Rebuilding the Q-ship; a 1964 Harley Davidson Sportster

[chuckthebeatertruck]

Renewing a S&S L Series Carburetor

One of the "trick" pieces on Gladys is an original GBL carburetor. These carbs were made from approximately 1972-76 and represent the second series of S&S carburetors (L = late; G = gas, B = 1-3/4 throat). There are also early carbs (no L) and later carbs (M GBL or M GAL - M = modified). The easiest way to think of these carbs is as a giant Linkert DC with some modifications and improvements.


What's cool about these carbs is that they work really well and are way less fussy than a DC Linkert. The GBL can be used on stock displacement bikes right up to the mid 70 inch mark. There are so few moving parts it is not even funny, and no rubber/plastic to wear out.

IMG_1786.jpg

BUT, and there's always a but . . . most of these carbs have suffered either really poor repairs or are just worn out after 40+ years. The wear is almost always in the alloy body of the carburetor at the throttle shaft. However, contrary to what many believe, the issue caused by a sloppy throttle shaft is NOT unmetered air. Instead, the throttle disc looses it's precise location in relation to the air bleeds in the roof of the carb. Often you will find bushings installed in these carbs using a "repair" kit -- and often the bushings are installed a few thousands the wrong way . . . and the carbs run poorly.

Add in worn throttle discs and the result is that the L series has gotten a reputation lately for being a carb that only works at WOT and is only purchased by hipsters that don't know better.

We say "boo-hoo" to both statements. L series carbs when properly repaired are LOVELY and work very well. So, how do you properly repair a carb for which no parts have been available for a couple of decades? Answer -- you make stuff.

In this case, my GBL became the prototype for making new throttle discs and fixtures for relocating the throttle shaft with precision.

Here's how Dr. Dick did it.

First step was to pull apart a bunch of GBLs that the Doc had on hand and take careful note of the throttle discs. With that done, a fixture was made up to allow the Doc to cut down readily available GAL thorttle discs (1-7/8 bore) to fit the GBL (1-3/4 bore)

image002.jpg

Next step was to fixture the carburetor and get both ends perfectly parallel for installation in the mill vice

image005.jpg

With the carb in the mill - the exact location in relation to the idle bleeds is found

image008.jpg

This carb had one axis correct on the throttle shaft bushing "repair," but the other was out by .040 and the throttle disc filed to suit! No wonder it was a MOFO.

image010.jpg

[chuckthebeatertruck]

Renewing an L Series Carb part 2

With the location of the centers established, a new hole was bored in the bushings

image012.jpg


A steel sleeve was turned to the OD of the new bore and the ID of the throttle shaft

image018.jpg

After being pressed in; the ID of the sleeve closed up a bit and the final diameter was set by honing to size

image019.jpg

Using ground carbide pins, the location was verified by gauging. In this case, the throttle shaft is represented by a .250 pin. A .750 pin (1/2 of bore diameter minus shaft diameter) is a snug fit and a .751 won't go -- indicating we have precise vertical alignment within .0005 -- and that's a lot better than off by .040 :-0

image021.jpg


The intermediate circuits were then checked, float set at 5/8, a 76 main jet installed, and the whole thing reassembled.


To say there is a "slight" difference in the before/after performance is an understatement. The bike did not want to start readily before the carb renewing -- now it lights clean off and the throttle response whilst riding is FANTASTIC.


So, before you give up on your L Series carb -- make sure someone hasn't screwed it up. if they have, and you're unable to tackle it yourself, Private Message me and I will put you in contact with the Doc to discuss a rebuild of your carb. I too will eventually offer this renewal service, but have to get a few under my belt and consult extensively with the Doc before I'm comfortable doing so.

[chuckthebeatertruck]

Summing it all up

As stated in the first posts on this thread; our goal was to show you what it takes to rebuild a basket case to a high standard and what goes into an old "hot rod" bike.

In total, this bike took a little over 270 hours from start to finish -- not including hours spent hunting parts, talking with vendors, and keeping organized.

We did miss our $10,000 budget. At the end of the day, our bills totaled up to $10,700. That's the total cost including tax, title, tags, and first year of insurance, oil and fuel for break in miles. Not too shabby for what amounts to a brand new antique motorcycle.

If we didn't hot rod this thing - the total bill would have been about $3,000-3500 less and more reasonable for a lot of enthusiasts than $10,000.

As for timing -- our goal was to finish the bike and share it with Old Sportster and K model Research Group (OKSRG) at the 2020 Davenport show. We are proud to say that instead of showing off Gladys at Davenport this weekend -- we're out riding her around the prairie.

