Tuesday, October 24, 2017

Benly BIG-time Blow-up

The whole CB92-CA95 (US models) line up was a lightning bolt into the hearts of other manufacturers during the late 1950s and early 1960s. These tiny electric-start 124-154cc twins would spin up to 10,000 rpms, when pressed or just doddle along at traffic speeds without complaint, in most cases.

The engines were 360 degree firing parallel twins with a short 41mm stroke, two-valve cylinder head design operated by an OHC camshaft. The camchain ran off of the left end of the crankshaft sprocket, over a guide roller and through an adjustable camchain tensioner. The 4 speed transmission was a normal return change design, operated by a 5-plate light-pull clutch assembly. The back of the clutch basket was machined with an eccentric that connected a flat steel arm which worked a push-pull plunger inside a cast steel oil pump body. The whole design was elegantly simple, but built with precision and quality materials.

A weakness in the design of the crankshaft became evident early-on which required several different crankshaft changes. The main differences started with the 1959-90 crankshafts which had a larger center main bearing than the two on each end. Initially, all three bearings were located in the engine cases with the “half a piston ring” retainer.

In 1961, the center main bearing was reduced in size to match the other two and the bearing was located with a dowel pin to keep the oil feed holes in alignment. The end bearings were still retained with half rings, which allowed them to rotate in the cases at very high rpms.

 The engine cases were also slotted to help facilitate oil moving over to the crankshaft in one design. On early CB92s the engine case slot (at the rear of the cylinder base) lead to a low cut window in the back of the left cylinder liner to help lubricate the piston skirt, which apparently was not getting enough oil. This lack of lubrication caused piston seizures and subsequent crankshaft failures when the connecting rods were brought to a halt suddenly.

The 1962 crankshafts used pinned bearings all the way across. The rotor side main bearing of the crankshaft was a caged-roller style with a thrust washer, while the transmission drive gear end used a ball bearing. Some versions had a larger diameter end bearing, as well.

The current repair bike in the shop was a low-miles, mostly original, 1961 CA95 Benly with engine number #59. The owner had owned it for a few years and put less than 1,000 miles on it but the engine began to smoke a lot from one side, so the top end was freshened up with .50 pistons/rings and a valve job. Right away, the owner reported that the bike lacked the snap of his other Benly 150 and wouldn’t go faster than 45 mph in top gear. In the middle of a ride, the engine slowed down, seized up and began to make some unpleasant knocking/rattling noises when restarted.

Having owned a handful of CB92s, plus a 150 Benly or two, I was fairly comfortable doing a forensic exam report on the engine after it was transferred to my custody at a VJMC event held in Solvang, CA in Sept 2017.  Once it was on the bench the engine was removed and the teardown/inspection process proceeded. The drained oil had a lot of bright shiny particles in it, which is never a good sign. The ignition timing was checked to eliminate that as a source of the piston seizures, due to detonation from too much spark advance. The top cylinder head cover was removed to view the valve train which all seemed to be in good condition, but still flecked with brass-colored particles here and there.
The cylinder head was lifted off and set aside for further inspection. The cylinder bores seemed to be relatively unharmed, apart from some superficial wear/scratching. Once the cylinder block was lifted off, the serious damage began to be revealed. There were a couple of missing pieces from the bottom of the left piston skirt and both pistons showed signs of repeated piston seizures in the past. When the left side connecting rod was checked, there was noticeable up/down play which shouldn’t be there at all.  A few pieces of the piston skirt were lying just below the pistons on the crankcase bosses and more of that shiny brass particle debris became evident.

The camchain, stator and left side engine covers were all removed to facilitate engine case separation. When the clutch cover was removed on the right side, a stream of contaminated oil gushed out of the hidden nooks and crannies all laden with large amounts of brass filings and particles. There seemed to be more brass residue than was possible from just one rod bearing failing, but that was about all that could be seen as a source. The entire engine was contaminated with metal particles, including the oil filter. The clutch was removed, along with the oil pump which is driven off the back of the clutch basket eccentric.

The engine cases split without too much difficulty and more metal debris poured out onto the work bench. The transmission gears seemed to be in good condition; however one of the locating dowels for the transmission bushings was half bent to one side, due to a shearing force on the transmission shaft. There were more questions than answers, but one thing was for sure; the crankshaft and pistons would have to be replaced!

Continuing the repairs…

A lot of cleaning has taken place to expunge all of the little brass particles that were circulating around inside the engine before it finally expired. The oil filter, which spins on the end of the crankshaft, was packed with colored bits of course. Top and bottom cases were cleaned thoroughly as were all the clutch parts and transmission gearsets. Rather than replacing the hollow dowel pins on the transmission shaft bushings, a couple of 6mm crankshaft roller bearings were ground down to a matching length and secured into the top case half, which was lying on its back awaiting newly freshened parts.

