What Is A Good Safety Factor For Threaded Rod

[bostonman]

McMaster-Carr lists stainless steal 3/8" 24tpi threaded rod as having a tensile strength of 70,000psi.

Without a safety factor, I'm calculating it can support a weight of 6,148lbs. Obviously this is the maximum, so I'm wondering if a typical safety factor exists.  From reading, I'm seeing 50%, so the calculation would be based on 35,000psi, but wondering if this isn't enough of a safety margin.

[bdunham7]

Is that the yield strength or ultimate tensile strength?  Can you link the description?

The selection of a safety factor would depend on the use and I suppose you'd have to consult engineering tables from whatever field governs your use.  Keep in mind that if whatever it is that you are fastening moves around or if the force has any shear or bending component, you'll likely need to increase that safety factor.  A clamping load where the fastener is very stable is a different application than suspending something, like hanging equipment from Uni-strut or something like that.

[CatalinaWOW]

Safety factors are chosen to cover all of the unknowns.  In some situations 5% is the right answer. 

Your situation has too many unknowns to use safety factors that small, but far fewer unknowns than many other situations.  You have an environment with limited temperature range, relatively low rates of thermal change, (hopefully) no sensible water, corrosive chemicals, limited cyclic loads and the like.  Vibration loads and accelerations are either small are perhaps neglectable.  Earthquakes are vanishingly rare in your location, and if the odds go crazy and you have a big one, it is unlikely that your bookshelf will be the first thing in the building to fail.  New England hasn't implemented earthquake resistance building codes or required retrofit like earthquake prone regions have.  Probably your biggest unknown is in how you calculate the cross sectional area of the threaded rod.  I don't believe MC specifies the exact thread go-no go tolerances for these rods and there is a bit of voodoo in how the load shares down the spiral ramp of the rod.  They only have a nominal diameter and thread pitch.

Remember that in your calculations you have hidden safety factors (estimates of the book load etc.) and at least the last time you visited this were worst casing it by assuming that each of the multiple support rods had to support most or all of the load because of uneven load sharing.

I would personally be comfortable with a 50% margin in your application. 

All of these words apply to the real need for a safety factor.  If you have gotten a building permit for your project and need to have the design approved by the building department the only answer is what the building department requires, whether that is 50% of 1500%.

[totalnoob]

McMaster-Carr lists stainless steal 3/8" 24tpi threaded rod as having a tensile strength of 70,000psi.

Without a safety factor, I'm calculating it can support a weight of 6,148lbs. Obviously this is the maximum, so I'm wondering if a typical safety factor exists.  From reading, I'm seeing 50%, so the calculation would be based on 35,000psi, but wondering if this isn't enough of a safety margin.

I work in facilities and deal with Structural Engineers all the time. They use a 4:1 factor of safety (FOS) for the majority of the times they design something structural (fall protection systems, I believe, get a higher FOS) and the FOS is based on the Yield Strength, not the Tensile Strength. As soon as the steel starts to noticeably deform, its considered failed because you don't want to wait for it to completely break. Not knowing what you are designing for, my only advice is to use the yield strength and build your FOS off of that. 6K# is a LOT of weight and if it falls down, it's going to be a bad day.

[thm_w]

First google result looks like a good starting point: https://www.engineeringtoolbox.com/loads-hanging-rods-d_1341.html
https://www.garagejournal.com/forum/threads/holding-power-weight-of-threaded-rod.373215/

[bostonman]

Maybe I should have included more detail. I assumed the question was adequate enough to assume it was a vertical load without any anomalies, but I should have included more detail without assuming.

To elaborate a bit, my question was loosely discussed in another thread I had a question about regarding something else that deviated towards threaded rod. The threaded rod is being used to support a bookcase I built that will be against a wall, but I know if I offer more details, this thread may deviate towards others having a different way of doing my project rather than the mechanical question in hand.

Basically some of the threaded rod will be in the attic (so assume 25 degrees F to 125 degrees F throughout the year) while the rest will be at room temperature year round. No other stresses except vertical (maybe some natural vibration) will be on the threaded rod.

The link provided is one I saw before, however, I missed the footnote where it states it's based on 12k PSI with a safety factor of 25% (down to 9k PSI). The link states 3/8" threaded rod (which I'm using) is 600lbs based on 25% safety factor. The threaded rod on McMaster-Carr (as I initially stated) is 70k PSI.

My calculations based on the 9k PSI come out to be about 800lbs; not 600 such as the website states. Maybe they are using less threads than 24. Either way, it's not that important, but I guess based on the website, it's safe to take 25% of 70k PSI and use that as a weight?

[CatalinaWOW]

My calculations based on the 9k PSI come out to be about 800lbs; not 600 such as the website states. Maybe they are using less threads than 24. Either way, it's not that important, but I guess based on the website, it's safe to take 25% of 70k PSI and use that as a weight?

