Illustrated tour of the pre-war Douglas o.h.v. crankshaft.
For those that have been playing about with the old Dougies for a while, but have not yet had personal experience with the pre-war o.h.v. machines, the following visual tour of the crankshaft might be of interest. In time you are bound to hear stories about the unique assembly procedure for the o.h.v. pre-war engines; or how once worn out or damaged they can not be rebuilt. And it is more or less true!
Recently I was finishing up the assembly of a client’s crankshaft, and took the opportunity to take some photos illustrating the sequence. Then some other pictures from my collection, illustrating other problems that may be encountered should you be blessed with such a creature. The crankshaft shown is for a Dirt Track model, essentially the last and strongest of the 82mm stroke crankshafts Douglas made. The connecting rods shown are new manufacture
(See advert), and not original Douglas components.
The pre-war o.h.v. Douglas crankshaft is a two-throw crank, made of a single forging as seen above. This presented a challenge to assembling the roller bearing big end bearings, which Douglas overcame in an ingenious way, subject of several patents.
The first image shows the flywheel side connecting rod and counterweight installed, and the crankshaft is ready for the timing side connecting rod.
As the big end eye is not split, to allow an uninterrupted race for the bearing rollers to run on, the rod needs to be threaded onto the crank over the throws. As you can see, the size of the throw and the diameter of the big end eye are optimized to their respective limits, it just passes!
And so then around the second bend.
The rod is now in its final position, but the big end bearing is not installed.
Here we see the connecting rod roughly centered on the crankpin, ready to receive the big end bearing.
The big end bearing is made up of two aluminum cages, slotted to carry ten Ø1/4x1/4” rollers in each half. This is slid sideways past the haunch of the throw in to position.
Once in, it is pushed around to the inside of the throw to make way for the second half of the cage carrying the ten remaining rollers.
This in turn is introduced. And here a pause while it is explained one of the limitations of this design. Obviously in a new or refurbished big end bearing you want no perceptible radial play in the bearing. In reality this equates to a 0.001” running clearance. This just allows you to slide the cage and its rollers sideways into place. In time, wear occurs. A worn connecting rod can be re-honed, but as there is no separate race one can only go a little bit before the surface hardness is compromised. Worn rollers can be replaced with new oversized rollers, assuming you find them these days. That leaves a worn crankpin, which can be re-ground. BUT. You can only reduce the diameter but a mere 0.003” or so before you run into trouble on assembly.
That is because the haunch of the throw, shown as “A”, and the crankpin “B” are pretty near flush. If you were to grind the crankpin to a smaller radius, the rollers would have to ‘step down’ onto this surface as the cage was slid sideways into the connecting rod eye. The first half of the cage is not a problem, but the second half only has a thou clearance, and you can not get the rollers to enter. If it were close, you might drive the rollers in, but besides not doing the bearing any good; you will never again get it apart. You may be able to force the rollers down a step, but as the next image shows, once in place there is practically nothing to grab onto to pry the rollers back up over that step!
And here it is with the second cage in place. For removal you only have the low hub on the side of the cage to grab onto with your fingernails, to wiggle and coax the cage out. For added fun try it when everything is oily. Least you think the solution is to reduce the radius of the haunch of the throw, I might point out that you would also have to reduce the width of the throw as well. In turn, as the counterweight is a snug fit on this, new counterweights with undersize windows would need to be made. Then there would be the issue of installing those, as the width of the throw is not only the same as the crankpin, but it is also identical to the main shafts, which the counterweights have to slip over... Best not to go there.
As mentioned the counterweight fits snugly on the throw. There is a registration groove indicated by “A” above that locates it. This is a different crankshaft to that shown in the other illustrations, as I had already assembled that before thinking to take this and the next four images. Also shown at “B” are typical roller ‘tracking’ and chatter marks that makes one wish you could re-grind the crankpins more than just a light ‘lick’.
The DT crankshafts are drilled for oil, but part way through the evolution of the model the factory added the gash at “C” on the leading edge of the center web to introduce a little extra oil alongside the connecting rod.
Inside the counterweight we see the corresponding lip that fits the groove in the haunch of the throw.
At the main shaft end of the throw is a V groove. If it looks like it is not level, that is because it is cut on a 2º inclination.
