- Joined
- Jun 4, 2008
- Messages
- 3,291
- Reaction score
- 635
And to think this started with seeing if he could make the crankshaft.
Another Radial - this time 18 Cylinders
- Thread starter mayhugh1
- Start date
Help Support Home Model Engine Machinist Forum:
- Joined
- Jun 28, 2011
- Messages
- 238
- Reaction score
- 480
I'm thinking an R4360 !
- Joined
- Feb 17, 2014
- Messages
- 41
- Reaction score
- 6
I'm so speechless, every time reading this monster cylinder topic :bow: Thm:
... or a quarter scale Merlin V-12.
- Joined
- Jun 28, 2011
- Messages
- 238
- Reaction score
- 480
One more request, how about an Allison V-3420.
Pete
Pete
Last edited:
The 1/64" thick Aramid sheet gasket material that I ordered from MSC arrived, and while cutting out a new batch of exhaust flange gaskets I temperature tested some samples of this new material. MSC advertises its maximum temperature at 700F although it's not at all clear what that means. This material has a kid leather-like feel, but it cuts and tears like the auto gasket material I've been using. I did a web search, and what I have looks very much like a product called Durlon 8500.
For my test, I cut out some washers and sandwiched them between metal washers in a stack on a 10-32 bolt. I heated the compressed stack to 500F for an hour. Disappointingly, the result looked pretty much like my over-heated auto gasket material as it had become very brittle and dark in color. The test gaskets also stuck to the metal washers but not as badly as I had seen with the auto gasket material. I re-ran the test at 300F which is more realistic for my application, and in this test the material did not discolor, and it remained flexible. In fact, it looked so good that I repeated the test to make sure I had programmed the oven properly. The results were the same, and so I started the process of replacing the gaskets already in my engine. Evidently MSC's maximum temperature isn't a service temperature but, instead, a point just below where the material combusts. For future reference I repeated the test at 350F and 400F using two new batches of test washers. The 350F results still looked pretty good, but the material had turned too brittle for my use after an hour at 400F.
Assembling the cylinders onto the crankcase wasn't as bad as I was expecting. The engine stand was invaluable for accessing the studs, and it was satisfying to see all the complexity finally coming together. I was expecting poor access to the studs around the sump, but the biggest challenge turned out to be working around the oil feed tubes. There were about a dozen nuts out of the 164 that became mini-milestones during the assembly. I was eventually able to tighten the nuts on every stud of every cylinder, but to do it I had to make up a set of 3/16" crowfoot wrenches.
The strap wrench for working the stainless steel sleeves over top of the heat-shrunk Kynar worked well on all the intake tubes. It even worked well under the notch of the very last installed cylinder. While I was dealing with the sleeves I also made a couple extraction tools that fit under the clearance notch in the rear heads and allow me to push the relatively flimsy thin-wall sleeves off during disassembly. Hopefully, they won't have to be used very much, if ever. Cylinder clearances around the sump and oil feed tubes came out as expected, and I should be able to remove any single cylinder should it become necessary. Kynar has a fairly high shrink temperature, and so a reflective shroud was required around the nose of my small heat gun to focus enough heat on the splice.
I measured 17 in-lbs of torque needed to spin the crankshaft and overcome the friction created by the 54 piston rings. This is equivalent to a 1/3 horsepower loss at 1000 rpm. This frictional loss should decrease some over time, but additional losses will be added when the pushrods are installed.
