Solder dots around mounting hole - what are they for?

[eTobey]

Hi,

can anyone explain to me, what the purpose of these dots is? It is a mounting hole, and a screw will sit on them.

Cheers,

Toby

[Nominal Animal]

They're not dots, they're actually vias that happen to be filled with solder.

Their purpose is to strengthen the PCB, so that tightening the screw won't deform and/or crack the board.

[AndyBeez]

They also provide contact points. As this is ground, that looks like a chassis connection. Did the PCB have a metal washer or metal post here?

[wraper]

They're not dots, they're actually vias that happen to be filled with solder.

Their purpose is to strengthen the PCB, so that tightening the screw won't deform and/or crack the board.

There are not necessarily vias and actually placing vias under those is problematic as solder will escape to opposite side during reflow. They get crushed/worn instead of copper while tightening the screws and prevent screw from loosening as well. Generally it's used with self-locking screws which can easily destroy thin copper foil in a few screw-unscrew cycles.

[nctnico]

They're not dots, they're actually vias that happen to be filled with solder.

Their purpose is to strengthen the PCB, so that tightening the screw won't deform and/or crack the board.

No. Tin does nothing where it comes to adding mechanical strength. Wraper is 'on the money'. Putting vias under screws is a bad idea because they will deform and are likely to crack.  If you want to stitch ground planes together at a screw hole, the vias must be placed around the screw head but not so many that the strength of the board is weakened (like mouse-bites to seperate boards).

[thm_w]

I would guess these are not vias, just solderpaste as mentioned. Though some people do use vias as well.

I don't know if I buy the "vias increase strength" claim, without seeing some evidence presented.
The solder contact point claim, I see it working for those self-locking screws wraper posted. But for a regular smooth screw I can't see it being superior. I've had no issues with properly torqued screws directly on ENIG, for normal use cases.

I'm sure NASA has some recommendations for high vibration environments, if someone can find it.


https://electronics.stackexchange.com/questions/137394/mounting-hole-on-a-pcb (this one has an interesting point about using non-plated through holes to avoid metal shards)
https://electronics.stackexchange.com/questions/240092/why-do-some-pcbs-put-vias-around-mechanical-through-holes
https://electronics.stackexchange.com/questions/524791/why-do-some-pcbs-have-these-circular-rings-are-they-plated-through-hole-vias

[nctnico]

If you really want to get into the details to ensure good, long term contact, you need to take material compatibility into account as well. Probably most manufacturers / system assemblers won't care though for grounding purposes.

[wraper]

I've had no issues with properly torqued screws directly on ENIG, for normal use cases.

ENIG is not that common for mass produced devices except maybe smartphones and tablets. Usually it's either OSP or HASL depending on use case.

[thm_w]

I've had no issues with properly torqued screws directly on ENIG, for normal use cases.

ENIG is not that common for mass produced devices except maybe smartphones and tablets. Usually it's either OSP or HASL depending on use case.

So then for OSP you'd have to put solder dots as shown in OP to get proper connection.
HASL already has the tin there, wouldn't expect anything to be needed.

[Smokey]

I've had no issues with properly torqued screws directly on ENIG, for normal use cases.

ENIG is not that common for mass produced devices except maybe smartphones and tablets. Usually it's either OSP or HASL depending on use case.

Tell that to all the people using fine pitched parts that aren't tablets and smartphones.  You need to do a tradeoff analysis.  Is saving an extra $0.10 on the PCB worth having a higher % failure rate because of HASL assembly issues and tiny parts?

[wraper]

I've had no issues with properly torqued screws directly on ENIG, for normal use cases.

ENIG is not that common for mass produced devices except maybe smartphones and tablets. Usually it's either OSP or HASL depending on use case.

Tell that to all the people using fine pitched parts that aren't tablets and smartphones.  You need to do a tradeoff analysis.  Is saving an extra $0.10 on the PCB worth having a higher % failure rate because of HASL assembly issues and tiny parts?

Those people use OSP. And did you miss "depending on use case"?

[EPAIII]

So, no one really knows for sure.

[Smokey]

I've had no issues with properly torqued screws directly on ENIG, for normal use cases.

ENIG is not that common for mass produced devices except maybe smartphones and tablets. Usually it's either OSP or HASL depending on use case.

Tell that to all the people using fine pitched parts that aren't tablets and smartphones.  You need to do a tradeoff analysis.  Is saving an extra $0.10 on the PCB worth having a higher % failure rate because of HASL assembly issues and tiny parts?

Those people use OSP. And did you miss "depending on use case"?

I actually did miss that.   

