Engine Number Four

Part Two

The Electronic Governor


14 February 2011:

The first thing I needed to do was to see if I could use the head position voice coil assembly out of an old hard drive to work the throttle on the mixer.

Here's the setup for testing the head positioner.

I found that with drive current of about 1/2 amp, I could make the positioner move full travel against a spring that is strong enough to pull the throttle securely closed.


15 February 2011:

When I went out to the shop this morning, I found that the water jacket "O" ring was seeping a little coolant.  This was caused by the short hoses to the cooling tank putting a strain on the jacket, unseating the bottom "O" ring.  The solution was to pull the head and water jacket and make an 0.020 shim that fits into the groove in the lower end where the "O" ring seats.  This adds to the compression of the "O" rings.  After running the engine for about an hour, there was no leakage.

I've decided to use an optical interrupter to sense crankshaft speed.

Tachometer sensor.

I made an interrupter wheel out of a piece of 0.031" printed circuit substrate.  In order to make it opaque to the sensor, I painted it with Dykem (layout blue).  The interrupter uses an LED as an infrared source and a phototransistor as the sensor.

The phototransistor signal will be squared-up using a Schmitt Trigger integrated circuit.  The same I.C., since it contains six of the circuits, will also be used to generate a fixed length pulse that will be fed into a charge pump to generate the analog voltage from the pulses.  A charge pump gives a rising analog voltage when the pulse rate increases.  Once I get the charge pump done, I can work on the operational amplifier and voice coil drive circuit.


16 February 2011:

Spent most of the day fiddling with the circuit for the tachometer and governor.  I went through about four iterations, ending up with the following sketch.

Rough schematic of circuit.

I'm using 1/2 of a 14538B C-MOS dual one-shot to take the signal from the tachometer sensor and turn it into a fixed width pulse.  The pulse is steered two ways via the diodes.  One of the diodes feeds through a resistor and calibration pot to a 100 micro amp meter.  I will make a scale for it from zero RPM to 2,000 RPM.  I won't finish the meter until I've got the governor on the engine and can see how the engine will do at 2,000 RPM.

The other diode feeds the pulse to an R-C network that makes a varying DC voltage out of the pulses.  The voltage is higher the closer the pulses are together.  This varying voltage is fed to an operational amplifier ("op amp") used as a sloppy comparator.  The non-inverting input is varied to control the point at which the op amp output changes from low to high.

Since the op amp output does not reach the supply rails, a transistor is used to add a little more gain and to offset the output of the op amp.  This transistor collector will drive a power transistor which will, in turn drive the voice coil throttle actuator.

I use a 7805 5 volt regulator to keep the circuit stable.

Breadboard under test.

Above, you can see the typical "rat's nest" bread boarding setup.  Once the circuit is far enough along to test, it will be moved to the engine and tried, probably necessitating adjustments to the op amp gain and damping factor.


19 February 2011:

I (semi) finished the circuit and started the testing and optimization.


Electronic governor under test.                                  Throttle actuator made from hard disk head positioner.  

The throttle actuator is driven by a 2N2955 PNP power transistor that I added today.  After hooking everything up, the engine was started and immediately showed speed instability due to the governor circuit over sensitivity and under damping.  In my experience, most servos need a bit of tweaking to get them right and the initial design was a guess at starting parameters.

Several component changes were made and the circuit seems to work better but I did find out that my hoped-for 2,000 RPM top speed was a little on the high side.  I think a better top end will be around 1,500 RPM and that will be straining the intake valve a bit.  Also, I thought I could get it to idle down to about 200 RPM but again was a little on the optimistic side.  A more attainable idle will be on the order of 400 RPM.  It could be a little lower once carburetion and timing issues are worked out.

I'll modify the tachometer to show a top speed of 1,500 RPM and will change the pulse width of the charge pump to get full range of output within the new RPM range.

The carburetion issue is in getting the flutter choke and mixture needle optimized.  This may take a bit of redesign of the mixer.


