RR Kestrel Gilman Bearings

powerandpassion — PM sent.

Thanks,

Dave

Bump. Useful discussion I need to refer someone to.

And, if you want a boisterous fire-side pub discussion-topic for a long winter’s evening (that is guaranteed to drive any female company further away than they already were) how about this: why machine everything on the bearing to sub-thousandths of an inch to keep the oil in and then cut two dirty great chamfer tunnels along each split-line edge for the oil to pee out of?

Chicks dig tribology! I have rubbed against a few and tried different lubricants, in the name of science, and I encourage all youth serious about the topic to investigate and learn how they too can get a free running crankshaft!

In terms of inexplicable chamfers I recollect a conversation with a free talking engineer who dropped the mask of confident authority when discussing a particular engine out in the market. They had no idea about fixing foaming in the sump. The configuration had been set up in the Jurassic era by some forgotten authority and never questioned. Then they cut a window in and observed what was happening inside. There are so many unique conditions set up with each configuration that it was only old fashioned observation and experiment that resolved the issue. I guess that lots of different things have been tried and unpredictable results obtained that seem counter intuitive. Maybe a partial vacuum is set up that sucks lubricant in. Maybe the chamfer aligns with the earths magnetic field and sends a signal out to alien life forms, that then back channel Ghostbusters type goo, which is they way they communicate. I hope my suggestions are being useful!

I think the bearings are cut into two halves, clamped back together, and then line-bored (or broached probably).

The repair scheme for centricast RR Kestrel big end bearings allows this as a once off refurbishment. There is plenty of bearing meat. There is far less meat on the Kestrel Main bearings and even less meat on later silver -indium bearings for new fangled engines like Merlins, so the risk is the actual bearing material is removed at the clamping edges, if this is done. I still think the bearings were line bored as an intact circle, then split. I think that differential expansion, between a nickel chromium alloy crankshaft, and a plain carbon steel shell with copper-lead-silver coating would require some clearance. The bearing shell would need to expand without buckling. I guess this was part of the bearing maker’s ‘black art’, how much clearance to allow, and what sort of blade to use to split the shell to arrive at this clearance.

There is a mob reasonably close by who make bearings for racing engines, I will have to ask. If I am found floating face down in a river, they told me the secret…

Seafire, I have to acknowledge that the Gilman Bearing and the Allison Engine Co’s inventiveness, all God fearing Apple Pie American stuff, saved the Free World, and for this, we are in your debt, Sir.
Americans are also better at fitting bearings, because thermal transfer from a Texan gut resting against a crankcase allows the caps to slip in easier….:)
As an Australian, our only contribution is in pioneering the use of flip flops, part of the national costume, to drive sticking parts in, or to plug leaks..
You will always need an Aussie mechanic in flip flops somewhere in the hangar, in case you need to quickly cut an oil seal.

…and dropped on the floor by a mechanic attempting to fit a bearing while eating a donut (only Americans do this), or the modern equivalent, while taking a customer call on a mobile phone cradled against the ear…

I beg your pardon? As an American, I can attest (with no scientific backing) that the bloody customer can wait until I’ve finished my donut! (I wonder if jelly-filling or custard cream would make a better pre-lube while installing con-rods?)

I just stumbled across this thread, and while I didn’t read every bit, I found it very interesting. And with very slight modifications, I share your vision of heaven.

bob

What I still don’t get, because it is assumed knowledge in all the literature, is that a round shell bearing, split into two, cannot be all encompassing around a crankshaft journal. In other words, when you cut a line bored, round shell bearing into two halves, some dimension, the thickness of the cutting blade, must be lost from the circumference…

I think the bearings are cut into two halves, clamped back together, and then line-bored (or broached probably).

Also, speaking from experience of automotive engines long-ago, the bearing shells are bigger than the journal in the crankcase / bearing cap; torquing-up the cap-nuts puts a distinct ‘crush’ on the bearing shells and makes the shells conform to the perfectly round shape of the journal! It also means any measurement of bearing shells not subject to this ‘crush’ is meaningless; I used to use ‘Plastigage’ to get the final bearing / journal clearance (and to selectively-fit different sized bearing shells for the best clearance). I wonder if you can even still get it?

Modern bearing shells have a distinct chamfer along the split-line to avoid any risk of ‘peening’ the soft bearing material over when the bearing ‘crush’ comes on; not to protect the crank so much but to avoid the peened-over edge wiping the oil-film off the crankshaft and leading to an almost instant bearing overheat and failure!