A temporary oil tank was jury-rigged and, after inserting some rubber plugs in the lower tappet bores, I used a drill to spin the engine and exercise the oiling system while flushing two quarts of oil through the engine. A borescope inserted into the front and rear tappet bores showed all the cams were being well oiled even with the drill running at very low speed. With the drill running at 600 rpm and gravity feeding the full 1/4" oil line to the engine's pressure pump I was easily able to get ahead of the scavenger pump and flood the crankcase. The oil pumps in this engine are similar to the ones in my H-9, and so an oil feed regulator similar to the one I made for my H-9 will eventually be required when the engine is actually running under its own power. After flushing the oil through the engine, I cleaned some ring debris from the magnetic drain plug in the sump. The next step is to finish the pushrods and install the rockers for the cylinder compression tests. Terry
For my test, I cut out some washers and sandwiched them between metal washers in a stack on a 10-32 bolt. I heated the compressed stack to 500F for an hour. Disappointingly, the result looked pretty much like my over-heated auto gasket material as it had become very brittle and dark in color. The test gaskets also stuck to the metal washers but not as badly as I had seen with the auto gasket material. I re-ran the test at 300F which is more realistic for my application, and in this test the material did not discolor, and it remained flexible. In fact, it looked so good that I repeated the test to make sure I had programmed the oven properly. The results were the same, and so I started the process of replacing the gaskets already in my engine. Evidently MSC's maximum temperature isn't a service temperature but, instead, a point just below where the material combusts. For future reference I repeated the test at 350F and 400F using two new batches of test washers. The 350F results still looked pretty good, but the material had turned too brittle for my use after an hour at 400F.
Assembling the cylinders onto the crankcase wasn't as bad as I was expecting. The engine stand was invaluable for accessing the studs, and it was satisfying to see all the complexity finally coming together. I was expecting poor access to the studs around the sump, but the biggest challenge turned out to be working around the oil feed tubes. There were about a dozen nuts out of the 164 that became mini-milestones during the assembly. I was eventually able to tighten the nuts on every stud of every cylinder, but to do it I had to make up a set of 3/16" crowfoot wrenches.
The strap wrench for working the stainless steel sleeves over top of the heat-shrunk Kynar worked well on all the intake tubes. It even worked well under the notch of the very last installed cylinder. While I was dealing with the sleeves I also made a couple extraction tools that fit under the clearance notch in the rear heads and allow me to push the relatively flimsy thin-wall sleeves off during disassembly. Hopefully, they won't have to be used very much, if ever. Cylinder clearances around the sump and oil feed tubes came out as expected, and I should be able to remove any single cylinder should it become necessary. Kynar has a fairly high shrink temperature, and so a reflective shroud was required around the nose of my small heat gun to focus enough heat on the splice.
I measured 17 in-lbs of torque needed to spin the crankshaft and overcome the friction created by the 54 piston rings. This is equivalent to a 1/3 horsepower loss at 1000 rpm. This frictional loss should decrease some over time, but additional losses will be added when the pushrods are installed.
A temporary oil tank was jury-rigged and, after inserting some rubber plugs in the lower tappet bores, I used a drill to spin the engine and exercise the oiling system while flushing two quarts of oil through the engine. A borescope inserted into the front and rear tappet bores showed all the cams were being well oiled even with the drill running at very low speed. With the drill running at 600 rpm and gravity feeding the full 1/4" oil line to the engine's pressure pump I was easily able to get ahead of the scavenger pump and flood the crankcase. The oil pumps in this engine are similar to the ones in my H-9, and so an oil feed regulator similar to the one I made for my H-9 will eventually be required when the engine is actually running under its own power. After flushing the oil through the engine, I cleaned some ring debris from the magnetic drain plug in the sump. The next step is to finish the pushrods and install the rockers for the cylinder compression tests. Terry
- Joined
- Jun 24, 2010
- Messages
- 2,425
- Reaction score
- 959
Woah-Wait-What? The Kynar is heat shrink tubing right? (clear stuff you show with heat gun & reflector). So is the external SS sleeve pre-positioned on the straight segment up near the cylinder & then you slide it back over heat shrink? Is the intention to do this hot so the shrink tubing OD kind of bonds to SS sleeve or maybe some in-between adhesive step? Now what about disassembly? sacrificially cut & peel off the tubing sandwich? Very spiffy end result look!
The strap wrench for working the stainless steel sleeves over top of the heat-shrunk Kynar worked well on all the intake tubes. .. I also made a couple extraction tools that fit under the clearance notch in the rear heads and allow me to push the relatively flimsy thin-wall sleeves off during disassembly... Kynar has a fairly high shrink temperature, and so a reflective shroud was required around the nose of my small heat gun to focus enough heat on the splice.
