it in a good way. Also, force/speed/acceleration is not scalar but a vector, and can be applied in 3 dimensions. In addition; "shock" implies a very short trauma, not extended periods of time (milliseconds vs seconds).

Honestly I wouldn't really know how to "reliably" test it. You do have those vibration tables, but we often only use those to see if screws & board assemblies remain fixed during transport/shaky equipment usecases. Maybe there are similar pieces of test equipment for shock.

Or you can just overengineer the crap out of it
Not sure, but I think large (especially tall components) are better to be fixed through hole than SMT. I still prefer through hole USB connectors over SMT USB connectors for that reason.
I've also had some large SMT aluminium caps "fall" of a PCB once when the board felt on the ground. So I guess those are also a worry point.

[Simon]

So boss asks me to design a circuit, I do that, Nice little SMD board just a micro and power supply to it. Now he wants it to withstand 500G shocks. Hm I think soic and 1206 packages are ok, what worries me are some though hole connectors and the 4mm diameter electrolytic capacitors that are the only things with any real mass versus their footprint.
Any idea's. He is thinking of potting it which will be messy. As I can do wthout the capacitors I am in favor of starting again and using 1206 ceramics maybe a few in parallel.

Nice Do you have any facilities to test how your stuff deals with the shock ? Otherwise it might be a bit pity to find out that the homebrewed shock-proof design fails miserably in reality

Nope no testing at all, we don't usually do shock. I think the whole thing will be going out for testing

[Simon]

Or you can just overengineer the crap out of it

That is what we usually end up doing hence we will pot the whole thing so that nothing goes anywhere and screw through the potting.

[ConKbot]

Duration matters a lot.  500G for 1us? 1ms? 1s? for 1uS, you could get away with good design, for 1S, obviously you'd need potting. Is this a single impulse shock? Vibration? "500 G" by itself is almost a worthless spec imo.  If you can, pin them down for more information.

The nasa hand soldering standard (search online for a pdf) gives some insight into some shock/vibration resistant work. Leaded components are your friend (SMD and TH) as the leads provide some stress relief to keep from stressing the solder joint. 

Also remember the cube sqare law.  Your component volume (and mass) will go down with the cube of the size if you scale a component down, and the cross section of the solder joint (pad size) will only drop with the square, so your mass per mm^2 on the joints drops as you go smaller. 

Havent tried it, but I'd suspect its a lot harder to use shock and rip off a board one of the flexible terminated 0402 capacitor than a 1206 on the same pcb.

[Simon]

yea I agree 500g on 3 axises does not tell us alot but unfortunately 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.

[kizzap]

testing should be easy...hit it with a hammer

[DaveW]

My day job involves around 1,000g-2000g half sine shocks, so these might be overkill. Having said that,

If you must use larger components, glue dot them down to the board
Don't mount any components near to the edge of the board (within around 3mm). Neither should any components be mounted too close to supports, ie. where the board is fixed to the enclosure
If you're using larger ceramic components, they can crack under shock. But with lowish levels of shock, you'll probably be OK with 1206
The flexiterm components don't seem to help much with shock-this tends to useful with mismatched TCs

Are you going through vibe as well? If so, your resonance points in one plane will be at lambda/4, lambda/2 & 3*lambda/4-avoid mounting any large components across these lines

[Simon]

testing should be easy...hit it with a hammer

well I think my boss said something about hitting it with a cricket bat but he may have been joking

[poorchava]

I'd watch out with surface mount ceramic caps. The laminate is not very stiff, which means it can easily deform with 500G applied to it. And the longer the component is in the direction of bending, the greater stress it experiences. And ceramic caps are not very resistant to stress... They simply crack. I would refrain from using anything which is not on some for of leads. Leaded SMD packages are ok, but chip capacitors and resistors are not necessarily a good thing, that is unless you have the pcb screwed to a stiff backing plate of some sort.

[Simon]

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

[KedasProbe]

The way it is mounted to the 500G object can also seriously reduce the G forces on the components. Like a low pass filter.

[Simon]

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