Shrink fits for the model engineers workshop?

There seems to be quite a lot of information on the web for shrink fits. However very little on the temperature differences between the two parts. If I want to fit a sleeve to my Jacobs chuck (which had a split outer sleeve) what difference in diameters would I need (thous per inch)?

OK, first one has to define the lowest temperature for the split sleeve and how hot can ! get the new sleeve? In the days before retirement I would have dunked the split outer sleeve in liquid nitrogen and job done.

I am now limited to a domestic freezer and gas stove (or even a propane torch!). so how many thou per inch of diameter should I allow? Plenty of info on the net, but I suspect the temperature differential is somewhat greater than I can manage!

I can afford to louse up the operation, as the chuck now has an intact donor outer sleeve. But the split sleeve is sitting on the bench and seems to be challenging me to experiment!

If anyone has done this type of job within the limited temperature differential that is available to us. I would love to hear what thous per inch was used in coming to the negative size difference for the two components.

Andrew.

If you look at the relevant limits and fits standards to establish the required size differences then jus use what heat and cooling you have all should be well. I think you are overthinking it, use of heat and cold merely makes the fitting process easier, ie a force fit with the right temperature difference will drop in place. I have used professionally nitrogen for fitting bushes in holes and heatiers of various forms to apply bearings to shafts.

Very approximately, for steel, 100 deg. C differential gives 1 thou. per inch expansion difference. This seems achievable with domestic equipment but the crunch will be the need to mate the parts extremely quickly as the temperature difference will reduce rapidly as the items are slid together.

Edited By Clive Brown 1 on 17/07/2023 20:52:45

My advice is: Don't. Shrink fits is hard on small parts, they loose heat too fast. Anything short of a sausage-in-a-hangar fit will surely bind halfway down the joint and seize permanently.

At least take the time to work out a reliable method in advance. As for the details, consult this page: https://www.engineeringtoolbox.com/linear-expansion-coefficients-d_95.html

If this is any help.

When I made my Comparator, the 6.125" diameter x 1" thick, cast iron base was reamed 1", because that was the largest reamer available.

Having measured the resullting bore, the 1.125" column was turned down to be a thou bigger thann the bore.

The coumn spent the night in the freezer, so should have been at -18. The cast iron base was in the oven for two hours at Gas Mark 6.

The column was removed frrom the freezer and wrapped in a towel while the base was removed from the oven and placed on a piece of plywood on the kitchen floor.

The column was unwrapped, and instantly became covered in frost. It was a loose fit in the hot cast iron base. As time pssed the frost disappered up wards from the base.

Once evrything was at room temperature, the column was, and has remained, a tight, immovable, fit in the base

Which was the object of the exercise.

REcently did something similar with a a "plug" tdestined to fill a 26.5 mm hole in an embryo backplate for a 80 mm chuck. Timescales were shorter, so the pkug requiured some assistance to enter squarely intomn the bore, but withstood subsequent machining and boring and tapping 3/4 BSW. So the 0.001" interefernce fit must be sufficient.

Howard

I have "1 thou per inch dia" in my head from somewhere as a "general engineering" interference fit.

I come across interference fits mainly in two applications - One is for pressed up cranks in engines (mainly JAP V twins), here the interference typically used is 4 thou, some folk push this up to 5 thou, this is on shafts of around 1 1/2" diameter depending on the application. The shafts are case hardened and ground, the crank discs are surface hardened - tuftriding for example. The hole in the crank needs a bit of a lead and the shaft needs a bit of a radius - they are almost always pressed together cold with lubrication (Castrol R seems popular) . Failure to surface harden the crank leads to the shaft broaching it's way through with the loss of the interference fit.

The other application is the fitting of valve seat inserts in cylinder heads. Interference fits here vary, Serdi (manufacturer of valve seats and associated tooling) seem to use 6 thou interference. Their tooling system aims to give the 6 thou interference by design. It is not uncommon then to see people hammer the valve seat into place (see YouTube) . The standard Serdi valve seats have a generous nose radius to help them in and a mandrel to guide them is also used, fitting into the valve guide. People working on our pre war JAPs tend to ease off on the interference fit, down to 3 thou or maybe 4. Valve seats falling out can be a problem, but there are various factors in play here.