We hope you enjoyed this series; learned a couple of new things; and are inspired to tackle your own goal/project.

Happy Riding!

[chuckthebeatertruck]

And, we have one more installment for this series on valve stem protrusion and valve train geometry. It's taken Dr. Dick and me about six months to write the article and bounce it back and forth for editing. We're almost done -- honest :-)

It's a long article . . . about 12 pages so far!

[chuckthebeatertruck]

An Electric Foot (Roller Starter) to help light the fire

After nearly 30 years of kicking over hard to kick bikes, I decided it was time to find an easier way. Adding electric starters is one option for many bikes, but not for tin cover, magneto fired sportsters.

Enter the roller starter (aka as paddock or pit starters). These generally take two forms: 1) starters that act on the crank and 2) starters that act on the rear wheel to “bump” the bike to life. Because all my bikes have rear wheels – I went with option #2 as the most useful.

For at least the last decade I’ve seen advertisements for Doc Z “solo starters” and so I checked them out first. I quickly learned that pre-made roller starters are at least $800 and units with good reputations are over $1000. There’s also a small cottage industry of guys making these for others – but the price is always over $750. And, many use 12v starter motors – so you still have to add batteries. That pushes the price up and doesn’t make it as “portable” as 12v suggests.

As a result, I decided to build my own – and make it operate on household voltage. Here’s why:
1) I can more easily get commercial 120v motors for my needs for free or little charge
2) No batteries to maintain or lug around
3) The starter is for tuning purposes, first start situations, and most importantly – first cold morning start. Once warmed up – the bikes are easy to kick. So, I only need it in the shop.

With that decided, we started scrounging for scrap and buying what we needed. All measurements are in imperial units. Here is my materials list:
2 – 2x2x24 square tubing
2- 2x2x36 square tubing
2 – 3x3x10 square tubing
1 – 3x3x6 square tubing
1 – ¼ x 4 x 10 steel plate
1 – 1/8 x 1-1/2x5 steel strap
4 – 5/8 bore cast iron pillow blocks (3,500 rpm)
2 – 12 inch long by 5/8 OD keyed jackshafts (go kart parts)
3 – 20 tooth #40/41 chain 5/8 bore keyed sprockets
1 – 10 tooth #40/41 chain 5/8 bore keyed sprocket
2 – 5/8 bore keyed hub connectors
5 – 3/16 by 1 key ways
4 – 3 inch OD by 1-1/2 steel slugs
2 – 3-1/4 inch OD by 7 inch DOM tubing
2 – 12 inch lengths of 3/8-16 all thread
6 – ¼-20 by 1.5 inch machine screws
10 – 3/8-16 nuts and washers
1 – Foot control switch (15 amp minimum)
1 – 7/9 inch 15amp angle grinder (6000 rpm)
1 – 5 foot section of #40 chain and 2 master links
1 – 5 foot section of 4 inch wide 3M grip tape
1 - #40 chain tension wheel


Here is where I purchased stuff in Summer 2020:
• Grinder motor, all thread, nuts, bolts and grip tape – Menards
• Pillow blocks, #40 chain, tension wheel, and master links – Amazon
• Foot controller – Harbor Freight (borrowed from a ¼ hp wood carver)
• Jackshafts, sprockets, keys, and hub connectors – BMI Karts
• Structural steel – scrounged
• DOM tubing and slugs (cut to order) – Speedy Metals, Appleton, WI

Buying all this stuff off the shelf is about $400. Less if you scrounge. You’ll also need a welder and some way to bore the slugs for the jack shafts. I used my lathe to make all the parts.
This particular set of materials gives you a rear wheel rpm on a standard 18 inch tire that equals about 30 mph. For a stock bike, this is more speed than needed. But, on a big inch, magneto fired stroker, we need the extra wheel speed to light the bike.

Here’s how it works. The 15amp grinder spins at 6000 rpm. This is about 3hp in mechanical terms. More than enough to get the job done. But, 6000 is way too fast. We gear this down with the sprockets and chain at a 2:1 ratio. This gives a roller speed of 3000 rpm. The diameter of the rollers was chosen to give a rear wheel rpm of 375-400 rpm on all my bikes. The math is just calculating surface speeds similar to using rpm to calculate true mph.

Because some of you may want to calculate the roller diameter for a different tire; here’s the basic math.

A 16/18 inch rear tire averages 26 inches in diameter.
30 mph target speed = 44 feet per second
44 feet per second times 60 seconds = 2,640 feet per minute

26 inch tire diameter times Pi (3.1416) = 81.6816 inches traveled per revolution (aka circumference)
81.6816 inches divided by 12 = 6.81 feet

2,640 feet per minute target speed divided by 6.81 feet per revolution = 388 rpm

Therefore, our target rear wheel speed is roughly 375-400 rpm.