And, if you want a boisterous fire-side pub discussion-topic for a long winter’s evening (that is guaranteed to drive any female company further away than they already were) how about this: why machine everything on the bearing to sub-thousandths of an inch to keep the oil in and then cut two dirty great chamfer tunnels along each split-line edge for the oil to pee out of?

Incidentally the connecting-rods are numbered so you don’t mix the rods and caps up (as the journals are reamed / broached together as a set); it shouldn’t matter where a rod or a piston goes in the engine so long as each component is weighed and balanced within limits, and the bores aren’t different sizes (which they sometimes are)!

The exciting bit is that the CSIR, charged with research and development, then started a program of refining techniques for the making of copper lead (and other) bearings, so there is a cook book around….

Have found the CSIR bearings literature, and have now stood upon the Everest of vintage aeroplane engine bearings literature, and the view is good. Bearings from various Japanese radial engines are examined as well as techniques for casting Allison, Kestrel, Merlin and PW radial crankshaft bearings. Time to get in the kitchen and rattle some pots and pans.

What I still don’t get, because it is assumedknowledge in all the literature, is that a round shell bearing, split into two, cannot be all encompassing around a crankshaft journal. In other words, when you cut a line bored, round shell bearing into two halves, some dimension, the thickness of the cutting blade, must be lost from the circumference. I assume a very fine blade was used to split the round shell. So when you clamp the two halves back together, you get an egg shape, not a full circle. That’s assuming that the two half shell ends meet, or is there some float between the two? One piece of literature talks about the desirability of nothaving a pure circle, in that it forces the bearing into better contact with the crankshaft journal. By splitting the circular bearing, you would also release stresses imparted during the high temperature bearing casting process, so the halves would open up to a certain extent. I cannot image these halves, within their packaging, also holding their shape as they were rattled around from factory to workshop, and dropped on the floor by a mechanic attempting to fit a bearing while eating a donut (only Americans do this), or the modern equivalent, while taking a customer call on a mobile phone cradled against the ear…

So the important dimension is the outside dimension of the steel backed bearing, that integrates accurately with the crankcase, that clamps an egg shaped bearing shell around a crankshaft journal, me thinks..

Very interesting, thanks for the update.

No worries, pleasure.
Without taking away from the courage of Lindbergh crossing the Atlantic in 1927 and Kingsford Smith et al crossing the Pacific in 1928, both it transpires, used Gilman bearings in their ‘dependable Wright Whirlwind’ engines, so this forgotten technical innovation was perhaps critical to the success of some of the most iconic air transport achievements of the 20th century. No doubt these aviators searched for the ‘inside dope’ on any innovation that would enable success. The Gilman bearing also allowed engines in the 1930’s to move through an exponential development of more horsepower per lb of engine weight that brought us to the 1,000+ HP V12s and radials of WW2. For all this Gilman is a largely forgotten character and there is very little mention of him in Allison engine histories, and I have yet to find a photo of him. He must have been a very self effacing man.

[/ATTACH]

Rolls Royce Kestrel engines used Gilman bearings for the conrod big ends

Here is a wonderful chart found stuck on the wall of the workshop at Shuttleworth, which confirms the “Big end bearings were made by the Centricast process, using lead bronze on steel shells” and “main bearings were made by the Allison static cast process”.

[ATTACH=CONFIG]243609[/ATTACH]

So we have pretty much figured out he big end bearings for the conrods, what about the Main bearings for the crankshaft?

Now the crankshaft main bearings are backed by aluminium ‘blocks’, which, given aluminium’s lower melting point, could not cope with the 1050 degrees C the steel backed bearings were subjected to. So now we need to investigate whether the “Allison static cast process” was a fancy combination of steel – lead/copper-alumiunium casting.

Very interesting, thanks for the update.

In the Kestrel rod in the photograph there was a thin sheen of oil on the contact faces when the rod was separated, so no doubt at 3000 RPM the conrod bolts are stretching and, in a fractional sense, this becomes an articulated joint. Now I am starting to think about the extravagant conrod shell bolts and their metallurgy that are a feature of Kestrel and Merlin, that have alternately thin and thick shanks, and a keyhole and pin on the head arrangement to prevent rotation. Rotation could only occur if there was a loss of clamping, consistent with stretch. I thought the thin & thick shank was light weighting but now I am thinking that this allows a factor of controlled stretch.

A wise old man told me about how the axle shafts on his 1930’s British racing car kept shearing, and this was a problem with similar cars in the car club. More powerful engines slotted in and drive trains that ‘could not take it’. Replacing an axle shaft was a real pain, worth groaning about with other club members at the bar.

