Interference Fits – rule of thumb?

[norm nortonParticipant @normnorton75434]

I want to permanently fix a model locomotive cast iron valve sleeve of 1.25″ diameter, with a thick wall, in a cast iron cylinder block. I think the standard practice is to make this a cold force fit. The surfaces must fit well or steam will leak between the sleeve and cylinder block.

One option is to use loctite but I don’t like this as it will wipe away from critical joints and flow to fill many ports. It also softens when exposed to pressurised steam temperatures of 160 degC, although the cylinder block will perhaps not get that hot.

I have just suffered a small disaster trying to be clever. I knew exactly the diameter of the bore from plug gauges, and made the liner 0.002″ oversize. If heated to 200 degC the cylinder block bore will increase a shade more than 0.002″ and so I did this and had a 2 ton press ready to force the liner in. It slid nicely for the first inch, then stiffened and stopped at 2 inches insertion, stopping short by one inch. Oh dear. 2 tons will not move it either way now, even when reheated to 300 degC.

I have learnt a lesson. The heat transfer from block to liner was too rapid to allow the fitting. Never try that again! I will carefully bore out the stuck liner and make a new one.

My question is, what is the rule of thumb for cold cylindrical interference fits? Looking on the web there are complex calculators that need to know the surface friction coefficients. Does anyone have knowledge of doing this? I have a 3″ long thick wall cast iron tube of 1.25″ diameter and up to 2 tons available to push it in. My gut feel tells me to look for 0.0005″ of interference (half a thou) but should I go for less lest it all gets stuck again?

Norm

[HopperParticipant @hopper]

Shop floor rule of thumb for press fits is one thou per inch of diameter. So you were perhaps a little tight but not drastically so. Just a tight press fit! Crush fit would be 4 to 6 thou per inch diameter. (Think new Harley pressed-in crankpin to flywheel fit.)

But all is not lost. A 2 ton press is nothing. Well almost nothing. A standard home workshop hydraulic press around here is 12 ton, and a decent commercial grade hydraulic press is 20 ton or even 50 ton. Find someone with a decent press and press it the rest of the way in cold. (Providing your casting is not so fragile the tight press fit might crack it?? In that case, get the 12 ton press to press it out and reduce the diameter a smidge.)

If you can’t access a decent press, there is a lot can be done with thoughtful use of a four pound hammer and brass drift if job is supported on a good solid steel base or large vice etc. This is actually a preferred method for a heat shrink fit as you attempted, because you can whack it in FAST before the heat transfers and expands the cold bit.

And for future reference, no need to be so shy about Loctite. The high temp grade works well. And it does not “wipe away” on press fits. It lurks in the surface striations and pores of the metal and that is sufficient to do the job. Loctite literature confirms this. Needing a gap for the Loctite to fill up is internet mythology. Likewise, it will not block up ports etc as it is anaerobic so the air in the ports keeps it liquid so it can be cleared out easily. If you were to go with half a thou interference, I would definitely add Loctite of suitable grade.

It is also important to have a good chamfer on the leading edge of the item being pressed in, and on the entrance to hole you are pressing it into. Makes a remarkable difference.  Adding a lubricant such as oil can help too, but Loctite acts as a lubricant when still wet so is my preferred choice.

[Clive FosterParticipant @clivefoster55965]

Given the apparent prevalence of stuck liner problems when relatively inexperienced folk attempt this sort of job in the usual manner using parallel sleeves and bores I do wonder if the self gauging taper method of producing interference fits might be a reliable alternative.

For self gauging taper fits the two components are made with a taper on diameter of twice the desired interference fit over the full mating length. Using one component as master, usually easiest if it’s the male one, the second one is finish machined so that the master enters “exactly” half the length of the taper section. The taper angle is invariably very small so the size can easily be adjusted by longitudinal shifting of the cutting tool to make the cut a bit deeper all the way along. The longitudinal adjustment is pretty insensitive so, providing your taper cutting method is reliably repeatable it’s quite easy to end up with an adequately accurate pair of components.

Almost certainly easier than producing a known, accurate, 2 thou or so variation between parallel male and female parts using typical Model Engineer or Home Workshop Guy/Gal equipment. I suspect that for most folk the actuality is closer to “some interference fit” rather than “exactly what the design called for”. Which objectively matters little so long as it works.