[Smokey]

To get back on topic, It's pretty obvious the OP found PCB crop circles:

[Nominal Animal]

I see vias placed around mounting holes in most larger boards I have.  Structurally, they do seem to help avoid cracking due to the static forces exerted by the fastening screw: instead of behaving like a laminate, the region around the hole acts more like an uniform solid.

On PC motherboards, self-locking screws are typically used, but on all I've seen –– many –– the "dots" have been vias.  As the mounting hole distance is quite large (100mm or more), and forces relatively large (inserting connectors, expansion cards, memory modules take quite a bit of force, relatively speaking), the mounting holes get stressed laterally, in the plane of the board.  Without the vias, the board could crack from the mounting hole.  While the screws are not supposed to be very tight, just snug, overtightening a self-locking screw without the hole strengthened with vias could delaminate the board due to the excess pressure while twisting, too.

There is always a good chance I've understood the situation wrong, or that the vias around mounting holes are there just because "everyone does it this way", but I don't think either is the case here.

Edited to add: While many PCB design guides talk about guard vias around plated holes used for grounding (the board to chassis or similar, when screwed in), on many boards I have the mounting holes' guard rings are isolated and not grounded.

[eTobey]

They're not dots, they're actually vias that happen to be filled with solder.

How can you say that they are vias, when you actually dont know? Because they are not vias!

They also provide contact points. As this is ground, that looks like a chassis connection. Did the PCB have a metal washer or metal post here?

It had neither. Just a screw. On the other side to the chassis, there is copper.

Those two big vias connect to the other side with copper that is around the mounting hole.

I hate, that so many people do it because "everybody does it". This mentality is purely dangerous for society (as history proves) and does prevent improvement. Also its kinda stupid, doing something you dont know why.

If you say, that it would strengthen the PCB, then compare the properties of copper/solder vs. those of the PCB. Holes also rarely strengthen something. Protecting the copper layer or securing the screw, sounds more appropriate to me. I can also think of, that it could equalize any height differences.

[Nominal Animal]

How can you say that they are vias, when you actually dont know?

Because every time I've seen such "dots" –– on several dozens of different boards from PC and server motherboards, SBCs, and routers –– they have always been vias and not "solder dots".  A solder dot would make very little sense, as a crush washer would be better, simpler and cheaper.  (The simplest crush washer is simply a thin annealed copper washer when electrical connectivity is desired, and a teflon or nylon washer when not.)

Holes also rarely strengthen something.

Wrong.  Holes are an extremely important feature that can make a structure stronger than if it were solid.

The key is mechanical properties.  PCBs are laminates: layers of thin conductive metal and bulk material; for FR4, glass fiber reinforced epoxy.
When the screw is tightened, the glass fiber reinforced epoxy (or comparable substrate material) will contract along the axis of the screw, and enlarge in the plane of the PCB.  Without the holes, the expansion can cause fractures –– because the conductive copper layers break before they stretch ––, even delaminate the structure, and break the board.  The via holes provide room for expansion of the epoxy material, allowing local deformation around the mounting hole, without crack formation; plus the plating material acts a bit like glue helping keep the layers from delaminating.

[DavidAlfa]

They provide mechanical relief (Act as a cushion when tightening), protecting the copper layer while also enhancing the electrical connection to the chassis, as these solder dots will squeeze and make a much better contact than plain copper.
Anyways, modern computers's nonsense aesthetics like to paint the screws and the case, so they actually make no contact at all.

[Nominal Animal]

Just as a data point:

If you have a standard 1.6mm thick PCB, and you use a M3 or #4-40 mounting screw, starting from just snug or making a contact but exerting no force, a single quarter turn reduces the PCB thickness by 7.8% and 10%, respectively (pitches being 0.5mm and 1"/40).  Decide for yourself if you think that matters or not.

[nctnico]

And how much torque is needed to turn a screw that much? Won't it snap off before that? There is a reason why there are torque specifications and not rotation specifications for fasteners (unless you are using stretching bolts but that is a different story).

There is an interesting discussion here:
https://www.reddit.com/r/engineering/comments/12csy9j/setting_torque_values_based_on_clamped_material/

[Nominal Animal]

And how much torque is needed to turn a screw that much? Won't it snap off before that?

For FR4, no.  It just deforms.  I don't have a torque wrench to measure, but it's what I can do by hand when extending the screwdriver in the same direction as my arm to M3 or #4-40 with a Philips head, i.e. using my wrist only.

It's easy to verify for yourself, no need to take my word for it.

[nctnico]

And how are you sure you aren't stretching the screw + thread? The link I provided shows some good starting points for going from compressive strength versus tensile strength of a screw.

Going by the 350N/mm2 number from that link, an M3 screw with a head diameter of 6mm and a hole size of 3.2mm should stay below a total compressive force of 7000N (20mm2 * 350N) before the FR4 gets damaged. A mild steel (4.6 class) M3 screw is rated for a tensile strength of 2100N. IOW: You'll snap a mild steel M3 screw long before damaging FR4.