He was an aircraft inspector in real life. So he pulled out the axle shafts from his car and machined down the shanks of the shafts, thinner in the middle, original thickness at both ends. This meant that the shear load, instead of concentrating at the weakest point of a shaft of standard thickness, would transfer to the longer, thinner portion of the shaft. Because there was ‘more’ thinner portion, the shear force spread out over it, causing it to twist, but not break. So he never had another breakage again. He became famous for never breaking an axle shaft on Hill Climbs. Everyone in the club begged him to explain how he had solved the problem. Surely he had found shafts made of some exotic metal, or made thicker shafts. He held off long enough until one day he explained, ‘machine your shafts thinner’. Everybody laughed and considered it a good joke, and kept on breaking their axle shafts.

In respect of the manufacturing of CuPb bearings for RR Kestrel engines the following description from 1938 seems to match well :

Handbook of Aeronautics Vol II Aero Engines Design and Practice Third Edit 1938

Lead Bronze Bearings pg 26

The bearing shells are generally lined with the lead bronze by means of the centrifuging process. In this process the shell is manufactured with one blank end and left a rough machine finish. The shell is then mounted on a spindle either vertically, in which case the top is left open, or horizontally with the open end sealed by welding and rotated in a furnace, enough lead and copper in the correct proportions having been put into the shell to make a lining about one eighth inch thick. The temperature to which the shell is heated is in the order of 1050 degrees C and the speed of rotation is dependent upon the diameter of the shell, but for a normal size, say 3 inches in diameter, the rate is 800 RPM.

A small amount of phosphorus , about 0.05%, is generally added to the lead copper content and this acts as a deoxidant, having a cleansing affect on the metal and interface.

After reaching the desired temperature the shell is cooled either by air blast or oil quenching whilst the same rotational rate is maintained. The rate of cooling is important as it governs the nature of the dispersion of lead in the copper. These metals do not alloy, but merely mix and they segregate with slow static cooling. It is therefore important that the cooling conditions be such that the maximum amount of dispersion of lead in copper is obtained.

Here is a youtube of ‘centricast’ bronze casting, which incorporates what appears to be red phosphorus and water chilling, as per the process described in 1938 :

https://www.youtube.com/watch?v=rTcq39TOEjs

[/I]Method of Making Bearings US patent 2130461
The bearing resulting from the practice of the method herein set forth is made the subject matter of another application No. 575,117, filed November 14, 1931, as a division of this application.
NORMAN H. GILMAN.[/I]

Just to solidify the chronology of the adaptation of the Gilman bearing the following extract from the SECRET ” Historical Summary of the Royal Aircraft Establishment 1918 -1948″ by CF Caunter Report AERO.2150A Nov 1949 details, under Metallurgical developments for 1931 :

“Following an examination of Allison steel backed bearings, experiments on methods of producing lead bronze bearings in steel shells were initiated. The method of pouring molten lead bronze into rotating steel containers and allowing the metal to solidify in situ was developed… The preparation of bearing shells on these lines was demonstrated to interested firms, and a period of vigorous development followed in connection with aero engine applications”

So we have confirmation of technology transfer from Allison to the British aircraft engine manufacturing industry in 1931, consistent with Gilman updating his original patent in 1931.

So I would guess that the Bristol Pegasus, as a development of the Bristol Jupiter, went to Gilman bearings, and the Nappier Dagger too. This is an engine technology that persisted in designs from 1931 to 1938, and as few of the engines remain from this era to be rebuilt, fairly obscure today. Still you never know when you might need to rebuild a Napier Dagger…

The ‘locking tangs’ were small areas bent into the thin steel backing for the bearings; I certainly wouldn’t go so far as to call it a forging process. The bearing material was (probably) added after the steel backing was bent.

I just came cross this thread? Does anybody know or have any details in a publication of how the locking tangs on the early shells were incorporated into the steel portion of the bearing shell? I would have thought that if forged – as described above, then the tangs would have been formed at this time and then coated. Were any of these tangs welded on the back of the shell, or was this something fitted to the ground engines (ie Meteor/Sea Griffon)?

FB

In respect of the manufacturing of CuPb bearings for RR Kestrel engines the following description from 1938 seems to match well :

Handbook of Aeronautics Vol II Aero Engines Design and Practice Third Edit 1938

Lead Bronze Bearings pg 26

The development of greater power and higher speed in modern internal combustion engines has brought about a demand for bearing metals capable of working at higher temperatures and pressures than is possible with the usual type of white metal bearings.
To meet this demand, bearings lined with an aggregate consisting of lead dispersed in a matrix of copper or copper alloy are being developed.

