To pin, or not to pin

[AlanWParticipant @alanw96569]

Should I back up high strength loctite joint by pinning?

[AlanWParticipant @alanw96569]

I have joined a cylindrical CR2 ER32 collet extension onto a shaft using Loctite 638, the high-strength version. My dilemma is whether to trust the joint or drill through the diameter and fit a roll pin. I have previously extended a pulley boss using standard Loctite only and that is still giving good service on a 3/4 hp motor. If that failed it would be straight forward to rectify but failure in this instance would risk irretrievably damaging the shaft that would be difficult for me to replicate (hence the need to join the two items).

The specified shear strength of 638 is very good but just how good is very good? Use is for milling under the power of the aforementioned 3/4hp motor via speed reduction pulleys

The 32mm o/d tubular shaft has been bored concentric to 21mm id and sleeved (using 638) to reduce the bore diameter to fit a 5/8" shaft. I am now concerned that the 50mm long x 5/8 diameter wetted joint area is not sufficient. But will pinning weaken the 5/8" diameter shaft unduly?

Sorry about the mixed measurement systems. Can't be avoided.

Any advice appreciated.

Alan

[not done it yetParticipant @notdoneityet]

If you search, you will find an identical thread starter only about 2 months ago.

search ‘pin’ as the other one is slightly different, grammar-wise.

Edited By not done it yet on 12/11/2019 17:37:45

[AlanWParticipant @alanw96569]

Thanks NDIY, I missed that. Not much help though, with all the conflicting opinions. I think I will trust the Loctite, after all, I have made absolutely sure of the fit and cleanliness before joining.

[old martParticipant @oldmart]

The racks pinion shaft on the museums Smart & Brown model A apron was very badly worn on the od particularly next to the gear where it ran through a bronze bush. The bush was easy to replace, but the shaft and gear was not. I turned up a new shaft with a reduced diameter where the gear went, and bored the old shaft to the same size deeper than the width of the gear. The gear was then parted off the old shaft and Loctited onto the new shaft with 638, after thoroughly degreasing the parts. The joint was about 5/8" diameter by 1/2" wide. After the rest of the wear in the apron was attended to and it was re assembled, there was a suspicious amount of slippage. The Loctite wasn't enough. Mike found some 1/16" diameter needle rollers, and he drilled the joint axially in three places 120 degrees apart and Loctited them in place after cutting them to length. Fortunately there was enough thickness of metal to do this and it has been ok ever since.

Your joint length is nearly 4 times as long as the one I had made which slipped, which is a good sign. If you do decide to use a pin, then fitting a properly reamed taper pin would be much better than a roll pin.                                The strength of the joint is increased if the surface finish of the components is not too smooth, a ground finish in one half of the joint would not be as good as if both were being turned.

Edited By old mart on 12/11/2019 19:33:56

Edited By old mart on 12/11/2019 19:45:55

[Martin ConnellyParticipant @martinconnelly55370]

There is always the risk that you will machine something that gets hot and the heat transfers to the joint. That is when you will find out how well Loctite works when it is hot.

Martin C

[Ron LadenParticipant @ronladen17547]

Posted by AlanW on 12/11/2019 16:29:23:

failure in this instance would risk irretrievably damaging the shaft that would be difficult for me to replicate

Alan, I would have thought that answers your question……pin it to be certain.

[AlanWParticipant @alanw96569]

Thanks folks. I abraded the shaft well before joining and the finish of my bored sleeve would not have been acceptable if it were in view. I am still leaning towards pinning so the comment re taper pin v roll pin is interesting.

Alan

[VicParticipant @vic]

Posted by Martin Connelly on 12/11/2019 19:43:24:

There is always the risk that you will machine something that gets hot and the heat transfers to the joint. That is when you will find out how well Loctite works when it is hot.

Martin C

Good point Martin!

[AlanWParticipant @alanw96569]

Martin C and Vic,

I wouldn't work any machine that hard!

Edited By AlanW on 13/11/2019 16:50:43

[Adrian 2Participant @adrian2]

A few years back to reassure myself of the strength of Loctite I assembled a large metric (20mm) nut and bolt using high strength thread lock.

Bolt head in vice ring spanner on nut and try to undo. Very impressed it will not let go. Cleanliness is all important. It also takes considerable heat to break down.

Have faith.

Adrian.

[AlanWParticipant @alanw96569]

Adrian,

That is reassuring and the 638 is supposed to be resistant to high temperatures.

Alan

[old martParticipant @oldmart]

If you can keep the temperature below 300C the Loctite will be ok. Two smaller taper pins spaced evenly in the 50mm joint and at right angles to one another would retain most of the original strength of the joint. 2.5mm pins would be a good size.

