1/3 Scale Ford 289 Hi-Po

The front mounted distributor is driven from the camshaft through a helical gear set. Specialized equipment is used to manufacture commercial helical gears, but usable one-offs can be made in a home shop using conventional or, if necessary, shop-made gear cutters.

Involute cutters are designed to cut accurate (although not necessarily perfect) tooth profiles square to the face of the cutter. This isn't how helical gear teeth mate, and a mismatch occurs that grows with the angle of the helix. For the same DP, the pitch diameter of a helical gear is greater than the pitch diameter of a spur gear by a factor of 1/cos(A) where A is the helix angle. That is, PD = N/DP for a spur gear, and PD = [1/cos(A)] * N/PD for a helical gear where N is the number of teeth. The remainder of the spur gear equations apply so long as the PD is modified accordingly.

When cutting a helical gear with a conventional involute cutter, common shop practice is to use a cutter designed for a somewhat greater tooth count. The usual multiplier is 1/[cos(A)]^3. For example, to cut a 16 tooth 48DP 45 degree helical gear, a #3 48DP involute gear cutter intended for 45 teeth would be used.

Early modeling which focused on coming up with a faithful reproduction of the engine's front end eventually arrived at a suitable solution based upon a 48 DP involute cutter that I owned. The solution was a pair of 16 tooth 45 degree helical gears with a center-to-center spacing of .470". If a suitable 48 DP solution hadn't been possible, a custom DP button cutter would have been machined as was necessary for the gear set in my Ford Inline six build.

The biggest drawback to machining helical gears in the home-shop has been the need to coordinate the movement of the cutter along and around the longitudinal axis of the gear blank. Chuck Fellows solved this problem for us several years ago when he published his design of a fixture capable of coordinating these moves on a lathe or on a manual mill. In my case I used a four axis setup on my Tormach with the rotary positioned under the spindle at a 45 degree angle.

The form factors of the two gears are not the same (different shaft i.d.'s, lengths, and end features). A number of identical brass and steel blanks were prepared whose ends would be finished after cutting the teeth. The camshaft gear was machined from steel while brass was used for the distributor gear. The gear cutter was mounted in the spindle in a commercial holder, but a custom mandrel was needed to hold the gear blanks while staying out of the way of the gear holder. A shop-made alignment tool was also needed to align the rotary to the cutter.

One of the photos is a screenshot of the scratch-written g-code which is heavily commented with descriptions of the cutting parameters. For anyone interested, my earlier Inline six build on this forum contains more information.

The machining of the very first pair of gears went without a hitch, and a measurement of their spacing matched the target pitch diameter. The next step was to verify their fit inside the block.

With the design of only the lower half of the distributor completed, a dummy lower section was run with the camshaft to test the gear set inside the block. The gears' end features were finish machined, and a reducer was pressed into the brass gear to accommodate the distributor shaft. The gears turned smoothly inside the block with no tight spots and a backlash that appeared to be some 2-3 degrees.

The two tenon'd camshaft sections were then permanently joined together inside the steel drive gear with Loctite 638. Fortunately, I remembered to capture the front ball bearing and its retainer between the two camshaft sections during the assembly. This was needed because both the gear and cam lobes are too big to pass through the inner race of the front ball bearing which otherwise couldn't be installed in the block.

I wasn't expecting the very first pair of gears to wind up inside the engine. A precise measurement of the spacing inside the block had been difficult to make beforehand, and I was prepared to do some tweaking. Before taking down the machining setup the left over blanks were turned into helicals for future use. - Terry

Terry,
Unexpected Left-Overs are for your next engines. Just thinking ahead.
Great work on the engine, yes I am still following.
Cheers
Andrew

Before beginning work on the distributor, a few checks and loose ends involving final assembly of the camshaft and crankshaft sent me down a rabbit hole.

A few thousandths interference between the camshaft hub and the timing cover was discovered and cleared up with a slight counter bore inside the timing cover. This bit of interference wasn't unexpected, but then a much more serious problem showed up between the bottom of the intake manifold and the shoulder bolts in the lifter assemblies. I was able to salvage the lifter assemblies by turning them around so the heads of the shoulder bolts face outward rather than inward toward the manifold. However, the bolt heads didn't clear the block fillets located just above the lifter bores and had to be re-machined into smaller hemispherical heads The bottom edges of the pairing straps also had to be re-worked. The shoulder bolts were finally secured in the lifters with Loctite 248 thread locker. The wax stick eliminated wicking that might have locked the straps to the bolts.