[eTobey]

A solder dot would make very little sense, as a crush washer would be better, simpler and cheaper.

Thats just wrong, as the dot comes with no additional cost. You cant go cheaper than for free ;-)

When the screw is tightened, the glass fiber reinforced epoxy (or comparable substrate material) will contract along the axis of the screw, and enlarge in the plane of the PCB.  Without the holes, the expansion can cause fractures –– because the conductive copper layers break before they stretch ––, even delaminate the structure, and break the board.  The via holes provide room for expansion of the epoxy material, allowing local deformation around the mounting hole, without crack formation; plus the plating material acts a bit like glue helping keep the layers from delaminating.

I may make some cookies, to see if holes really add structural strength ;-) Or maybe i just handle them appropriately and eat them, to not make a mess.

I just found this picture in google:
My guess is, it should make contact with the screw, but the soldermask has to stay (dunno why).

[eTobey]

If you have a standard 1.6mm thick PCB, and you use a M3 or #4-40 mounting screw, starting from just snug or making a contact but exerting no force, a single quarter turn reduces the PCB thickness by 7.8% and 10%, respectively (pitches being 0.5mm and 1"/40).

Arent those vias just underneath the screwhead? Wouldnt they then become rather useless?

[Nominal Animal]

If you have a standard 1.6mm thick PCB, and you use a M3 or #4-40 mounting screw, starting from just snug or making a contact but exerting no force, a single quarter turn reduces the PCB thickness by 7.8% and 10%, respectively (pitches being 0.5mm and 1"/40).

Arent those vias just underneath the screwhead? Wouldnt they then become rather useless?

Yes.  No, because the epoxy material doesn't compress, it deforms.  Think of clay, play-doh, or dough, and not foam.  Instead of expanding in every direction in the plane of the PCB and potentially causing the damage I described (especially to inner conductive layers up to an inch away, say), most of the expansion happens into the via holes, with the plating material acting a bit like glue, helping avoid delamination there.

The reason it makes the PCB stronger is that it allows larger and more varied dynamic forces to elastically deform the PCB near the mounting hole without propagating the deformation further; keeping the deformation localized there.

Like I said, there is no need to take my word for it: this is easy to test with an unneeded board.  If this is something that really interests someone, they could design a multilayer board with comb-like capacitive stress measuring structures and various types of mounting holes placed symmetrically near such structures, and test it.  One could manufacture a set of five for just a few dollars at JLCPCB even with six layer boards right now.  My claim is simple: with vias within the screw head area, at any given screw torque the PCB stress/deformation is minimized, compared to similar mounting holes without vias, or with solder dots.

The image you showed of small "pads" around the mounting hole is new to me; I've never seen that in real life.  But plenty of the via kind.  I'm not an EE myself, but my background is in physics (mostly HPC and computational materials physics; not this kind of mechanical stress stuff, but atomic level stuff) and before that, in custom full-stack development.

[eTobey]

The reason it makes the PCB stronger is that it allows larger and more varied dynamic forces to elastically deform the PCB near the mounting hole without propagating the deformation further; keeping the deformation localized there.

Are there actually some theories or research about this, or is it just your idea of how it works? I really am surprised, that there is not much information about this aout there.

On my first picture, its easy visible, that the solder did not get squashed too far, and i cant see how this would affect the pcb at all. What i can think of is, that in the early days, the copper ring got detached from the PCB, so they put solder dots on it. And then there was multilayer and vias...

[Nominal Animal]

The reason it makes the PCB stronger is that it allows larger and more varied dynamic forces to elastically deform the PCB near the mounting hole without propagating the deformation further; keeping the deformation localized there.

Are there actually some theories or research about this

Sure.  Talk to any mechanical engineer with related experience.

[wraper]

The reason it makes the PCB stronger is that it allows larger and more varied dynamic forces to elastically deform the PCB near the mounting hole without propagating the deformation further; keeping the deformation localized there.

Are there actually some theories or research about this

Sure.  Talk to any mechanical engineer with related experience.

First you say a load of bull about vias underneath solder dots, then send to talk to mechanical engineer. Very informative argument

[wraper]

Here is PCB from Phillips TV. Via stitching in between of solder dots but not underneath them. Also you don't want to deform material where holes are located. They can become nice fracture starting points. Arguing that holes in compression area will somehow make the material stronger is pure nonsense. Not to say do not forget FR-4 is not a homogenous material but a composite - epoxy impregnated glass fiber. By creating holes you interrupt glass fibers which bear the strength in material.

[2N3055]

Just to confirm what was stated above, FR4 will plastically deform under pressure. An elastic washer or elastic material washer is recommended.

