Solder dots around mounting hole - what are they for?

[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.