20 February 2011:

Last night, I had a brainstorm concerning the instability of the governor.  I hadn't put an oscilloscope on the output of the optical interrupter acting as the tachometer sensor.  When I did so, I found that the signal going to the one-shot was not going rail-to-rail as it should.  The problem was traced to poor geometry between the tach wheel and the interrupter.  A bit of Dremel tool whittling on the sensor mount and turning off about 1/16" from the diameter of the tach wheel fixed the problem.

I went ahead and changed the tachometer meter scale to 1,500 RPM full scale and tweaked the circuit to accommodate the lower speed range.  When I mounted it up and started the engine (by the way, I can now start it by flipping the flywheel), it was much more stable.  After fiddling with the circuit, I think the remaining instability is caused by a mixer problem.  Tweaking on the flutter choke and fuel needle helped but I will still have to address the issue.

I ran the engine for a couple of hours today and, with the governor, I can load it up using the rag-on-the-flywheel-rim dynamometer.  I can really cook the rag with this engine.  It also acts like a real engine in that, when the load is quickly removed and the governor snaps the throttle shut, it backfires ....... Whoopee!  Also discovered that when the engine is run over 1,500 RPM, it blows oil out of the breather, probably caused by a bit too much oil in the crankcase.  Better, though to have a little sloppiness outside than a lot of damage inside!

I suppose the next thing on the agenda should be neatening up the throttle actuator then it's on to the automatic spark advance.  In running the engine, I think it needs quite a wide range of ignition timing.  It appears that it will start and run pretty well with ignition occurring about 15 degrees BTDC.  It will probably run at speed (1,500 RPM or so) with about another 10 degrees or so of lead because it labors a bit as it is now.  For easier starting and good idle, I need ignition to occur closer to TDC, so about 25 to 30 degrees lead after about 750 RPM should be in the ballpark.


22 February 2011:

Today, after wasting about three hours fixin' my el-cheapo band saw (a freebie), I made the housing for the throttle actuator.


Box laid-out.                                               Box soldered together.                                Plexiglass lid for box.

I had to make an enclosure for the hard drive head position actuator (now the throttle actuator) so the magnets in it didn't pick up any more filings and metallic junk.  The housing is made from some copper clad printed circuit board material I have laying around.  After cutting the parts, it was assembled by soldering the parts together using the copper.  Then, I made a plexiglass cover so I can watch it work.

Actuator mounted on mixer and on engine.


23 February 2011:

The newly configured governor is working well.  If I set it at 1,050 RPM with no load other than "wind milling" the permanent magnet DC motor, it only drops about 25 RPM with a 260 watt load.


Engine running with 260 watt load.

Since it takes about 760 watts to make an electrical horsepower, or about 0.34 HP.  Considering the approximate efficiency of the motor/generator at around 50%, the engine seems to be putting out about 0.68 HP.  While the load test was going on (about 30 minutes), I noticed that the throttle was only about 1/4 open so I think that, with a good dyno test, it may make as much as two horsepower.

Also, while the test was going on, I put a black paint spot on the head and used my infrared thermometer (Harbor Freight) to find that the temperature of the engine was running about 185 degrees.  Not bad, even if I do say so myself!

So far, total running time of the engine is about four hours.  No oil used but I do notice a bit of blue smoke when blipping the throttle after a couple of minutes of idling.  This may be because I removed the expander spring from under he oil ring to improve oiling.  If that smoke I saw today is as bad as it gets, it will do. 

The next thing I need to do is work on the automatic advance circuit.  To get the engine to run over 1,000 RPM without laboring, I've had to advance the spark to the point that, if cranked relatively slowly, it will kick back.  I'm going to have to put the degree tape back on and find out exactly how much advance is on it now.

I'm planning on posting one last video when I've got everything finished.  That could be a while.


10 March 2011:

I got back to the project today.  I got the Hall-Effect transistor timing positioner made-up and mounted plus, I got what I think is the final automatic spark advance circuit drawn and breadboard tested.