Next, the front crank seal was installed in the timing cover. The cover was originally bored for a 5/8" o.d. lip seal (5/16" shaft diameter), but this i.d./o.d. combination turned out to be something of a rare bird. After awaiting delivery for a few months, the cover was rebored for a more commonly available 3/4" seal, but then problems occurred with its installation. I couldn't get the seal to slip over the shaft in the proper facing direction without damaging it, and after ruining two seals the third one was installed facing backwards. The crank's rear oil seal and its retainer were installed without issue, and the mains were torqued for (hopefully) the last time.

The timing sprockets and roller chain were then installed. The small drive sprocket is keyed to the crankshaft while the larger driven sprocket slips over the camshaft's hub where it's secured with three SHCS's. The sprocket's though-holes are slotted, and the valve timing is set by rotating the camshaft inside it before tightening the screws. With the witness marks on the sprockets and camshaft hub aligned as shown in the photo the number one cylinder is nearly at TDC, and its intake valve is just beginning to open.

Integrated into the alternator's lower mounting bracket is a timing pointer to be used to time the distributor. This bracket and the keyed crank damper were temporarily installed. With the crankshaft rotated to TDC of cylinder number one's power stroke, a temporary timing mark was penciled on the damper adjacent to the timing pointer. The damper was then moved to the mill where the TDC mark and a series of five degree BTDC timing marks were engraved on its periphery.

Work can finally start on the distributor. - Terry

mayhugh1 said:

View attachment 154368View attachment 154369

Click to expand...

Hi Terry !
How do you make them work smoothly ?
Again, Every detail is perfect .

Terry, not sure about the arrangement of your crankshaft that caused problems with the installation of the oil seal, but going back it looks as though you had to get it over a shoulder. If that is correct, and in the unlikely event that you had not thought of it, could you have used a temporary sleeve, butted against the shoulder, having a diameter a trace larger than the shoulder and a tapered or rounded outer end to slip the seal lip over?

Charles Lamont said:

Terry, not sure about the arrangement of your crankshaft that caused problems with the installation of the oil seal, but going back it looks as though you had to get it over a shoulder. If that is correct, and in the unlikely event that you had not thought of it, could you have used a temporary sleeve, butted against the shoulder, having a diameter a trace larger than the shoulder and a tapered or rounded outer end to slip the seal lip over?

Click to expand...

Charles, there wasn't any step, but turning a chamfer on the end of the shaft would probably have helped a lot. Can't help but think the dimensions of the seal were off for some reason. The o.d. measured .820" (I had to actually bore the recess to .780" to press it into the timing cover, and a gage pin measurement showed the i.d. to be .303". - Terry

minh-thanh said:

Hi Terry !
How do you make them work smoothly ?
Again, Every detail is perfect .

Click to expand...

minh-tanh,
Not sure what you're asking, but the lifter bores are .002" larger than the lifter bodies, and one of the holes in each strap is slotted to allow the vertical motion. The shoulder bolts were machined so there's enough clearance for the strap to freely move with the bolts bottomed out. - Terry

The full-size engine's distributor was fully mechanical with extended-dwell dual points. The model's distributor will be electronic and based upon a Hall effect device. Trigger possibilities include eight magnets mounted on a non-ferrous disk or a single stationary magnet mounted behind an eight-hole 'shutter' disk. Either can be made to work well, and I've used both. Closely mounted magnets can become weakened over time due to the resulting opposing fields, but shutter disks require more work. A single magnet trigger was selected for the 289's distributor with the plan being to hide the sensor inside a faux vacuum advance mounted on the front of the distributor.

The components making up the distributor's rotating assembly is shown in the photos. The distributor shaft is supported by a pair of ball bearings inside the distributor's body. An outside third bearing is there to stabilize the lower end of the distributor against the counter torque of the oil pump. Although it's becoming less likely that an oil pump will be used, the third bearing was retained.