[AlanWParticipant @alanw96569]

That's useful thanks, Old Mart. Another method I had considered, that shouldn't compromise the strength of the shaft, is to drill three radial holes to just dimple the inner shaft and tap for grub screws. Any views on that suggestion?

[Anonymous]

Posted by old mart on 13/11/2019 19:29:09:

If you can keep the temperature below 300C the Loctite will be ok.

Where did that come from? Looking at the datasheet for 638 it is down to about 35% of initial strength at 200°C. Extrapolating the curve (always a dangerous thing to do) it crosses the zero axis at about 250°C. I wouldn't trust 638 at 300°C.

Andrew

[not done it yetParticipant @notdoneityet]

Posted by AlanW on 14/11/2019 15:34:38:

That's useful thanks, Old Mart. Another method I had considered, that shouldn't compromise the strength of the shaft, is to drill three radial holes to just dimple the inner shaft and tap for grub screws. Any views on that suggestion?

Why ‘dimple’ it? You could drill in, thread, use high tensile fixings and loctite those, too.

[old martParticipant @oldmart]

I had to get a joint apart and it was still a tough job even at 300C, you should try it some time, the data sheet errs on the side of safety, with litigation in mind.

Grub screws are fine if there is a decent wall thickness for the threads.

Edited By old mart on 14/11/2019 19:05:06

[AlanWParticipant @alanw96569]

Wall thickness is just over 8mm. Should be plenty.

NDIY, I don't want to drill into the 5/8 shaft for fear of creating local weakness.

To put the temperature concerns to bed. This job is further a development of an adaptation to a Pollard bench drill for milling. It has only ever been capable of light milling and has never been worked hard enough for the chuck to get warm, let alone up to 300 or even 200deg C. Even if I had a 'proper' mill, I would be looking for 'something wrong' if the tool chuck started to get that hot.

The reason for this mod? Originally, I made a ER32 collet chuck to fit the existing No2 Jacobs taper but this gave quite an overhang from the bottom bearing = lack of rigidity + loss of headroom when drilling. This mod reduces that overhang by around 25mm and, by fitting a secondary bearing housing via a spigot into the existing bearing recess, allows a much larger bottom bearing to be used (32mm id).

Alan

[old martParticipant @oldmart]

That is a similar reasoning that I came up with when I decided to increase the spindle of the Tom Senior mill from MT2 to R8. The bottom bearing is increased from 1" bore to 35mm bore.

There are several ready made small mill lower parts which have the X and Y axes on the market which are potentially very useful when making a mill. I found this one, but there are larger ones about.

**LINK**

Here are some pictures of the two halves of the spindle which was bonded with Loctite 620. No pins were needed as the halves also screwed together.

Edited By old mart on 15/11/2019 16:43:44

[AlanWParticipant @alanw96569]

Hi Old Mart.

I'm using the largest of the x-y tables available from ARC, the same one used on their Super X1L mill; it's been OK for my uses. I'm also fitting stops while the machine is stripped for the spindle mod.

Unfortunately threading and screwing my parts together wasn't an option. Nice to know that someone else has carried out a similar mod; I don't feel such a loon now!

Many would say "Don't bother to mod a bench drill for milling" (and so would I, now) but having already invested a lot of time in the adaptation, I feel it would all have been a waste if I gave up now for the sake of a few tweaks. It does all I need it for so long as I am patient when working steel.

Alan

[old martParticipant @oldmart]

You must post photos as you go. It looks like it will be a nice machine when it is finished.

[AlanWParticipant @alanw96569]

The machine has been finished and in use for getting on for seven years now and this is just a development (I never gave up chasing rainbows). The original adaptation was the subject of an article in MEW five or six years ago. I would quote the issue number but can't find the on-line index any more. I can put some photos up, though, when it is back in use. I promised Neil an article on further development but that is so long ago now that I doubt if it would still be of interest.

Alan

[Michael GilliganParticipant @michaelgilligan61133]

Alan

Personally … I would have no hesitation in using 638 without pinning

I have fitted a new nose to a lathe spindle with 638 and then machined the Myford thread onto it without any problem.

MichaelG.

[AlanWParticipant @alanw96569]

At last, some experience to back up the theory. I take it, then, that it is withstanding turning jobs where a lot of torque is present. In view of that, Michael, I may just reassemble and give it a try.

Alan

[Andrew EntwistleParticipant @andrewentwistle]

Hi Alan,

Another data point; seven years ago I secured the 1/2" throw crank on my power hacksaw to the 5/8" diameter crankshaft (both MS) using 638 only, with no issues.

Andrew.

[Author]

Posts