We choose a roller diameter of 3.25 inches to check our speeds.
Circumference is 3.25 times Pi (3.1416) = 10.2102 inches

Output (tire speed) divided by Input (roller speed) gives our ratio
81.6816 divided by 10.2102 = 8

400 rpm times 8 = 3200 rpm
375 rpm times 8 = 3000 rpm

So, we now know that a 3.25 inch diameter roller spinning at 3000 rpm will rotate an 18 inch rear tire at 375 rpm or just below 30 mph true speed.

Because we are using an input motor spinning at 6000 rpm; we know that with a simple 2:1 reduction we can reach our target speed of 3000 rpm on the rollers. The easiest way to accomplish this was to use a 10 tooth sprocket on the input motor and 20 tooth sprocket on the rollers.

Using the same math and formulas, you can easily calculate your desired rear wheel speed for a variety of combinations.

With that done, we moved on to making the actual rollers. The frame is the easy part. You simply tack all your pieces together and then go for heavy beads. We chose to stack the 2” tubing in order to elevate the “rear” roller. Talking to other people revealed this is the biggest improvement in designs. We also learned that a center to center distance of between 12 and 13 inches is about ideal for cradling a tire. We went with 12.5 inches to split the difference.

To make the rollers, I turned slugs of 3 inch OD by 1.5 inch thick 1016 steel to be press fits in the DOM tubing used for the rollers. On the “drive” side for the main roller, we made this a .004 interference fit and used a 20 ton hydraulic press to seat slug. While in the lathe, the slugs were center bored 5/8 for the axle shafts. The slugs were then positioned and tack welded to the tubing.

To transfer power to the tubing (and make things easily replaceable) I used hub connectors. These are nothing more than big sleeves with three holes in them. To easily find our positions, I slipped the hubs over the shafts and installed them in the rollers. I center punched the holes, drilled and tapped them ¼-20. Now, the shafts can be secured to the hub and the hub to the roller. If anything gets damaged, standard go kart parts can replace everything in minutes.

The rollers ride in cast iron pillow blocks. These are secured to the 2” tubing by 3/8-16 studs so the pillow blocks can move fore/aft to tension the chains. I drilled and tapped 3/8 holes for the pillow blocks and then screwed in all thread. Each all thread chunk also got one tack weld to hold it to the frame during assembly.

The grinder motor conveniently comes with three 3/8-16 threaded holes for the grip handle. We used these to mount the motor to a plate welded in the rear. We also used a big old hose clamp to provide “tension” for setting chain slack. The sprockets are all held in place with keyways and set screws. So, this is a bolt together, bolt apart affair outside the frame and every part is replaceable independently in case of damage or wear.

After using the rollers for a few weeks, I added a chain tensioner to the primary drive chain. This is nothing more than a chain wheel that presses up on the chain to give it tension. It helps to keep the chain from trying to jump when we dump the clutch.

The last step was adding grip tape to the rollers. We simply wound strips of 3M safety grip tape around the rollers.

The reason for using 3” square tubing at the front was so that when I put a ¾ inch plywood ramp in front of the starter the actual “bump” to get over is only 1 inch. This makes it a lot easier to use the starter by yourself.

With all that said and done, I put my 1964 XLCH on the rollers for their first test. If they could start a 78 inch magneto fired stroker – they can start most anything.

Indeed, the starters worked as intended. I put Gladys in 4th gear, spun the rear wheel up, dumped the clutch and enjoyed the sound of fury.
The main thing I learned in using roller starters is to be patient. Let the rollers really get the rear wheel spinning and the bike lights instantly. If I rush it; it laughs at me. So, I found that if I start the rollers, count to 5 and then drop the clutch it starts first time without any fuss. I like this system so much, I’m pretty much not kicking anything these days for first starts.

It did take me about 15 hours to machine, weld, and assemble everything. I know why these things cost $800-1000 minimum – that’s about what it would cost me to make them if I paid myself. My real cost was $250. I got much of the steel for free and I used a bunch of rebates at Menards to get the grinder for free. Most of the $250 I spent was for shipping on heavy steel items and having to buy a new spool of welding wire because I ran out. I also got some really good deals on karting parts, which lowered the cost as well.

IMG_5179.jpg

IMG_5180.jpg

IMG_5181.jpg

IMG_5185.jpg

IMG_5187.jpg

[chuckthebeatertruck]

And, one more tidbit - making intermediate jets for L series carbs.