Maybe you should use a margin of 5-10 times, based on this sentence alone.  You do not use the strength in PSI (either yield or ultimate) as a weight.  You must find the effective cross sectional area of the rod (which is approximately the diameter across the base of the the threads) and multiply that by the PSI rating to get the load the rod can carry.  But that isn't the only problem in the sentence.  When I go to McMaster Carr's web catalog I find three kinds of steel rod, high strength, medium strength and low strength.  They state that the high strength rod has a strength of 170ksi and that the medium strength rod is about 25% lower and that the lows strength rod is about half the medium strength rod.  They don't say if this is yield or ultimate.  But there is no way any of these values would result in a 9k psi number.

If you are unsure how to calculate the strength, or simply can't be bothered to put the numbers together properly you must massively overbuild to cover unknowns and errors like these. 

To avoid calculation you might refer to standard loading tables for bolts such as the following:

https://www.engineeringtoolbox.com/us-bolts-tensile-proof-load-d_2066.html

The strengths for grade 2, grade 5 and grade 8 bolts roughly correspond to the strengths referenced by MC for their three grades of threaded rod.

[bostonman]

but there is no way any of these values would result in a 9k psi number.

If I understood correctly, I think you didn't know where I got the 9k PSI from. I was referring to this website: https://www.engineeringtoolbox.com/loads-hanging-rods-d_1341.html

If you are unsure how to calculate the strength, or simply can't be bothered to put the numbers together properly you must massively overbuild to cover unknowns and errors like these. 

I couldn't agree more. I am trying to calculate it rather than not be bothered because I like to know the "why" behind numbers, and I only need both rods to hold 1200lbs. At the moment, I'm only supporting about 450lbs, but this will probably increase over time; and even 1200lbs is an exaggerated number.

So even if one rod can "only" support 1200lbs, then two rods (the total number I'm using) will support 2400lbs. That's twice my exaggerated weight.

If I'm doing the math correctly, and use a ridiculously low number of 25% of the 70k PSI listed on McMaster, I calculate a weight of 1,537lbs each rod can support. Two rods would be almost 3,100lbs which is over twice my exaggerated weight and far less than listed, so I should be more than safe.

My math is basically using the formula in the attached.

[Doctorandus_P]

42

[bostonman]

Am I missing a joke or something? What is 42?

[geggi1]

You can probably use the same safety factor as on chain lifting gear that is 4.

[isometrik]

42: c.f. The Hitchhiker’s Guide to the Galaxy

[bostonman]

Oh okay, sorry, I never read the book, so I didn't get the reference.

I'm uncertain if this topic is fully closed. It seems I have more than enough margin for error, however, without a real number. If 4 is a good error factor, then this means I should base the rod on 17,500PSI (70k divided by 4) which still puts me well over my already over estimated weight.

To deviate a bit, does a similar formula apply to nuts? The threaded rod will be suspended using nuts, so I'm wondering if I need to double up on the nuts or use coupling nuts since they have more thread.

[mendip_discovery]

Remeber the threaded rod is only part of the problem.

What rating is the threaded thing it screws into?

You might find the McMaster is referring to the bar load, aka a 3/8 of steel will take 70k load.

The threads and the nuts will have a difference depending on how many turns are in the nut.

[thm_w]

The nuts will be strong enough if they are fully threaded: https://www.crestfasteners.com/images/pdf/catalogs/general%20info/Proof%20Load%20Charts.pdf
Of course there is no harm in using two, then that would also reduce the chance of them coming loose, or use loctite, whatever.

Oh okay, sorry, I never read the book, so I didn't get the reference.

Don't worry, it wasn't funny or appropriate for this thread.

[bostonman]

Wow, from that PDF, it looks like the nuts are stronger than the rod. As mentioned, I may even use coupling nuts to grab lots of thread, but, in either case, the nuts will be more than fully threaded.

[CatalinaWOW]

I suggest that you read the Wikipedia article on "Factor of Safety".  There is no singular correct answer for your question. 

In general you are in very good shape.  But the answer really depends on how you feel about it, how your insurance company feels about it, and possibly how local code enforcement feels about it.  If you are going to bring insurance or code agencies into the question your calculations are meaningless.  Those agencies will require an analysis be a registered PE.

[bostonman]

There is no singular correct answer for your question.

I fully agree with you and know this. It's like anything else, the unexpected can always happen which is why disasters occur.

Normally I don't even like to use x2 as a safety number, I'll go x4 or x5, but it depends on the job. In any case, I have a habit of going overboard when it comes to safety factors, but, I'd rather be overboard than under or marginal.

In my case, a single threaded rod (assuming the load is balanced and not bending the rod) would be enough to support double the already overestimated maximum potential weight. So I am in very good shape with two rods.

