THREADS LAID BARE – HOW TO MEASURE
In this two-part feature Peter describes how to measure and identify any thread form on any bolt, nut,threaded rod or hole without removal or dipping into drawers of fasteners – it’s all about a clever piece of scientific software, and it’s free...
So, you’ve trawled the dealer’s websites, checked out ebay (risky, but perhaps housing a bargain…), maybe an auction or two, and taken the plunge: a shiny new (to you) steam locomotive sits proudly in your fledgling workshop and you gaze adoringly at its beautifully-made motion, its shiny brasswork and its slightly tired, but still gleaming, paintwork and dream of sunny days puffing around your local club track.
You look carefully and notice that one of the nuts is missing from the crosshead and a section of shiny thread protrudes, challenging you to make it complete. You look around the bench, the floor, the boot of your car, but the elusive nut remains just that, elusive. You conclude that it probably wasn’t there in the first place and you hadn’t noticed in the thrill of taking delivery of your pride and joy. “It’s just a nut – I can replace it”, you say to yourself, “how hard can it be?”
The answer is, of course, ‘not very’, assuming that you know what thread-type the elusive nut requires. That is where the fun starts…
Identity Crisis
We’ve probably all been there at some time or other, perhaps not with a steam locomotive but some piece of equipment we’ve needed to repair or refurbish. How do we identify the correct thread type in order to make or buy a nut, bolt or length of studding with which to effect a repair?
I was confronted with this problem when I bought my first loco, a Super Simplex – I’ve still got it and it’s a real gem (Photo 1). It wasn’t a missing crosshead nut, but we’ll come on to that later. Experienced model engineers reading this will smugly point to a cupboard full of threadgauges (Photo 2) and a full set of drawings for the loco, plus dozens of drawers of every conceivable type and size of nut, bolt and washer from the early bronze age to the present.
The newly-minted enthusiast, with a bench, a few screwdrivers and a lot of optimism, on the other hand, isn’t in this position. In my case, I had a pretty well-equipped general workshop, but not a model engineer’s one and not the first clue about ME (model engineer) threads, for starters.
But surely the drawings will provide all of the answers? You did obtain a set of drawings with the loco didn’t you? They may, if the builder (who may well not have been the vendor) built the loco exactly to them. Many locos are ‘based upon’ a set of drawings, but often materials and parts to hand (or still available) are substituted for those specified by the drawings. In particular, nuts and bolts are often substituted with nearequivalents – perhaps modern metric examples replacing BA (British Association) parts, for instance, since some BA sizes are getting harder to source. In such cases, the drawings will be of limited use.
In my case, all those years ago, it was a troublesome boiler fitting – something I clearly didn’t want to get wrong! The consequences of mangling the threads on a boiler bush didn’t bear thinking about. Although I now know it is recoverable, it is still something very much to be avoided.
Thread gauges
You’ve just spent thousands on your loco; surely a few tens of pounds invested in every set of thread gauges you can lay your hands on makes sense? The simple answer is of course, yes, however thread gauges may not be available (or easily available) for all types. I’m not aware of anyone who produces thread gauges for ME threads, for example, although since these are based on Whitworth threads, using thread gauges for this thread type is an option. The only thread gauges I have are metric.
Thread gauges are also difficult to use to identify which thread type you have. If you know that a particular thread is, for example, metric, then it is relatively easy to use a set of metric thread gauges to find out precisely which ‘M’ thread it is (as we will see).
If, however, you have no idea what the thread type is, trying various gauges may not yield a convincing answer, just a narrower list of possibilities. For example, 9BA and M2 threads are very similar in most regards and a couple of sets of thread gauges would likely struggle to tell them apart.
Thread gauges are most useful, in my view, where the thread/bolt is of a known standard (such as metric) but the thread pitch (TPI) is unknown. Taking the example of a metric bolt (Photo 3); measuring its diameter (even if it wasn’t known to be a metric bolt prior to measuring) makes it clear that this is likely to be an M5 bolt. Note that for a genuine bolt (as opposed to a machine screw), it is usually better to measure the unthreaded part (Photo 4) although, in this case, it makes very little difference (~0.02mm). Whilst this diameter should read 5.00mm, bolts are not (generally) precisionengineered parts, unless specifically made so, and will typically read a little under their stated diameter.
