Snapping taps

[Enough!Participant @enough]

Posted by Michael Walters on 21/03/2016 13:32:57:

i don't have a pillar tool and i'm not keen on machine tapping into blind holes.

You don't really need a pillar-tool to properly support/guide the tap-wrench during hand tapping. If you have a mill or a bench/floor drill you can use one of these. Put the guide in the drill-chuck or collet and it gives all the support you need. You can pull the tommy-bar completely to one side if you want without putting side-loads on the tap.

You don't even have to buy one. I bought my first few in several sizes but I've since converted existing tap-wrenches and made additional guides. It's a simple job. I do virtually all my tapping this way and haven't broken a tap in years. I use them on the lathe too.

Of course, you still need the right-sized hole

[Harold Hall 1Participant @haroldhall1]

Thanks Neil for your comments regarding Tubal Cain's value for 8BA but unfortunately I did not make my comments clear. Rather than talking about actual values I was commenting on his general principle to use drill sizes larger than were quoted in most tables and that 60 to 70% thread depth was adequate for most situations, even quoting less than 60% for some.

When I was producing my "Metalworkers Data Book" I calculated the drill sizes myself on that basis for all the common, and less common, threads.

On a wider range of data, I found a number of errors in Tubal Cain's "Model Engineer's Handbook" as I did in other publications, even British Standards. Unfortunately, in the attempt to get the project finished, as it had become very time consuming, I did not make a note of the errors found.

Harold

[Clive HartlandParticipant @clivehartland94829]

I use a simple calculation, the, 'K' factor of 0.8. Multiply the thread diameter by 0.8 and I use that size as the tapping drill size. An example for M4, 4 x 0.8 = 3.2 drill.

Clive

[John FieldingParticipant @johnfielding34086]

Hi Clive,

Unfortunately your formula is wrong.

The standard method is to subtract the thread pitch from the diameter to obtain the maximum thread depth. Hence M3 x 0.5 is 2.5mm and a M4 x 0.7 is 3.3. But as Prof Chaddock pointed out years ago this gives 95% thread depth which is not warranted in many cases. For soft materials such as aluminium about 80% depth is adequate for 90% of the potential tensile strength. For harder material such as mild steel then as little as 60% thread depth is more than adequate. So a drill of 3.3mm to 3.6mm is the range for a M4 x 0.7mm thread.

[HopperParticipant @hopper]

Another check before starting to tap the drilled hole is to stick the taper tap in the hole and observe how far in it goes and how deep the tap's teeth are at that point. For a 75% thread depth, you want to make sure the tap will slide straight in until the tap teeth are about 1/4 of full depth.

If you stick the tap in the hole and it only goes in up to the first row of teeth or so, the hole is too small.

Small holes is recognized widely as the number one cause of broken taps, so it's worth checking as you poke the tap in.

[MWParticipant @mw27036]

I too am with the clive hartland school of logic obviously but clearly the census is that i am wrong and i'm not about to argue against it as i did find that before the tap broke, as chris evans said from working in his die making game, when it starts to bind on the back up movement, chuck that little so and so away, good for neither man or beast at that point.

As for the "never broken a tap", was that all the times you were counting or just the selective memory kicking in. (it's also said as though it's never about to happen, i would touch wood, or metal if you prefer) I also notice the brand thing too, I use presto taps too! nothing more annoying than breaking a premium tap.

So theres a good likelihood i've been trying to remove too much material with a 0.7 pitch.

Speaking of breaking taps, i've broken tap wrenches too, i dont use the cast body ones subsequently! Strangely the same rule doesnt apply to die holders. Maybe it's got something to do with not even a brute like me can break a piece of metal thick as a die.

Thanks,

Michael W

 

Edited By Michael Walters on 22/03/2016 08:33:50

[Ian S CParticipant @iansc]

Don't worry, I managed to break a 1/2" UNC die, I think it was a 1' button die, 13 tpi is fairly hard going, and I was trying to rush things along, still managed to finish cutting the thread.

Ian S C

[HollowpointParticipant @hollowpoint]

Posted by Michael Walters on 22/03/2016 08:19:27:

As for the "never broken a tap", was that all the times you were counting or just the selective memory kicking in. (it's also said as though it's never about to happen, i would touch wood, or metal if you prefer) I also notice the brand thing too, I use presto taps too! nothing more annoying than breaking a premium tap.

