Internal Suspension Girder Forks - Page 3

(Found out that the situation at Executive Choppers was just a problem with lag-time during a switch over to a new web hosting company. The shop is in full operation).

The early Harman forks did indeed have some problems as we touched on in the previous pages but to be fair most of these problems are exactly the same as you see on so-called 'reversed Springers' where the springs are mounted on the tube legs just above the rockers. foremost among these situations is the issue of not being able to equalize the load between the left and right springs. On a conventional Springer the sprung leg spring bridge acts more or less as an equalizer between the two sets of springs. On a Spirder or Reverse Springer you don't have that bridge so the wheel loads are transferred somewhat haphazardly to two independently working springs which contributes to bent, twisted or broken rockers. From my perspective this is the single biggest drawback to the Spirder front-ends but there isn't a thing you can do to eliminate this issue since it's intrinsic to the basic design so we have to live with it. People have tried adding keyways on the axle and rockers and this helped a little. Some folks went to a single, thick (.75") rocker one each side but it didn't do anything except look bad. The current Harman/Holland rocker design is about as good as you can get so not much need in trying to reinvent the wheel.

The second problem is not being able to have what we usually call 'rebound' springs. On a conventional Springer you have two sets of springs acting in cooperation with one another to handle both compression and extension loads transferred from the rockers. This creates a somewhat 'self-damping' effect much like a modern day shock absorber. Ironically the old original ISG from the 1900's had rebound springs as part of the standard design. This dual spring system can be used on a Harman type ISG so I'm not sure why it wasn't part of the original design to begin with.

The third problem is a lack of basic suspension travel. Most of the Harman and Harman clone forks have only been designed to provide a little over 1.5-inches of suspension 'travel'. This isn't much even by Chopper standards. I had a Harman on a bike I once owned and my only complaint about that front end was that it 'bottomed-out' on even the slightest road irregularity. If you hit a good sized hole in the highway the rocker springs would hit home plate and then you'd see the forks flex pretty radically. I actually stripped the threads out of a rod end and bent a rocker on that bike one day. Stiffer springs might have helped a little bit. I've always thought that these forks were 'under-sprung' to begin with but my forks were old originals and I don't know if Harman and Holland made changes to the springs over the years. Actually most of the Harman and clone forks I had experience with would flex far more than I'm customarily comfortable with but I've never seen or even heard about a complete front end failure so they were plenty strong enough to have survived the test of time.

Some riders stiffened up the forks by welding in short sections of tubing between the front and back legs about midway down the forks and this really does help a lot. I've also see a cross-brace between the front fork tubes, much like a hydraulic fork bridge/stiffener. Johns own bike had a Maltese cross welded in between the front tubes and I've always wondered if he did that to act as a reinforcement or if it was just for appearance. I've already mentioned that the old original .058" wall tubing was upgraded to .063" wall at some point in time.

The fourth problem with Harman style Spirders is that you can't get access to the suspension components. maybe this should have been listed as problem number one. A very long time ago we identified most of the issues with these forks and came to the conclusion that there were some basic 'conditions' we would be faced with. 

Most new forks work just fine for a few years and then things begin to happen. If you ride the bike everyday for a lot of miles the system seems to work forever with out any problems so long as you lube the springs every now and then. If you 'weekend' the bike, especially in wet or dirty conditions the suspension will eventually begin to fail, usually on only one side of the forks. If you don't ride the bike much at all, like most 'bar-hoppers', both sides of the suspension system will fail in only three to five years from an accumulation of rust and the springs taking a 'set' after being under static load in one position for a long period of time.

I personally think, based upon my experiences that most of the original Harman forks had pretty 'cheap' springs installed to begin with. If you're buying a set of Harman forks at a swap meet have the seller remove the axle and then cycle each rocker independently. More often than not you'll find that one rocker has little, if any,  'spring resistance' at all while the other rocker is almost unmovable as it should be. Failure of one spring either due to loosing temper, taking a 'set' or being broken is very common on these forks.