We have now, fairly convincingly, deduced that we have an M5 bolt, however which type of M5 bolt do we have? There are two thread pitches in (fairly) common use for M5: 0.8mm and 0.9mm (we will discuss and measure thread pitches in due course – for now, it is sufficient to note that they are different).
This is where a set of thread gauges comes in most useful, in my experience. It is relatively easy to tell these two thread pitches apart (see Photo 5 and Photo 6) and thereby deduce the full specification of the bolt. In this case it is an M5 x 0.8mm threaded bolt. The only remaining parameters to specify are the head type, bolt length and thread length; again, we will return to these later.
It is evident, from Photo 5 and
Photo 6, that even when a bolt is mostly ‘known’ (for example its standard, such as metric), and the likely thread parameters are known due to a given size of bolt only being manufactured in a couple of thread pitches, it is still not that easy to judge with absolute conviction that the right thread has been uncovered using a set of thread gauges (particularly if you are a complete beginner or your eyesight is not what it used to be...).
It becomes even more difficult if the bolt is completely unknown – which type of thread gauges should be tried? What happens if more than one standard/gauge looks like a plausible match? Techniques to help answer these questions are outlined below.
Nut Job
A first option is, of course, simply to try various candidate nuts (or bolts, as relevant) on the thread, until one seems to fit convincingly. This is fine so long as: a) You either have a vast selection of nuts and bolts (those new to the hobby probably don’t)... b) ...or you have a similarly vast selection of taps and dies (Photo 7), with which to make candidate nuts or threaded rod for trying out. Again, those new to the hobby probably don’t. c) And the end of the thread (and, indeed, the whole thread) isn’t
“Materials and parts to hand are substituted for those on the drawings...”
damaged – if it is, even the correct nut (say) will either be stiff or not want to thread on at all.
Unless you know for certain which thread you are dealing with, applying a large amount of force to a nut (with the wrong thread) in an attempt to ‘clean’ or re-form the threads may well spell disaster for the bolt (say). If this is
Loctited-in to the motion, for example, or would require disassembly of half of the loco in order to replace, then it is best to be very sure of having the right thread before applying such force (or running the correct tap/die, as appropriate, along the thread).
If you either don’t have access to the required nuts/bolts/thread gauges or are nervous of breaking your precious new loco (I still remember
“There exists a bewildering array of thread types...”
such feelings!), what do you do? The answer is to apply a little technology, as will now be described. And the best thing? It’s all free…!
Thread Types
There exists a bewildering array of thread types (the list in Table 1 is far from exhaustive) and a good number may be found on model locomotives – likely candidates are highlighted in bold in the table. Even within a standard, there are differing thread pitches for the same diameter of thread (as seen above with metric M, threads).
At least the thread taper generally used on the threads themselves is typically fairly standard across the most commonly-used types, at 55 or 60 degrees (with BA being a notable exception, at 47.5 degrees). These two angles are close enough that they may sometimes be interchanged, with caution. I know of a few model engineers who don’t worry about this difference when thread-cutting on a lathe, for example.
ME or Whitworth?
Whilst ME threads are based upon Whitworth threads, they typically differ in their thread depth and, almost always, in their TPI vs diameter. For example, a 40 tpi ME thread should have a thread depth of 0.016-inch irrespective of the thread diameter, which is much smaller than the equivalent Whitworth threaddepth in all cases but 1/8-inch, where they are equal (as it is the only diameter of Whitworth thread which boasts 40 tpi).
The coarser ME thread, 32 tpi, has a thread depth based upon the BSF standard and again, the only thread density at which the two types are equal is the smallest size defined for BSF (3/32-inch). This diameter does not exist as an ME thread, however, making identification easy.