        Actuator assembly.                                                              Actuator assembly on engine.

I'm not sure the timing actuator is in it's final form yet.  I will go ahead and test it but I think it may be necessary to make a housing for it like the one for the throttle actuator.

Presently, I have it set for 10 degrees ATDC for retard and 30 degrees BTDC for advanced.  I can adjust the total swing of the actuator to modify the amount of timing change required.  You can see the screw (the one that also holds the return spring) in the photos.  It contacts the little bumper on the back of the actuator.

I had to remove the flywheel to mount the timing actuator and, when I did, I found that the flywheel had spun about a quarter turn.  I think this happened the last time I ran the engine and the ignition timing went whacko causing the engine to backfire hard at speed.  This time, I put the flywheel on with a little help from my friend Locktite.


11 March 2011:

Did some tweaking on the circuit then clamped the breadboard (a.k.a. "rat's nest") circuit to the engine table.  After several more tweaks, I had it running with the automatic spark advance.  After adjusting the timing to a point where it runs nicely, I shut it down and checked to see where the timing is.

WOW!  The retarded timing is about the same as I originally set it, about 8 degrees ATDC but the advanced timing is really early for this engine.  The timing switches over to advanced at about 750 RPM and it is right on 50 degrees BTDC!  The engine runs well at 1,500 RPM, the speed I've got the tach set to top-out at, but it wants to run faster.  I will set the limit on the governor to about 1,600 RPM because I don't really want to spin that flywheel any faster.  

Also, when it is running at those speeds, it blows some oil droplets out of the breather.  At least, I know it's oiling well.

Next, I will accurately schematic the breadboard with it's modifications, put the circuit into CAD then build the working circuit, semi-ready for prime time.


15 March 2011:

I've been working on the final version of the circuit and building the wire wrap board.


Wire wrap board in progress.

Since the circuit is kinda complex for building a soldered-wire final version, I decided to use up some of my old wire wrap stuff.  One thing I now remember about that particular brand of wire wrap is that sometimes and for no real reason, the wire doesn't make contact with the posts despite 12 to 15 turns.  I verified every node with a 50 milliamp test light and found three nodes that had bad connections, causing me to have to rip them out and do them over.  I think I've got all the paths good now.

The next thing to do is to run the electrical tests of the actual circuit and make corrections as necessary.  After it's working well, I'll figure out an enclosure, clean up all the sensor and actuator cabling and mount it on the engine stand.

Once the engine's running okay on the final circuits, I'll make another flick and post it to YouTube.


18 March 2011:

Well, today I finished the electronics and, although they need some tweaks, everything works except the ignition.  That part is problematic.   Mixing logic and power circuits on one board is fraught with unintended consequences, including power transients caused by switching the primary of the ignition coil.

When I got everything else working all right, I wired up the coil and hooked everything up to the engine.  When I turned the power on, everything was fine until the first time the coil was supposed to fire.  What happened was there was a power glitch.  Two transistors and the 5 volt regulator were cooked, letting out all the magic smoke.  Rather than go through the process of chasing down the culprit, I simply disabled that part of the circuit, replaced the toasted components and used one of my solid-state ignition modules.

Showing the modified cooling tank and the controller.

I took the engine outside and shot some video but don't know if I want to publish it on YouTube yet.  I've got to look at the raw footage and see if I can edit it into something coherent.  Also, the governor wanted to hunt and I think I need to stiffen-up the intake valve spring.  It did easily pull 360 watts worth of light bulbs, probably close to a horsepower.

While shooting video, it started running flakey with the governor hunting really bad.  I believe I may have to blow out the fuel line and get a new spark plug.

Stay tuned.  If I get decent video, I'll upload it soon and leave a link here.

AND here it is!


22 March 2011:

I've gotten around to formalizing the schematic of the controls for Engine Number Four.

Control schematic for Engine Number Four.