Both the distributor's body and the magnet's housing were machined from 6061. A stacked pair of 1/8" neodymium magnets were pressed into the housing which is secured to the body with a pair of stainless SHCS's. The south pole of the magnet pair faces toward the outside of the distributor.

The shutter disk was machined from 12L14. Soft steels are suitable for this application since hard alloys can retain magnetism acquired from the field of a nearby magnet. Hall devices are triggered by flux emanating from the south pole of a magnet, but nearby ferrous material can shunt some of it away requiring the sensor to be located very close to the magnet. The operating gap between the sensor and the magnet can be increased by using a disk with minimum material around its open shutters. If the disk is made too thin however, the closed shutters can become saturated with flux and may appear open. Triggering is affected by a number of complex factors, and optimizing it requires testing and design iteration.

The shutter openings in the disk's outer perimeter were machined using a horizontal rotary to ensure their placement accuracy. Excess disk material was also removed from behind the shutters. The shutter disk is attached to the distributor shaft through a stainless steel bearing spacer using three SHCS's. The shutter disk's mounting holes are slotted to allow fine tuning the shutter locations so the vacuum advance of the finally timed distributor can be kept oriented toward the front of the engine.

The helical drive gear is pinned to the shaft with press-fitted dowel pin. The bearing spacer which sandwiches the shaft between the bearings inside the distributor body is secured to the top of the shaft with a grub screw tightened against a milled flat.

Today's automobiles contain lots of Hall sensors. There's an ever changing long list of parts to choose from since hall part numbers go obsolete pretty quickly. Thanks to the auto industry they're cheap, and I purchased a number of TT Electronics OH090U's a few years ago that I used in my inline six's distributor.

Testing for a reliable running gap was done by running the rotating assembly against one of the OH090U's cabled to a test board. The sensor turned on with a whopping 3/8 inch air gap but had trouble turning off. Excessive flux from the stacked magnet pair appeared to be saturating the shutter disk. A new mount was machined for a single 1/8" magnet. This change reduced the maximum gap to .20" which should be adequate for the planned packaging for the sensor. The sensor didn't appear to have any problems turning off, but its operation will be more thoroughly tested after a few sensors have been packaged.

Still left to do are the hall's packaging and the cap and rotor. - Terry

Terry,
You might find that you have excess dwell time with the size of your shutter window. The diameter of the magnets I use is .125 and my windows are just slightly larger and my 302 will Rev cleanly to 7800 rpm. (Not free rev) I use a Dave Sage ignition module with a universal 6-12 volt motorcycle coil running on 6 volts. It will also run on an S&S but doesn't seem to be quite as clean.

Terry:
PM me if you want a module.

Dave, George,
Thanks for the responses. I'm planning to use one of Roy's magnum CDI modules, and so dwell shouldn't be an issue. I won't be be able to rev to 7800 rpm, but that's ok. I suspect my rpm will ultimately be limited by my intake valves. - Terry

Terry

mayhugh1 said:

Dave, George,
Thanks for the responses. I'm planning to use one of Roy's magnum CDI modules, and so dwell shouldn't be an issue. I won't be be able to rev to 7800 rpm, but that's ok. I suspect my rpm will ultimately be limited by my intake valves. - Terry

Terry

Click to expand...

Terry,
Wonderful work! Do you already have the module from Roy? As I have just bought his last one and he advised that there will be no more .
Regards
Mark

gadabout said:

Terry,
Wonderful work! Do you already have the module from Roy? As I have just bought his last one and he advised that there will be no more .
Regards
Mark

Click to expand...

Mark,
Yes, I have the module. It's a spare that I kept on hand since I've used so many of these in my other engines. That being said, Roy had one last PCB that he's making up for me to replace my spare. - Terry

The Hall sensor was potted (using JB Weld) inside the bracket that attaches the non-functional vacuum advance to the distributor body. A failed sensor would be tough to replace, and so a batch of brackets were made up. There's little room inside them for the sensors, and so Kapton tape was used to cover the soldered joints. Soldering required care and a tiny iron under plenty of magnification, but once I had a technique that seemed to be working I made up a lifetime supply of sensors.