If you mount a heatsink through PCB, you are in world of trouble if you don't provide spring to keep constant pressure on heatsink. Heat makes creep even worse..

Solder dots are there to provide electrical contact.
Via stitching is sometimes provided to ensure good GND low impedance contact for inner layers to protective case..

I don't see copper adding much mechanical strength to glass fiber composite. It might help alleviate some creep on surface. But copper is quite soft...

[Nominal Animal]

The reason it makes the PCB stronger is that it allows larger and more varied dynamic forces to elastically deform the PCB near the mounting hole without propagating the deformation further; keeping the deformation localized there.

Are there actually some theories or research about this

Sure.  Talk to any mechanical engineer with related experience.

First you say a load of bull about vias underneath solder dots, then send to talk to mechanical engineer. Very informative argument

Bull?  [Citation needed].

Also, why does your following post then support my argument?  Can't you make up your mind about whether I'm talking bull, or are you confused as usual?

[wraper]

The reason it makes the PCB stronger is that it allows larger and more varied dynamic forces to elastically deform the PCB near the mounting hole without propagating the deformation further; keeping the deformation localized there.

Are there actually some theories or research about this

Sure.  Talk to any mechanical engineer with related experience.

First you say a load of bull about vias underneath solder dots, then send to talk to mechanical engineer. Very informative argument

Bull?  [Citation needed].

Also, why does your following post then support my argument?  Can't you make up your mind about whether I'm talking bull, or are you confused as usual?

It supports none of what you said. There are no vias under solder dots, I scraped off one of them to show it. Via stitching is in places in between of the dots and do not get compressed by the screw. Looks the same as via stiching all along GND plane edges which helps preventing unintended antennas. As of citations, you so far provided exactly zero which would even slightly support your claims.

[Nominal Animal]

Here is a more detailed reply, in case my previous one seems snarky.  (It wasn't intended as such, as you'll see.)

Are there actually some theories or research about this

This is governed by mechanical physics.  The glass fiber-reinforced epoxy (especially FR4) is a viscoelastic substance, and its behaviour (especially under thermomechanical stress, i.e. expanding and shrinking when heated and cooled) is widely researched, but there do not seem to be any peer-reviewed published research as to what kind of mounting hole patterns yield the best results against damage due to mechanical forces.

(In particular, the way through holes are plated with copper seems to yield ductile copper in vias and PTHs; as opposed to the non-ductile copper foil in the conductive layers, as described in some of the articles you can find using Google Scholar.)

Thus, for a scientific explanation and references, you do need to go to a mechanical engineer with the experience and (simulation) tools in this area: a comparison model of a board with two through holes, one with vias, the other without, and a few runs of different forces acting on the holes (the same force for each hole, of course), simulated using finite element analysis software and existing FR4 models (especially the epoxy is tricky to get right), will show the exact situation.
(My claim: With vias, the stress is concentrated very close to the hole and the vias, and drops off much quicker as a function of distance from the mounting hole, compared to without.)

As to vias helping strenghten mounting holes, this came up here at EEVBlog forums a decade ago, when steaky1212 found that suggestion by Tom Hausherr (PCB Libraries author) on the Mentors website (Wayback Machine archived page).  Quoting that blog:

There are two primary reasons for adding vias to the supported mounting hole. The first was to insure that if the screw threads stripped the copper plating from the main hole that the vias would still provide adequate ground connections. The second reason was for additional support to prevent the PCB from crushing when too much torque was used to tighten the nut.

So no, it is not my idea.  The explanation of how it works I wrote above is based on my understanding of the mechanics of the situation, trying to sketch out the mechanisms involved.  I do now realize that effort was silly: I should simply have linked to that Wayback Machine blog article, because argumentum ab auctoritate is more acceptable to most members here than reasoning based on physics.

[eTobey]

Sure.  Talk to any mechanical engineer with related experience.

That does not answer my question.

And any stress would IMHO just move around the hole. For your theory to work, it should be a groove around the hole. But that itself would cause to form a crack... And the mounting hole itself would be enough space for the material to go on compression.

[Nominal Animal]

For your theory to work, it should be a groove around the hole.

No, because the glass fiber-reinforced epoxy material (between the copper layers) is viscoelastic.  It is obvious you're not going to trust anything I claim, so the suggestion to talk to a suitable engineer is sensible.

[Siwastaja]

So no, it is not my idea.

Certainly not. To me, this was always common sense; the idea of material expanding under pressure, and holes adding places where to expand to, minimizing the risk of nearby copper tracks stretching to breakage, has been totally intuitive to me, to the point I never thought about it much and just copied the usual pattern of just adding those vias; I can confirm having seen it in gazillon of PCBs. It's worth noting though that filling the vias with solder would compromise this property and probably cause more problems instead, so that would explain why wraper sees the "solder dots" and vias at different places.