If you want to follow my description of the circuit, you may want to click on the image above then copy the Jpeg file to your printer or print the page.



On the upper left-hand side of the page is a common 5 volt regulator to provide the logic with clean power.  Note that there are two voltages used in the circuit.  Where a wireline does not connect them to their respective sources, the voltage will be shown inside the Vcc bubbles.


I think the best bet is to start on the lower left-hand side of the schematic where you find the optical isolator for the tachometer.  The wheel that it senses has eight holes, giving 8 pulses per revolution of the crankshaft.  R3 and C5 are a low pass filter that eliminates any glitches in the tach signal.  

The filtered tach signal goes up to the non-inverting input of one-half of the MC14538 re-triggerable multivibrator.  C2 and R4 set the pulse duration of the one-shot.  The output pulse string is fed to three diodes.  D2 goes through R5 and R6 (RPM calibration pot) to a 100 microamp meter that is scaled 0 to 1500 RPM.  This circuit comprises a run-of-the-mill charge pump tachometer.

For your information, the tach signal ended up being a nice round 20 pulses/second for each 150 RPM.  20 pulses/second = 150 RPM and 200 pulses/second = 1500 RPM.


Okay, back to D1.  D1 is also part of a charge pump which turns the pulses into a DC voltage that is dependent on the number of pulses per second that is presented to it.  R7, C3 and R8 comprise the rest of the filter.  The varying DC voltage is fed to the inverting input of the op-amp

(A 1458 dual op-amp) through R9 which, in combination with R14, determine the gain of the op-amp.  The op-amp gain is very high, causing it to work somewhat like a poor comparator.  

The non-inverting input of the op-amp receives it's signal from the potentiometer R11 which is part of a voltage divider consisting of R10, R11 and R12.  As the RPM rises, the voltage at the inverting input of the op-amp rises.  When the voltage on the inverting input rises to slightly above that of the non-inverting input, the output of the op-amp goes from a high state to a low state.

The op-amp output goes through the R15, R16 network and NPN transistor Q1 to invert the op-amp output and to level shift the signal from the +5 volt line to the +12 volt line.  The collector of Q1 goes through R18 to the base of PNP power transistor Q2.  The emitter of Q2 has as it's load, the hard disk head positioner coil which operates the throttle.  When the engine is running slower than the speed control is set, Q2 is turned on, opening the throttle.  The converse is true.

I plan on making some modifications to the throttle servo circuit because it seems to be overly sensitive, making the engine run erratically.  If carburetion and a poor spark plug aren't causing the instability, I will lower the gain of the op-amp by decreasing the value of the feedback resistor, R14, and, maybe, put a very small value of capacitance across the feedback resistor to slow the response.


Following the line through D3 takes us to the timing control.  R19, C6 and R20 are another charge pump filter that feeds the non-inverting input of the other half of the op-amp.  R25 and R26 pull the non-inverting input of the op-amp to about 2.5 volts.  The inverting input of the op-amp is fed by the output of another voltage divider that sets the RPM point where the op-amp (connected more like a true voltage comparator) output changes from a low state to a high state.  The output is fed through D4 to take advantage of the bandgap of 0.6 volts, thence through R27 and R28 to shift the turn-on voltage of Q3 which uses the hard disk head positioner coil as a load.  This hard disk head positioner swings back and forth to change the angle at which the Hall-Effect transistor ignition sensor switches.on.


This part still has problems but I will show it in it's present imperfect form.

The Hall-Effect ignition trigger open collector output is pulled up by R35.  R30 and C1 form a low-pass filter to eliminate any glitches in the cable.  The filtered trigger signal goes to the inverting input of the second half of the  MC14538 re-triggerable multivibrator.  Since the collector of the Hall Effect goes low when the camshaft mounted magnet approaches, a 1 millisecond negative-going pulse is generated at this time.  R32 and R36 feed the base of NPN transistor Q4 which, having it's emitter connected to the +12 volt line, acts as both a level shifter and a driver for MOSFET transistor Q5 which drives the ignition coil primary.  C11 isn't really needed for a hot spark because the MOSFET turns off (generating the spark) much faster than the magnetic field of the coil core can collapse.  I put the capacitor C11 in the circuit to try to eliminate the instability that was present in the circuit.  

As of right now, I'm using a stand-alone solid-state ignition module until I get the built-in ignition circuit sorted out.

Anyhow - that's the story up to now.  I have some other things that need doing so it may be a few days before I get back to Engine Number Four.  To find out if I've posted anything new, go to my home page and look beneath the photo of the project, noting the date of it's last update.


26 March 2011:

Today, I worked on the ignition and found several problems.  One had to do with letting the "Magic Smoke" out of the ignition coil when the output transistor shorted.

Gee, the smoke all leaked out!

It doesn't take long to build up a lot of heat with 12 volts across a 3 Ohm coil.  This is the first of these "mystery" coils I've destroyed and now I'm down to my last one.

What I think happened was, while I was corralling the glitches that were plaguing the circuit, and while the power was turned on, the transistor failed.  In the ignition circuit, the coil's only supposed to be energized for one millisecond (one thousandth of a second) so, after a few seconds, the coil turned into a heater.

After replacing the transistor, I found that I really didn't need the 5 volt regulator so I took it out and the circuit (except for the ignition coil itself) is now all powered by filtered 12 volts.  If anybody's interested, I'll publish the revised schematic here.

Anyway, after putting a 4 amp light bulb in series with the coil to limit the current to a non-fry value, I had the ignition working fine.  Took the engine outside and ran it for about two hours.  At near the end of the last tank of fuel, I decided to belt the motor/generator back up and run it with a 360 watt load.  It took the load fine but after about a minute, guess what.  The ignition failed again, probably due to the added stress of firing the plug under a load.

Oh, well - back to the ol' 'scope.  Prolly somethin' simple.


3 April 2011:

Well, I did find out why the ignition failed last time.  Because it ran fine with no changes after the engine had a chance to cool-off, I figured the problem was too much clearance between the magnet and the Hall Effect sensor.  I've found that they lose some sensitivity as the chip temperature rises.  Reducing the clearance to around 0.015" has solved the problem completely and in the last few days, I've probably put another two or three hours on the engine.

On 31 March, I decided that the engine exhaust was too loud with the little muffler on it so I drew-up a better one and started on it.

The sheet metal is some left-over 0.016" steel roofing material and the rest of the parts are out of the junk.  Overall, the muffler is about 13 inches long and about three inches in diameter.

First attempt at an end cap. 

Since I ain't no freakin' sheet metal guy, the first attempt at making the end caps was a notable flop.  I didn't have enough clearance (0.015") between the punch (left side, left photo) and die (right side, left photo).  There ended-up not being enough space for the steel to go, it sheared  at the die line as you can see.

Successful making of end caps and fitting them into the muffler body.

I increased the clearance to 0.030" and the punch and die worked fine.  I intentionally cut the blanks larger than I needed because I didn't know exactly how deep I could make the caps.  I quit pressing when the caps were about 1/4" deep and before they sheared again.  The ugly stuff sticking up after pressing was turned off with the lathe.

The body of the muffler was cut out of the same roofing steel and a finger joint (or whatever it's called) was made so the caps would fit tightly inside.  On one cap, the fit was good.  On the other one, I had to fight it for a couple of hours and finally used the press to coax it into the body.  I don't trust myself welding material that thin so I used a bunch of small sheet metal screws to secure them to the body.

There was no calculation done on the volume of the muffler body but I did calculate the area of the exhaust pipe.  Then I selected a drill bit that would make a hole about one 22nd of the area of the inside diameter of the exhaust pipe.  Inside the muffler at the inlet and outlet are 5 inch long pieces of thin-wall tubing with a cap welded over one end.  I random-space drilled 24 holes with the selected drill in each of the lengths of tubing.  The tubing is welded to the inside ends of the inlet and outlet 1/2" NPT pipes. 

I punched clearance holes for the 1/2 NPT inlet and outlet pipes and threaded bushings on either side of the caps as you can see in the left-hand photo below.


Finished muffler (sans paint) on the engine.

It took a lot of fiddling to get the muffler assembled and, all the while, I was thinking that it would probably be a really poor silencer.  Boy, was I wrong!  When I tested it on the engine, I couldn't believe how much of the bark had been removed from the exhaust.  It's now just a gentle thumping sound and the loudest sound the engine makes now is the snorting of the intake valve and flutter choke.  The next loudest sound is the breather valve opening and closing.

The muffler is a bit on the flimsy side and shakes a bit when the engine is running so, to keep the inlet end cap from fatiguing, I took a couple of old wire coat hangers and made angle braces from the top of the muffler to the cooling tank mount (not shown).

Wow!  Almost ready for Prime Time!


4 April 2011:

I painted the muffler yesterday and today, I ran the engine to cure the paint.  While it was running with the generator powering the lights, I thought, "Hmmm, the light bulbs brighten a lot when the engine fires, even at 1,000 RPM.  Maybe I ought to take the suggestion of my Aussie friend, Denis Basson, and increase the mass of the flywheel."

Rooting in the steel pile, I found a 14 inch circle of 1/2 inch thick steel plate.  This is about two inches bigger than my mill can handle so I did some whittling.

Band-sawing diameter.                                               Finished!                                            Ready to bolt on and turn.

There are some boring aspects of these projects.  One of them is sawing 1/2" steel!  It took about two hours to get it done.

I then drilled and bored a hole part way through the center to press-in a pin that mates with the bore of the iron flywheel.  This will insure that the additional wheel is centered.  I also drilled two 3/8 inch holes (not shown) in the additional wheel that screw into the puller holes in the original wheel.

My plan is to mount the sandwiched wheels in the mill spindle using the 3/4 inch pilot you see in the original wheel.  I will mount a large carbide cutter in the mill vise, then turn the O.D. of the additional wheel to match that of the original wheel.


5 April 2011:

The flywheel addition is finished and tested.

Turning O.D. of the flywheel addition.

I pressed a mandrel into the bore of the existing flywheel then, after drilling the mounting holes in the addition, bolted it to the existing wheel.  It was than mounted in the mill using a 3/4" collet.

I used one of the big carbide tool bits mounted in the mill vise to turn the O.D. of the addition.  I locked the quill and raised and lowered the knee to feed the tool.

After removing the wheel from the mill, I witness marked the addition and separated it from the flywheel.  The pilot was pressed out of the flywheel and it was then mounted back on the engine.

To make the addition look nice, I used my belt sander to semi-polish the steel.  Mounted back on the iron flywheel, it was ready to go.

Finished product.

After starting the engine, the first thing I noticed was that it idles much more smoothly.  There's still an issue with the fit of the throttle plate in the mixer causing it to want to idle fast but, once the ignition  timing goes to retard, it idles nicely at about 450 RPM.  After the throttle plate wears-in some, I think the engine will idle down to around 300 RPM.

The only thing that bothers me is that there is a slight wobble in the flywheel.  I think it is caused by a slightly loose fit of the crankshaft in the flywheel bore.  Since I used Loktite to set the wheel on the shaft, I'm not worried about it coming off - it's just frustrating to have it wobble.

Also, the governor is a lot more stable with only the occasional throttle blip, caused by some issue with the mixer.  All in all, it was worth the trouble.


11 April 2011:

I've been fiddling with the controls and now have the engine running pretty nicely.  I could be convinced to post another video of it, running at about optimum.  We'll see.  Since the run time was less than an hour with the original gas tank, I made another one.

The new gas tank.

The new tank is 3" high by 4" deep by 6" long inside.  This gives exactly 1.246753 quarts (1.179869 Liter) more or less.  I can now run the engine 'til I get bored listening to it run without it running out of gas.


In it's near final form (30 April 2011).

Here it is shown running with a 240 watt load at 900 RPM.  Not even breathing hard.  If you click on the photo then click on the image, you will see an enlarged view and may be able to see that the throttle is just cracked.

Here's The Final Movie


04 October 2011:

I decided to revisit the project and make it so I could lift it into the car trunk or trailer to take it to shows.

First thing I did was take off the generator and lights to save weight.  Then the cooling tank went back on the shelf and I used what scraps of copper I had to make a radiator.  Since I wanted the radiator short, I cut two pieces of the old air conditioner coil to make a two row radiator.

Engine Number Four with radiator, ready for showing.

I had a plastic fan blade from a dead microwave oven so I used some scraps and a couple of little ball bearings to make the fan shaft.  A piece of PVC was turned into one pulley and a piece of laminate flooring was made into the lower pulley.  An "O" ring from somewhere is the belt.

I tested the setup and found that, running the engine at 1200 RPM long enough for the water temperature to stabilize, it was running about 185 degrees (F).  Under load, I think it will probably boil a little but I'm not planning on any heavy work for this one.

Testing came to a halt when the governor started acting wonky.  Some detective work found the throttle butterfly was loose on the shaft so that was fixed.


24 August 2014:

After running the engine for about three years, it developed a bit of runout in the flywheel.  After I added the washing machine motor/generator to it, the wobble got a lot worse.  Today, I disassembled the engine (the WHOLE THING had to be taken apart!).

Crankshaft, right end is bent.

If you look carefully at the photo above, you will also see that the setscrews have slipped and buggered up the shaft.  I will try to repair the crank by removing the bent end of the shaft and replacing it with a steel shaft that is larger in diameter and has a keyway.  Because I'm increasing the diameter of the shaft, I rooted in my bearings and found a metric bearing with an I.D. of 25mm or 0.985".  


         Shaft sawed off.                                            Boring crank cheek for new shaft.

When the end of the shaft was sawed-off, a stub was retained to use to get that end of the shaft centered.  After carefully centering the shaft, the cheek was drilled all the way through to 1/4".  Then it was bored out to 0.874" down to about 0.150" from breaking through between the cheeks.

Parts ready to assemble.

A new shaft end was made of some 1" diameter shafting.  One end was turned to 0.875" to be pressed into the crank cheek.  The other end was turned to 0.985" and a keyway was milled.


        Assembled crankshaft.                                                                 Note bolt and suspenders bolt.

The new shaft was coated with bearing-set Loktite and pressed into the cheek then the bearing was pressed on with a light fit.  After all that, a 1/4-28 bolt was Loktited into the end of the shaft as insurance against it working out.  The other end of the shaft will be drilled and tapped 5/16-18 for a bolt to hold the flywheel on.

Yet to do is boring the side plate and retainer for the larger bearing then reassembling the engine.

After that, the flywheel will be reworked to accomodate the new shaft.  At that point, I will be ready for either re-mounting the engine on the test stand or moving it to the butt-buggy it's intended for.


30 August 2014:

Over the past few days, I've been making some improvements to other parts of the engine.  

New cam on right.

In the photo, what looks like a defect is actually a slight indentation that was the chomp mark of a dull blade in the steelyard's shear.  It is only a few thousandths deep and does not go into the track of the cam follower.  The reason it's there is that's the only piece of 3/8" steel I had.

I've increased the exhaust duration from about 180 degrees to 220 degrees.  It opens the exhaust valve at 145 degrees past TDC on the power stroke and closes it at 5 degrees past TDC on the intake stroke.  This ought to make the engine breathe easier and, since it will not be required to idle really slow (the speed range is between about 450 RPM and 1,000 RPM), it will make better power.

I'm also re-working the cast iton flywheel hub to make it stronger and less prone to flexing.  This is taking some time because there's a lot of turning and boring required.