The strap wrench for working the stainless steel sleeves over top of the heat-shrunk Kynar worked well on all the intake tubes. .. I also made a couple extraction tools that fit under the clearance notch in the rear heads and allow me to push the relatively flimsy thin-wall sleeves off during disassembly... Kynar has a fairly high shrink temperature, and so a reflective shroud was required around the nose of my small heat gun to focus enough heat on the splice.
I started a set of pushrods last year in order to get some of the high volume work done before the Texas heat set in. At that time I was able to machine only one end of the rods since the head design wasn't yet complete, and I was unsure of the final length. There's a total of 14 stacked machined parts with tolerances that affect this length, and so there is also a lash adjustment on the end of each rocker arm. To determine the exact final length, I made up a variable length pushrod and trial-fitted it in six different positions around the engine. This empirically determined length came out to within .010" of my design value, and fortunately the length was correct for all six positions.
After determining the pushrod length, I made a fixture to cut my batch of semi-finished rods to their final length + .003" using a slitting saw and fixture on my mill. The second end was then turned to a hemisphere using my 9x20 lathe and the CAM program I developed last year for the other end.
Although the 3/32" diameter pushrods look at home on my H-9, they're a bit too skinny looking in front of the massive heads on this engine. In my scrap collection I found a bag of 3"-6" drops of hardened hypodermic stock with the proper i.d. to slip over my rods to increase their o.d.'s to nearly 1/8". Unfortunately, both ends of the tubes had been crushed flat and would not fit into my lathe's collet chuck so I first rough cut the ends using an abrasive disk in a fixture on my mill, and then I finished them in the lathe. I heat treated the sleeves, not for any metallurgical reason, but for the dark color that the heat imparted to them. One of the photos shows a side-by-side comparison of the sleeved and un-sleeved rods.
With the pushrods completed, I was anxious to install a pair of rocker arm assemblies and run a first compression test on one of the cylinders to sanity check my work up to this point. Unfortunately, my compression gage would not fit between the rocker assemblies on the front row of cylinders due to an interference with the gage's pressure release button. I was able, though, to make a test on one of the rear cylinders thanks to the head geometry differences between the front and rear row cylinders.
The next problem I ran into was created by my jury-rigged oil tank. Without proper flow control, the engine quickly flooded with oil when the crankshaft was spun fast enough to get a consistent compression test. I rigged up a temporary flow-controller using a cheap aquarium air flow regulator purchased from a local pet supply store. Then I reluctantly added an extension to my compression gage to get the release button above the rocker assemblies. This was a lot easier said than done, though, and I spent an entire day chasing leaks in my gage created by the modification. I also had to make a special box wrench to access the lock nuts on the lash adjusters. The few tiny commercial wrenches I had on hand were too awkward to use in the rocker arm slots in the rocker boxes.
During my H-9 build, I estimated the engine's expected compression test pressures based on the simple formula:
Ptest = (CR)(Pman),
where CR is the static compression ratio, and Pman is the engine's manifold pressure. Pman is 14.7psi for a normally aspirated engine.
For my H-9's compression ratio of 5, the expected pressure was (5)(14.7)=74 psi, and all its cylinders measured to within 10% of this value just after the engine was first assembled.
This simple formula, though, does not take into account the cylinder's volumetric efficiency or the adiabatic heating of the compressed air charge during the test. A more complete expression is:
Ptest = (Veff)(CR^1.33)(Pman),
where Veff is the volumetric efficiency and is typically 0.6 during a cranking test. The 1.33 is a specific heat term to account for the self-heating of the compressed air in the combustion chamber during the test.
When the H-9's compression ratio is plugged into this equation:
Ptest = (.6)(5^1.33)(14.7) = 75 psi.
The two correction terms in the more complete expression work to offset one another, and for low CR's they nearly cancel. This is not the case, though, for higher CR's found in full-size engines or in model competition engines where the full equation is more applicable. For example, with a compression ratio of 10 the simple equation gives 147 psi while the more complete expression gives 189 psi.
The H-9 compression ratio of 5 that I've been using is considerably lower than the advertised 6.7 for that engine. I believe the difference is due to the fact that the cylinder's thread relief clearance as well as the head gasket thickness were not accounted for in the original calculation. Due to the small combustion chamber volume, the compression ratio is very sensitive to these relatively small errors. I eventually verified my H-9 calculation with a burette test on one of my spare H-9 heads. Since I used the same combustion chamber geometry for my T-18 heads, I'm assuming the compression ratio of this engine is 5 as well.
