Britbike forum Forums British Motorcycle specific Brands Ariel forum 1928 Ariel Model C

Oh, well. It might be of some use finding something close after you measure the current valves. Widening the search a little, some Ford and Chev valves are very close if the Ariel has 3/8" stems.

That one went straight to the pool room.

I had thought that you already had something in mind, if not then it may be worth contacting G&S valves. When I spoke to them they were very helpful and if nothing suitable is available they will make bespoke valves in small quantities. I think they quoted me for a batch of 2, 6 or 10 for my 20F. However they list valves for lots of bikes in their vintage motorcycle range so they may find something close to what you need in their existing catalogue. See here:http://www.gsvalves.co.uk/catalogues.html

If you look at option of modifying a modern diesel valve then the problem I had was that the valve catalogues are all published in pdf format with differing formatting throughout. The catalogues are huge so in the absence of being able to convert easily to a searchable database it is a long tedious job looking for valves close to that you need. ( I found some Caterpillar ones close to what I need but it took some time to find them)

John

Oh, I forgot to mention, when I spoke to G&S they didn't have valves for my 20F but had made them previously for someone else so they did have a drawing which they emailed me a pdf print of. If they have done valves for your Ariel previously they may have a drawing which they might share with you which may be of help if you decide to make enquiries closer to home.

John

I'm afraid you guys are a few weeks ahead of me so please save those thoughts for later. There are still quite a few hours between now and when I'll have the bottom end assembled (rebalance crank, install bearing, install and line bore bushes, Cr plate and regrind cam spindles, ...). Back to plating:

The bearing OD is 2.5" x 3/4" for a surface area of 5.89 in.2. According to the Caswell instructions for Cu, a current of 0.07 A/in.2 will plate 0.001"/hr. (for an increase in diameter of 0.002"), which would mean a current of 0.412 A. This turned out not to be quite correct, but more on that later.

I followed the plating instructions more closely than most people might, but not to publication-level standards. Instructions say to "try to keep the anode at least 3" away from the article being plated and no more than 9" away." The issue with this is the bigger the container the more plating solution that is required to fill it to at least 3/4". With the Cu anode wrapped around the inner edge of a 7-1/4" glass container the bearing was 2-1/2" away and I had enough solution to just get to the top of the bearing. The instructions say you should try for 3", not that you must, so that's the container I used.

Not enough Cu sheet came with the kit I bought to cover the ~23" circumference of the container so I supplemented it with two 1"-wide strips of OFHC Cu to fill in the gaps and completely surround the bearing. I placed the bearing in the center of the container and preheated those items on a hot plate, without solution as yet, to roughly the 110 oF called for in the instructions. I then added the solution and waited until it was up to temperature. I did it in this order to minimize the time the Kapton would be in contact with the solution. Once the solution was up to temperature I turned on the power supply and adjusted the current for 0.41 A. Over the course of the next hour the temperature varied between 117 and 104 oF as I fiddled with the hot plate settings. I agitated the solution manually every few minutes by stirring with the thermometer. For what it's worth, the instructions say the plating voltage will be "2 volts approx" and I measured 1.74 V.

The Kapton tape itself was fine, but not the glue. I saw it lifting at the edges so placed a block of Teflon over the bearing to keep the tape flat. I'll try to find something better to use if I need to plate another bearing, and certainly before I hard chrome the cam spindles.

After an hour I removed the bearing, rinsed it, and measured the diameter. It should have increased by 0.002" according to the information in the Caswell manual, but it had increased by just under half that, 0.0008". I put new tape on the bearing and placed it back in the solution for another hour and a quarter at a ~25% higher current adjusted for the measured deposition rate during the first hour. At the end of the second plating session the diameter had increased by the total of 0.002" I wanted. The plating is uniform, varying by no more than ~0.0002" in my various measurements.

All reliable information on hydrogen embrittlement emphasizes the importance of having the least possible delay between plating and heat treating. So, after removing the tape and brass wire electrode around its circumference (which left a very narrow unplated line) I put it in the oven. No matter what it had been ~2-1/2 hrs. since hydrogen started infusing itself into the steel.

----------- Sidebar ------------
Nb. Most people will want to skip past this sidebar. I'm writing it more to document it for myself than for any one else.

You might think that it would be easy to find the answer to the question, at what temperature and for how long do I need to heat an electroplated part, but it's not. Far from it. This was a case of, the more I read, the less I felt I knew.