You can 'rebuild' a set of original forks but you have to do a 'forensic' cutting on the 'patch' that covers the spring rod hole. I've rebuilt three sets without any problems with the exception of finding new springs that matched the originals so I've had to have springs custom made.

Now to hit the 'plus' side of things. Almost all Harmans are designed to run with close to zero trail or a wee bit of positive trail, around 2-inches, so they behave much like Sugar Bear's Springers exhibiting virtually no flop at all. As a result they take a little time to get used to after riding with a more conventional fork set for any length of time but the handling is typically excellent. These are very sweet riding front ends on the right bike.

The downside to that last statement is that these are very sweet riding front ends on 'certain' bikes. Since both the geometry and spring characteristics are more or less 'hard-wired' into the design you'll find that the forks are very sensitive to weight distribution, rake, trail, and about everything else you can come up with that effects handling characteristics. The same set of forks installed on two different bikes will behave differently, sometimes significantly differently so let the buyer beware.

Well so much for the typical long drawn out bunch of text about the why's and where fore's about these forks so let's try to build some.

If you're a purists type of person you'll probably try to duplicate the later originals and use the 1"o.d x .063 wall chromo. More power to you if you have access to the gear needed to work with this alloy. 4130 steel is typically sold in the 'annealed' state, meaning that you can work it into shape and then have it hardened with heat treatment and then stress relieved once you're finished doing whatever it is you might have been doing. 4130 tubing is typically sold in the 'normalized' state meaning that's it been pre-stress relieved so you can at least bend it without having it fracture. Most folks who actually make stuff from 4130 will tell you that it needs to be 'gas' welded but the 'modern' guys will tell you that you can Tig it. This is somewhat correct but if you Tig 4130 you actually create a very very brittle weld and a sharply defined HAZ that can lead to stress fractures a lot faster than the same joint done with gas. If you do insist on using Tig the 'experts' recommend that you use E7052 rod so that the weld itself is somewhat softer than the adjacent material. No matter how you weld it you still have to go through a stress relieving process if you want to have an assembly that stays together. The 'experts' on 4130 are divided over whether or not you really need to do a complete stress relieving on thin-wall chromo and the general consensus and compromise is to just stress relieve the welded connections.

I gave up working with 4130 well over 30 years ago as it simply wasn't worth the hassle dealing with it. A far better alloy for those folks who insist on minimizing weight would be one of the Reynolds alloys which has all the advantages of 4130 but none of the drawbacks. Talking about this stuff only opens a bag of worms as everybody and their brother has an opinion about tubing specifications. Plain old mild steel DOM is about a thousand times easier to work with and weld but you do need to have a slightly thicker wall thickness or larger diameter section to get equal strength over 4130. For bend resistance however it doesn't make any difference at all what alloy you use as all mild steels including 4130 have the same resistance to bending and for Girder type forks this is all that matters. There are some good technical articles posted on the web concerning computer based structural analysis on various girders if you want to do some research on the subject.

The measurements that we'll be using in this article are taken from my notes made somewhere between 1966 and 1973 as near as I can remember. These may or may not have been taken from John's forks. Back in those days there were a lot of forks being made by a bunch of guys who worked on and off with John and did their own stuff on the side. From what I can see as I now redraft them in AutoCAD these are good enough to get you started. I hope you understand that we're not trying to build a museum replica or a historically exact reproduction. I do have copies of Harman's work notes but we won't be publishing those. If you have a problem with my sketches then go find some better ones.

At first glance these forks look like they're pretty complicated to build to various lengths but in reality they are extremely simple. All of the bends are identical no matter how long or short the forks need to be. The angles of the tubing copes need to be changed slightly based upon lengths but this is easy to do. When I was helping to build these at the old Vegas shop we have a massive drill press that had been jury rigged to run horizontally and this was our coping/miter machine. Our bender was an ancient Hossfeld set up with two dies stacked on top of one another so we could bend two identical tubes at one time. We could have easily bent and coped 20 sets of legs per day but we usually only did one or two per month. Our bread and butter was Girders, sissy bars and seats and more seats and even more seats. I learned how to do seats there but I still hate it.