Bolt types
When specifying a bolt (or a machine screw), the following parameters are generally required:
1) Diameter – maximum outer diameter of the threads, in other words any sample bolt will fit through a hole of this size
2) Length – it is not always obvious how this is specified, see below
3) Threaded-length (for bolts)
4) Thread-type/standard (BA, M, ME, BSW and such like)
5) Head style (and sometimes diameter as well, if this differs from that specified in the relevant standard) 6) Head type, for example slotted, cross-head, posidrive or such.
At this point, it is worth distinguishing between a bolt and a machine screw. The former is only partially-threaded, in other words the end section is threaded but there will remain a section closer to the head which is plain and unthreaded. A machine screw, on the other hand, will be threaded right up to its head.
Photo 8 shows a selection of commonly-available bolt and machine-screw types – most of the examples in the photo are metric, although one is BA, can you guess which? (answer at the end). Working from left to right, these are:
1) Hexagonal-head (or hex-head) machine screw. Note that the term
‘hex head’ is sometimes (confusingly) used to refer to Allen-headed bolts
2) Cap-head Allen (or ‘hexagon socket’) bolt. Most cap-head bolts have Allen-key compatible heads
3) Flange-headed machine screw.
4) Countersunk machine screw – these are available with a variety of head types, such as Allen, slot, Torx and various types of cross-head
5) Cheese-head machine screw (with a slotted head)
6) Pan-head machine screw (this one is almost a halfway house between a true pan-head and a cheese-head)
7) A more conventional pan-head, slotted, bolt
8) Coach bolt
There are a couple of other common types which I didn’t fall across in my rummage around for the photo...
Raised-countersunk – as its name
l suggests, this is a countersunk-style bolt/machine screw with a raised rounded head
Button (Allen) head – similar to a
l pan-head, but with a more rounded head (the profile forming part of the arc of a circle). Almost always designed for use with Allen keys.
There are other styles based upon these (such as types with both cross-head slots and hexagonal heads – often found on Jubilee-clips for example) and some more obscure types sufficiently rare to be ignored here. I’ve probably forgotten one or
two ‘obvious’ types – if so, I’m sure someone will write in to remind me!
Specifying Length
This is not obvious, or at least it wasn’t obvious to me when I started out. Those of you who served an apprenticeship and spent a career bolting things together will smile at my erstwhile naivety, however a career in the radio and microwave industry didn’t involve the specification of too many fasteners...
For most types of bolt, the length is defined from the furthest extent of the thread at the bottom of the bolt, to the underside of the bolt head. In other words, the length does not include the head. So far, so simple. When dealing with countersunk heads, however, the specified length includes the head!
The best way I have found to think about this (and it’s obvious when you do think about it) is that the length of a bolt is defined from the bottom of its thread to the upper-surface of the material into which it is inserted, when fully inserted and tightened. Based upon this definition, the fact that the length of a countersunk bolt includes its head, whereas that of a normal bolt (such as Hex or pan head) excludes its head, makes perfect sense.
Introducing Imagej
What is Imagej? It is a free, opensource, scientific image analysis software package originally written to perform tasks such as counting or measuring cells on a microscope slide, analysing growths in a petri dish and the like. It is a very powerful package and is useful for many aspects of model engineering. But despite its power, it is very easy to use (for the simple things we will be doing) – in effect, we are using a Le Mans supercar to pootle to the shops.
Examples of areas in which Imagej could prove useful to a model engineer include:
1) Measuring thread-pitch and counting threads-per-inch (TPI), as discussed in this article
2) Accurately measuring small apertures, too small for a bore gauge to be used (or for any aperture, in the event that you don’t yet have a set of bore gauges – I’m still a little lacking in this area)
3) Accurately measuring loco components which are awkward to measure by conventional means (using callipers, micrometers and such like), even whilst they are still attached to the loco!
4) Accurately measuring components on full-size locos, in preparation for modelling. These would obviously still be attached to the loco, in most cases.