Edit: Actually, the purpose of this thread was to double (and maybe triple) check my numbers (ignoring the fact there isn't one exact safety factor number). I knew 5/16" threaded rod would be more than strong, I increased it to 3/8", went overboard on the maximum potential weight it may see, used double that number when I looked for adequate size rod, and got additional confirmation on the math. I believe this is a very good way to add to the safety margin.

Had I got 90% replies stating 3/8" is marginal, then I would have had a different point of view.

[bostonman]

To close the loop on this topic (and the other topic I had about possible issues mixing stainless steel with pressure treated wood), I've basically completed this project.

The bookcase itself (wood, hardware, even the paint) was weighed and got about 105lbs. Recently I obtained a free grams scale (it goes for around $1200) and thought I'd be cute by weighing all the hardware. Even took an empty can of clear stain (the paint used on the wood), used that as the tare weight, and weighed a full can of paint so I knew (within reason) the weight of paint.

In any case, the 3/8" threaded rod is approximately 10" from the sides and centered front to back on each shelf. When I placed the rod through all the holes, the rod slid in and out without an issue indicating the holes were all in line.

The threaded rod goes through the ceiling, and through a 2" x 6" piece of non-PT wood that lays across several attic joists (obviously with washers/nuts on the threaded rod to hold it up). The attic joists at this point sit on the load bearing wall distributing the load downwards to the lally columns in the basement. Currently I'm letting the bookcase acclimate and settle, but, going forward, I'm laying a 4" x 4" (non-PT) wood on top of the 2" x 6" (the threaded rod will be fed through the 4 x 4) and place nuts on top. Using the 4 x 4 on top of the 2 x 6 will allow double support protection, but also reduce any wood bending (I knew a 2 x6 wasn't strong enough, and a solo 4 x 4 may bend slightly over time).  Also, I'll use a bracket to hold the bookcase to the wall so the bookcase can't be pulled outwards or side-to-side. In the X and Z direction (side-to-side and outwards from the wall - if those are the X and Z axis) won't have any force on it; unless someone comes along and yanks on it.

It came out exactly as I envisioned it years back and it came out quite square. I'm uncertain how pros do this kind of work, or, even how store bought furniture is so straight.

In this case I'm dealing wit 50" length wood (11.75" wide), 1" thick. One piece was warped as a cup (on an exaggerated level) front-to-back, the width of the wood varied by 1/16", sometimes the square would be square when compared to one side of the bookcase and underneath the shelf, but wasn't square when compared to the same side, but on the top side of the shelf. This told me the shelf could be warped, or the side panel warped, or both.

Store bought furniture always seems "perfect", wood working TV shows always seem to have perfectly straight/square pieces, etc...

I began thinking, I could "rip" the wood on a table saw to straighten the sides, but then I thought: if the wood is warped, say an S shape (on an exaggerated level), when it's fed through the table saw, it's going to follow the path of the S as it rides the guard, thus giving you a smaller (less wide) piece, but still the same S shape.

If you even manage to get the S shape worked out, it can still be warped as a cup (front to back or side to side), or have a radial twist.

So I don't get how wood workers get it so "perfect". In the case of my bookcase, every shelf is leveled, the shelves are all fairly square with respect to the side panels, and some slight rocking if the side panels are pushed one at a time into the wall indicating the middle of the shelves are resting against the wall with a space on the outer sizes (this is an easy fix - I'll add a few rubber bumpers on the rear along with the bracket that will hold it against the wall).

As "perfect" as mine is, I know it's not pro and/or store bought quality, so I can't understand how pros do it.

[CatalinaWOW]

Congratulations on your completion!  Hopefully it will give you many years of service. 

There are lots of tricks to the pros getting straight furniture.

1.  Buy better wood.  It comes in grades and obviously goes up in price as the quality goes up.
2.  Either buy it dried straight (again at premium prices), or straighten it yourself by stickering.  You take wet wood and stack it with wooden spacers and weight it down so that it is forced straight as it dries.  It will tend to stay in this condition.
3.  Buy oversize wood and mill it straight before cutting to size.  There are tricks to this also.  One example- most table saws have a guide slot running parallel to the blade. You can make a sled that has features to follow the track and then clamp your board to the sled.  Alternatively you can find one straight board (perhaps ripped from a sheet of plywood) and stick it you your crooked board with double sided tape, small brads or screws.  Saw the edge of the crooked board with the straight board against the saw fence.  Either of these will get one edge straight and you go from there. 
4.  Buy dimensionally accurate and stable wood (high ply count hardwood plywood for example) and build key parts of the furniture from that and use it to force the "natural" wood straight.

And much more.  All this is pretty painful building one or two items in a home shop with a limited tool set.  It gets a bit easier in a production line environment with lots of room and equipment, but still takes a lot of time and work.