I have been experimenting fitting valve seats in a pre war head and used the Serdi 6 thou interference, but used temperature difference to fit them - I am not a fan of the lump hammer. The CI head went into the oven at top wack, then the local AI man came around with his liquid nitrogen to cool the seat down. With that delta t the seats dropped in, only a light tap with the hammer was used to make sure the insert was properly seated. The old CI head survived the heavy interference without cracking.

Ian

Andrew,

If it is of any help and reassurance, in your earlier post I mentioned that I had seen a successful repair to a multipoint Britool socket with a sleeve over it to close a full length split.

My friend Richard who did that work was a time served engineer from Parsons Turbines and being a mean old sod he would have used whatever came to hand such as a piece of water pipe to make the sleeve! I don't suppose there was much sophistication put into the fitting job either, maybe some heat and a quick bashing together on an anvil top with a thick walled collar under a big hammer. I don't think for one moment he would put his mind to the degree of interference fit other than to bore a short lead into the collar that just fitted the socket and finish the rest of the bore to 1 thou under that.

Regards

Brian

I just overhauled a Jacobs #36 20mm chuck. I pushed the outer sleeve off using a 6" vise along with a collar the correct size I machined for the purpose. After cleaning/derusting & oiling I pushed it back on using the same vise.

I have overhauled quite a few Jacobs chucks & have never seen the need to use any form of heating/freezing to get them to fit together.

regards

I think the point was that the sleeve was split and this was a repair

One thou per inch is the rule of thumb, but here you are shrinking one thin sleeve onto another. The inner one will reduce in diameter, and since that itself is then an interference fit on the rest of the chuck, I'd go for a lesser interference. After all, you're not trying to transmit any load through the joint, just trying to stop the crack opening up, so the requirement is just that it doesn't drop off in normal use. If you are unsure of your ability to measure and turn to small limits, let loctite be your friend.

As I understand it, this sleeve is intended to repair a Jacobs chuck with a split outer shell. In turn, the outer shell needs to be sufficiently tight on the internal split threaded actuating ring to prevent slip. I have had Jacobs chucks where these parts slipped when being fully tightened despite not being split! I doubt the repair will be successful unless the sleeve is substantial and a tight press fit.

Thanks everyone,

Malcolm, I have had the same experience with a Jacobs chuck. I think, on reflection, that brazing would be unlikely to work and I suspect a decent interference fit would be needed for the outer sleeve to do the job.

Andrew.

Tried posting this a few minutes ago and it disappeared, apologies if it ends up doubled-up.

Just completed my first attempt at shrink fits, putting steel rims on cast iron wheel centres for a model loco. Wheel centres were 67mm dia, bored the rims to 0.1mm undersize, dropped the centres in the domestic freezer overnight, then this morning put the rims in the oven at 220°C for about 45 minutes, took them out and the centres just smoothly dropped in under their own weight, all except one . It is now stuck about 4mm from home with a slight misalignment. Don't want to press it the rest of the way as I don't have a press and doing it in the vice may put too much stress on the fairly delicate centres.

Thinking I could pop the assembly onto a loose fitting shaft, and rotate it slowly in front of the gas torch to see if I can heat the rim up enough to allow the centre to seat properly with some light pressure.

Any other suggestions?

For smallish parts, you may find a heat-shrink-tubing gun quicker and more efficient to do the heating bit.

(My first post on this disappeared too).

Assuming mild steel rim on a cast iron wheel, there will be next to no differential expansion between them by heating. The stuck shrink fit parts will also be in sufficiently good thermal contact that they heat up in unison (unless you can attack the rim with a large jet oxyacetylene torch, which might just work if it doesn’t melt first).

There is really no alternative to resorting to press fitting to complete assembly. What this needs is tooling (parallel thickness spacers to directly support the offending parts, and a press or vice that ensures parallel and axial motion. With a sound setup there is a good prospect of successful recovery.

If all else fails hacksaw radially through the tyre to relieve the stress, then it will come off easily. If you're careful you won't mark the wheel, even if you do a small mark won't matter.