So, my GBL came to me with only one intermediate jet - a number 31 (.031 outlet hole). Problem is that I was experiencing a mid throttle stumble caused by poor jetting. Compounding the issue is that S&S hasn't offered intermediate jets in a few decades. They do come up for sale; but you're never really sure if someone screwed with the holes . . .

To get around this, I just made some up. The main jets out of any L series, Super B, E, or G will fit directly in the intermediate hole on an L series. The tubes themselves are .125 by .016 originally . . and we can get .125 by .014 brass tubing right from Amazon without fuss.

So, ordered a dozen main jets and a few feet of brass tubing. After whipping out my 61-80 index drill set, we were ready to make some new tubes.

To do so, I simply soldered the main jet closed, then mounted the jet in the lathe and bored it .125 for the tubing. Tubing was then soldered into place and cut to length. The main bleed hole drilled and the intermediate jet drilled. This allowed me to create a selection of jets from .0292 to .040 in minutes.

I placed an .033 tube in the GBL and Gladys woke up instantly. Yippee.

So, if you need intermediate jets and can't make your own (or don't want to make your own) drop me a private message. I can make tubes for you in any size from 28 on up for about $20/tube.

IMG_5190.jpg

IMG_5191.jpg

IMG_5192.jpg

[chuckthebeatertruck]

Just how does my valve train work . . . and is valve stem protrusion really an issue?

During the course of putting together the heads on my 1964 XLCH stroker, I ran into a rather common issue of being over the factory service manual (FSM) limit for valve stem protrusion. Originally, this article was going to focus only on the protrusion issue and how to deal with it.

As we talked through the pros and cons, our technical editor, Dr. Dick, reminded me that the protrusion is the very end of the system. A common problem for all mechanics if focusing on “issues” in isolation and not as the sum of all the parts. To understand what that FSM spec really means, and whether we have an issue, one has to take a more in-depth look at the entire valve train from the pinon shaft up.

So, that’s what we’re going to do here – run through how the valve train works to transmit power from the crankshaft to the valve. The goal isn’t to give you a degree in mechanical engineering or debate the merits of the design. It is to help you visualize how the components work, observe where you may have issues with your own bike, and plot out reliable, long-term solutions. Then, we’ll run you through one way to deal with stem protrusion.

Let’s start with the order of power transmission. After the piston fires; it transmits its power to the crank pin. The job of the crank pin is to transform reciprocating (up and down) motion into rotary (circle) motion. That rotary motion is what we use to drive the valve train on the right or “pinion” side of the motor. In this case, our order of power transmission is:
1. Pinion Shaft
2. Pinion Gear
3. Cam gear
4. Cam lobe
5. Tappet roller
6. Tappet body
7. Tappet adjuster screw
8. Push rod
9. Rocker Arm at push rod end
10. Rocker Arm at Valve Stem end
11. Valve Stem (whose motion is resisted by the valve springs)

As you can see, the valve stem (and therefore stem protrusion) is the last in line. Let’s look at each component.

All motion in the valve train starts with the crank’s pinion shaft. The pinion shaft transmits the crankshaft’s rotation to the four cam gears via a pinion gear. A pinion gear is simply the smaller of a driven gear. The pinion gear is indexed to the pinion shaft by a series of four splines – with one spline being the “master” and larger than the others so the pinion only fits one way. On stock bikes the gear is generally a light press fit on these splines. However, on high mileage bikes and/or bikes running more extreme cams, it is not unusual to find the pinion gear loose on the shaft. As a rule of thumb, we want a gear with as little wear as possible that fits snugly to a light press fit on the shaft at room temperature.

Cam gears themselves rarely give trouble. Instead, lobes tend to be an issue. This is because of how the roller tappet actually works. The tappets aren’t in a perfect position and they don’t move at a constant speed. They are off to the side a bit in comparison to the cam lobe centerline. This is a by-product of the design and places huge side thrust on the cam as it transitions from the base circle to the actual lift. In short, as the lift starts to happen, the tappet roller suddenly has to speed up at the same time is starts encountering resistance from the valve springs. It goes from a slow lazy circle to a short radius as it comes up on the lift. This speeds the tappet roller up very fast and places a lot of momentary load right on the cam’s edge. You can see the results most clearly on more “square profile” cams like many Sifton grinds – the cam lobe corner gets eaten away from the pressure.

Hopefully it also makes sense why HD switched from bronze crankcase cam bushes used in the early ironheads (1957/58) and replaced them with needle bearings. Put simply, the bearing that sees that momentary pressure spike is the crankcase cam bearing. The needle bearing can handle this high pressure situation better than the bronze bush. The other end of the cam doesn’t see the same extreme pressure – and so bushes continued to be used.