Each potted sensor was tested during its construction using the distributor's rotating assembly and a test board driving an led indicator. The signature of the Hall effect's trigger was captured with an oscilloscope while manually spinning the distributor. Dwell, taken directly from the waveform's duty cycle, is important for a points or transistor type ignition because it's a measurement of the coil's charging time. The scope trace shows the Hall device turning ON (output going LOW) for 15 degrees and turning OFF for 30 degrees of each 45 degree firing cycle. The signature is a complex function of the sensor, the air gap, the geometry of the shutter disk, and any nearby ferrous metal. In this assembly a nearly ideal 30 degrees of dwell would be available with an appropriately designed transistor ignition, but since a CDI will be used, it isn't particularly significant. The ac coupled CDI will be triggered by the rising edge of the waveform, create a very short spark, and then have nearly the whole 45 degrees to recharge its capacitor. The sensor cables were left long and will be trimmed later.

The distributor cap was modeled after Ford's original molded black resin cap. The closest matching material I had on hand was black Delrin which really isn't recommended for high voltage use. Its electrical properties aren't specified by manufacturers because of the carbon black used to color it. Although I don't recommend them, this will be my second black Delrin cap. The cylinder numbers were engraved near their corresponding plug towers as a reminder of the firing order.

Machined brass electrodes were pressed into the towers after the cap's exterior was machined but before its i.d. was turned. The electrodes themselves were drilled and reamed for Molex connector pins that will eventually terminate the plug wires inside the tower boots. An additional .010" hole was drilled through each electrode to allow air to escape during its press-fit.

The plug towers straddle the outer perimeter of the cap's external surface as they do in the full-size distributor, and this created a real Delrin work-holding headache. Total machining time was about four hours that didn't include the first cap that was scrapped because of a .010" concentricity error between its inner and outer diameters. An alignment ring machined into the bottom of the cap fits snugly into a recess in the body to ensure consistent assembly. The second cap wound up with a more acceptable .002" error, but if I were making another one I'd use a different machining sequence. The first step would be to pre-install the electrodes in a carefully prepared blank, and then bore the i.d. using a 4-jaw chuck. After mounting the workpiece on a shop-made expandable mandrel, the cap's outside surfaces would then be machined.

The rotor was machined from white Delrin which doesn't have issues with high voltage. Its brass tip rotates within a couple thousandths of the tower electrodes while making electrical contact with the center tower electrode through a beryllium copper spring. The rotor's tip is extended a bit in the direction of rotation (counterclockwise) to ensure it's under the appropriate tower electrode even when the timing is advanced. The rotor is secured to a flat on the distributor shaft with a grub screw.

This completes work on the distributor except for its high voltage testing. Once the ignition is up and running the distributor will be thoroughly tested using the plug wires and a bank of spark plugs. - Terry

Before starting work on the carburetor, high voltage testing was performed on the distributor. All my distributors are finally checked in a test set made up of one of Roy Sholl's first generation CDI's and a bank of wide-gapped spark plugs.

The plug wires were made up so they could be included in the testing. The distributor ends were terminated in right angle boots, but the wires were left long with stripped-back ends that were wrapped around the terminals of the test spark plugs. Boots for the Viper plugs will be made up later.

The distributor boots are modified Vacu-tite 47408 automotive vacuum fittings purchased from a local automotive parts store. The i.d.'s of the large ends were opened up slightly with a reamer for a better fit over the cap towers. They were also shortened a bit.

Molex connector pins (part number?) were soldered on the ends of the wires inside the distributor boots. The tower electrodes were designed for snug fits of these pins. The plug wires are lengths of .126" diameter 20 kV silicone high voltage stranded wire.

With the distributor temporarily in place on the engine and the high voltage switched off, the rotor was locked to the distributor shaft with its grub screw while sitting directly under tower number one and coincident with the test set's led firing indication of cylinder number one's TDC. After alignment the distributor was removed from the engine and spun by hand in a dark room with the high voltage switched on to look/listen for signs of plug misfires. The distributor was also spun at high speed for a few minutes in a battery powered drill.

An occasional misfire was detected but misdiagnosed as an inconsistent turn-off of the Hall sensor. Thinking the magnet was still too strong, yet another magnet assembly was machined holding a magnet that was the same diameter but half the thickness of the one under test. This magnet assembly (eventually scrapped) turned out to be too weak, and the potted sensor had great difficulty turning it on. The actual problem turned out to be a problem with a connector in the test set. After replacing the connector and re-installing the original magnet assembly there appeared to be no more issues.