Because the compression gage adds a portion of its own volume to the combustion chamber during the test, the raw readings need to be corrected for the additional volume below the gage's check valve. I had to modify my gage's correction factor when I added the extension needed to clear my rocker assemblies. Another issue that one has to be aware of when making compression tests on small model engines is excessive oil in the combustion chamber. An innocent looking puddle of oil on the top of a piston can raise the compression ratio of that cylinder significantly. My new engine stand allowed me to rotate the engine while I was installing the rocker arm assemblies and pushrods. When I started making the compression tests I wasn't being careful to return the engine to its full upright position before making each test. Since the oil return channels in the bottom of the crankcase were no longer at the lowest point in the engine, the crankcase partially filled with oil causing certain pistons, depending upon their locations, to pump oil into their cylinders. When I ran compression tests on those cylinders I found the pressures to be higher than expected. In fact, on one cylinder the excess oil had raised the compression high enough to damage the meter on my gage and it had to be replaced. I learned it was necessary to soak up the oil in those cylinders with Q-tips before running compression checks.
Obtaining absolute numbers in a compression test is a pretty dicey exercise. It might not even be all that meaningful, especially in a model engine where the combustion chamber is so tiny, and so many small effects can contribute errors. In any event, with a target value of 75psi, my compression test results were:
#1=65psi, #3=75psi , #5=67psi, #7=61psi,
#9=80psi, #11=80psi, #13=80psi, #15=68psi, #17=77psi,
#2=72psi, #4=68psi , #6=70psi, #8=71psi,
#10=81psi, #12=77psi, #14=64psi , #16=67psi, #18=65psi.
The high readings on the three very bottom cylinders (9, 10, 11) are likely a result of these cylinders, behind their pistons, being filled with crankcase oil during the test. These three or four tablespoons of oil in these inverted cylinders insured the rings were absolutely sealed. The fact that the readings were 80psi instead of 75psi probably indicates an error in my gage's calibration or correction factor. This oil drainage problem is one of the 'features' of a radial. Under power, the scavenger pump and windage keep these areas from filling with oil. But when the engine is shut down, crankcase oil drains back into these areas and contributes to the billowing smoke typically seen on start-up.
I don't plan to do any 'motoring' to 'break-in' the rings. The compression test results show the rings are sealing pretty well as installed. After doing a lot of thinking about this 'motoring-in stuff', I've come to my own conclusion that it's not a good idea.
At best, I feel it does nothing useful because there's no combustion pressure to push the rings hard against the cylinder walls which is required to properly seat them. Even with the pistons working against their cylinders' own compression only a fraction of the required pressure is produced compared to what's generated when the engine is running under power.
At worst, without this high pressure, motoring a new set of rings in a new cylinder might cause their surfaces to glaze over and then delay or even inhibit break-in. In my mind, it's a bit analogous to a cutting tool rubbing against a workpiece when the chip load is too low.
I reluctantly ran two quarts of oil through this engine only because I felt I had to thoroughly verify my new and unproven oiling system before the engine was allowed to run under its own power when an oil failure could be catastrophic. It took about 30 minutes of total run time spread over several days to flush the two quarts of oil through the engine. If I had thought out the assembly process a little more, the engine would have been flushed without the rings installed. The torque needed to spin the crankshaft changed from 17 in-lbs to 15 in-lbs as a result of the flush, and the decrease was likely due to ring wear. Hopefully, it will turn out to be beneficial to the break-in. With all the pushrods but none of the spark plugs installed, the torque required to spin the crankshaft remained at 15 in-lbs. This indicates the frictional losses related to the cams are insignificant when compared to those of the 54 piston rings.
The next and final steps in the engine assembly is to install and time the distributors and to come up with a spark plug harness. I'm still about a month away from attempting to start the engine, though. I still have to design and build the final mounting and control components including a firewall, fuel and oil tanks, fuel pump, tach, throttle control, and electrical wiring. - Terry
After determining the pushrod length, I made a fixture to cut my batch of semi-finished rods to their final length + .003" using a slitting saw and fixture on my mill. The second end was then turned to a hemisphere using my 9x20 lathe and the CAM program I developed last year for the other end.