This is to receive a tapered cotter driven through a cross drilling in the counterweight, that wedges the counterweight onto the throw.
This is a view of the inside aspect without the connecting rod, showing that the counterweight hangs over the inside face of the throw. The center web is at the top. The counter bore in the face of the counterweight receives the hub that protrudes from the side of the big end cage. The floor of this counter bore near enough lines up with the inside face of the throw. However since the cage is a 180º segment, it can span the opening in the counterweight. This stops the cage from migrating sideways.
This counterweight has had its lip built up with hard-face weld and machined back to reduce the end clearance on the connecting rod. The counterweights are just mild steel and were not originally hardened.
Here the counterweight is being introduced onto the throw. They need to be a snug fit, else they will shuffle about and work the cotter loose no mater how tight it is driven in.
And with the counterweight fully home. The flare on the counterweight lines up with the eye of the connecting rod.
And here we see another view of it. Note that the slot in the counterweight is slightly longer than the throw. This allows the lip to slide over the haunch till it drops into the groove. This leave a gap (A) at the main shaft end, but it is of no consequence. Also the throw is recessed into the counterweight. It is eased back at the crankpin end; if it were not the connecting rod would be unable to be threaded around the bend onto the crankpin.
As the counterweight itself can not be shifted, having one fixed position, the axial clearance of the rod is set by adding or removing material from the cheek of the counterweight. This means assembling and disassembling the crank several times. But no one said owning a pre-war o.h.v. Dougie was all beer and skittles! Actually a few beers can help one's temper while working on one!
The counterweights are stamped on the inside face “T” and “F” for timing and flywheel sides respectively. They also have a number, which originally would match the number stamped on the crankshaft center web periphery. Crankshafts were batch assembled, fitted, and balanced in the Douglas experimental department. In fact the whole engine was built there before being sent out to the main factory for assembly into a chassis. But by now it is rare to find matching numbers as counterweights have been hopelessly mixed up and interchanged over the years.
The completed crankshaft assembly, rods tucked in. Here you can see why these engines required short skirt pistons.
And in its full 82mm long stroke glory! This same assembly was used in the 500 & 600cc, and made to order 750cc size engines with the simple expedient of changing the bore size.
Other faults to look for. Any crankshaft that has been used in competition is probably sprung to a slight extent. That caused by excessive rpm is in the plane of the crank throws and relatively easy to setup in the press to straighten. Setting up is easy, by that is not to say straightening the crank is easy! Even more difficult is correcting twist; from perhaps a catastrophic seizure. These can be diabolical to rectify. The crankshafts are incredibly springy, and you have to press them some 0.040” before they take a set. It is a long process of trying to sneak up on ‘dead-true’, as you do not want to over correct and end up with it bent in the opposite direction. All this bending reduces the fatigue life of the crankshaft.
Here can be seen a crankpin that has been friction sawn nearly through alongside the center web, and the remaining bit broken off. Note the fibrous nature of the original core material, which follows the serpintine shape of the crank. They are incredibly tough buggers, or they would not have lasted as long as they did. But they are not impervious.
Any crankshaft being refurbished should be as a matter of course Magnafluxed to inspect for cracks. I would be very much surprised if you did not find any! Not to worry, they all are like that! Most will be heat cracks from when the crankshaft was originally ground, either from pushing the removal rate or more likely using a dull grinding wheel.
Very typical are these radial cracks on the face of the center web, indicated by orange Magnaflux residue pointed out by “A”. Using the side of the grinding wheel to ‘face’ the web as the factory did is prone to burning the metal and creating these cracks. The cracks are typically very shallow, just a few thousands of an inch deep, and are nothing to worry about (not that you can do anything about it.) After all if it were going to break it probably would have done so by now. The exception would be if they go all the way in and touch the radius around the crankpin, which is a highly stressed area.
“B” points out a groove worn by the side of the connecting rod big end eye. If the big end bearing cage gets twisted or ‘wracked’ it will tend to work the rod hard over to one side or the other. Perversely it always seems to be into the center web rather than the replaceable counterweight! However these grooves can often be repaired.
Another place micro cracks might be reveled are on the flywheel taper. If they are axial, they are further examples of heat cracks during grinding. Any circumferal or spiral cracks are cause for alarm.