I think DavidAlfa said it well, too. I really don't understand why you are getting so much backlash. To me all you write is obvious.

[nctnico]

That is more like copying something everybody else is doing rather than looking at material engineering.

This page shows a stress / strain curve for some FR4 material:
https://www.researchgate.net/figure/Stress-strain-curve-for-FR4-PCB-material_fig4_4058600

This page explains what such a curve means:
https://www.xometry.com/resources/3d-printing/stress-strain-curve/

So the strain is the amount deformation according to the stress level. The curve shows a maximum deformation of 1.8% at the point of failure which does not look like the properties of a material that 'flows'. The curve for FR4 resembles the curve shown for a brittle material. Which is logical if you look at what FR4 is made of: glass fibers glued together with epoxy resin. When you put pressure on FR4, you are pulling on the inner glass fibers which will snap at some point. As wraper already noted: If you are putting vias under a screw head, these vias are breaking the weave pattern and thus weakening the board at that point.

The only reason I see for putting vias under a screw head is to ensure contact between the screw and all ground layers in a board without needing too much space. For a mild steel M3 screw this is probably just fine (assuming there are only a few vias) as you can't really damage a board with a mild steel M3 screw (see my previous post for the calculation). For bigger screws it might be wise to place the vias outside the area of the screw head in order to prevent cracking the vias and increase strength of the board under the screw head.

[Nominal Animal]

This page shows a stress / strain curve for some FR4 material:
https://www.researchgate.net/figure/Stress-strain-curve-for-FR4-PCB-material_fig4_4058600

Read that paper! That curve is derived from impact tests, not from bending or stretching.  For a viscous or viscoelastic material, the stress is directly proportional to the rate of strain, and not to strain itself like for elastic materials, for small deformations.  Thus, bending/stretching differs wildly from dropping and impact behaviour.

You'll get better understanding of the topic from e.g. Characterization of the material properties of two FR4 printed circuit board laminates by Haugan and Dalsjø; the full PDF article is open access.

[SiliconWizard]

The only reason I see for putting vias under a screw head is to ensure contact between the screw and all ground layers in a board without needing too much space.

Yeah I think so too. Not convinced about the benefits against stress.
Now using several "dots" instead of a solid pad probably helps preventing the screw from getting loose over time to some degree.

[Siwastaja]

Especially when it comes to PCB design, we all "copy" things we don't exactly and perfectly understand. If we didn't, no one could do anything (the classical trap of perfectionism). There's still a long way to cargo cult engineering, which means copying basically everything with very little understanding about anything to the point of creating an unworkable design.

So far, I'm unconvinced about the scientific evidence of anything regarding those holes. Half of the people here are saying they make the PCB stronger (specifically against copper track breakage) and others say they make it weaker (by cutting the glass fibers of the FR4 itself). If the latter is the case, I wonder why having them is so common in professionally designed products*. On the other hand, not having the vias seems quite common, too; then the question is, are they "left out" because the designer thinks they are just unnecessary, or because they think they are detrimental? A big difference.

*) and ground plane stitching is not the explanation; for that, the holes could be further away, not under the screw head

Probably this is one of the cases where it usually does not matter either way.

[nctnico]

It is an interesting subject. I would like to see some numbers about how much FR4 'flows' though. I have not been able to find those -yet-. I doubt it is a large amount though otherwise screws would get lose over time by themselves. And the 'flow' has to be elastic otherwise you'd get a permanent dimple in the board (I don't recall seeing that ever). Keep in mind that the screw itself acts as a spring because it extends when tightened.

I have seen quite a few high current applications where wire lugs or busbars are bolted to an FR4 board directly. So it looks like this is accepted as a reliable method to get a lot of current in & out of a board.

[nctnico]

You'll get better understanding of the topic from e.g. Characterization of the material properties of two FR4 printed circuit board laminates by Haugan and Dalsjø; the full PDF article is open access.

Sorry, but like many other reports, this is about the 'glass transition phase' at high temperatures to research the effect of the soldering and the effect of high temperatures on board stiffness. It doesn't provide any information about effects of putting local compression stress on FR4. Also the fact that the report was written as part of a summer internship makes it just another collection of datapoints.

[Nominal Animal]

You'll get better understanding of the topic from e.g. Characterization of the material properties of two FR4 printed circuit board laminates by Haugan and Dalsjø; the full PDF article is open access.

Sorry, but like many other reports, this is about the 'glass transition phase' at high temperatures to research the effect of the soldering and the effect of high temperatures on board stiffness. It doesn't provide any information about effects of putting local compression stress on FR4. Also the fact that the report was written as part of a summer internship makes it just another collection of datapoints.