The crankshaft replacement really only had used parts options, as new crankshafts are NLA in much of the world. A $135 (delivered) used crankshaft from an eBay seller came with the rotor and left cover as the seller was unable to pop the rotor off the end of the crankshaft during teardown.  After measuring all the crankshaft main bearings, it was discovered that Honda had increased the bearing size on one end. The old crankshaft mains were in good condition so the end main bearings were switched from the old crankshaft to the new(er) 1964 part. That swap was successful and the updated crankshaft was placed into the cases. Although the engine had been freshened up within the past couple of years, new seals were ordered along with a fresh gasket kit. The oil pump was cleaned and inspected for damage, then the small ball spring replaced, along with a new o-ring for the pump’s side plate.

Fortunately, the cylinder bores (bored to .25 oversize) cleaned up to like-new condition, so fresh pistons and rings will have a nice new home to work within. A correct slotted base gasket was discovered for sale on eBay, so the oil channel will have a maximum flow rate to the center main bearing. The original base gasket was the later model flat version, normally used with engines which do not have the oil channel, so there might have been some flow restriction at that point.
The cylinder head had little debris, apart from the inside of the camshaft oil feed hole. The cam lobes and rocker arms were unharmed. Spray cleaner was used to check for any valve leakage, but none was found, so the valves, which had been replaced earlier, were left undisturbed.

Still looking for the “smoking gun” that caused this disaster, the carburetor was inspected for jet selection. The main jet had been increased one jet size; however this carburetor has “power jet” enrichening circuits which are controlled by two tiny air/fuel correction jets. With information gained from the Facebook Vintage Japanese Bike Club subgroup, which is focused upon the CB92 (Super sport version of the CA95 Benly), a member from Sweden offered up a specification page from an ancient manual showing that the jets for the power jet system were #130 in both locations. When the fuel jet was removed from the carb body, its size was not marked at all. The air corrector jet at the top rear of the carburetor body was stamped #130. Using a metric drill bit set, I determined that the inner jet was a #150 size.

Honda seems to have determined that these jets were not to be altered, so no references to the part number or sizes can be found on current online microfiches or even in my 1966 CB92 parts manual. Digging deeply into arcane parts references books, a part number was discovered that might have been correct for these jets. The ones listed in the book were not #130 size, according to the part number suffix, but substituting the end numbers with 130 got me closer to a source for them. While an eBay search for “air jets” revealed similar jets which fit CR250 motocrossers and GL1000 Gold wings, they are all ISO thread pitch jets. A 1961 CA95 Benly is all JIS thread pitch built.
Doing a web search for the desired part number revealed that it had been superseded and used in G30 Honda generators and early Honda 360-600cc twin cylinder cars. 

Fortunately, I live nearby a Honda 600 car guru, who has tons of used parts. I ran the part numbers and descriptions past him and he came up with a #120 air jet from an AZ600 carburetor! The jets matched up dimensionally and on the thread pitch, so I gratefully brought it home and reamed it out to #130 size using tapered jet reamers. Right or wrong, at least the carburetor was now close to specs for that power jet system.

The previous mechanic had contacted me, asking about my findings, as well as giving an overall view of what he had done and found previously in this engine. He mentioned that the carburetor needed some massaging to get the slide to stop sticking and I noticed that the carb flange was somewhat warped, even now. Obviously any air leaks at the intake flange are potential causes for piston seizures due to lean mixtures causing piston overheating.

Honda’s specifications for spark advance are puzzling as they show a range from 34 to 47 degrees before top dead center. In comparison, a CA160 has a 40 degree maximum spark advance range. I chose to squeeze the stopper ends of the advancer closer together in a vise, so that the maximum spark advance will be somewhat reduced. Over-advancing spark causes piston overheating, usually resulting in piston seizures, particularly when the piston clearance is only about one and a half thousandths of an inch.

Newly installed pistons and rings on a fresh rebore need some “break-in” time, which I suspect didn’t happen previously. Honda usually specs about 500-600 miles of easy riding before extended full-throttle is used. The engine’s last repair work didn’t include case splitting, only a bath in a solvent tank to remove thick layers of sludge from the bottom of the cases. When the cases were split, the sealant wasn’t familiar looking to me and someone had painted the number 78 inside the case halves! One can only guess as to whether the 78 marked the year when it was taken apart or some mechanic’s ID number or something totally different!

Having looked at each and every part of this engine assembly, I have confidence that it can become fully functional as designed, given a proper break-in period and judicious checking of the oil levels, ignition timing, carb jetting and use of the correct heat range spark plugs. The old spark plugs were not present in the engine, as a pair of C6HA plugs were loosely screwed into the cylinder head. Honda specs for the small 10mm plugs, used on the 1960-61 engines is a C7HA plug, which is a heat range cooler than what was supplied.  I will fire the bike up and ride it gently on the C6HS plugs then switch to C7HSA plugs for extended riding during initial break-in runs.