Although the 3/32" diameter pushrods look at home on my H-9, they're a bit too skinny looking in front of the massive heads on this engine. In my scrap collection I found a bag of 3"-6" drops of hardened hypodermic stock with the proper i.d. to slip over my rods to increase their o.d.'s to nearly 1/8". Unfortunately, both ends of the tubes had been crushed flat and would not fit into my lathe's collet chuck so I first rough cut the ends using an abrasive disk in a fixture on my mill, and then I finished them in the lathe. I heat treated the sleeves, not for any metallurgical reason, but for the dark color that the heat imparted to them. One of the photos shows a side-by-side comparison of the sleeved and un-sleeved rods.
With the pushrods completed, I was anxious to install a pair of rocker arm assemblies and run a first compression test on one of the cylinders to sanity check my work up to this point. Unfortunately, my compression gage would not fit between the rocker assemblies on the front row of cylinders due to an interference with the gage's pressure release button. I was able, though, to make a test on one of the rear cylinders thanks to the head geometry differences between the front and rear row cylinders.
The next problem I ran into was created by my jury-rigged oil tank. Without proper flow control, the engine quickly flooded with oil when the crankshaft was spun fast enough to get a consistent compression test. I rigged up a temporary flow-controller using a cheap aquarium air flow regulator purchased from a local pet supply store. Then I reluctantly added an extension to my compression gage to get the release button above the rocker assemblies. This was a lot easier said than done, though, and I spent an entire day chasing leaks in my gage created by the modification. I also had to make a special box wrench to access the lock nuts on the lash adjusters. The few tiny commercial wrenches I had on hand were too awkward to use in the rocker arm slots in the rocker boxes.
During my H-9 build, I estimated the engine's expected compression test pressures based on the simple formula:
Ptest = (CR)(Pman),
where CR is the static compression ratio, and Pman is the engine's manifold pressure. Pman is 14.7psi for a normally aspirated engine.
For my H-9's compression ratio of 5, the expected pressure was (5)(14.7)=74 psi, and all its cylinders measured to within 10% of this value just after the engine was first assembled.
This simple formula, though, does not take into account the cylinder's volumetric efficiency or the adiabatic heating of the compressed air charge during the test. A more complete expression is:
Ptest = (Veff)(CR^1.33)(Pman),
where Veff is the volumetric efficiency and is typically 0.6 during a cranking test. The 1.33 is a specific heat term to account for the self-heating of the compressed air in the combustion chamber during the test.
When the H-9's compression ratio is plugged into this equation:
Ptest = (.6)(5^1.33)(14.7) = 75 psi.
The two correction terms in the more complete expression work to offset one another, and for low CR's they nearly cancel. This is not the case, though, for higher CR's found in full-size engines or in model competition engines where the full equation is more applicable. For example, with a compression ratio of 10 the simple equation gives 147 psi while the more complete expression gives 189 psi.
The H-9 compression ratio of 5 that I've been using is considerably lower than the advertised 6.7 for that engine. I believe the difference is due to the fact that the cylinder's thread relief clearance as well as the head gasket thickness were not accounted for in the original calculation. Due to the small combustion chamber volume, the compression ratio is very sensitive to these relatively small errors. I eventually verified my H-9 calculation with a burette test on one of my spare H-9 heads. Since I used the same combustion chamber geometry for my T-18 heads, I'm assuming the compression ratio of this engine is 5 as well.
Because the compression gage adds a portion of its own volume to the combustion chamber during the test, the raw readings need to be corrected for the additional volume below the gage's check valve. I had to modify my gage's correction factor when I added the extension needed to clear my rocker assemblies. Another issue that one has to be aware of when making compression tests on small model engines is excessive oil in the combustion chamber. An innocent looking puddle of oil on the top of a piston can raise the compression ratio of that cylinder significantly. My new engine stand allowed me to rotate the engine while I was installing the rocker arm assemblies and pushrods. When I started making the compression tests I wasn't being careful to return the engine to its full upright position before making each test. Since the oil return channels in the bottom of the crankcase were no longer at the lowest point in the engine, the crankcase partially filled with oil causing certain pistons, depending upon their locations, to pump oil into their cylinders. When I ran compression tests on those cylinders I found the pressures to be higher than expected. In fact, on one cylinder the excess oil had raised the compression high enough to damage the meter on my gage and it had to be replaced. I learned it was necessary to soak up the oil in those cylinders with Q-tips before running compression checks.