We’ve now made it all the way to the tappet. The name of the game in any valve train is stiffness. In this regard, the tappet bodies are not an issue, but the adjustment screw is a bit of a problem. It’s not exactly the stiffest component in the system – and you’ll note that on almost all ironheads you wind up with the screws pretty far out. No matter how tightly you lock down the adjuster nut; the tappet screw itself remains long and flexible. The reason for this has to do with the fact we need to be able to install and deinstall the pushrods in situ. To do this, the push rod must be short enough to clear the tappet cup for removal. This results in a long adjuster screw. Between tappet slop and this flexibility – you get misalignment in the valve train.

One of the tricks is to run longer pushrods and grind away most of one flat on the tappet cup so you can slip the rod in from the side. It works to keep the adjuster screws short, especially on bikes running stroker plates or extra tall cylinders. If you ever wondered why you saw fat ½ inch or tapered pushrods that seemed way too long for an IH but were all wrong for Big Twins . . . now you know what you might have been staring at. We actually have the ½ inch thick rods on Chuck’s 64CH. They are a quarter inch longer and they do reduce the height of the adjuster screw by a considerable amount.

But, there’s no free lunch.

Believe it or not, the pushrod does not move straight up and down. Instead, it moves in a small arc forward and aft as it travels up and down. At the tappet end, this doesn’t cause much issue. At the rocker end, too much movement results in: 1) rubbing of the pushrod on the cover and leaks or 2) over rotation of the rocker arm with the possibility of the p-rod popping out. If you have a fat ½ inch pushrod, you can see how situation 1 is made worse.

And, here’s where we pause again. If you’ve been fighting pushrod tube leaks – check whether the pushrod is striking the cover at the upper end. If it is – find out why. Also, be advised the FSM and parts book showed the lower cover upside down on 900 ironheads. The flat end seats in the cork at the tappet, the funnel end points up. Here’s why: when you fire up your motor it starts growing taller. As the heat builds, the motor can grow as much as .060-.080. This means the pushrods and the pushrod tubes have to expand and contract. To do so, the tubes are made in two pieces and designed to telescope. The middle cork gets pushed into the funnel end of the lower tube by the cup spring – which then acts as a type of compression seal. HD put a lot of thought into these bikes. They aren’t simple “tractors.” Tip number 2 – almost all current cork seals – including NOS Harley, are too big an OD for the middle seal. Take the time to grind it a bit so it doesn’t touch the cup and it will seal better to the lower tube. I thought this was odd too . . . and then I tried it on my 59 XLH and the pushrod covers have never been so dry.

NOW we are finally all the way to the rocker arm. As we just learned, the pushrod travels in an arc and we can alter its length as well as its diameter. Look at the rocker and you’ll see it has a cup in it to capture the pushrod male end. The more extreme the angle between the pushrod and the rocker – the more the cup experiences side loading vs. linear motion.

No free lunch.

On the other side of the equation; the radius of the rocker pad is such that it hits the valve stem and literally slides across it. The “goal” is for the rocker pad to stay roughly centered in the stem at full lift. If the rocker pad over rotates, the very tip of it rides off center of the valve stem and gets torn up. This problem is made worse by two situations: 1) stem protrusion below the minimum 1.375 and 2) high lift cams. In the case of short stems – the rocker has to rotate a greater amount just to make contact with the stem. This places the initial action on the valve stem outboard on the rocker pad and also places the push rod very high – which means a cranked out adjuster screw and more slop. In the case of high lift cams, the valve literally gets pushed open further, which means the rocker pad has to travel further. This has a similar effect to too short a stem as we approach full lift – the rocker over rotates.

To correct either situation, we want to adjust the geometry to start the sliding motion closer to the pivot point of the rocker pad. In essence, we want to move the initial contact point towards the body of the rocker and away for the tip of the pad. You can do this in multiple ways, from regrinding the rocker pad radius to altering the stem protrusion.

With all of this in mind, you can start to visualize the whole valve train and how each component works together and is affected by changes in another component. Once you do this, and think about the last paragraph you discover valve stem protrusion is one way to adjust the geometry.