The rest of the potted sensors were tested, but out of a total of six, two were scrapped for inconsistent triggering. Even though there wasn't much room for the sensors inside their potted housings, two had evidently cured misaligned. Rechecking both with the original double magnet assembly showed they turned on easily, but still had difficulties turning off.

The cables of the four sensor assemblies that passed testing were terminated in Futaba J male connectors. The connector pins were soldered and the rear ends of the plug shells filled with epoxy to protect the connections from engine fluids.

A typical high voltage waveform was captured by capacitively coupling a scope probe to the wire in the center tower of the distributor cap. The duration of the firing event is extremely short compared with what would be expected from a transistorized or Kettering ignition. CDI's typically compensate for the low energies expected from narrow sparks by generating higher voltages and operating at higher spark currents. Below a few thousand rpm, full-size CDI's also take advantage of the brief capacitor charging times and generate multiple sparks for even higher energy burns. - Terry

Terry, (first putting on my asbestos flame suit ) am puzzled by the choice of CDI, neither D.K.Grimm nor anyone else has been able to come up with any compelling reason or evidence that it works better than LDI, either way you charge up a device (either capacitor or inductor) with a certain amount of energy and then let that go through a coil and some of it appears at the spark gap. So IMHO CDI offers additional complexity with no additional advantage (most even believe that given equal amounts of energy at the start the CDI delivers less to the spark, but my reading of D.K.'s experiments is that that seems un-verified, at least to this date).
Maybe you have some evidence that LDI doesn't work, or maybe you just want to be authentic because the 289 Hi-Po came with CDI, or ...?...

GreenTwin said:

I have heard of CDI systems, but am ignorant of LDI systems, and can't find anything on them on the net.
Can somebody give me the sort story on what LDI ignition systems are?
.

Click to expand...

Just a guess.... but..... L is the standard abbreviation for a inductor in electronics. Rather than a capacitor discharging I am guessing it works on "inductive kick" principle using a inductor.

peterl95124 said:

Terry, (first putting on my asbestos flame suit ) am puzzled by the choice of CDI, neither D.K.Grimm nor anyone else has been able to come up with any compelling reason or evidence that it works better than LDI, either way you charge up a device (either capacitor or inductor) with a certain amount of energy and then let that go through a coil and some of it appears at the spark gap. So IMHO CDI offers additional complexity with no additional advantage (most even believe that given equal amounts of energy at the start the CDI delivers less to the spark, but my reading of D.K.'s experiments is that that seems un-verified, at least to this date).
Maybe you have some evidence that LDI doesn't work, or maybe you just want to be authentic because the 289 Hi-Po came with CDI, or ...?...

Click to expand...

It's a reasonable question and one that I re-ask on every build. The answer is that with a CDI I don't have to deal with hiding a big coil. I've enclosed a photo showing a size comparison between a couple model engine coils and a CDI. I used the Exciter coil in a couple engines along with a TIM transistorized ignition. That particular coil used to be available from Jerry Howell, and although it was a bit larger than I'd liked, my long term testing showed it was robust and reliable. As far as I can tell, even though Jerry was buying them from elsewhere they haven't been available for nearly a decade.

The smaller coil held a lot of promise. I purchased a half dozen of them at a NAMES show several years ago, but my testing showed they were short lived. The four coils I tested died within thirty minutes.

Other possibilities include motorcycle or automotive COP coils, and others are using these and building their models over wood boxes where they can be hidden. I'm just not a wood box guy though. I'd much rather machine a reasonable facsimile of a scaled down MSD housing where I can hide a CDI. - Terry

Terry:
I'm with you.
Even though I make my own coil drivers I too don't like the large (COP) coils and the necessity to hide them. So I've been using Roy's CDI modules (originals as you show). But alas, they are not available any more. I guess it's about time to design my own. I have a few ideas. I just have to find the motivation to follow through.

If you do that Dave, see if you can use the miniature coils Roy used to use. The physical size is great if trying to hide it and the little buggers will jump a half inch gap.