When you look at all that goes into good furniture the high cost isn't quite as hard to swallow.

[bostonman]

I imagine the pros buy their wood at high quality stores. Mine wasn't HD or Lowe's, it was purchased at a place that sells higher end wood, however, it's not to say they get the best quality that has been given proper attention to keep straight (or buy straight).

You're correct, more room, larger power tools, etc... would make obtaining perfection easier. That guy Tom on Ask This Old House always seems to get perfectly straight wood without anything more than using a table saw. He just drops a piece of wood on a table saw not much better than mine, slides it through, measures, drills, screws, and the entire thing looks perfect. My assumption is the wood is prepared well in advance along with the possibility of things not being as square as they appear; maybe they just scrap the projects after completion because they aren't really square/straight.

Even when I built a jig for the router to cut dados, I couldn't get the jig square no matter what I did (I honestly spent about two-weeks off and on trying to make a jig along with trying to get a good depth). Once I got the jig "square", I put the square on three sizes, all perfect, and the fourth side was off which annoyed me. Once I laid it on the side panel to cut the dado, the side panels would cause the jig to be off even further. At one point I took the jig I built and broke it in pieces just so I could start over.

I'm sure the pros use planers, have better guides on their table saw, more room to work, larger tables, etc... It's just amazing that computer desks aren't warped, or become warped, house furniture is nice and square, etc...  Even a basic computer desk that's relatively cheap seems to hold its own over time.

I know the wood used is probably that cheap wood pressed together from wood chips (forgot the name) with paneling glued over it, but still, I've seen that same wood at HD and never found a piece that wasn't warped (although I'm sure it's not purchased at HD by the computer desk makers).

[jpanhalt]

I am not sure exactly what you are asking.  The usual tool for getting a straight edge on surface-finished wood is a jointer.  Once you have one straight edge, you can run it through a table saw to get the other edge straight and parallel. There are other ways to get a straight edge with just a table saw, some of which attach a straight edge to the board for running against the fence.  Several YouTube videos are on that subject.  A planer is not a very good way to edge wood or to get it straight.  It simply gives parallel surfaces.

Mills have a special saw with cleats that garb the wood and feed it straight without a fence and give a straight edge.  They do tend to leave marks on the wood, which is why with rough sawn wood, you get an edge first before doing the surface.  It's called a line rip saw or something like that.

[bostonman]

They were basically rhetorical questions by thinking out loud how much effort goes into making stuff square and flat.

After looking at videos on using a jointer, it looks like that's a good tool to use when striving for perfection.

[jpanhalt]

They are nice and are one of the more "dangerous" fools in a shop, as some people try to touch the spinning blades.  That's how I got my second, much larger and older Powermatic from a school shop.  The longer the infeed and outfeed tables, the better.  You can sort of make do with roller supports and so forth.  Thay are also good for rabbeting.

[bostonman]

as some people try to touch the spinning blades

Wow.... that's scary.

Seems like another safety issue is the wood kicking should it hit a hard spot on the wood (such as a knot).

[bostonman]

I have two more threaded rod questions.

How does the tensile strength change when the rod is bent? If say the rod was mounted vertically on a bracket that was 6" away from a wall and the object it supports is 8" away, this would cause the rod to bend. I'm not looking for an exact answer, but curious how the force(s) on the rod changes. I'm sure a longer rod would be more forgiving whereas a short one would probably snap.

My other question is just basically what happens to a long length of threaded rod if someone tries to bend it; just one bend, not back and forth which would certainly cause it to break? Does it eventually kink, snap, or take the bent shape? It obviously depends on the arc, so guess maybe I'm referring to a length of several feet where one end is in a vice and the other end getting the force so maybe the object is to have both ends touching.

My experience with bolts and screws is they just bend, but usually they are too short for attempting to bend them to such a degree.

The rod I had was 96" (3/8" x 24) and too expensive to experiment with.

[jpanhalt]

If you bend a threaded rod, you will mostly see difficulty moving a nut over the bend area.  Threaded fasteners are very sensitive to distortion of the threads in either part.

Mild steel is pretty forgiving to being bent over a sufficient radius.  That radius varies with thickness.  Tables are available for most common metals, like mild steel, most stainless alloys (like 303), and most aluminum alloys.

[bostonman]

If you bend a threaded rod, you will mostly see difficulty moving a nut over the bend area.  Threaded connectors are very sensitive to distortion of the threads in either part.

I wasn't referring to whether a nut can be threaded, it was more of a general question of strength degrading on the rod.

[jpanhalt]

Obviously a threaded rod has "creases" in it, and that weakens it.  Bending will probably weaken it more.  Will that affect your use?  I doubt it.  Just FYI, critical threaded rods in aircraft are often/usually specified to be "rolled thread."  They are quite strong, but I can't imagine you need to go to that bother.