[chuckthebeatertruck]

Valve Stem Protrustion Part 2

So, you might be asking why is there a specification in the Factory Shop Manual if you can adjust this stuff? That’s a simpler tale to tell. The FSM is meant as a companion book to a field mechanic. It isn’t designed for the DIY home mechanic and it isn’t designed for the high performance machinist. It is designed to help a professional motorcycle mechanic properly restore safe and reliable operation to what was, at the time, a relatively new motorcycle. The FSM spec for safe range on the valve stem protrusion ensures that the field tech can easily check new or gently used parts and assemble them with near 100% assurance they will function reliably as the factory intended. Those specs aren’t meant for dealing with .500 lift cams, 60 year old clapped out heads, or weird aftermarket valve stem lengths. So, when you start dealing with non-stock stuff – you need to get educated and you need to open your mind to new possibilities.

We are now back at the original question: How much valve stem protrusion is too much valve stem protrusion. The answer as we’ve seen is: it depends.

If you’re building a stock bike with stock stuff – go with the FSM limits. If the valve stems exceed the limits; then you need to read the following and decide if you really want to do this – or if you’d rather find another set of heads – or just have seats installed.

The stem protrusion by itself is a byproduct of many things – most notably the actual combustion chamber shape of an ironhead. It is a perfect hemisphere (see the below diagram of a 1970 and up intake valve). This means that the edge of the valve closest to the head gasket surface is a different distance from the valve center line than the inner edge. As you cut the valve seats, you’ll soon discover you are sinking the valve on the outside a considerable depth compared to the inside edges near the center of the chamber. By the time you have a good seat, the valve is now much deeper in the head than you would expect with other head designs. This means the stem is now “protruding” further from the guide. Similarly, if you install a larger valve in the head – then you need a longer stem to keep the same protrusion because the valve sits further away on the chamber radius. This problem is made worse by folks approaching these heads like cars and doing a “three angle” valve job without thought. Look at the below diagram of a 1970 and up intake valve and this problem starts becoming apparent.

TOP CUT PROB-11024_1.jpg

Look at the diagram again . . . if you top cut the seat – look where that top cut intersects with the chamber wall. It’s generally a better idea to worry less about the “angles” and more about a smooth profile from the 45 degree seat to the chamber. Note, we said generally – not always. This is why it’s important to take your heads to someone who understands not just Harleys, but the weirdness of ironheads.

To help the field tech determine when the seats were sunk too deep for reliable operation, Harley issued a valve stem protrusion specification in the FSM. For stock valve lengths on a stock motor with stock cams, this is 1.375 to 1.420 inches. This measurement, contrary to popular belief, has more to do with spring pressure than with valve train geometry.

In essence, as the protrusion grows, the spring installed height increases, seat pressure drops and positive control of the valve at high speeds becomes less positive. A thinking mechanic can restore spring pressure by shimming . . . but the long stems remain and may/may not cause an interference with the rocker arms.

That’s one part of the equation – the other is the fact that Harley offered extra length valves for XLR models. Yes, you read that correctly. Harley offered really big, really long valves for the racing model. So, from new – the XLR motors wound up using a stem protrusion that was near or over the FSM limit for “regular” ironheads. Yet, these are the “race” motors and expected to turn 7000 rpm. Obviously the “geometry” isn’t destroyed by the extra valve stem length. There’s way more going on.

Let’s get to it.

Keep in mind R = race department. These were not regular production items though you could order them through an HD dealer. In essence, all the development for racing was done in the classic “shed in the back” vs. as part of the overall HD enterprise. This means that NOS R parts are not always the best quality – so don’t rush out and buy stuff just because it has an R in the part number.

You will find R parts all over the place at meets . . . here is a whole box of them from summer 2019.

box of valves.jpg

The XLR was developed by some pretty interesting personalities – including Tom Sifton who is rumored to have played a big role in the valve train and cam department. One of the first things that happened was for the team to shove big valves in the intake to take advantage of the hottest cam they could. This had the effect of moving the valve further out on the hemi chamber radius – which in turn shortened the valve stem protrusion. To get a good spring pressure – longer valves were fit to give a more “normal” installed height. Hence, R length valves which are longer than stock. Understand, this had to do with spring pressure and spring pack height as much as anything else.

These valves had a secondary purpose post 1965. Sometime in ‘65, alloy top collars came out for the valves. This helped reduce valve train weight, but introduced a new problem – broken top collars. Basically, the alloy collars were copies of the steel collars they replaced. They simply weren’t strong enough. So, the team re-designed them, making them thicker and stronger. To make them reliable, they fit new “double length” keys that grip the valve stem all the way to the top of the stem – not just at the bottom like standard keys. You can see this in the below picture:

underside of sifton collar and keepers.jpg

sifton collar and keepers assembeled.jpg

sifton collar and keepers.jpg


This had two effects: 1) the now thick collars and keys were at risk of contacting the rocker arm, and 2) the thicker collars further compressed the spring pack, which meant that longer valves were needed to give the correct installed height.