Obtaining absolute numbers in a compression test is a pretty dicey exercise. It might not even be all that meaningful, especially in a model engine where the combustion chamber is so tiny, and so many small effects can contribute errors. In any event, with a target value of 75psi, my compression test results were:
#1=65psi, #3=75psi , #5=67psi, #7=61psi,
#9=80psi, #11=80psi, #13=80psi, #15=68psi, #17=77psi,
#2=72psi, #4=68psi , #6=70psi, #8=71psi,
#10=81psi, #12=77psi, #14=64psi , #16=67psi, #18=65psi.
The high readings on the three very bottom cylinders (9, 10, 11) are likely a result of these cylinders, behind their pistons, being filled with crankcase oil during the test. These three or four tablespoons of oil in these inverted cylinders insured the rings were absolutely sealed. The fact that the readings were 80psi instead of 75psi probably indicates an error in my gage's calibration or correction factor. This oil drainage problem is one of the 'features' of a radial. Under power, the scavenger pump and windage keep these areas from filling with oil. But when the engine is shut down, crankcase oil drains back into these areas and contributes to the billowing smoke typically seen on start-up.
I don't plan to do any 'motoring' to 'break-in' the rings. The compression test results show the rings are sealing pretty well as installed. After doing a lot of thinking about this 'motoring-in stuff', I've come to my own conclusion that it's not a good idea.
At best, I feel it does nothing useful because there's no combustion pressure to push the rings hard against the cylinder walls which is required to properly seat them. Even with the pistons working against their cylinders' own compression only a fraction of the required pressure is produced compared to what's generated when the engine is running under power.
At worst, without this high pressure, motoring a new set of rings in a new cylinder might cause their surfaces to glaze over and then delay or even inhibit break-in. In my mind, it's a bit analogous to a cutting tool rubbing against a workpiece when the chip load is too low.
I reluctantly ran two quarts of oil through this engine only because I felt I had to thoroughly verify my new and unproven oiling system before the engine was allowed to run under its own power when an oil failure could be catastrophic. It took about 30 minutes of total run time spread over several days to flush the two quarts of oil through the engine. If I had thought out the assembly process a little more, the engine would have been flushed without the rings installed. The torque needed to spin the crankshaft changed from 17 in-lbs to 15 in-lbs as a result of the flush, and the decrease was likely due to ring wear. Hopefully, it will turn out to be beneficial to the break-in. With all the pushrods but none of the spark plugs installed, the torque required to spin the crankshaft remained at 15 in-lbs. This indicates the frictional losses related to the cams are insignificant when compared to those of the 54 piston rings.
The next and final steps in the engine assembly is to install and time the distributors and to come up with a spark plug harness. I'm still about a month away from attempting to start the engine, though. I still have to design and build the final mounting and control components including a firewall, fuel and oil tanks, fuel pump, tach, throttle control, and electrical wiring. - Terry
Swifty
Well-Known Member
It's getting to be a real work of art, you certainly are a very skilled worker. Even people with no idea of the machining skill involved should be impressed with the beauty of the engine.
Paul.
Paul.
holy cow does that look nice.
This build should impress anyone! I am a Tool and Die-maker (Injection mold maker) with over 40 years experience and I am blown away with this build.
Thanks all for your comments. - Terry
The engine was temporarily moved from the rotisserie to its vertical stand so the rear cover could be removed and the distributors installed. The distributor gears were installed after the distributors are inserted into the rear cover. Their mesh with the crankshaft driving gear was verified several months ago before any components were added to the crankshaft and when any binding could be detected. The backlash was re-checked, though, and measured just over one crankshaft degree. Distributor backlash typically isn't a problem since IC engines run in only one direction, and all timing measurements and adjustments can made after rotating the crankshaft in that same direction. Once back on the rotisserie and with the transparent distributor caps installed, the engine was ready to have its distributors timed. Since the two distributor gears are on opposite sides of the crankshaft drive gear, the distributors rotate in opposite directions just to make things even more interesting. As a reminder, the rotors' directions are engraved on the tops of the caps.