Like I wrote earlier, I cannot find any peer reviewed articles about mounting stresses on FR4 or similar materials for different styles of mounting holes.  For someone who does not even recognize the difference between elastic and viscoelastic materials, that PDF is a good starter on how these things are measured, giving you a better understanding of the topic at hand, instead of picking individual sentences or images at random that you think bolsters your argument.

If you are hoping for a good, possibly peer-reviewed article on FR4 mounting stresses, there do not seem to be any; only "anecdotal" evidence from PCB designers having done stuff for decades, and "understanding" based on materials physics (viscoelasticity) and mechanical engineering (controlling and localizing strain under various stresses using physical structures, i.e. shape).

The next best thing, in my opinion, would be a Finite Element Analysis with a few different forces, both on the mounting hole itself (corresponding to forces in the PCB plane) and thus with multiple different hole models, and on the PCB (perpendicular) with different holes symmetrically placed around the stress point in typical rectangular configuration.  This yields images where the resulting strain is easy to see; compare to this one[/ULR].
(That [URL=https://www.researchgate.net/publication/337968594_Analysis_of_Printed_Circuit_Boards_strains_using_finite_element_analysis_and_digital_image_correlation]publication is open access, comparing digital image analysis of real boards to finite element simulations.  So even that one is just an example of the type of such investigation, and does not answer the question at hand.)

A set of boards with strain rosettes and different mounting hole styles for measuring the actual deformation would be easy to manufacture, but even with minimum trace widths, you'd need some high-resolution test equipment to measure.  You'd also want to have a good microscope to examine the results of a crush test, say with thin angular (not perpendicular nor tangent to the hole) traces near the screw head just to make observing the effects easier.
Using minimally thin test traces in inner layers close enough (too close!) to the mounting holes should show breakage at high enough deformations, but I suspect one one would have to test batches of boards, dozens of mounting holes of each type, to get any kind of meaningful results.
Even if I did such a test and posted the results here, most of you would not believe me anyway; and it would be utterly wasted effort for everyone.

Thus, the only real possibility is for someone else to investigate and report.

I could be wrong here, even if I don't think so.  My own opinion is based on relatively simple physics I've explained already, but seems to be countered with "I feel you're wrong" -type truthiness without any basis, which annoys me to no end.  I'm done arguing from simple physics principles against vague gut feelings.

[Siwastaja]

I have seen quite a few high current applications where wire lugs or busbars are bolted to an FR4 board directly. So it looks like this is accepted as a reliable method to get a lot of current in & out of a board.

This is also something you equally often hear described as unreliable; even something you should "never" do. What are the exact conditions where it would fail, beats me. I'm sure it's OK when not overtightened but what's the maximum acceptable torque on typical materials? I guess no one really knows. For the record, I have done that, too, and never had problems, but I still feel a bit uncomfortable knowing that maybe someone who repairs the thing seriously overtightens the bolts to the point of causing irreversible damage to FR4, beyond the elastic range, and with say 8mm bolt with 13mm hex nut this could really happen; compare it to a, say, M2.5 screw with PH0 head used to mount a PCB to a plastic case, it would be impossible to overtighten that.

[thm_w]

I have seen quite a few high current applications where wire lugs or busbars are bolted to an FR4 board directly. So it looks like this is accepted as a reliable method to get a lot of current in & out of a board.

This is also something you equally often hear described as unreliable; even something you should "never" do. What are the exact conditions where it would fail, beats me. I'm sure it's OK when not overtightened but what's the maximum acceptable torque on typical materials? I guess no one really knows. For the record, I have done that, too, and never had problems, but I still feel a bit uncomfortable knowing that maybe someone who repairs the thing seriously overtightens the bolts to the point of causing irreversible damage to FR4, beyond the elastic range, and with say 8mm bolt with 13mm hex nut this could really happen; compare it to a, say, M2.5 screw with PH0 head used to mount a PCB to a plastic case, it would be impossible to overtighten that.

Probably should never do in a very high reliability product (military, avionics). Otherwise, most applications I would expect to be OK, I agree. Creep could come into play at high temps and pressures, I don't know what a normal SS 304 M3 or M4 screw would apply pressure wise.. you'd have to do some calculations. The max allowable torque for A2-70: M3 1Nm, and M4 2.6Nm. They will shear off around twice that. 3Nm already feels like too much for the M4 on FR4.