Hydrogen embrittlement is a very complex subject with the exact mechanism responsible still not understood after more than a century of study. To summarize a number of sources ranging from the Caswell manual ("The possibility... is remote"), to the web (" A simple hydrogen bake out cycle can be performed to reduce the risk of hydrogen damage ... Caution: over-tempering or softening of the steel can occur,..." to recent peer reviewed scientific publications ("...the mechanisms by which hydrogen embrittlement occur and the suitable means for its prevention are yet to be fully established."), basically, the required heat treatment depends on the exact composition and previous processing of the steel because that determines the diffusion constant of the hydrogen through the lattice.

Based on quite a bit of reading, originally I planned to use 375 oF to drive the hydrogen out but I subsequently found one reference to bearing steels that quoted the lower 275 oF figure. Looking further, technical information in the FAG catalog says bearings that will operate above 300 oF need special heat treatment. That refers to operating temperature, not temperature when static, so looking further I found that ball bearing balls and races are hardened at high temperature and then annealed at 300 oF to temper them. So, I decided to use the 275 oF value.

The issue with temperature is that removing the hydrogen is an exponentially activated diffusion process so a lower temperature is much less effective than a longer time. Dropping from 375 oF to 275 oF reduces the absolute temperature by a factor of 464 Kelvin/ 408 Kelvin = 1.14, but to compensate the time has to be increased by a factor of exp(1.14) = 3.12. So, I simply have to leave it in the oven for 3.12x longer than, well, longer than what? I couldn't find a reliable source for the time for a bearing steel.

I found that bearing races have a hardness of 58-62 Rockwell C, and the ASTM document B 850-98 (2009) Standard Guide for Post-Coating Treatments of Steel for Reducing the Risk of Hydrogen Embrittlement has a table of heat treatment times at 375 oF for hardness values from 31 up to 51. Extrapolating those values to ~61 gives a time of ~30 hours. Multiplying by 3.12x gives ~94 hrs. Based on this quasi-scientific analysis, I'll be heat treating the Ariel's bearing at 275 oF for roughly 4 days.
---------------- end sidebar ---------------

The bearing has been in the oven at 275 oF for ~20 hours as I upload this so it has another 3 days to cook.

I've been looking forward to this electrochemical bit.
The project when doing my degree, was on the solubility of hydrogen in various alloys.
This was the late 1970's.
The idea in our minds at the time was to store hydrogen in a safe form for use in vehicles.
Our favourite alloy at the time was of titanium/iron. The lattice was distorted such that it "welcomed" hydrogen interstitially to lower its energy state.
Its welcome was so strong, that with a pressure of a few atm, it would hold more H than you could fit in the same volume of liquid H2.
It only required a little lowering of pressure and/or modest warming, and the fuel was released in a controlled way.
It was even demonstrated in the mid 80's in a BBC programme called Horizon, where efforts were made to explode/ignite this lump of concentrated hydrogen.
It could not explode or combust in any significant manner, because the hydrogen could not diffuse rapidly enough to the surface.

This is where I draw a parallel to your hydrogen embrittlement situation.

The hydrogen ions produced by electrolysis are attracted toward the same target as the copper ions.
The huge copper ions can only stick in the rough surface of the steel.
However, hydrogen ions (protons) are small enough to burrow in between any metal atoms, disrupting the ordered strength of the crystal lattice, inducing stress.
You are right to describe this as diffusion, it is like osmosis in a way, but it is the solute (hydrogen) which is migrating.
The rate at which hydrogen can diffuse through a metal like bearing steel could only be tiny, nanometres per month for arguments sake, especially as you have no potential drawing inwards.
In a nutshell, I think any hydrogen will be limited to the outer 2,3 or 4 atomic layers, and as such easily driven off by your heat treatment. I'd be doubtful that it would matter anyway. If it were near to a bearing surface, it may matter more?

As yet, my intterest in the 70's hasn't spawned any fruit in the way intended, though oddly the solubility of hydrogen in various alloys has become relevant in for example, NIMH cells.

Going off topic again,
Koan,
That is rivitingly interesting.
So what stops the H ions migrating through the lattice then forming H2 molecules which rip the lattice apart under external stress ?
As an undergraduate I worked with Prof Muir on his project determining hydrogen gas precipitation at the root tip of cracks in maraging steels, for which he was awarded an ACTA Metallurgia.
I was several years after the award when he was widening the investigations into the same mechanisms which lead on to the brand new field of fracture toughness.

With relatively low concentrations of H2 generation we could get cracks to rip through steels at phenomenal rates.

Everything you need to know about hydrogen-powered vehicles you can learn in 1 min. at:

https://www.youtube.com/watch?v=jH-mhZLuGRk

and about hydrogen embrittlement in electroplated films in 20 sec. at:

https://www.youtube.com/watch?v=u0nqcpHHImo

Although I was being facetious with the first video, the "hydrogen economy" hasn't lived up to the promise it seemed to have when proposed nearly 50 years ago.

And atomic fusion has not quite lived up the expectations either, apparently we are just 10 years or so from cracking it which is where we were when I was an undergraduate , perhaps they are Pluto years.

The hydrogen economy is here right now but like a lot of things that challenge the existing economic structure, those who suffer finanacially from it are doing their utmost to delay it.
Hydrogen welding is now done industrially and is rapidly knocking acetylene out as the preferred fuel. BOC ( CIG here) has done all it could to prevent this happening in Oz & I am sure the gas suppliers in the USA have done the same.
Hydrogen generators are readily available but I have not seen any offered by major gas companies, then again I have not been looking all that hard either.
When the government slapped an enviromental levly on acetylene down here doubling it's price I did have look but decided it was not worth the effort at this point in time .

Back in the 90's UNSW had pilot electryolys plants working but funding got cut, no doubt due to the influence of both the coal industry & supported by the royalty seeking individual State governments.
Several plants were set up at near commercial scale in various Pacific Islands as part of the foreign aid budget because they could not get research funding and AFAIK they are still running but the people I knew who were involved with this are long out of my social circle.

We seem to be drifting again, hope the life is teathered securely to the Ariel .

On the subject of hydrogen enbrittlement, I can not think of any work done on it under compressive loading, only tensile & shear, but mostly tensile .
While the mechanism for hydrogen migration througn a lattice under tensile strain should be the same as migrating under compressive strain I really could not see your bearing fracturing during use.
As it will be running at elevated temperatures I would imagine that hydrogen gas would continue to be exuded by the bearing during use.

Thin films and monocrystaline fibres are great for research purposes and in particular make the maths accessible ( buy don't ask me to do it now ) the actual bulk mechanical properties are somewhat different in real life applications.

I have a newspaper clipping from the mid-1970s in my office when, while visiting a research facility, the then-Secretary of the Department of Energy commented that it appeared producing energy from fusion research might be faster if we just burned the money. Although it is especially noticeable when it happens with major research projects, the nature of research is that the outcome is judged promising enough at the outset to pursue, but by no means can it be guaranteed.

With the bearing slow roasting for two more days I returned to the flywheel. The first order of business was to double-check that all drillings in the flywheels and crankpin were clean. Then back on the rollers to see what needed to be done to balance it to be 65% with the Omega piston as I had determined the original balance factor to be.

My notes show the weight of the complete Omega piston assembly (piston, rings, gudgeon pin, and clips) was 435.0 +/-0.5 g, but I reweighed it to be sure and got the identical result. I had machined and installed a new bush for the small end and, when the crankshaft was apart, was able to accurately weigh the small and big ends to be 267.0+/-0.5 g and 291.0+/-0.5 g, respectively, vs. the 267.5+/-1 g of the small end I had previously measured in situ with the old bush.

Up until now I had been checking the balance using a pair of bearings with the same OD but with 7/8" and 1" IDs to level the crankshaft on the wheels. However, the bearings are narrow enough that they can "walk" off the rollers after enough rotations so I took the time to machine two sleeves from 1-1/4" OD material, reamed 7/8" and 1". This was much more satisfactory.

After spending quite a bit of time I'm not convinced I can reliably determine the balance to better than +/- 2.5 grams hanging from the connecting rod, which is equivalent to 1.1 g at the edge of a flywheel. According to the specs I should be able to get to ~0.1 g but that wasn't possible (due, it appears, to residual friction of the big end rollers). Anyway, 2.5 g more than 193 g definitely was a bit too much, and 2.5 g less definitely was a bit too little. As will be seen from the final calculation, being able to balance it better than this wouldn't make a difference in my conclusion.

Plugging the measurements into a reduced form of an earlier post results in the following if I use the crankshaft as-is:

Balance Factor = balance weight + small end weight / piston weight + small end weight

Balance weight:

-- weight to balance crankshaft in its present form = 193.0 +/-2.5 g

Small end weight:

-- Current small end weight = 267.0 +/-0.5 g

Piston weight:

The "piston weight" is that of the complete assembly of piston, gudgeon pin, circlips and rings.

-- weight of aftermarket +60 Omega piston assembly = 435.0 +/-0.5 g

Balance factor with Omega piston:

193.0 + 267.0 / 435.0 + 267.0 = 460.0 / 702.0 = 65.5 +/-0.4%

Thanks to a happy coincidence, within experimental uncertainty this is identical to the 65% I determined the original factory balance factor to be (actually, I found I found 64.8+/-1%). So, I'm going to leave the crankshaft as-is and bore the cylinder for the Omega piston.

If instead I used the Gandini piston as-is the balance factor would be 58.7%.

Thanks for the videos! I've always been amazed that some survivors ran away from that, I can only imagine that the updraft brought cool air in just long enough.
I actually found the 2nd vid fascinating, I know I must get a life!
It is all about stress fracture, and I have no problem with this, and WM20's info. Under stress, any flaw has a tendency to open, increasing access for hydrogen. This is not what is going on during electroplating a quality ground surface.
Plating using one of the "semi-noble" metals, employing an appropriately low current, will not result in significant H+ production in the 1st place.
The only dangerous possible penetrations would be where crystalline imperfections already exist at the surface, but as it isn't a stressful situation?
H can and will creep into a metal lattice through the surface, this is a very different phenomenon to penetration through fractures, and where the penetration is at the already most vulnerable point. There is also little opportunity for H2 (vastly larger than H) to form within the lattice, where space is available there is already structural vacancy.
WM20, Why would you think that adsorption would be unchanged with compression/tension? When H is adsorbed into a metal, the volume increases slightly, of course. The ease with which this can happen is affected by how you squeeze or stretch the sponge.
Your talk of H2 formation applies as you said, at the fracture point in a tensile situation.

If the electroplating process were 100% efficient all I would have to do would be to count the number of electrons that entered the solution (current x time) and I would know the exact number of Cu atoms deposited. However, it isn't 100% efficient as evidenced by the visible bubbles of hydrogen gas when I plated the bearing. This happens because some of the electrons are hijacked by the H+ ions to allow the formation of H2 gas (the chemistry is a little more complicated but this captures the essence). "Low current" wouldn't help since, say, one-tenth the current would just take ten times as long. The total amount of hydrogen produced would be the same.

Don't you love it when things work out!
You must be living right.
Good job.


Kevin

.

If there were no metal ions in the electrolyte (eg if it were just water or sulphuric acid), then the flow of +ve ions to the cathode could only be H+.
For electroplating, the electrolyte contains the desired metal ions, which will be replenished by the anode as they are deposited.
In this case, both Cu++ and H+ ions are available.
Cu++ are more readily made at low potential. However, their migration is much slower than H+.
So if the current exceeds what Cu++ migration can support, the excess will be performed by H+.
The optimum recommended current will be a compromise between reasonable plating time and tolerable hydrogen production.
There is no doubt that increasing current will result in greater hydrogen produced in proportion to copper deposited.

I can't remember... how was the cathode attached to the bearing race ?
And did it interfere, in any way, with the depositing of the copper ?
.

I owe it all to clean living... no, wait, obviously there must be another explanation...

I'd hate to think how many hours it took for me to simply determine the goal was 65%. Measurements, calculations, tracking down original Hepolite pistons, etc. If only the 65% figure had been documented some place it could have saved all that time. That's why I'm documenting what I'm doing so the one other Black Ariel owner in the world who might come along in the future won't be starting from scratch. The same thing can be said for your thread on the AMCA forum. If I owned a pre-'20s Indian I would be very grateful for the detail you've been providing.

This is straying a bit off topic. I'll just leave it for now by saying that it is known that all plated coatings on all steels are susceptible to hydrogen embrittlement to some degree, with the problem more severe for higher strength steels, requiring a suitable bakeout as quickly as possible after the plating is finished. I'll return to electroplating when I deal with depositing hard chrome on the worn spindles of my camshaft.

I don't have the brass wire I used handy at the moment but its diameter was about that of typical safety wire (~0.030"). I wrapped it once around the circumference of the outer race and twisted it tight with safety wire pliers. Although I used brass wire, stainless would have been fine. The wire masked the area under it but that only reduced the coverage by ~0.03" out of the 0.75" width of the race (~4%). I could have devised something more complicated to avoid that strip but, since I was doing this to provide an interference fit, not for a cosmetic coating, the appearance doesn't matter nor should the loss of 4% of the gripping force.

I checked again this morning and it's still cooking away with another ~18 hours to go.

Thanks,

What was process for choosing copper as the plating material , as opposed to say... nickel ?
Ease of plating,
Toxicity of electrolyte , cost ?
.

Less "research" went into my choice of Cu than into some of the other things I do. Any metal I could deposit to a thickness of 0.001" would have been fine for this purpose. For that matter, if it would have been possible to wrap the bearing with 0.001" steel or brass shim stock and press it into place without destroying the shim stock, that approach would have been fine as well.

That said, I've done plating in the past so I know Cu is pretty well behaved. Also, one of the reasons it's used as the first layer in "triple chrome" plating is because it adheres well to steel. For my purposes Ni probably would have deposited just as well onto the bearing, but I "knew" Cu would deposit in a uniform coating, vs. being "pretty sure" Ni would.

Another factor in my choice of Cu is it can be plated on pot metal. In the back of my mind is the idea to sometime experiment with using it to build up worn carburetor components. I don't have anything in particular in mind right now, but having the Cu solution in hand means I could try it without any (additional) expense. If it worked to reclaim whatever pot metal component I tried it on, great, and if it didn't, it wouldn't have cost anything.

Nickel plating is notorious for hydrogen embritlement and most of the research in the early days was done on nickel plate.

At the interface of the bearing & the plating solution if hydrogen is generated then it will be drawn into the metal by the loose electrons which is the nature of the metallic bond.
Once in the structure it can & will travel quite freely as it is a very small atom ( well proton actually ) happily sharing electrons so will migrate along paths of slightly higher negative valence or get pushed around by the positative valence of the steel's atoms.

However like Koan said, if the plating current is slow there should be next to no H disolving ino the steel.

We did not look at H being adsorbed from the atmosphere.
We were looking at H that was already in the lattice ( a lot more than you would think ) migrating to the crack tip.
What actually forces the steel atoms to separate and thus the crack to grow, is not the applied stress shattering bonds but the H2 produced internally shattering the bonds.
This is why the research was so important .
We were only looking at the H2 evolution inside a metallic grain and not grain boundry H2 generation, which is the prime process in hydrogen embrittlement.
You get substantial amount of H2 at the grain boundries thus creating a lot of stress so substantially less load than would normally be required to initate failure cause a grain foundry failure.
And while not being 100% accurate , brittle failures in steels are grain boundry failures.
The test wires were made from vacuum remelted steel and the volume of H that gets liberated from the melt pool on the first remelt was really astounding. ( well it surprised me anyway ).

Hydrogen
I disagree. While reducing the current reduces the rate of hydrogen production, it also reduces the deposition rate of the Cu so the part has to stay in the solution exposed to hydrogen ions proportionally longer. The time a part can be exposed before the ingress of hydrogen is irreversible despite subsequent heating is in single digits of hours so longer has potentially severe consequences.

For me to go further down this hydrogenated side road by supplying references to the technical literature would require time I need to use for making progress on this rebuild so I'm stepping away from electroplating for now. If there are further H posts please don't take my lack of response to them as tacit agreement.

Crankshaft
Because there seemed to be a bit of stiction in the rod when I was measuring the balance factor I installed the crank in the lathe, put it in back gear, and spun it at ~60 rpm for two minutes while holding the rod away from the ways. I used plenty of oil on the crankpin, and felt no resistance at all. I then took it out of back gear to let the chuck spin free and spun the crank by hand by pumping the rod. Again, no resistance or rough spots.

With the crank now reacquainted with its task of spinning I put it back on the rollers where I found I now had more sensitivity. The balance weight is now at the low end of the range in my earlier post, i.e. the balance factor even closer to 65.0%. However, hard as it was to fight OCD, I forced myself to accept good enough as good enough and not to spend any more time determining the weight even more accurately. There are many other tasks that are more important to spend my time on. So, I'm certifying the crankshaft as fit to return to service with an Omega piston.