A huge advantage of self gauging taper fits over parallel ones is ease of assembly. The taper makes things pretty much self aligning and full interference only occurs when the parts are fully home so the forces involved with cold insertion are much less. If the female component is expanded by heating the male pretty much drops right in, with perhaps a light push for the very last bit, so far less time for troublesome heat transfer effects to occur.

Clive

[norm nortonParticipant @normnorton75434]

Thank you Hopper, your cheerful and friendly response inspired me to have another go. The comment that 0.002″ interference might be a 10-20 ton press or big hammer job told me that it ought to move and come back out again.

I  made a brass drift and a ring for the cylinder block to sit on, on a heavy steel base, used my biggest club hammer (3 1/2 lb), and gave it some big swings. Once it started to budge it steadily moved more freely. About 50 strikes. Some scuff marks in the bore where the cast iron has picked up, and the insert will be skimmed to a 0.0005″ interference. I agree, some loctite will help lubricate and do no harm. Might help seal the scuffs as well. And I hear what you say about some decent chamfers, didn’t have those before.

Norm

[derek hall 1Participant @derekhall1]

When I used to work in the water industry and made new brass sacrificial sleeves for the pump shafts we used the rule of thumb 0.001 inch (a thou) per inch dia.

However you have to aim for this tolerance but it was not super critical as the brass sleeve was never pressed onto the shaft. The sleeve was heated up to expand it and then simply dropped in place and allowed to cool.

Now a couple of things, I am referring to steel shaft and brass sleeves around 3 inch diameter. I would be very careful in pressing into cast iron as cast iron is not happy being “stretched”, cast iron much prefers to be in compression and may well crack if you try and press a liner into a cast iron cylinder.

I would tend to look at loctite as Hopper mentioned above.

[Merddyn’s DadParticipant @merddynsdad]

In my youth, I assisted in the assembly of a electric motor rotor on to its shaft using our 600 ton wheel press.

This job was a ‘foriegner’ for a well know turbine factory in Rugby, the shaft was cooled in a makshift bath of liquid nitrogen, the rotor ‘warmed’ with a electric blanket (industrial sized).

My assessment as a ‘youth’ was that this lot could be easily be assembled by hand, I had checked the interferences as insignificant compared to what we were used to.

But no their engineers insisted it would need all of our 600 tons to make the asssembly, additionally, just to make sure one of their team emptied several very large bottles of loctite 601 (I think) onto the somewhat chilled armature shaft before the assembly.

The assembly, when their engineers had deceided everything was ready went without a hitch, our wheel press’s pressure gauge, I was watching it closely, bearly flickered.

Why they came to us I don’t know, but I never issued the ‘tonnage’ certificate ‘cos there was never anything to measure I was only the ‘youth’.

As an asside we had to have all the doors open so we wern’t asphixiated by the several lorry loads of liquid nitroged as it boiled away cooling the shaft.

BTW the interferance used routinely by us was 1.25 thou per inch + 1 thou.

 

[larry phelan 1Participant @larryphelan1]

My only experience of this kind of thing involved placing one part in my freezer overnight, heating the other part and bringing both together, quickly. Not very technical, but it worked.

[JAParticipant @ja]

There are industrial standards for fits. With a 1.25″ diameter shaft/bore, if you require a Medium Drive fit the interference should be between 0.0021″ and 0.0035″. For a Force fit, the extreme, it is 0.0029″ to 0.0039″. There is no indication of the force required.

I will post more tomorrow morning (I avoid posting anything, or sending emails, in the evening since my mind is slightly too lively).

JA

[HopperParticipant @hopper]

The charts in my Machinery’s Handbook come out to equal about one thou per inch for what they call a drive fit. Double that for a force fit. But it was published in 1948 based on experience submitted by a major engineering works. No doubt the computer boffins have worked out better ways since then.

[JAParticipant @ja]

Some years ago I found these charts/plots on the internet. I think they came from an ISO standard.

 

Chart 1 ranks “machining” operations in terms of accuracy. The lower the IT Grade the better the inherent accuracy (and finish etc).

Chart 2 suggests what is required for a particular fit. It is up to you to decided what you want (sorry for the double entry).

Note that a hole has a upper case letter and shaft lower case. The number is the expected IT Grade. If you have ever wondered what H7 stamped on a reamer meant, it will become apparent.