Bearing shells should be of mild or medium steel or 5% nickel steel. In general alloy steels are not suitable as the cooling from the high working temperatures upsets the heat treatment of the steel.

The bearing shells are generally lined with the lead bronze by means of the centrifuging process. In this process the shell is manufactured with one blank end and left a rough machine finish. The shell is then mounted on a spindle either vertically, in which case the top is left open, or horizontally with the open end sealed by welding and rotated in a furnace, enough lead and copper in the correct proportions having been put into the shell to make a lining about one eighth inch thick. The temperature to which the shell is heated is in the order of 1050 degrees C and the speed of rotation is dependent upon the diameter of the shell, but for a normal size, say 3 inches in diameter, the rate is 800 RPM.

A small amount of phosphorus , about 0.05%, is generally added to the lead copper content and this acts as a deoxidant, having a cleansing affect on the metal and interface.

After reaching the desired temperature the shell is cooled either by air blast or oil quenching whilst the same rotational rate is maintained. The rate of cooling is important as it governs the nature of the dispersion of lead in the copper. These metals do not alloy, but merely mix and they segregate with slow static cooling. It is therefore important that the cooling conditions be such that the maximum amount of dispersion of lead in copper is obtained.

Adhesion between the lead bronze lining and the steel shell can be mechanical, by suitable milling of the surface of the shell, but it is preferable to leave the shell a rough machine finish. Satisfactory adhesion is obtained by the use of this method and the tensile strength of the interface is greater than that of the lead bronze mixture itself.

The softening point of lead bronze depends on the melting point of lead, namely 327 degrees C. The softening point of white metal depends on the melting point of tin, namely 232 degrees C.

The use of centrifugal casting makes obvious a further rationale for the three piece RR Kestrel conrod design, although not for the later Merlin design with removable bearing shells.

The comment on use of mild or medium rather than alloy steel bearing shells due to engine (?) working temperatures affecting heat treatment challenges my earlier comments on the use of Nickel Chrome steel for bearing shells, although AP 1416 D Vol III Kestrel V Schedule of Spare Parts lists the appropriate parts (and conrods) as NS under material, I assume Nickel Steel.

WH Hatfield ( in charge of Firth Vickers research laboratories) Ferrous Metallurgy in Aeronautics from Aircraft Engineering magazine May 1935 pg 113 provides in great detail a list of steels and applications in the aviation of the day.

In this there is detailed only one 5% nickel type composition, being described in British Standards S4 (sheet), S 67 (billets & forgings) and S 83A (billets & forgings), all of identical chemical composition, moderate weldability.

Tempering on this composition is at 570 degrees C, I am not sure what temperature an engine can get to in the crankcase, but I guess this refers to the comment about not using heat treatable alloys in this application due to engine working temperatures.

Further appendices in Aircraft Engineering by the same author June 1935 parts are listed with suggested steel types.

Con rods : S11, S65, S81 (all NiCr)
Main bearing caps, covers & housings : S.70 (Med Carbon), S 11

I figure answering this question on what the bearing shells are made out of is best progressed by sampling the material.

I am getting so drawn into this that the only way to extinguish this out of my mind is to make up some bearings……

More

In the interests of completing the story, ordering my own confusion and promoting the cause of illuminati who love conrods I make the following update :
.

Because I am a dumbkopf I now make the following realizations care of the observation of a practical informer, in reference to the posted pictures of conrods :

1. Three piece RR conrod was due to the adoption of a steel backed & shell design that needed to be separate from the conrod shaft to enable external machining of the plain rod bearing face. Ahhhh! In the Kestrel the plain rod bearing face was cast in CuPb. In the Merlin and Griffin a steel backed CuPb then Pb-Indium bearing was fitted to the machined face of the heavy duty Nickel Chromium bearing support. All needed ‘outside’ machining access.

2. In the Allison V 1710 Gilman’s 1938 patent of steel backed CuPb bearing shells was adopted, allowing a conventional two piece conrod. Less metal, less machining, and from what all the old timers say, a very, very reliable motor.

3. I still think that an additional benefit of the RR approach was a doughnut shaped forging that was stronger than a conventional two piece conrod forging.