The image below illustrates a typical Harman style Spirder layout to handle forks from stock to 42-inches over. All forks of any length range are made from two basic parts, the front tubes and the rear tubes and of course you need a pair of each. The bends never change no matter what the fork length but the 'rotation' of the rear tube pivots around the upper tree bolt center depending upon the desired length. Naturally the length of the 'straights' changes but the you never have to mess around with 'special' bends.

This concept shows pure genius in the design implementation. I'm pretty good at design concepts but never in my wildest dreams could I have come up with this idea. Not in a million years. 

Now that I've had a chance to reverse engineer this fork style in Cad instead of just pencil sketches I think it's safe to say that Harman originally designed his system to best suit the bender that he was using which was most likely one of the old 'one-shot' type units used for pipe or conduit where one crank, ratchet, push, pull or plunge gave you 5-degrees of bend. Since he was using 4130 he most likely was getting about 3 to 5-degrees of spring back so this layout makes sense to me but how he came up with idea to begin with still amazes me. From what I've heard from guys who knew him in High School we didn't build any prototypes of this concept or at least they never saw any. I would not be to surprised, knowing his talent, that his first actual trial product ended up being exactly what he envisioned it to be in his minds eye.

Don't worry about not being able to read the text in this image as we'll get to larger scale drawings later in this article. For the time being all that is important is that you get some idea of the basic proportions and concepts of the design.

The bends in the lower end of the forks is more or less as shown in the next picture. I say more of less as I've seen these angles and lengths vary considerably from one set of forks to the next that I've measured. Keep in mind that a lot of the forks I've handled were not made by Harman and even some of the forks that I do know for sure are Harman's were slightly different. I suppose that this exhibit could best be described as a 'good average' of the measurements I've taken.

Again, as I mentioned earlier it appears as if the original intention was to keep all bends within a range of 5-degree increments. The lengths of the straight runs I've measured vary significantly from fork to fork. A lot of that variation is probably due to 'fit-up' when the copes are dressed for welding. I'm pretty good at this stuff but on these long shallow miters I'm happy with getting to within a quarter of inch of the plan dimensions with respect to exact lengths. It's far more important to get both 'sides' of the fork halves to match exactly.

A friend who has an old confirmed set of Harman forks has told me that on his set the 165-degree angle shown in my sketch measures 160-degrees but the 15-degree angle is the same meaning the bend on the rear leg is only 150-degrees (These are bender degree measurements and not plan measurements). We need to get more as-built measurements on verified Harman forks to determine exactly what these angles were supposed to be but what is shown will get you started on a project.

Here's some sketches of other configurations I've measured. Keep in mind that all of the Harman forks I've had in the past were pre-1973-74 vintage so I have no idea what changes if any were made after the 1974 modifications to the rockers.





It's pretty easy to see from comparing these sketches that you can significantly change the rocker geometry just by slight modification here at the lower ends of the tube legs. We've shown five different end arrangements with rockers ranging from 2 to about 3.5-inches between pivot points. Remember that the shorter the distance between points the more deflection will be seen in the spring rods. We'll take more about some of these dimensions and various rocker designs later on down the line.

Before going much further we need to talk about neck assemblies. As mentioned earlier most Harman type forks (with a few exceptions) are built to a very specific neck lengths. This refers to the overall length or height of the assembled steering stem/neck tube assembly including all bearings, spacers, dust shields, adjustment nuts, lock nuts and anything else that goes with your particular bike. This assembled length can vary considerably from American bikes to Imported bikes, from old bikes to new bikes and even between various models of the same bike. The reason this measurement needs to be precise is because the Harman style forks bolt to the sides of the trees. The distance between the holes for the mounting bolts is fixed so if the upper tree for instance sits 1/16" higher or lower than you expected then the forks holes won't line up properly. This is the reason you see so many Harman forks with the lower bolt hole hogged out or actually welded in and re-drilled in a new location.