You can see part of the Sifton aftermarket “solution” in the below photo for that rocker arm contact risk. The collar now has a chamfer for clearance, even though the keys are clear in the middle of the stack:

sifton solution.jpg

[chuckthebeatertruck]

Valve Stem Protrusion Part 3

Around 1970, the 1.75 diameter exhaust valve became available. This followed all the same parameters as we just discussed about the R length intake valve. These got quickly picked up on by the performance market. But, unlike the R length intakes, each major manufacturer offered 1.75 exhausts in different stem lengths! Today, only a couple of 1.75 exhausts are available new. So, you get what you get as our choices are limited these days. Basically, whatever Manley or Kibblewhite have on the shelf – or what we are willing to pay to have valves made to spec.

Let’s pause right here and talk about this “big” exhaust valve. On a stock 900 sporty it is totally and completely unnecessary. The valves are already a bit too big and going up to a 1.75 exhaust is like hunting elephant with a howitzer. However, as you start increasing the displacement the bigger valves start coming into their own. Once you hit the mid-70 inch range, they are a proven combination with 1.94 intakes. So, the big valves tend to be more of a hot rodder thing and are not recommended for stock bikes.

At this stage, you should have a decent understanding that the shape of the combustion chamber, the valve diameter, the seat depth, and the spring pressure are among the factors that influence the stem protrusion min/max specification in the Factory Shop Manual – and how that spec is related to the rest of the valve train.

And, here is where we exit stage left and enter into a new world of possibilities. As we discussed, the valve train is designed so that with stock cams the center of the rocker radius tip is roughly in the center of the valve stem at a 90 degree angle – which equates to roughly half the total lift. If you have higher lift cams, then the extra “lift” past 90 degrees sees the valve stem dancing on the very edge of the rocker arm. This is where problems start and valve control becomes less positive. In essence, we actually want to start the valve “push” deeper on the rocker pad radius “heel” for higher lift. That ensures when we reach full lift the rocker is able to control the valve stem.

One way to do this is to use longer valves or to sink the seats.

But we now have a protrusion exceeding the FSM. What does this mean? Well, it means we need to ensure no spring troubles and no contact between the rocker arm and the top collar. If the rocker contacts the collar, you run the risk of broken collars and a mess. So, you need to carefully check and make sure you have at least .060 clearance between the rocker and the collar at all times – and no contact between the rocker tip and the valve keys. It also means you have to be aware of how extreme the angle might be on the pushrod end – and take into account that like the pushrod, the valve also doesn’t open and close straight up and down as you’d think – it too moves in a small arc. That arc places side loads on the stem. So, as we get to higher and higher lifts, with more and more protrusion, we have to take into account many more variables than if we were just rebuilding a stock motor. Failure to think through all of this before you touch a wrench is one of the key reasons so many “high performance” motors are unreliable. If you take your time and think of the whole system, not just the component, some solutions become apparent for getting maximum reliability.


Let’s talk you through what all this meant on Chuck’s 64CH.


On this motor, we elected to use a set of early 1972 cylinder heads. These heads were only made in 1972 and the first part of 1973. They have the necessary larger counter bore to take the 3.25 alloy cylinders we are using, but are a 900 base pattern. The set we started with had seen at least one previous valve job – and the valve job was done by someone who likely had mostly automotive experience. This meant that when they did the 30 degree top cut – they really hogged out the exhausts. When Chuck went to install 1.75 exhausts, they wound up pretty deep.

How much worse: on our front head, the stem protrusion is 1.498 and 1.501 (in. and ex. respectfully) and on the rear head it is 1.470 and 1.468 (in and ex respectfully).

A lot of people would immediately say – add a shim between the rocker box and the head. Problem is that while it seems logical to do so – you actually make the problem worse. The rocker travels in the same arc and the valve stem moves away from the rocker. So, if you raise the pivot point by shimming the box – the arc of travel actually moves inboard! See the below illustration. It isn’t to scale, but it helps illustrate the effect. The blue line is a stock height arc – the red line is with the pivot point raised just 3/32 of an inch (.094) or just a bit more than our measurements would indicate we need to get into the middle of the FSM specification.

valve arc.jpg

Now that you can see and understand why rocker box shims won’t help . . . let’s talk about higher lift cams. An old rule of thumb is to increase protrusion by ½ the increase in lift. For the 64ch we are using Sifton cams with .436 total lift on both intake and exhaust. This works out to adding roughly .020 to the intakes and .030 to the exhausts (compared to stock P cams). This means our “max” protrusion without any further surgery would be around 1.440 on the intake and 1.450 on the exhausts. So, on this set of heads, we are somewhere between .030 and .050 over ideal length.