I’m not sure how long the bike was down since the blow-up, but the battery fluid levels were near the tops of the cells and measured voltage was 1.4v. An overnight charge brought it up to 6.66v initially, but started to fall to 6.4v later in the day. With only 6 volts available to power these bikes, you need everyone you can corral.

UP and RUNNING, but not ready yet

I was able to get the bike started up and ran it around briefly. Lingering problems exist in the fuel delivery and clutch. The engine ran well at low speeds but started to break up as throttle was increased. One issue was a bad spark plug cap connection on the right side. I tried jetting the bike down once to see if the oversized-main jet was too rich, but it got worse and responded to adding choke under power. I then took a jet reamer to the #100 main jet and opened it up to about 105 size. The bike has stock mufflers, but no baffles installed so jetting must be compensated for those changes, as well. The bike began to run much better and then the aftermarket clutch cable broke.

Both the clutch cable and throttle cables were too long and are old Dixie/Superior branded parts of less than wonderful quality. So, I needed to track down some correct cables for the bike, which were found on eBay; real ones this time, not reproductions. eBay sellers came up with actual OEM correct gray cables with grease fittings installed, plus the little tensioner springs were included.

The clutch drag continued, even with correct adjustment. I have to do research on whether the clutch pack wound up being too tall and that changed the angle of the clutch lifter arm and/or the clutch adjuster threads are worn out and we are losing pushrod travel because of that.

I had to replace the oil filter with a later model version due to the change in crankshaft designs. So far the engine sounds great, doesn't smoke and isn't making any unusual noises or leak oil anywhere.

 More work and more research and good results

The OEM Honda clutch and throttle cables came in quickly and were slowly replaced. Working on a Benly requires a bit of watchmaker and puzzle solver in order to fit all the small bits back together again. The cables must be routed through the pressed-steel chassis holes and meet up with their end point connections.

With the clutch cable installed, the clutch adjuster was checked for excessive play, but nothing out of the ordinary was observed. The throttle cable required the same feed through the chassis holes and then the cable end connected to the carburetor top components. The main jet was removed and opened up further, followed by another test ride. While the engine continued to show more power the clutch continued to show signs of dragging, despite all the cable adjustments being done correctly.

All that was left was to drain the oil and remove the clutch cover (after the exhaust system) and take a look at the clutch release function at the end point. Pulling the clutch lever in, at the handlebars, showed a restricted clutch release travel at the plate set, when the assembly was all together.  The feeling at the lever was that perhaps the clutch springs were coil-binding at full travel. Honda didn’t design that kind of problem into this model and this one shouldn’t be coming to a hard stop at the handlebar lever.

With the springs and spring plate removed the clutch travel seemed to increase. All the clutch plates were checked for thickness as one odd one was in the pack. Most of the correct plates measured out around 3.4mm, but one was at 3.0mm and a different type with thinner friction material. The PLATE A was a thick alloy piece, which was changed in later models to a different design along with a different center hub.  David Silver Spares came up with a correct friction plate, steel Plate A and outer pressure plate. The pressure plate showed some warping when the clutch pack was spun by hand. There were two low spots on the plate surface, as well.

The clutch spring retainer plate has a central hole, which is a stamped steel piece which is punched through creating an offset hole that matches the one on the outer clutch pressure plate. Putting the plate on with the offset outwards appears to cause some unnecessary binding of the plates when they meet in the middle. Turning the plate around so the offset nestles down next to the pressure plate’s center allows a more full travel of the plate set. With a correct set of friction plates, pressure plate and Plate A, the clutch should resume normal function and travel once again. A rough reassembly using the old parts showed increased travel/plate spread when the outer support plate was installed correctly. I discovered some Barnett Kevlar clutch friction plates in a box which fit the clutch hub on this bike, so it is getting treated to a hopefully bullet-proof clutch pack to carry it into the future.

When received, the steering head bearings felt notched and self-centering; so the bike just wanted to go straight ahead all the time. Removing the handlebars and mount allowed access to the top plate stem nut which was tightened with a great amount of force. Backing off the nut allowed the stem bearing nut to be loosened slightly and suddenly the steering felt transformed back to a normal feel once again.

The drained oil showed minute traces of stray flecks of brass and some fine powdery grains of steel being loosened and held in suspension by the Honda GN4 oil. The second cleaning of the filter showed almost no brass particles now. It appears that the updated, used crankshaft transplant has been successful, but not a completely easy task. Still, with the inventories of new crankshafts exhausted, selective use of the newer versions seems to be a viable alternative.

The final test rides showed a normal pull clutch lever feel, good power through the rpm range and a very slick feeling transmission function. The owner was very pleased with the results in the end.

This has been a thought-provoking and time-intensive project from start to finish. Some $600 worth of replacement parts have been purchased to date, along with a great deal of research time online for parts and hands-on labor to R&R the engine, tear it down, inspect and clean all the parts and order the correct replacement items to bring the bike back towards the original running condition. To paraphrase, “You learn the greatest lessons on a Honda!”

1 comment: