Fixing The Valve Train
The One Valve Engine
Rush back HOME
(Click on the pictures for a high resolution view)
5 September 2011:
Well, I've been doing some thinking (That, in itself, is a dangerous thing) and decided to scrap the oddball eccentric/ratchet valve drive and go with a roller chain drive and a conventional cam.
Of course the cam isn't what you'd call conventional for a 4-cycle engine. The duration of the one lobe is 210 degrees. I plan for the valve to open about 30 degrees before bottom dead center on the power stroke and close right at the bottom of the intake stroke. That should work.
Today, I did some whittling on the hub for the governor so it will accept a 15 tooth #25 (1/4" pitch) roller chain sprocket.
Governor hub, modified for sprocket. Governor hub assembly.
I turned the hub diameter to match the crankshaft diameter (0.750"). The sprocket will mount with it's hub facing the end of the crankshaft and plate facing the governor trunion. The pulley hub will be turned to 0.750(-) for a press fit on the governor hub and will be faced off so it protrudes 0.100" past the end of the governor hub. This is in order for it to slip over the end of the crankshaft to align the two shafts. There is an internal bolt hole in the governor hub that matches a threaded hole in the end of the crankshaft. A 1/4-20 socket head bolt goes inside the governor hub to attach the hub to the crankshaft.
I wasn't happy with the steel-on-steel bearing for the governor which ended up working rough so, since I needed space between the sprocket and the governor trunion for the chain to clear, I turned the diameter of the governor hub to the same 0.750" as for the sprocket and lightly pressed on a 1.000" O.D. brass bushing. I put the whole thing in the lathe and turned the O.D. past the 0.100" spacing shoulder to 0.950".
I then put the governor trunion in the lathe and bored out the I.D. to 0.9505" for a nice slick fit.
Making the cam.
The old cam assembly, after discarding the ratchet, was mounted in the lathe and the old cam was turned to a diameter of 1.250".
The cam blank is made of 0.250" hot rolled steel plate. The roughed-out square cam blank was mounted in the 4 jaw chuck and centered. The I.D. bore was drilled then bored to 1.249" for a press fit over the old camshaft.
A fixture was made for the rotary table and was used to hold the cam blank for turning to the final outer diameter.
I made use of my rotary table in the mill to make the cam. I don't know how real machinists do it, but I just mounted the cam blank with it's rise diameter turned to size on the rotary table using the made-up adapter.
The tangent of the cam blank was found. The carriage on the mill was moved off the cam tangent to it and then was traversed so the tool would cut 0.215" from the diameter. This is the lift plus clearance dimension.
While the mill was running, I moved the carriage to the previously found tangent point, machining the "fall" of the profile with the rotary table set to zero degrees. After that, I slowly rotated the table 150 degrees, removing material to make the low side of the cam. When 150 degrees was reached, I again moved the table in a tangent direction to the blank to make the "rise" of the cam. It worked like a champ - a LOT better than my old method of making the cam to a printer-plotted pattern glued to the cam blank. Oh, yes, 150 degrees of low side equals 210 degrees of high side or duration.
New cam on left and modified old hub on right. Hub pressed into new cam.
Now, all I have to do is wait for my sprockets to come in and I can see if it runs better.
7 September 2011:
Today, the sprockets arrived.
All I got done today was to unpack the sprockets and lay 'em up aside their mountings. The crankshaft sprocket (the smaller one on the right) is going to be fun to fit but it can be done.
8 September 2011:
The timing chain is done.
Sprockets bored, pressed on and faced. Lifter guide and lifter relieved for chain clearance.
After boring the I.D.'s of the sprockets, they were pressed on their respective shafts then faced-off. The cam sprocket was faced-off flush with the end of the bearing bore and the crankshaft sprocket was faced to allow it to slip 0.095" over the crankshaft stub for alignment. There is a bolt in the hollow governor hub to hold it to the crankshaft stub.
The next order of business was to relieve the lifter guide and lifter for clearance for the chain to pass. Since the chain lays over the end of the lifter, the original governor latch won't work. I plan to extend the fork and make a pin and slot arrangement on the lifter like the one on the Non-McVicherish engine.
Sprockets and chain in place. Tensioner added.
The nearest link on the chain that I could match up with the repair link caused it to be a bit on the loose side. That problem was solved by making a copolymer plastic idler roller that is on an eccentric for adjustment. That item will need a little modification because the eccentric turns as I tighten the 10-32 screw. I may just mill a couple of flats on the end cap to hang onto it with a Crescent wrench while tightening the bolt.
I'll have to finagle the spark saver to somewhere other than where it was on the ratchet drive and make an arm that will contact the little toggle lever to relieve the fuel pump during latch-up so fuel is not delivered and wasted.
After the governor latch is worked out, the next thing to tackle is the fuel pump cam follower. It will run on the valve cam but will be timed about 240 (crank) degrees after the valve opens. That will make the injection point at about 30 degrees after the beginning of the intake stroke when there should be some velocity in the transfer port.
10 September 2011:
Today, it ran again. This time, after a lot of fiddling, it ran for a longer time on it's own than before but there are still fuel pump issues. At one point, it drove the drive motor and latched so it does make power when it wants to.
Governor latching arrangement. Latch in place.
An 1/8" deep slot was milled in the valve lifter. This aligns with the hole in the lifter guide. The pin is a hard steel 1/8" dowel pin with a sleeve pressed over one end. A spring goes around the pin to allow it to retract from the slot in the lifter.
A piece of spring brass was bent to act as a buffer for when the governor goes into the latch position before the valve opens. This keeps from straining the governor yoke and trunion.
The valve train, fuel pump drive and governor latch. The engine as of today.
The entire governor, fuel bypass and fuel pump drive arrangement is shown. The fuel pump arm was cut and angled so a small ball bearing could ride on the cam to operate the pump. Timing is about 45 degrees into the intake stroke.
Note the high-tech rubber washer to hold the bypass valve shut. The engine doesn't want to run when it's free so I think I need to revisit the return spring.
The governor leaf springs aren't strong enough to allow the engine to run at a speed at which it's happy so, temporarily, I tied a rubber band between the governor weights as you can see in the photo above right. Tomorrow, I'll drill the arms for the weights and put a tension spring there to help. The rubber band, although technically viable isn't what you'd call a permanent thing.
I also have to hook up the spark saver.
Maybe tomorrow, I can get the fuel pump valve issues sorted out. I got some very fine alumina lapping dust and I'll try that as a final lap. Once I get the fuel pump working right, the engine should run for extended periods.
11 September 2011:
Well, today I think I can call The One Valve Engine finished. There are always a few tweaks to do after this point but it has now run for as long as 30 minutes and can be started by spinning the flywheel with a leather strap.
What I did today was to re-lap the pop-off valve in the injector using the fine alumina dust and mineral spirits. I also installed a stronger return spring for the injector valve and drilled a spring seat in the head of the valve poppet so the spring will be centered on the valve head. It seems to work better now, although I still think there are some minor leakage issues because the engine will occasionally misfire and it seems to me that the pump stroke is awfully long for the amount of fuel it's pumping. This is probably due to gas bubbles from the leaking injector valve compressing when the pump cycles, making it pump a smaller volume.
I also made a new, longer (and straighter) fuel pump plunger. This eliminated the seal leakage and sticking. Also, the spark saver was hooked up and a tension spring was installed between the governor flyballs.
Anyway, I made a new video of the engine today and here it is on YouTube:
16 September 2011:
Yesterday, I fiddled with the injector, trying to get it to stop leaking so the engine would run better. No luck there.
Today, in an attempt to make a poppet valve that wouldn't leak, I made a totally new injector.
The new injector.
The body is a piece of 0.750" leadloy bar stock. The poppet valve stem is made from a piece of 0.093" stainless with a 1/4" head shrunk on the stem. Instead of a push spring as in the original injector, I have arranged a pull spring inside the cap. As usual, I'm having a devil of a time to get the valve to quit leaking but, at least, it's big enough for me to see the leakage path. More lapping it and it should seat fine.
With valves this small, it's hard to get the compromise between a compound coarse enough to cut without galling and one fine enough to remove any metal at all.
16 September 2011:
After lapping the injector poppet another time, I've finally gotten it to seal. I also decreased the strength of the spring so the pump doesn't have to make such a high pressure to crack the valve. Also, with the decreased pressure, the pump bypass valve works reasonably well.
I also rooted around for something to make a see-through fuel tank. This is because, if it runs out of fuel, the whole fuel system has to be purged of air, which is a pain. With a visible fuel level, hopefully I can remember to add fuel before it runs out. Also, I can see the fuel pump bypass flow. What I found was a big oiler. I got it with some other oilers I bought a couple of years ago. I wondered what to do with it and now I know.
Sporting the new fuel tank (naphtha fuel), the new injector and the data plate.
You will note the Mickey Mouse way I've got the bypass line rigged to the fill lid of the tank. I did this temporarily and, since I took the photos, I've drilled and tapped the top of the oiler for 1/8 NPT and now have a proper fitting and spigot so I can see the return flow.
I ran the engine for about an hour and a half off and on and am getting a feel for the adjustments. When it's running well (more of the time, lately) it will hit once then coast for about 12 revolutions and do this steadily, running at about 750 RPM.
I also figured out how to adjust the fuel pump stroke (fuel mixture). When the engine is not loaded, it's very difficult to find the sweet spot but I found that when I put a rag on the flywheel to load it to where it hits every time, the mixture can be leaned 'til it hits every other time then richened until it fires every time.
A couple of things went haywire and made the engine quit. One of them had happened several times. That problem was the pump arm slipping on the cam follower shaft. I fixed it by changing the clamping screw from a 6-32 to an 8-32. That allows more grip on the shaft.
Then, the flyball spring (between the governor balls) broke, making the engine try to run so slow it would quit.
Then, the solder joint on the pump inlet valve nut cracked, causing a small fuel leak and intermittent sucking of an air bubble in the pump. I got that fixed then, the return spring on the fuel pump plunger broke, causing the engine to run erratically. I'll fix that tomorrow. Just another bit of de-bugging a new design.
There's one thing that I didn't think about when I designed this engine. It will run fairly well in reverse. This was discovered when starting it, when, a couple of times it backfired and started running backwards. I think with some tweaks, it could be made to run equally well in either direction. Now that I think about it, I can see how that can happen.
The valve is open for one turn and closed for one turn. The fuel is injected about midway in the intake stroke and spark occurs a little before TDC. If the engine tries to run backwards, the intake stroke becomes the exhaust stroke and the exhaust stroke becomes the intake stroke. Fuel injection occurs relatively late in the exhaust stroke so most of the charge stays in the cylinder. Interesting!
26 September 2011:
Mechanically, I think I've got it ready for showing.
Ready for prime time.
I made a table that I can carry the engine on in the car then, by adding four EMT tubing legs, it will stand about 30 inches off of the ground.
There are only a few issues to eventually address. One of them is that there must be an air bubble trapped in the lower part of the injector. The fuel pump stroke is much greater than it needs to be so I think the bubble is being compressed until the cracking pressure is reached. I think this causes the occasional "hiccup" when it's running.
There might be an issue with the fuel (naphtha) itself. When the engine is first started, it runs very steadily but once warmed up, it starts acting odd on occasion. I'm able to make it run better for a while by, while it's running, cracking the bleeder screw on the pump. There is always a little air (gas) bubble that blows out. After "burping" it, it will run better for a while until I think the fuel boils again.
10 November 2011:
I've spent a lot of time trying to find the source of the air bubbles that appear in the fuel pump. After trying a number of "fixes", I decided to do the pragmatic thing.
What I did was to build a small brass block that fits on the fuel pump bleeder screw, the purpose of which is to allow a little fuel/air to flow through the bleeder port and pass back to the tank.
This device has a spring loaded check valve and a needle valve. I'll try to explain how it works. The cap screw screws vertically into the bleeder port of the fuel pump. Since only a very small volume needs to pass through the bleeder, I rely on flow up the thread clearance of the bleeder port. There are two sealing washers, one between the pump and the block and another between the screw head and the block.
The hole that the check valve is in necks down past the ball seat and ends-up connecting with the clearance hole to the bleeder port screw. The check valve spring is adjusted so it cracks at a slightly lower pressure than the poppet in the injector.
From the check valve passage, another hole is drilled vertically from the top of the block. This is threaded partway then drilled with a smaller bit at the seat for the needle valve until it intersects with the check valve bore. When fuel/air is pushed past the check valve, it is metered through the needle valve.
Another hole is drilled above and parallel to the check valve bore and intersects the needle valve bore. It is threaded 6-32 for a short distance and the end of the 1/8" copper tubing is likewise threaded. The 1/8" tube is screwed into the block then soldered in place.
Note that I'm using a tiny string of Teflon plumbing tape on the threads of the check valve screw and the needle valve screw to seal against leakage. So far, this appears to be working fine.
Automatic Bleeder mounted on the fuel pump.
In the above photo, you can see the automatic bleeder block mounted on the fuel pump. Adjustments are made with the engine running. The starting adjustments have the metering needle valve open about 1/4 turn and the check valve screwed-in to where the spring is totally compressed.
The check valve screw is backed-off until fuel/air flow is noted in the tubing back to the fuel tank. The pump stroke must be lengthened to compensate for the lost fuel in the "controlled leak". The needle valve is then closed until fuel/air flow is just enough to make an air bubble appear every few seconds. It seems that this adjustment is best when about one drop of fuel/air is returned to the tank on each stroke of the pump.
Air bubbles being bled-off in the fuel returning to the tank.
I ran the engine for about an hour after the temperature stabilized, making further adjustments of fuel pump stroke, bypass valve and bleeder needle. After I got the adjustments dialed-in, the engine ran for a solid half-hour with no tweaking.
At some points, the hit-miss cycle was very steady but did change some caused, I think, by varying amounts of air traveling to the automatic bleeder. When a bubble or two forms in the pump and while moving toward the bleeder needle; the fuel pump cannot deliver as much fuel to the injector as it can when there's no air in the system so the mixture leans out and richens as the engine runs and this causes the hit-miss cadence to change.
12 November 2011:
Yesterday, I ran the engine continuously. Occasionally, I held a rag to the flywheel and let the engine run at full power for a couple of minutes. It ran steadily under load but, interestingly, tended to run rich when firing every time.
After about an hour of running, it got to running poorly then quit. It appeared that the temperature had risen in the injector to where the fuel boiled and this is what made it quit. Priming didn't help.
Today, I primed the fuel system and the engine fired right off (after I richened the mixture by about a turn and a half on the pump stroke nut). After running for a couple of minutes, I could reset the fuel pump stroke and the engine ran fine.
On the theory that the fuel was boiling in the injector, I removed the injector and put in a stronger poppet spring, raising the cracking pressure. The engine again fan fine for about an hour until it did the same thing as before. Right after it quit, I stuck a thermocouple in the hopper water and read 183F (84C) and the temperature at the base of the injector read 207F (97C).
At this point, I am at a loss as to what to do to make the engine run for over an hour at a time.
4 September 2012:
After thinking about it for a while, I decided to re-work the fuel pump and injector. The injector would occasionally get air or vapor bubbles in it, causing the engine to run poorly or quit. The only option was to loosen the screw atop the injector and work the pump to expel the air.
Also, when the governor was latched, I had the pump bypass fuel directly from the pump to the tank. This didn't help the vapor problem in the injector so I decided to make the governor latchout bypass happen at the injector. While I was at it, I increased the fuel pump plunger diameter from 0.100" to 0.156".
To make the injector work at higher pressure, I installed a stronger spring. This was done to try to help stop the fuel from boiling and also make the bypass work better.
The redesigned governor latch duel bypass.
In the left-hand photo above, you can see the brass bypass valve atop the injector. The horizontal arm connects the valve stem of the bypass to the governor arm. When the engine is below governed speed, the governor arm is toward the viewer, forcing the bypass valve against it's seat. The point at which the valve opens and closes is adjusted by the screw and locknut at the right-hand end of the horizontal valve arm.
When the engine reaches latch-out speed, the governor arm has moved away from the viewer, causing the bypass valve to open, allowing fuel to be pumped through the injector and back to the fuel tank.
The little cam ramp to the original bypass has been removed, keeping this valve closed all the time. If I really don't need it, I will simply replace the spring with a screw that will hold the valve firmly closed.
The engine was run today and, although some tweaking is still needed, it does run better - at least until it gets hot. Since I did so much fiddling with it today, about the time I came in from the shop, it had decided to quit running. I noticed a lot of vapor in the fuel bypass line at the time so, either the injector valve is leaking or the fuel still boils in the injector.
when the engine is running n
Cracked rocker arm
Preparing to cut out the second one. ormally, it fires once for every twelve revolutions and does this relatively steadily. At the same time, a stream of tiny bubbles appear in the bypass line. When the bypass valve opens, larger bubbles are moved to the line along with a volume of fuel. I'm not sure whether these bubbles are due to the injector valve leaking or air getting into the system from somewhere else.
I'll tweak on it some more in the coming days but I think I've just about developed it as far as it will go.
5 September 2012:
I spent the day fiddling with the injector bypass valve. I changed it from a pushrod forcing a ball against the seat to a needle valve. It seems to work better but whenever the engine latches up and the bypass opens, it takes a couple of strokes for the engine to fire once unlatched. Then it fires a couple or three times to catch-up. If I hold the bypass valve closed, the engine runs like a clock, hitting once after unlatching and coasting about 12 revolutions but you can smell the unburned fuel blowing out the port.
If I put a load on the engine so it never gets to governed speed, it will fire every time for an extended period so I'm sure the fuel mixture is correct. Releasing the load so the latch and bypass valve work makes the engine go back to missing a couple of licks then firing two or three times before latching again.
Then, the brass spring that works the cam follower latch broke. I replaced it with a bent piece of music wire and it seems to work fine.
I had to take off the vertical arm that goes to the bypass valve arm because it was getting very close to the fuel pump rocker bolt. It was also verrrrry close to the timing chain when latched-out. I removed metal from the appropriate places and everything's hunky-dory now.
Another thing I did was to remove the spring from the original bypass valve and put a screw in that forces it shut. If, after a couple more runs, I'm satisfied I won't need it, I'll remove the whole assembly and make a pointed plug to go in the hole and seal the valve seat.
Here's a photo of it running today with all modifications so far.
The little engine just a-doin it's thing.
All in all, today the One Valve Engine ran about three hours. The starting procedure is fairly simple. First, air must be bled out of the injector (it just materializes there for some reason I don't understand). This is done by holding the governor weights out which opens the bypass valve. Giving the fuel pump a few strokes soon has the air out of the injector and solid fuel going back to the tank. Then, the governor weights are released and the pump is given a couple more good strokes to force some fuel into the combustion chamber. A couple of flips of the outboard flywheel and off it goes.
At this point, I don't know if there's any way to get it to run right when the bypass is working. As it is, I'm not ashamed to show it.
Rush back HOME
BOY! This is fun!