There is one other readily visible fault with these crankshafts that typically renders it unusable, and that is a bad flywheel taper. These always were a bit marginal in surface area, and as the horse power increased, the situation only got worse. If you have a good taper, make sure it fits the flywheel perfectly and keep the nut dead tight. And check it several times during the course of the season to make sure it stays tight!
If you do not, they tend to fret the taper, as seen in this example. Also they are prone to chipping out along the edge of the keyseat, as the taper is case hardened too. This example exhibits fretting, metal transfer, galling (when the key sheared), and chipping. Actually the chipping is minor, I have seen far worse.
It is not uncommon, if the flywheel becomes slightly loose for it to hammer back and forth on the key, and a spiral crack to develop from the top corner of the Woodruff key seat. Magnaflux was not required on this example! Woodruff keys are not particularly brilliant if trying to avoid stress concentrations. At this point the crankshaft is in imminent danger of breaking, allowing the flywheel to go belting off in a kinetic catastrophe. And it is not an old legend, it really does happen.
If the fretting is not severe the taper can be re-lapped or a light pass taken with a grinder. As it is a shallow taper it does not take much before the flywheel moves sideway sufficient to foul the crankcase. Then remedial action needs to be taken with the flywheel hub, like a re-sleeve. The taper need not be completely cleaned up. It is more important that it be a lap fit on the flywheel. You should be able to dry wring the flywheel onto the taper such that it sticks tight on its own accord, and requires a tap with a mallet to jar it loose. The first taper shown above was successfully cleaned up to about 90%.
Welding up the tapers has been tried, but the crankshaft is a heat treated 5% nickel alloy that is unforgiving. Eventually the shaft breaks anyway at the end of the weld zone. Also the taper surface is more susceptible to fretting in a soft condition, and is difficult to keep tight. Hard chromium plating has been tried, but besides the low coefficient of friction, a taper so repaired broke right where the plating stopped after two seasons use.
The spiral crack in the example above was precipitated by the taper being built up with weld (using nickel rod) and then the hammering of the poorly fitting taper.
Here we have a crankpin where the surface is starting to break up; the term used is brinell. As the case hardness depth is only about 0.040”, the pressure of the rollers traversing the surface ‘kneads’ and massages the hard skin over the relatively softer core. (Much like what happens to asphalt at the traffic light where large trucks brake heavily.) In time this breaks up the interface between the two and flakes of the case break away leaving pits. These particles are naturally very hard, and on their way to the sump get run over by the rollers and crushed, adding more dents and dings to the rollers, connecting rod, and crankpin. The crankpin shown is probably just a little too far gone to reclaim by a light grind. Unfortunately this is all too common, as the severe quench process Douglas used resulted in a very hard case and tough core, but a rather shallow case depth with abrupt transition.
Connecting rods are not immune from brinelling either, especially where the load is the heaviest. This rod is scrap.
The other problem afflicting crankpins is forgetting to oil them. Either on the earlier engines forgetting to give the hand pump a stroke every now and then, or on the mechanically lubricated version forgetting to put oil in the tank or open the drip valve on the sight glass. Note the face of the adjacent center web is also torn up.
There is no current viable repair for this kind of crankpin damage. Perhaps some day a brave soul will attempt to weld one of these crankshafts up. But like the crank taper, it will be quite an involved process to have it done properly. Spray welding will not work (hard thin skin is bad, remember?) The crankpins need to be machined down a 1/8” to remove the carbon enriched surface, then built up with a nickel chromium alloy similar to the composition of the crankshaft (akin to SAE 9310 or En 39c.) After rough machining this needs to be case hardened again. That means heat treating the whole crank, and distortion is bound to occur, and a straightening session will be required. Then finish grinding. As the main shaft journals accuracy and flywheel taper will likely be compromised in all this heating and quenching, they too will need cutting down and welding up so that they can be reground at the same time. All this is technically possible to do, albeit very expensive. The unknown is how the core grain and fiber structure is going to interact with the weld zone, especially at the highly stressed transition between throws, crankpins, and center web. It will probably be all right, but who wants to be the first to try it in their engine?
Douglas Kephart, Glen Mills, PA, USA
© Jan 2005