Edited By Michael Walters on 22/03/2016 08:33:50

"Never" might be being a bit economical with the truth. I have broken taps before. XD But I honestly can't remember the last time!

[John FieldingParticipant @johnfielding34086]

Hi Ian SC,

That's nothing!

When I was an appie we had a gorilla in the workshop who one day broke a 1 1/2-inch Whitworth die in a handheld die stocks. When the others asked how managed it he sheepishly replied it was a bit tight going on and so he used 4-pound hammer to hit the end of the die stock handle. Needless to say he had to pay for the die and the bench vice which broke under the punishment!

[Anonymous]

For the internal thread the key material characteristric is the shear strength. Generally the external thread will fail in tension across the core area before the internal thread fails in shear. Even in aluminium thread depths of around 60% are fine. Hand taps tend to extrude the material slightly, so for ductile materials the thread depth is slightly greater than might be assumed from the tapping drill. Extrusion is another reason for not drilling at or near core size; the extruded material then has nowhere to go, so the tap binds. Sound familiar?

Some years ago I did an experiment with M4 threads in aluminium (6082) mated with high tensile (grade 12.9) socket head cap screws. I drilled holes in the aluminium at 3.4, 3.5 and 3.6mm and tapped them M4. The high tensile bolt broke in the thread drilled 3.6mm before the thread stripped.

I seem to be in a minority of one in that I rarely need to use tapping fluid for hand tapping. If nothing else it saves the palaver of cleaning out the tapped holes; especially useful when studs are to be Loctited in place.

Andrew

[AjohnwParticipant @ajohnw51620]

This has cropped up in Brian's stirling engine thread. I'd strongly suggest that people measure both the tap and drill diameters a simple aspect that never seems to occur to prof this and that.

It's also interesting on metric to check what tap core clearance is obtained when the dia – pitch is used to determine the tapping drill on small sizes which is where breakages usually occur. I was warned the Zeiss tables can be worth thinking about on imperial at all sizes as well even as long as these have been about. These all also assume on the face if it that things are exactly the correct size. They seldom are and drills tend to give different sizes in different materials.

Brian's thread came up with another – only tap to twice the diameter. I usually think about it the other way – engagement lengths, 1 to 1 1/2 diameters. I don't always stick to that which will reduce the strength of fixings but if I go over it I know that extra care will be needed as the size goes down.

Ordinary taps should be reversed to break the chip every 1/3 of turn or less. Where possible the drilled hole should go well past the threaded part.

Another favourite for breaking fine taps is entering them at an angle that doesn't match the angle of the hole. This is why the little ME pillar drill and taping and staking kit came about. I have used drilling machines where the quill is dead square to the table but never had one at home. Some older me articles suggested using a drill plate in the lathe to get round that problem – assuming that will be well aligned. The plates are too small these days.

John

–

[AjohnwParticipant @ajohnw51620]

Posted by Hollowpoint on 22/03/2016 10:16:54:

Posted by Michael Walters on 22/03/2016 08:19:27:

As for the "never broken a tap", was that all the times you were counting or just the selective memory kicking in. (it's also said as though it's never about to happen, i would touch wood, or metal if you prefer) I also notice the brand thing too, I use presto taps too! nothing more annoying than breaking a premium tap.

Edited By Michael Walters on 22/03/2016 08:33:50

"Never" might be being a bit economical with the truth. I have broken taps before. XD But I honestly can't remember the last time!

I truly can't recollect EVER breaking one but also can't think of every going below M2 or that sort of size in other standards.

I did buy a really cheap UNF set once and only used 1 tap before they all went in the bin. They need to be sharp.

John

–

[Clive HartlandParticipant @clivehartland94829]

I would like to add that I use the 3 set taps of the type where the No. 1 tap cuts about a 1/3rd thread depth and No.2 cuts about 2/3rds and the a finishing tap. From my work with LEICA everything had to be as they said, no half measures. The only time I deviated was when working with Stainless Steel items and then bought taps specially for S/steel.

Clive

[Harold Hall 1Participant @haroldhall1]

Thanks Andrew, that is very interesting. From my tables 3.4mm drill results in a 70% depth thread and 3.5mm very close to 60%, so with 3.6mm I would assume around 50%. Could you give us an approximate idea as to the depth of thread. This leading to the advice to those unsure, that one should err on a smaller drill for threads that are less in length than the threads diameter. Even here I would not go for more than 80%, normally no more than 70%.