PCB PEM inserts are commonly used on high end gear, which rely on the strength of the PCB to stay in place. They mention that the hole for swaged (press fit) does not need to be plated, so plating probably does not affect strength. They don't mention vias near the hole, just board edge distance. I'm sure if you asked them about putting a ton of vias around the hole they would recommend against it.

https://www.hindawi.com/journals/ijps/2022/9733138/
https://4spepublications.onlinelibrary.wiley.com/doi/10.1002/pc.10130
https://www.researchgate.net/publication/268209268_Creep_Behavior_in_Fiber-Reinforced_Epoxy_DGEBA_Composites

[KE5FX]

It is an interesting subject. I would like to see some numbers about how much FR4 'flows' though. I have not been able to find those -yet-. I doubt it is a large amount though otherwise screws would get lose over time by themselves.

It's hard to generalize about "FR4" other than to say that it is made of some sort of fire-retardant material with an E_r somewhere near 4.  But I've seen a lot of TO-3 mounting screws that have loosened over time.  Far too many to think that most PCB laminates -- or more likely, prepregs -- don't cold-flow to some extent.

[nctnico]

It is an interesting subject. I would like to see some numbers about how much FR4 'flows' though. I have not been able to find those -yet-. I doubt it is a large amount though otherwise screws would get lose over time by themselves.

It's hard to generalize about "FR4" other than to say that it is made of some sort of fire-retardant material with an E_r somewhere near 4.  But I've seen a lot of TO-3 mounting screws that have loosened over time.  Far too many to think that most PCB laminates -- or more likely, prepregs -- don't cold-flow to some extent.

Interesting observation / good data point for sure. But it makes me question whether temperature and temperature cycling also have an influence on the screws getting lose. Or put differently: how can we be sure the flow of the laminate is the only factor at play in such a situation?

[eTobey]

But I've seen a lot of TO-3 mounting screws that have loosened over time.  Far too many to think that most PCB laminates -- or more likely, prepregs -- don't cold-flow to some extent.

If you have mostly seen TO-3 screws coming lose, then it might be, that those joints were a bit different. Also take into consideration, that every contact surface pair will set (? - i dont know the exact translation) and reduce the tension of the joint.

Anyone has a good connection to Dave? He could make some tests with his vibration device!

[2N3055]

It is an interesting subject. I would like to see some numbers about how much FR4 'flows' though. I have not been able to find those -yet-. I doubt it is a large amount though otherwise screws would get lose over time by themselves.

It's hard to generalize about "FR4" other than to say that it is made of some sort of fire-retardant material with an E_r somewhere near 4.  But I've seen a lot of TO-3 mounting screws that have loosened over time.  Far too many to think that most PCB laminates -- or more likely, prepregs -- don't cold-flow to some extent.

Interesting observation / good data point for sure. But it makes me question whether temperature and temperature cycling also have an influence on the screws getting lose. Or put differently: how can we be sure the flow of the laminate is the only factor at play in such a situation?

They do (temp and vibration) There was a Motorola AN on the topic.
I troubleshooted a batch of PSU boards with this problem long time ago..

[VK3DRB]

Having solder on these vias is pure stupidity.

Solder will deform over time with pressure and the screw head may well have have a poor electrical contact or come loose. This is the exact reason you should never, NEVER , solder wires prior to inserting them in the screw holes where there is fire or other safety risks, such as mains wires on the back of a general power outlet.

The area on the PCB should be ENIG that is bigger than than the screw head diameter, and applicable creepage and clearances must be observed. For PE contacts, relevant regulatory standards must apply.

Don't ever assume PCB's laid out by big "professional" corporations do a good job.

[VK3DRB]

It is an interesting subject. I would like to see some numbers about how much FR4 'flows' though. I have not been able to find those -yet-. I doubt it is a large amount though otherwise screws would get lose over time by themselves.

It's hard to generalize about "FR4" other than to say that it is made of some sort of fire-retardant material with an E_r somewhere near 4.  But I've seen a lot of TO-3 mounting screws that have loosened over time.  Far too many to think that most PCB laminates -- or more likely, prepregs -- don't cold-flow to some extent.

That's because they are mounted poorly. Heat/cold cycles can loosen screws that don't have appropriate spring washers on them over time. Loctite can also prevent movement.

[wraper]

That's because they are mounted poorly. Heat/cold cycles can loosen screws that don't have appropriate spring washers on them over time. Loctite can also prevent movement.

Spring washers are among the most useless things that can be used for prevent screw from loosening though. https://ntrs.nasa.gov/citations/19900009424

The typical lockwasher serves as a spring while the bolt is being tightened. However, the washer is normally flat by the time the bolt is fully torqued. At this time it is equivalent to a solid flat washer, and its locking ability is nonexistent. In summary, a lockwasher of this type is useless for locking.

[thm_w]

Having solder on these vias is pure stupidity.

Its not vias, its just a pad, and it could be necessary in this case as the board appears to be OSP, as we discussed above.