4. I am entirely wrong in thinking that the three piece RR conrod was an articulated joint, if I accept the conventional wisdom that the clamping forces of the conrod bearing retention bolts were so specified to prevent this from occurring. I still observe that the Kestrel had a castellated nut and cotter pin at one end and a retention lug at the other end, when a conventional hex head bolt may have sufficed, as it does for other designs. I accept that no part of an aeroplane has more stress or energy imparted to it than the conrod, and the RR conrod is whippet thin. Something in the design allowed it to be so. I really need to sit at the grave of Mr Royce at midnight and see what whispers I can hear rising in the mist. No paper clip passed Mr Royce’s desk without being interrogated and improved. Why did the Merlin persist with a three piece conrod when the Gilman influenced Allison went for something simpler ?

Also now that I have seen a DB 601 conrod I now understand why there was such a fuss made over taking out the Nazi bearing industry and Swedish bearing supply. The DB engines were a far more complex and expensive design concept than the Gilman idea, no doubt strong and reliable, big Bavarian, beer drinking conrods. All these engines, generating the same horsepower bands, but utterly different in their guts.

I guess I am starting to see some of the Schneider racing experience feeding into the RR design mind and the entrepreneurial filching of good design ideas from other sources that is the mark of a dynamic, hungry organization. If 1935 RR was alive today they would be making solar chlorophyll powered turbines to win the 2015 Schneider Solar Challenge…

Ahh the fun I can have with an old conrod, and the stories it can tell !

Thank you powerandpassion. Amazing stuff.

Manufacturing bearings

That’s a very interesting bit of gen, thanks for putting it up.

In the interests of completing the story, ordering my own confusion and promoting the cause of illuminati who love conrods I make the following update :

1923 – Gilman lodges patent for cast in CuPb bearings, granted 1928
1930 – As RR “F” engine resolves into Kestrel, Gilman cast in bearings adopted by RR.
1938 – Gilman lodges new patent for steel backed shell bearings, these bearings incorporated into Allison V 1710, British RR Merlin, Packard Merlin, RR Griffon.

Only the Kestrel had cast in bearings that cannot be removed. Apparently Charles Lindbergh crossed the Atlantic on reliable Gilman cast in bearings in 1928 so early Wright Cyclone ( Bristol Jupiter ?) used cast in bearings.

Unique (?) to RR and entirely separate to Gilman’s innovation is the three piece conrod. Now I am starting to think that this is a RR innovation that had a specific purpose. In the Kestrel rod in the photograph there was a thin sheen of oil on the contact faces when the rod was separated, so no doubt at 3000 RPM the conrod bolts are stretching and, in a fractional sense, this becomes an articulated joint. Now I am starting to think about the extravagant conrod shell bolts and their metallurgy that are a feature of Kestrel and Merlin, that have alternately thin and thick shanks, and a keyhole and pin on the head arrangement to prevent rotation. Rotation could only occur if there was a loss of clamping, consistent with stretch. I thought the thin & thick shank was light weighting but now I am thinking that this allows a factor of controlled stretch. Perhaps RR, struggling with the destruction of whitemetal bearings in “R” & “F” engines came up with the concept of fractionally articulating conrods as a control measure. Certainly it is a unique feature which seems to travel through to Griffon.

There are some pictures of Allison, Merlin & Griffon conrods stolen from enginehistory.org

In my perambulations through the world of conrod bearing design I came across a very helpful and knowledgeable gentleman who made the following notes on Griffon/Gilman bearing shell manufacture :

You may find the following extra notes of some use regarding the actual manufacturing process, as done in Rolls-Royce in 1954 when they were still building a few “Griffon” engines. They are taken straight from the diary I kept while passing through the works as a graduate trainee.

“Lead-Bronze Foundry
Steel-back bearings are cast direct into forged outer casings, with a tin inner sleeve swaged in to hold molten bronze. Bushes are cast into tins. Throughout engine the Cu-Pb basic is reinforced with tin to give required strength. Sn varies from 1% on mains to 10% on little end bushes and exhaust valve guides. Tins, etc, are dipped in molten flux before pouring, and molten pour is quenched to prevent settling of mixture & to improve strength.”

Another note reads:-
“Stress relieving
Eg. steel shells of bearings are heated at 200 degrees C for 3 hours”.

I recollect that, after making the bearings they were checked for circularity on a jig and, if not right, given a few taps with a mallet by an experienced employee for correction!

Regards, Derek Taulbut.

In trying to understand what “tin inner sleeve swaged in to hold molten bronze” meant things got a bit fuzzy, after all this occurred when ERII was still a virgin. I would welcome any recollections from anybody that ever made these bearings and in return will apologize for linking the Sn with HM.