31 August 2014:

Actually, the flywheel is cast iron.  The hub is steel.  It is made from two pieces of steel.  That is because the hunk of steel I had was about 0.200" too thin to do the job.  I made an insert for the small end that is pressed into place.

Flywheel with new nub in place.

In the photo, you can see the partially completed broach bushing I've made.  The slot needs to be milled.  It seems that every time I have a keyway to broach, there isn't a bushing in the broach set that will work so I have to make another one.


3 September 2014:

Yesterday, I finished the broach bushing and broached the keyway in the flywheel.  When I went to mount it on the engine, I discovered that I'd forgotten a small thing.  The new hub on the flywheel interfered with the cam. Since the hub was pressed onto the wheel and my lathe is not nearly big enough to chuck the flywheel to turn the hub down, I had to resort to the 'ol rotary table in the mill trick.

Shrinking the flywheel hub to clear the cam.

I spent better than half the day getting the hub down to size and it now clears the cam and main bearing cover screws.

 The new flywheel hub.

The flywheel hub is pretty much a press fit onto the crankshaft.  With the key and the end bolt holding it on, I don't think it's going to slip!  The washer for the flywheel bolt is stepped for a slip fit into the hub.  The O.D. of the washer is a slip fit with the disc to center it.  The rotor shaft has a short pilot that fits into the small diameter hole in the disc and the whole thing clamps together with a couple of 5/16" cap screws.


The flywheel, mounting bolt and the step washer to locate the disc.                                            Locating hole in disc.                                 

The next part was changing the dimensions of the stator mounting frame to bring it closer to the flywheel.  This is to make the whole generator stiffer.

After all that, it was time to indicate the stator hub and stator.  To do that, I had to make a 10" long rod for my indicator and a plate to bolt to the end of the rotor shaft for the magnetic base of the indicator.


Indicator mount.                                                     Centering stator hub.                                                            Centering stator.

Centering the stator hub and stator was time consuming and I almost got a crick in my neck checking the dial.  When finished, I had the hub within 0.001" of dead center.  I quit on the stator when I got it to within 0.002".  I think the poles of the stator aren't all on exactly the same circle and that makes it really difficult to bump it to center..

The valves are re-lapped and back in and the head and water jacket are back on.

I've had some thoughts on the controller and will consider doing some modifications to see if I can get the speed/Voltage governor to work better.  At idle, the Voltage is about 35 Volts.  Under median load, the Voltage is about 24 and under full load, the Voltage is about 20.  I think I can change some components in the circuitry to get it to regulate better.  I know, though, that The Law of Unintended Consequences will probably take effect.

Another modification I think I'll incorporate will be a switch that forces the throttle to go to full idle stop and force the timing to go to retard.  That's another thing that might be fraught with demons.

Tomorrow, I should have the engine running again.


4 September 2014:

Truth in advertising and all that.  The engien is running again and here is the promised video.

Here's the flick of the engine starting, running and generating.

The increased cam angle (from a duration of 180 degrees to a duration of 220 degrees) doesn't seem to change the operating characteristics of the engine much.  It may idle a bit more roughly but that could just be the governor hunting.  At full load, the engine still has some throttle left so I consider the combination of engine and alternator to be workable.

One thing of note is that the motor/alternator has a sound all it's own that you may be able to hear in parts of the video.  I think it comes from the cogging of he magnets and armature poles.

I ran the engine for a total of about two hours.  While under full load, the temperature of the coolant in the top tank of the radiator peaked at 210 degrees.  After reducing the load to about 1/4th of full load, the temperature slowly fell to about 185 degrees, so I think it has an adequate amount of cooling.

In a few weeks, I'll make up my mind to either modify the 30-60 tractor (I still have the electric drive motor, motor frame and sprockets) or use the little scooter axle and build a totally new butt-buggy.


8 September 2014:

Since I've finished the engine modifications and most of the testing, I'm moving the narrative to The Washing Machine Motor Project page, where it will be applied to a butt-bugy.


 [email protected]