4

I recently had surface mount inductor fall off my board due to vibration fatigue. The lead actually broke off on the component where it enters the gray material. https://www.digikey.com/en/products/detail/bourns-inc/SRP1265A-150M/4876713

SMD inductor

This particular inductor is quite heavy, so I am looking for a lighter version, but wondering if there is another option for ruggedizing the heavy components on my board. It is in a very rough/vibey environment, and it needs to be heatsinked to the housing so I can't rubber mount it.

Is there an epoxy that is commonly used? Or a method of board design or component selection to prevent issues like this? I did some research and I'm unclear if you put epoxy under the component or over the top after reflow. Thanks guys.

Transistor
  • 168,990
  • 12
  • 186
  • 385
kobra20
  • 51
  • 4
  • 2
    whoa you broke off an SMD? was board flex involved? – Pete W Jan 04 '21 at 21:34
  • If it must be mounted to a heat sink that is then mounted to the chassis or something like that, I'd say increase the rigidity of the PCB attachment to that (2-4 additional PCB mount points immediately surrounding the vulnerable part). The usual methods like glue gun, cable tie, retaining washer etc, are for "free standing" items. – Pete W Jan 04 '21 at 21:45
  • 1
    If you securely attach it to the housing then it is more likely to suffer stress, if only due to differential thermal expansion of the housing and the PCB (also, any physical shock to the housing would also be transmitted directly to the inductor). Have you considered using a thermal pad between the inductor and the housing? – Andrew Morton Jan 04 '21 at 21:59
  • There is staking compound which is usually a silicone caulk-like adhesive. It's only applied between the side of the part and the board so you should be able to get good thermal contact to the top of the package. If you haven't finalized the PCB you might be able to split the inductance between two or more smaller parts. – vir Jan 04 '21 at 22:07
  • I'm with Andrew. Soft thermal pad to the cooling plate. Also check WHY you need cooling. I had almost the same Bourns component, where selecting a different type of winding lowered the heat loss considerably. On automotive and military PCBs I have worked with sometimes high (long) capacitors were glued together. But these low and flat inductors were never glued. Maybe have a look at how the satellite builders do it? Usually I think they have the whole PCB mounting dampened. – Dejvid_no1 Jan 04 '21 at 22:07
  • Maybe you need another mounting point near the inductor to prevent board flex. In high vibe environments the more mounting points the merrier! You can also do staking, which your assembly house should know how to do. – Aaron Jan 04 '21 at 22:18
  • Neutral-cure (non-acetic acid) RTV silicone is used for staking. Epoxy is too stiff. And actually, most silicones are too soft. The silicone made specifically for this purpose costs an arm and a leg. – DKNguyen Jan 04 '21 at 22:20
  • So it sounds like board flex is a major contributor? That's interesting because it's mounted to a flat aluminum surface with 100% surface contact and shouldn't really be moving around. With regards to the heat sinking - it's not the inductor that actually gets hot, it's the nearby IC, and a board-mounted thermistor that measures housing temperature by conduction to the housing. I suppose I could rubber mount it, use a thermal pad for the IC, and then use a thermistor on wire leads to the housing. It would just be a lot easier to have it all contained to one board. – kobra20 Jan 04 '21 at 22:27
  • For high vibration, reduce board flex and bond parts to board. Thicker boards, more mounting points, ribs to add support in the middle. Use fillet bonds on the sides without leads. Bonding material has to be slightly flexible so it won't crack. – Mattman944 Jan 04 '21 at 22:42
  • Was this for automotive use or aerospace? We used to test to 15g vibe 100 g shock – Tony Stewart EE75 Jan 04 '21 at 23:03
  • Automotive - motorcycle actually. Mounted to a chassis component that saw engine vibration transferred to it. So vibrations range from 15 to 1000Hz depending on engine configuration. – kobra20 Jan 04 '21 at 23:09

1 Answers1

3

The industry standard for adhering heavy parts from Vibe fatigue is polyurethane adhesive. Often you see it in white but it comes in various low VOC densities and colours but has a long cure time so they may be using UV Cured now.

Look for specs similar or better to this:

Shear Strength: 145 psi
Tensile Strength: 580 psi
Shore Hardness: 55 Shore A
enter image description here 50% strength 24 h Curing time should not pose a serious problem after final assembly.

Examine any PC PSU and you will see this PU adhesive. I do not have any Asian sources where it is mainly used. So I use architectural subfloor adhesive which is very strong plastic yet never brittle and sticks to all component surfaces. It doesn't take much to secure the part from vibration around the base unless it is also a tall heavy part.

Le Pages Fast Grab PL Polyurethane comes in 2 tube sizes for hand pumps in any hardware store.

RTV Silicone is too pliable for SMT but may be adequate for THT.

enter image description here

up to 1000 psi strength ( depends on substrate)

Tony Stewart EE75
  • 1
  • 3
  • 54
  • 182
  • Most silicone is too pliable, but there is much stiffer silicone marketed specifically for PCB components. You won't find it in a hardware store. And it usually comes in a tiny tube because it is so expensive. RTV162, for example. is Shore A 35. – DKNguyen Jan 04 '21 at 22:44
  • 35 Shore A is soft. I'm talking about medium hard >50 – Tony Stewart EE75 Jan 04 '21 at 22:50
  • this one is only 40 3M Scotch-Seal 540 gray polyurethane adhesive sealant is compatible with plastics, frp, smc, metals and wood materials with a 24 hr cure time. Delivers great performance with tensile strength of 300 psi – Tony Stewart EE75 Jan 04 '21 at 22:51
  • this one is white 55 Shore A https://www.rshughes.com/p/3M-560-Polyurethane-Adhesive-Sealant-White-Paste-10-5-Fl-Oz-Cartridge-49056/021200_49056/ Le Pages PL is more like 70 to 100. Although the old high VOC PU were the best and dried instantly and also stop subfloor squeeks too – Tony Stewart EE75 Jan 04 '21 at 22:53
  • I think there's a point where it gets too hard though. Dont know what that point is though. – DKNguyen Jan 04 '21 at 22:56
  • the point is when it breaks internally due to stress/strain like epoxy is too brittle and poor adhesion yet Sumutomo epoxy is perfect for sealing gas/ moisture tight and rated for moisture proof below freezing, whereas clear LED epoxy is not. Basically it must be many times stiffer than FR4 but not as brittle as epoxy – Tony Stewart EE75 Jan 04 '21 at 22:58
  • I would think the stiffer the better as long as you don't get a fracture? – kobra20 Jan 04 '21 at 23:10
  • @kobra20 Shrinkage while curing is an issue. – DKNguyen Jan 04 '21 at 23:50
  • Yes @kobra20 and PU is the right material. I once used it to repair a thin CPU plastic base that broke in the clamp. No shrinkage issues – Tony Stewart EE75 Jan 05 '21 at 00:21
  • @TonyStewartSunnyskyguyEE75 So do I just build by board normally then "pot" it in this PU material? Or am I to carefully put it underneath the component before soldering? – kobra20 Jan 05 '21 at 01:45
  • 1
    @kobra20 just add around the joints between top and PCB as thick as you like, but that reduces repair ability – Tony Stewart EE75 Jan 05 '21 at 02:40