That's because they are mounted poorly. Heat/cold cycles can loosen screws that don't have appropriate spring washers on them over time. Loctite can also prevent movement.

Spring washers are among the most useless things that can be used for prevent screw from loosening though. https://ntrs.nasa.gov/citations/19900009424

The typical lockwasher serves as a spring while the bolt is being tightened. However, the washer is normally flat by the time the bolt is fully torqued. At this time it is equivalent to a solid flat washer, and its locking ability is nonexistent. In summary, a lockwasher of this type is useless for locking.

They work better than a plain washer, that NASA quote is a bit deceiving. They still don't work that well though (skip to 1:40):

[SiliconWizard]

That's because they are mounted poorly. Heat/cold cycles can loosen screws that don't have appropriate spring washers on them over time. Loctite can also prevent movement.

Spring washers are among the most useless things that can be used for prevent screw from loosening though. https://ntrs.nasa.gov/citations/19900009424

Nice reading. Well recommended.

[VK3DRB]

That's because they are mounted poorly. Heat/cold cycles can loosen screws that don't have appropriate spring washers on them over time. Loctite can also prevent movement.

Spring washers are among the most useless things that can be used for prevent screw from loosening though. https://ntrs.nasa.gov/citations/19900009424

Nice reading. Well recommended.

Good doco. They mentioned nyloc nuts which have the greatest efficacy. It is a safety requirement in Australia that earthed chassis terminals have a nyloc nut securing the earth wire to the chassis and NO star washer between the lug and the chassis. I have also used nylocs on circuit boards where something must frame earthed.

Another point is that over time, metallic bonds are formed on threads and other mating metals over time, making them more diffucult to remove or shake loose.

A bloke came around here to buy my 30-year old 7' x 4' trailer earlier this year. The trailer had an oversized ball, which prevented theft. So he had to remove the 2 inch ball from his 4WD and put my 2.5 inch ball on. Trouble is the old ball had been left untouched in his 4WD for 25 years. To remove it, no amount of WD40 could shift it. I did not want to use a blow torch near a fuel tank or the greasy underside of his vehicle. A large pipe wrench with an extension handle could not budge it. What worked was putting the pipe wrench on the nut and backing the 4WD into a large power pole. The heavy duty pipe wrench handle was actually flexing. Eventually it released with an almighty BANG. Looking at the threads, I don't think rust played a big part. It was more likely the metallic bonds in the thread which had formed over 25 years. Anyone who has left metallic gauge blocks up against each other for a few weeks will know how strong these bonds can become.

[wraper]

They work better than a plain washer, that NASA quote is a bit deceiving. They still don't work that well though (skip to 1:40)

Initially it lost tension almost as fast as with plain washer, it's only total loss of tension that took a bit more time. So it does not prevent unscrewing any better than plain washer, it just delays total failure.

[Someone]

It is a safety requirement in Australia that earthed chassis terminals have a nyloc nut securing the earth wire to the chassis and NO star washer between the lug and the chassis.

Citation of that requirement? Doesnt seem to appear in the usual suspects.

[thm_w]

They work better than a plain washer, that NASA quote is a bit deceiving. They still don't work that well though (skip to 1:40)

Initially it lost tension almost as fast as with plain washer, it's only total loss of tension that took a bit more time. So it does not prevent unscrewing any better than plain washer, it just delays total failure.

Delaying failure means its better than a plain washer... Everything in the test other than the nordlock delayed failure, including the nyloc.

Of course this test was optimized to some level of vibration, and in other scenarios you won't have as much as this test, and the split washer may survive while a plain washer would fail.

[nctnico]

IMHO the Nord-lock video is doctored. Just look at a car and how many bolts are put into the engine block alone without a washer and they won't run lose when torqued (and some bolts do not require a lot of torque; like 40Nm to 60Nm ) to spec. It is just a matter of good engineering. What kind of worried me on my previous car where the 6x M6 bolts holding the crank shaft pulley in place while needing like 40Nm of torque on each bolt. They never ran lose though.

[thm_w]

Its an official test: DIN 65151, the bolt is hit directly with shear load vibration. They show initial preload is 75Kn, I guess that is M12 bolt? then it would be 180Nm of tightening torque.

If you have a setup with enough clamping force to prevent sliding movement, the bolt shouldn't easily loosen: https://www.boltscience.com/pages/vibloose.htm
In the case of your crank shaft pulley, I assume its balanced well, and the forces are opposing each other, that the shear load on those bolts is minimal. But yeah, not all automotive fasteners will have loctite or washers at all, surprising to me too.

Here is another example:

[eTobey]

I found a PCB, where there where these solderblobs around the mounting hole. And it was covered in a silicone film. That looks like they are used to pierce through the silicone  (hopefully) to make contact. Thats not really a great way to do it IMHO.