I found the only way to maintain sanity during the timing process was to treat the engine as I had originally envisioned it: that is, two separate engines on the same crankshaft and each with its own distributor with its own firing order. Standing at the rear of the engine and looking forward, the right-hand distributor controls the 'fire and skip' firing order 1-5-9-13-17-3-7-11-15 of the rear row of cylinders which are driven from the rod assembly on the rear crankshaft throw. Cylinder number one is physically located at the very top of the engine, and TDC of its firing stroke was located using a dial indicator in its spark plug port.
The distributor was lifted up from the rear cover to disengage its driven gear so the center of the rotor could be aligned as closely as possible to the number one cylinder tower. The TDC mark on the distributor body was kept in alignment with the stationary timing pointer on the engine. After getting the best rotor possible alignment, the hold-down setscrew was then retightened.
Because there are only 20 teeth on the distributor driven gear, the resolution of this adjustment is only 18 distributor degrees. To get the rotor more accurately positioned, the distributor cap must be removed to get access to a tiny setscrew that holds the rotor in place on the trigger disk. This setscrew tightens into a slot on the trigger disk so the two components are keyed together. The rotor was then removed to get access to a SHCS that secures the trigger disk to the shaft. The SHCS was loosened, the trigger disk rotated in the proper direction, and then the SHCS was re-tightened. The rotor was slipped back onto the trigger disk, its setscrew tightened, and its position was rechecked. The process was repeated until the centerline of the rotor was accurately aligned to the #1 tower.
After this step is completed, the firing event timing can be set. The distributor's trigger cable was then connected to an ignition module. The module was powered up with the high voltage switched off so its trigger LED could be used as a firing indicator without actually generating a spark. Even though the ignition is a CDI, the LED operation is designed to mimic a points-type ignition. That is, LED ON indicates 'coil' charging, and the plug fires when the LED turns OFF. A degree wheel was attached to the crankshaft, and the TDC dial indicator in the #1 cylinder plug port was used to zero the wheel. The crankshaft was then slowly rotated CCW until the LED turned ON and then OFF, and the firing error with respect to TDC was recorded.
The Hall sensor is mounted on its own disk inside the distributor body under the trigger disk and can be rotated with respect to it in order to zero out the timing error. This was done by loosening a setscrew on the underside of the distributor housing and rotating it and the internal sensor disk in the crescent slot until the LED turned OFF precisely at TDC.
The whole process was then repeated for the front row of cylinders using the left hand distributor. In this case, the timing was performed with respect to TDC of the firing stroke of the bottom-most cylinder #12, and the firing order for this distributor is 12-16-2-6-10-14-18-4-8.
In normal use, the timing of each distributor can be individually advanced by loosening the distributor hold-down screw and rotating its body. The scribed timing marks on the distributor housing are 10 crankshaft degrees apart.
After running through the combined firing order and being satisfied with the distributor timing the transparent distributor caps were replaced with the blue Delrin caps.
One of the photos shows the results of a solid tantalum capacitor that exploded inside one of my ignition modules while I was working with the distributor timing. The electronic components that I used are from 30 year old stock, but it's possible that I caused the explosion by inadvertently reversing the supply wiring. I was too fatigued at the time to know for sure. Anyway, the ignition module cleaned up OK, and the damage was limited to just the single capacitor.