On a general point regarding published information, it has always concerned me, coming from a non metalworking environment, how information is listed as exact. The drill for this thread IS "ab", the speed for machining 20mm diameter mild steel IS "cdef" and yes I have seen tables listing speeds to four significant figures. For me it is a case of , the drill for this thread CAN BE "ab", see notes at the end of the table.

Harold

[Howard LewisParticipant @howardlewis46836]

By not clearing the swarf, to my shame, I broke a M10 Tap!

Wherever possible, I use either a home made pillar tool (probably the handle is too long, but been lucky so far), or in the Mill/Drill a spring loaded device (Can't remember the name, Tap Tru?)

Am sure that bending causes many broken taps. If the Tap binds, and feels "springy", I always back out, clean and relubricate before returning to the task.

These methods have reduced my Tap purchases greatly. For lubrication, mostly Bacon fat is used, silly when there is great big tin of Rocol STD on the shelf! But in the old days they used tallow,'cos it was all that was available then.

Howard

[Anonymous]

Posted by Harold Hall 1 on 22/03/2016 12:03:28:

Thanks Andrew, that is very interesting. From my tables 3.4mm drill results in a 70% depth thread and 3.5mm very close to 60%, so with 3.6mm I would assume around 50%. Could you give us an approximate idea as to the depth of thread. This leading to the advice to those unsure, that one should err on a smaller drill for threads that are less in length than the threads diameter. Even here I would not go for more than 80%, normally no more than 70%.

I'm not sure which thread depth is being referred to? I'll assume the depth of the hole. Although my tests were done thru hole the final application was for a blind hole. The screws were holding down a three phase bridge module, about the size of a paperback book. Each arm of the bridge consisted of silicon dies mounted on a ceramic substrate with power and signal connections via spring loaded contacts. When the module was bolted down to the heatsink the spring contacts compressed ensuring good electrical contact and also provided a force to hold the ceramic substrates down to the heatsink. The module needed to dissipate up to 2kW through the heatsink into a liquid cooled array of fins. The final hold down force was estimated in the datasheet at 9000N over 8 M4 screws. Due to the limited thickness of the heatsink and the distance needed to compress the module I only had two or three threads engaged at the start of compression. The datasheet gave a sequence in which to tighten the screws, one turn at a time, and a final torque value. I was concerned about stripping the threads, especially at the beginning when only a few were engaged, hence the experiments. In the event there was no problem, but it's best to avoid nasty surprises if possible.

I have read that only three or so thread pitches are actually taking load, as the screw stretches slightly as the load increases. So deep threads do not have a greater carrying capacity than shallower ones. Makes sense, although I haven't done any experiments to prove same.

Andrew

[Martin KyteParticipant @martinkyte99762]

"I have read that only three or so thread pitches are actually taking load, as the screw stretches slightly as the load increases. So deep threads do not have a greater carrying capacity than shallower ones. "

That is true Andrew which is why counter boring to just over 1 thread pitch gives you a stronger fixing by burying the threaded portion. I know its not possible on thinner sections but I always do it when I can. It looks tidier as you can put a tiny chamfer on the edge of the hole and the surface stays dead flat when you tap. If you look really closely at a new tapped hole straight into a block the surface deforms upwards slightly around the tapped hole which a quick touch with a file or abrasive will reveal especially in soft materials. For heatsink applications this can compromise the flatness contact surface although I tend to just do it generally because it looks nice.

regards Martin

[JasonBModerator @jasonb]

Posted by Andrew Johnston on 22/03/2016 23:22:21:

Due to the limited thickness of the heatsink and the distance needed to compress the module I only had two or three threads engaged at the start of compression. .

Andrew

Sounds like a good reason to use studs, they could have been screwed into the softer aluminium to the full depth available then the stronger nuts would have coped better with the short engagement to start with.

J

[HOWARDTParticipant @howardt]

As Martin has said there is no need to tap to deeply. There is a lot of information out there on calculated strength of threads. You only have to look at the thickness of a standard nut. The nut thickness takes into account a chamfer each end of the hole and a given number of full threads. I was always told that in steel you have the depth of full thread equal to the diameter and one and a half times for aluminium. Always chamfer the hole to allow for the radius under the screw head and to remove the burr raised during tapping. The blind hole depth depends on the type of tap used, if you are hand tapping with a bottom tap you can obviously get closer to the bottom of the hole than machine tapping. Have a look at Emuge tap catalogue there is lots of information which may help to analyse what your particular taps are capable of.