The last point is quite a powerful use of the package for a model engineer – most museums would take a dim view if you started climbing all over their precious exhibits with your ruler and tape measure, but are quite happy for you to take as many photos as you like.
A detailed discussion of items 2 and 3 is beyond the scope of this article, suffice to say that the quality of the photo is very important; not just its resolution, but the angle at which it is taken. Accurate results will only be obtained if the photo is taken square-on to the item. Views up to, or down to, the component in question will yield inaccurate answers.
One other point is that it helps enormously if at least one dimension in the photo is known (even to the extent of placing a ruler somewhere in the picture). This dimension can then be used to ‘calibrate’ the result.
As a final tip: if you can use a flat-bed scanner, then do so. This guarantees a 1:1 scaling and allows the resolution of the scanner to be used directly as a part of the measurement – we will discuss this in more detail in relation to threads, below. Perhaps I’ll get a chance to revisit the use of
Imagej on loco parts, in a future article, however for the present we’ll concentrate on thread-identification.
There are a number of variants of the package, but the one which is most commonly-recommended (and which I use) is called ‘Fiji’. It can be downloaded from Imagej’s website at https://imagej.net/fiji – this page contains download links for all of the main operating systems, including both 32 and 64-bit Windows systems, MACOS and Linux. I have only used the 64-bit Windows version, but I’m sure the others will work just as well.
Using the software
To demonstrate the use of this package, two similarly-sized machine screws but of differing standards were chosen from my (now reasonable) collection. For the purposes of this article, the images of these ‘unknown’ parts were obtained in two different ways: using a fairly high-resolution smartphone camera (12MP) with its flash turned on (Photo 9) and using a flat-bed computer scanner at 1200 dots-per-inch (DPI) (Photo 10).
The scanner settings used on the flatbed scanner are shown in Photo 11.
Note that care should be taken with the scanner approach, as it would be easy for heavy-handed use of the lid (such as dropping it onto the bolts) to result in scratches of even cracking of the scanner glass.
Photo 10 also illustrates another minor problem with the scanner approach: the ‘focus’ of the scanner is set to the upper surface of the glass,
“It would be somewhat challenging to scan a 7¼-inch gauge Britannia
on a domestic
flatbed scanner...”
thus making anything which is not actually touching the glass slightly out of focus. This can be seen, in particular, on the section of thread nearest to the head on the right-hand screw. This is not a huge issue, however, since any part of the thread may be used for measurement purposes and hence the bottom end of the thread, which is in focus, will prove adequate.
The key advantage of using a smartphone (or, indeed, any camera) is, however, its portability – it can be taken to the mystery bolt or thread on the loco, which can be photographed in-situ. In contrast, it would be somewhat challenging to scan a 7¼-inch gauge Britannia on a domestic flatbed scanner…
Start with the Diameter
The first step is simply to use a micrometer or callipers to measure the diameter of the unknown threads. If you have absolutely no idea of a bolt’s provenance, measurements will need to be taken in both metric and imperial units – one or other will probably give an initial clue as to which is most likely to be correct, although it is best not to jump to conclusions at this stage!
For the two machine screws shown in Photo 9 and Photo 10, these measurements are shown in Photo 12 and Photo 13 for screw #1 and Photo 14 and Photo 15 for screw #2. In case these photos are too small to read when reproduced in the magazine, the values are shown in Table 2.
The values shown in Table 2 are not very convincing either way, particularly when it comes to the imperial values. Remember that a measured thread diameter will generally be a little smaller than its ‘official’ designation.
With this in mind, screw #1 could be 2mm metric and screw #2 could be 2.2mm metric. Equally, screw #1 could be 5/64”-inch imperial and screw #2 could even be 3/32”-inch imperial, although this might be a bit of a stretch. None of these measurements is conclusive, although they do narrow the field somewhat, by ruling out larger (such as 2.5mm) or smaller (such as 1/16-inch) threads.
n Next month Peter shows how to use Imagej and a little Sherlock Holmesstyle deduction to work out the thread type on any bolt, nut or other threaded component, whether or not it is still attached to your loco.