Because our desire is to have a completely user friendly bike – custom length valves are out of the question. We want common, off the shelf items that we can replace in a matter of a day or two – vs months for custom parts. As a result, we have to solve this another way. And, that way is stupidly simple. We are going to place a blob of silly putty or playdoh on each valve collar, install the rocker boxes, adjust the pushrods, and gently turn the motor over through two complete revolutions. We then take the box off carefully and using a very sharp razor, cut the putty in half. You then measure the thickness of the putty remaining on the collar. If it is .060 or greater – leave it be. If it is less than .060 – you will need to take a fine stone and carefully grind the underside of the rockers. If the motor binds up before a full revolution . . . then you know you need clearance! If you are having trouble measuring the thickness; try using a stack of feeler gauges. Combine a set of gauges to be .060 and add a .005 gauge to the top. Place your stack next to the putty and slide over the .005 gauge. You’ll know pretty fast if you’re close.

If you decide to clearance, take no more material than necessary to reach .060. Be sure to do a neat job of it and to use a fine finishing stone for the final passes. It should be smooth with no gouges and no potential for stress risers. Then, clean everything like crazy, reassemble, and you’re back in business. It is also not a bad idea to dye the rocker pad and make sure of the contact pattern.

ground shafts.jpg

play doh 2.jpg

playdoh 1.jpg

[chuckthebeatertruck]

Valve Stem Protrusion Part 4 (Final Part)

But, wait – there’s more. When we are screwing around with protrusion – it also means we are messing around with spring pack installed height. The installed height is the amount needed to maintain good spring pressure. Too much height and pressure is lost. Too little height and you’re likely to slam the coils together and break things. Pretty much every spring maker lists an installed height. Problem is that this spec cannot cover all combinations of cams, valves, etc. To get around this, we install the spring pack to a given coil bind spec.

Just what is a coil bind spec? Well, that is the difference between the free length of the spring and its length when fully compressed with the coils touching one another. Free length is easy – just use your calipers to measure it in hand. Do it for both the outer and the inner spring. Then, compress a spring in your vise and measure the “bound” height. Do it for both springs.

Now, install your spring packs without shims and measure the installed height between the upper and lower collars. Subtract the “bound” height from your installed height. This gives you your total potential spring travel before coil bind.

With that known, we can get down to maths. We know our lift is .436 for the Q ship and that we want an .060-.090 safety margin between maximum lift and coil bind. We chose this margin because it is friendly to springs. If you pack them tighter than .060 you create a lot of heat and over time, that heat affects the springs. You can pack inner springs more tightly if you wish, but generally a bit looser is better for a regularly ridden street bike. Considering ironheads aren’t really winning grudge matches anymore; you might want to aim for reliability over all out power.

To illustrate, let’s look at one valve on the Q ship. We’ll use the front intake for the example. The spring free length is 1.875 and it hits coil bind at .790. The installed height without shims is 1.373. Here’s the math:

1.373 (installed height) - .436 (total lift) = .937 (remaining height)
.937 - .090 (safety margin) = .847
.847 - .790 (coil bind) = .057 excess height to coil bind.

So, in order to achieve our target height, we’d need to decrease the installed spring height by .057. This is most easily rounded up to .060. This means that with an .060 shim under the bottom collar – we will come right to our desired pressure. Oddly, it also means we’ll wind up right near the manufacturer’s recommend install height for these springs. At this stage, don’t forget to make sure your upper collar can go through the full travel plus .060 without hitting the top of the valve guide!

shims.jpg

[chuckthebeatertruck]

Throughout the rest of the autumn, I had a chance to keep breaking gladys in and tuning her. By the end of November, we had gotten the carb very, very close to being perfect in every gear. I am fighting a small amount of reversion right at 50/55 mph. I've got a couple of ways to deal with it and in the mean time have just enjoyed riding this bike.

IMG_5449.JPG

[chuckthebeatertruck]

Gladys just keeps rolling along . . poor bike loves to leave rubber strips and make me smile.

But, the range with the small tank is just too, well, small.

So, I painted up an original turtle tank to match.

IMG_6802.JPG

I also switched back to the S&S tear drop air cleaner with a KN filter. I was having a bit of trouble getting enough air flow through the old cleaner at WOT . . .and the S&S solved that nicely.


The final change yet to come is mounting a 49T rear sprocket. I had a very lightly used OEM drum and 49T sprocket fall into my hands for free . . . and I'm very much looking forward to adding a bit more leg to the bike. It can easily pull the gearing . . so why not give it a shot.