Howard.

[JAParticipant @ja]

Posted by JasonB on 23/03/2016 09:35:44:

Posted by Andrew Johnston on 22/03/2016 23:22:21:

Due to the limited thickness of the heatsink and the distance needed to compress the module I only had two or three threads engaged at the start of compression. .

Andrew

Sounds like a good reason to use studs, they could have been screwed into the softer aluminium to the full depth available then the stronger nuts would have coped better with the short engagement to start with.

J

Simple design rule:

1. Nut, washer and bolt – good and easy
2. Stud – OK
3. Screw into tapped hole – full of problems (but looks nice).

JA

[Anonymous]

Posted by JasonB on 23/03/2016 09:35:44:

Posted by Andrew Johnston on 22/03/2016 23:22:21:

Due to the limited thickness of the heatsink and the distance needed to compress the module I only had two or three threads engaged at the start of compression. .

Andrew

Sounds like a good reason to use studs, they could have been screwed into the softer aluminium to the full depth available then the stronger nuts would have coped better with the short engagement to start with.

Not really! First there is no room on module moulding for nuts, or for the means of tightening them. As it was we required special low profile head SHCS. Even if we did manage to use nuts the studs would have stuck out from the module once tightened down which would have compromised creepage and clearance, and may well have shorted a busbar to the chassis. Since the inverter output was 600V DC that could have been spectacular.

Andrew

[Martin ConnellyParticipant @martinconnelly55370]

British Standard BS3643 part 1has a table for metric thread engagement. The table has short, normal and long columns against rows which have a range of diameters and pitches. For an M2 x 0.4 thread short engagement is up to and including 1mm long, that is 2.5 pitches. M3 x 0.5 short engagement is up to and including 1.5mm, that is 3 pitches. This, presumably acceptable, short engagement matches Andrew's post regarding only 3 pitches taking the load. Long thread engagements for these two threads are over 3mm and over 4.5mm respectively. Anything between the long and short values is classed as normal engagement. I think the load and materials involved would determine which engagement would be specified if you were doing a drawing for someone else to follow.

The usual guide of one diameter of thread engagement does not always fall into the normal range in this table, for fine pitches one diameter can be classed as long engagement but it is never a short engagement even for coarse pitches.

Martin

[AjohnwParticipant @ajohnw51620]

Most of the 1 to 1 1/2 diameters thread engagement rule is down to pitch errors and that is for high precision screws and bolts and tapped holes too.

What happens in practice as the engagement gets longer is the the pitch errors accumulate so some threads can finish up taking very high loads and a lot of distortion as a result. They may even strip in the extreme. This sort of corrects the problem after a fashion but isn't a good idea really. What it all means in practice is that maximum holding strength will be achieved in the recommended range – the materials don't really matter when viewed this way.

I bought some high precision high tensile socket cap heads not all that long ago for holding things down on the mill and in the lathe at times – they cost rather a lot more than the usual run of mill stuff and seem to be hard to find. I asked for the equivalent of old unbreako's in a screw nut and bolt shop.

People talk about finer details and don't measure the tools they are using even to make the hole and add the thread.

John

–

[HOWARDTParticipant @howardt]

In industry theoretical calculations are based on using the standards for the fixings, i.e. unless the fixings are to the standards then the calculations do not apply. Also any additions specified such as lubrication. Where strength is critical then the fixing standards have to be specified for purchase and may include testing requirements. All this is a little OTT for the model engineer. We have to accept what we get, and in most cases it doesn't matter that the fixings are from some unknown source. As John said trying to get fixings to specified standards is not easy as the market is flooded with cheap imports with no traceability to source. I will stick to 1 to 1 1/2.

Howard

[John FieldingParticipant @johnfielding34086]

As the thread (pun intended) has drifted off into thread length engagement etc it is useful to look at the different thread types used in different materials.

For example, automotive applications where a cylinder head bolt or stud is inserted into an alloy casting. Metric threads are not really suitable where high tensile loads are required, but UNC and Whitworth are superior. I once had a discussion with one of our young mechanical engineering guys who insisted his VW car had metric threads throughout. I told him he was wrong, there are plenty of non-metric fasteners used on those models. I told him to pull out a cylinder head stud and measure it. Sure enough it was UNC but the top bit where the nut screwed on was metric fine.

Another screw type on cars which is not metric is self tapping screws, as far as I am aware there are no metric versions of these.

[Author]

Posts