The last photo shows three high voltage boot/clip possibilities for a CM-6 spark plug. The white boot is a commercially available Rcxel product that I've actually never used. The clip, but with more layers of shrink tubing than shown in the photo, is what I used on my H-9. Both commercial parts are available from S/S Machine and Engineering. The clip works well and looks OK on my H-9, but I wanted to use plug boots for this new engine. I don't like the looks of the large white Rcxel boot, so I came up with the third alternative shown in the photo. This boot is based on a commonly available automotive vacuum fitting. I'm currently tweaking a process to convert the vacuum fitting to a spark plug boot, and I'm torture testing some of the results. I plan to detail this later in a separate post. - Terry
I found the only way to maintain sanity during the timing process was to treat the engine as I had originally envisioned it: that is, two separate engines on the same crankshaft and each with its own distributor with its own firing order. Standing at the rear of the engine and looking forward, the right-hand distributor controls the 'fire and skip' firing order 1-5-9-13-17-3-7-11-15 of the rear row of cylinders which are driven from the rod assembly on the rear crankshaft throw. Cylinder number one is physically located at the very top of the engine, and TDC of its firing stroke was located using a dial indicator in its spark plug port.
The distributor was lifted up from the rear cover to disengage its driven gear so the center of the rotor could be aligned as closely as possible to the number one cylinder tower. The TDC mark on the distributor body was kept in alignment with the stationary timing pointer on the engine. After getting the best rotor possible alignment, the hold-down setscrew was then retightened.
Because there are only 20 teeth on the distributor driven gear, the resolution of this adjustment is only 18 distributor degrees. To get the rotor more accurately positioned, the distributor cap must be removed to get access to a tiny setscrew that holds the rotor in place on the trigger disk. This setscrew tightens into a slot on the trigger disk so the two components are keyed together. The rotor was then removed to get access to a SHCS that secures the trigger disk to the shaft. The SHCS was loosened, the trigger disk rotated in the proper direction, and then the SHCS was re-tightened. The rotor was slipped back onto the trigger disk, its setscrew tightened, and its position was rechecked. The process was repeated until the centerline of the rotor was accurately aligned to the #1 tower.
After this step is completed, the firing event timing can be set. The distributor's trigger cable was then connected to an ignition module. The module was powered up with the high voltage switched off so its trigger LED could be used as a firing indicator without actually generating a spark. Even though the ignition is a CDI, the LED operation is designed to mimic a points-type ignition. That is, LED ON indicates 'coil' charging, and the plug fires when the LED turns OFF. A degree wheel was attached to the crankshaft, and the TDC dial indicator in the #1 cylinder plug port was used to zero the wheel. The crankshaft was then slowly rotated CCW until the LED turned ON and then OFF, and the firing error with respect to TDC was recorded.
The Hall sensor is mounted on its own disk inside the distributor body under the trigger disk and can be rotated with respect to it in order to zero out the timing error. This was done by loosening a setscrew on the underside of the distributor housing and rotating it and the internal sensor disk in the crescent slot until the LED turned OFF precisely at TDC.
The whole process was then repeated for the front row of cylinders using the left hand distributor. In this case, the timing was performed with respect to TDC of the firing stroke of the bottom-most cylinder #12, and the firing order for this distributor is 12-16-2-6-10-14-18-4-8.
In normal use, the timing of each distributor can be individually advanced by loosening the distributor hold-down screw and rotating its body. The scribed timing marks on the distributor housing are 10 crankshaft degrees apart.
After running through the combined firing order and being satisfied with the distributor timing the transparent distributor caps were replaced with the blue Delrin caps.
One of the photos shows the results of a solid tantalum capacitor that exploded inside one of my ignition modules while I was working with the distributor timing. The electronic components that I used are from 30 year old stock, but it's possible that I caused the explosion by inadvertently reversing the supply wiring. I was too fatigued at the time to know for sure. Anyway, the ignition module cleaned up OK, and the damage was limited to just the single capacitor.
The last photo shows three high voltage boot/clip possibilities for a CM-6 spark plug. The white boot is a commercially available Rcxel product that I've actually never used. The clip, but with more layers of shrink tubing than shown in the photo, is what I used on my H-9. Both commercial parts are available from S/S Machine and Engineering. The clip works well and looks OK on my H-9, but I wanted to use plug boots for this new engine. I don't like the looks of the large white Rcxel boot, so I came up with the third alternative shown in the photo. This boot is based on a commonly available automotive vacuum fitting. I'm currently tweaking a process to convert the vacuum fitting to a spark plug boot, and I'm torture testing some of the results. I plan to detail this later in a separate post. - Terry