Bearing
This morning the drive-side bearing had racked up enough hours to rid itself of H so I turned off the oven, opened the door, and let it cool in air as it had during the tempering cycle when manufactured. The Cu coating had discolored but passed the rub-hard-with-thumb adhesion test. Also it's still 0.002" thick.

Since the bearing had been pretty hot for a long time I wanted to be sure the hardness had not been affected so I measured it and a companion RHP bearing I purchased at about the same time, although the latter with 'Standard" clearance. Also, the non-plated bearing came in a box labeled "RHP, member of the NSK Group" so it was manufactured more recently than the RHP I plated, and possibly not in the same facility.

All the information on bearing hardness seems to fall between 58 and 64 Rockwell C. First I checked my tester against a 49+/-1 Rockwell calibration block, and then tested both bearings at several locations on the outside edges of the inner and outer races. The newer RHP/NSK bearing consistently tested at 65 Rockwell, while the plated one came in a bit lower at 61-62. While these differences are (just) within the claimed +/-1.5 Rockwell absolute accuracy of my tester, the side-by-side comparison leaves no doubt the newer bearing is a bit harder. However, 61-62 is within what bearings should have and, since I didn't measure it before the heating cycle, the hardness certainly could be identical to the value it had at the start.

The next order of business will be to install the bearing in the case. Taking a low value for the thermal expansion of Al along with the value given for bearing steel, the differential thermal expansion between the bearing straight from the freezer and the case pulled from a pot of boiling water should be ~0.0025". So, the bearing should drop into place with very little urging and the friction not be an issue for the relatively soft Cu. Again, I decided on 0.002" as the "press fit" because BSA Service Sheet says this bearing should drop out of the case if it is heated in boiling water.

p.s. the brass wire I wrapped around the bearing to make electrical contact is 0.039" in diameter, although the track it left behind is only 0.015" wide, so only 2% of the surface was left unplated.

Great stuff MM, now that you have finished baking your cookie you are now starting the reassembly process so it seems to me that you have passed a key milestone. We are in mid (ish) January so I think you are on reasonable course to have the bike ready in plenty of time to test ride it enough and iron out all of the bugs.

Keep up the good work.

John

Knock wood that your good wishes haven't jinxed it...

My most recent around-the-world trip along with its subsequent jet lag and the holidays kept me from making much progress during December, but things are nicely moving forward again. So far... (see above paragraph).

I have a large galvanized tub (20" dia. x 10" deep) that I'm forbidden to use ever again on the barbeque to degrease stubborn engine parts. It's amazing how acrid the smell and how long it lingers... But, this time it was just to heat a clean case in water so that's what I used. I tied a string to the case to use to haul it out of the boiling water.

I had placed the bearing in a baggie in the freezer at ~0 oF an hour or so earlier and the IR thermometer said it already was down to 18 oF when I fired up the barbeque. To have everything ready I laid out a quadruple thickness of towel so the case would be insulated from the counter top and cool more slowly, a socket with 2.48" OD, a Cu hammer for additional encouragement if required, and a pair of thick gloves.

The better part of an hour later the bearing was at 8 oF when the water started boiling. I wrapped the bearing in a towel to carry outside, quickly had the case in place on its towel, and just as quickly had the bearing installed. However, it was good I had the hammer and socket ready because I needed them.

The case is now cool and the bearing spins freely but with no sign of clearance, which confirms the fact that, like with a Gold Star, a C3 bearing coupled with 0.002" press fit is the way to go. I wonder how tight the original CN bearing I bought would have been had I used it instead.

I had carefully cleaned the bearing of all oil nearly five days ago so I generously lubricated it with lathe spindle oil from a pump oil can I had handy. Examined closely under a microscope there is no sign of Cu anywhere around the interface between the bearing and housing, which would have been there had any peeled or scraped off the surface during the installation. I'm very happy with how this turned out. Fingers crossed that this attention to detail pays off in September.

There's still plenty to do but, as John wrote in the previous post, it's now almost ready to (slowly) start going back together. As soon as I make and line bore three bushes on the timing side the bottom end will be ready to assemble. I already have the jig needed to bolt the cases to the mill since I made it to clean up the distorted hole for the drive-side bearing a month or two ago (which is why I needed to plate the bearing oversize).