Chart 3 gives dimensions (metric). It looks complex but is quite simple. Taking an example:

For a 20mm diameter shaft with a slight interference fit we can chose H7 and k6 from Chart 2. We have chosen H7 because we have an H7 reamer. This should give a bore diameter between 19.979mm and 20.000mm from Chart 3. Again using Chart 3, k6 gives a shaft diameter between 20.002mm and 20.015mm. In other words an interference of between 0.002mm and 0.036mm. This should be achievable by turning.

 

There are a lot of questions here. The main one is “What relevance has this to model engineering? Surely it is only of use where interchangability of parts is requireded”. My reply is anything that is a useful guide is worth understanding. Also one can frequently meeting repeat operations such as fitting crankpins into the six wheels of a model locomotive. I am surprised that Tubal Cain never included this in his “Model Engineer’s Handbook”. I have a vision of Walshaw wondering if his half page on fits was adaquate.

JA

 

[Howard LewisParticipant @howardlewis46836]

FWIW, I agree with Hopper as 0.001″ per inch interference.

Ideally, using the Heat and Freeze technique makes things easy.

When I made my Comparator, the stel column spent the night in the freezer, (-18’C) and then the cast iron base (About1″ thick) spent hours in the oven at gas Mark 6 (circa 200’C)

The two slid togethher easily, the cast iron did not crack (The ouiter wall was 0.5″) and ther is no sign of anything moving.

Howard

[duncan webster 1Participant @duncanwebster1]

If the valve liner is trapped by the end cover then there is no load on it, all the interference fit is providing is a seal. This being the case I’d go for the lowest interference to give a light press fit, and I’d ever so slightly reduce the liner OD where it isn’t needed for sealing.

Chap in our club fitted cast iron liners to his cylinders with loctite, and his boiler works at 120psi, so steam is hotter than usual even before superheaters. It has done a lot of miles with no sign of the liners coming loose. I doubt loctite would block a port, if you’re worried blow through with air before it’s had time to set.

[JAParticipant @ja]

Loctite thread locking and most sealing compounds cure with the absence of air. It cannot go solid and block a port open to the air. Blowing the remaining liquid out, though, would be wise.

Saying the obvious, there are a lot of Loctites and clones out there. It always worth finding the correct product and its data sheet. Then follow the instructions.

JA

[larry phelan 1]

On larry phelan 1 Said:

and bringing both together, quickly.

YES the important part that gets forgotten

 

[Charles LamontParticipant @charleslamont71117]

Full size loco cylinder and valve liners are shrink fitted with nitrogen. A tube about 3ft x 20″ x 1/2″ thick has to be slid into the cylinder, correctly oriented for the port cutouts, in a matter of seconds. Long bars are fitted to two of the cover studs, about 45 degrees either side of bottom to sit the liner on and line it up before shoving it in.

[Earny49Participant @earny49]

Hi. This is my first post on the new forum so greeting all.

I picked up this topic because I am facing a similar but almost identical situation with my work on a 5 “ gauge Duchess to Michael Breeze design. I have bored the cylinders and the bores for the liners for the piston valves but shied away from making the liners for the very same problem as experienced by the original author.

The difference with the Duichess liners is that they are in two halves, pressed in from each end of the cylinder block. Without looking out the drawings, I think a press fit of one and a half thou is called for on a 1.25” dia sleeve. I am, not unduly worried about that on the inside cylinders as there is plenty of ‘ meat’ around the bore for the liner but on the outside cylinder there is only 5/32 thickness  of cast iron  left between the bore and the outside of the casting. So I have been putting off making and fitting the liners because I was afraid that the small dimension of cast iron may easily crack when force fitting the liner.

Reading Doug Hewsons build descriptions of his B1 design in ME I was relieved to see that he recommends using loctite for securing valve liners but I also was worried about the loctite building up and perhaps solidifying in the ports and clogging them up as has been discussed above.
So this topic has appeared very timely for me and would be interested in learning if anyone has used the loctite method on a Duchess to Breeze’s design. I was going to bring up this very topic in ‘Postbag’ in ME in the near future but will see what knowledge is generated by this thread.

Finally, forgive me I I do not reply again immediately as I am off on holiday tomorrow morning for 10 days and may not be able / have time to reply.

David E

r

[Author]

Posts