The clearance in the present timing side crankshaft bush varies between 0.0045" and 0.0057" so it definitely needs to be replaced. However, I'm a bit puzzled by my measurements at the moment.

Normally non-custom tooling is used to ream a bush to a standard dimension and the shaft made sufficiently undersize to allow clearance for the oil. The only comparable figure for a crankshaft I've found so far is Nicholson who gives 1.3735" as the standard diameter for the timing side of a BSA A10 crankshaft, which would mean a clearance of 0.0015" when new if the bush were reamed to a standard 1-3/8". Consistent with this, separately he gives a minimum of 0.0015" clearance when a new bush is installed. Also consistent with this, a pre-unit BSA gearbox layshaft is 0.002" smaller than the bush, which itself is reamed to a standard diameter. Further, Nicholson notes that if the clearance has increased to 0.003" the bush needs to be replaced.

As for the Ariel, the shaft projects ~1-3/4" from the flywheel. Other than the outermost ~1/2" where the wear is more (it tapers to 0.8738" at the very end), over most of the length the diameter varies from 0.8741" at the smallest to 0.8743" at the largest. This would result in clearances of only 0.0007"-0.0009" if the bush were reamed to a standard 0.8750". I could machine a new bush with any ID I wanted, but I'm puzzled why it doesn't appear to have been a standard size when new. Or, if it was a standard 7/8", why Ariel made the clearance so small. I'll give some more thought to this before I make the replacement bush.

Once again it sucks not living in North Korea, because if I did I would be ecstatic if I could get my hands on any kind of bronze in the next six months, rather than being forced to select between scores of different varieties (most of which could be delivered by Friday, if not tomorrow).

I still have enough of the PB1 tin bronze (ASTM 90700 or SAE 65 nearest U.S. equivalent) that chaterlea25 sent me a few months ago, that I used for the small end of the connecting rod, and it has a large enough diameter for the 1-3/8" "hat" on the end of the 1-1/8" OD bush, but I'm leaning toward using "high strength" 544 phosphor bronze for this crankshaft bush instead. If the bush in the rod doesn't distort with the PB1's 25 ksi yield strength, the one for the crankshaft should be fine with the 544's 50 ksi. However, the jobs the two bushes do are different.

Both bushes are subjected to the hammering of combustion (although the crankshaft bush shares it with the drive-side bearing), but the small end sees the gudgeon pin oscillate back and forth by only ~+/-20o so the surface speed of the pin over the bush is relatively small (pi x 13/16" x 40/360 x 5000 rpm = 120 ft./min.) whereas the crankshaft bush sees the spindle rotate at fairly high surface speeds (pi x 7/8" x 5000 rpm = 1100 ft./min.). Further, the small end bush has a clearance of only a few ten thousandths while that of the crankshaft bush is ten times larger because it relies on a hydrodynamic cushion of oil. Again, not living in North Korea forces me to read specifications to try to decide on the best bronze for the application.

Once I decide on the material for the bush and make it I'll need to shim the crankshaft for the correct end float when the crankcases are bolted together. Ariel uses shims for this purpose and a few of them were in place when I disassembled the engine. Draganfly supplies a set of these shims in a variety of thicknesses for only ~$10 including postage so, although I don't know if I will need them, I ordered a set.

Many of the TS bushes available in the UK for the BSA A65 are made from solid lead bronze (SAE 660). Apparently this Bronze has good hardness and strength qualities as well as good wear resistance, it also has excellent machining properties and anti-friction qualities. Additionally I note that SRM sell TS bushes for the A65 made from phosphor bronze (PB1) which is suitable for bearings having medium to high loads and speeds and good resistance to impact loading or pounding.

I have no idea whether either of these bronzes are suitable for your application but to help you decide, there's a long article on selecting bronze bearing materials, see This Link

As you are using unfiltered oil, noting that it is total loss, you need to include embedability in your list of properties when selecting your bush material. Use a material with low embedability and any hard debris build up on the surface will convert your bush into a grinding wheel. Lead bronzes can suffer from this hence Vandervell's use of soft lead/indium overlay plating and Glaciers soft lead/tin overlay, these acted as sinks for debris as they are so soft its kept below the surface but kept the strength of the leaded bronze subtrate plus were more resistant to acids from combustion. That's not an option here but watch out for the temptation of using the highest fatigue strength.