The photo above shows some of the parts of a typical neck assembly. One the right is the dust shield and two types of bearings for the lower end. One bearing set is the 'sealed' type and the other is conventional. The difference in height between these two different types of bearing is a healthy sixteenth of an inch. On the left is the upper bearing, dust shield and several different types of adjustment nuts plus the stem cap nut. The style and height of the bearing preload nut and jam nut or lock washer varies in dimension by up to a good quarter of an inch, one-half inch in some cases. It's pretty easy to see that your stem assembly can end up being .5" shorter or .75" longer than 'design' specifications depending on what mix of components you end up using. A few days ago I was in a shop and had a chance to measure six different H-D choppers and found that the distance between the trees varied from 6.875 to 8.125-inches. All of these bikes were fitted with H-D wide glide front ends. Just trust me when I tell you that you really do need to accurately measure for your particular project and not rely on 'published' or so-called 'factory' dimensions.

The sketch below shows a hypothetical neck assembly complete with upper and lower trees. Substitute the dimensions shown here with your own unique measurements or just use these for the initial design development and layout of your forks. You can fine tune things as you go along. There is nothing wrong with using shim washers if needed so you can make a stack longer but you can't make it shorter.

The neck shown here is raked to 45-degrees. I do almost all of my design work based upon this angle and adjust backwards or forewords as needed for forks having different rakes. That's just a personal habit and it makes no difference at all how you decide to layout your project.

What is important to note however that the location of the upper fork mounting bolt hole, where it gets attached to the upper tree, is what I call the master control point for all subsequent measurements of distances and angles. this point will never change so it makes a good foundation for all of your work.

I still use 1-inch thick steel trees on my forks. Some folks use 1.125 or 1.25-inch steel and I've even seen 1.5-inch aluminum used. This doesn't make any difference so long as you account for the changes in dimensions needed to accommodate the thicker material.

Also shown in the sketch is the 1.5-inch 'offset' I almost always use on forks of any kind. You can change this if you want to but I've found that this 'number' works really well. You can change trail numbers by changing the location of the bolt holes in the trees, or even changing the offset if need be but we'll talk about trail adjustment later on. For the time being we'll just assume that the fork mounting bolt holes are equally spaced 1.5-inches out from the centerline of the stem and that the front fork leg runs parallel to the stem center line extended to the ground point. We'll also assume that the distance between the fork mounting holes is 8.5-inches. Knowing this exact dimension at the start of the project is much better than just guessing at it but for now we'll just use an estimate.

Also notice that I've drawn in a brown colored arc segment that intersects the center point of the lower bolt hole. This is just to serve as a reminder that the forks will 'swing' through an arc, rotating around the upper bolt if you start making changes in fork angles The dimension of this arc will always be the same dimension you calculated earlier, like the 8.5-inches we're using in this example. You can visualize that if you start to make even relatively minor rotational changes you won't have enough 'meat' in the lower tree to drill the mounting hole. For this reason many builders use a 'thicker' lower tree and/or drill the lower mount hole slightly off center. You'll often see lower trees with several different mounting holes drilled to allow for adjustment in trail.

While we're talking about these mounting holes it'll be a good time to see how some of the old forks and trees and were set up. I took these pictures from the web and a few were sent in by board members. Keep in mind that there are slight differences in forks depending upon their vintage.

The shot below shows an old set of Harman forks. The integral handlebars were bent beyond repair so they were removed and the tube ends capped of. Notice how small the upper bolt hole sleeve is. This is just a piece of 3/4" x .125 wall tubing that uses a 1/2"-13 shoulder bolt into the upper tree.


Also note that the welds aren't the greatest in the world but back in the day this represented 'top of the line' work compared to what a lot of shops were putting out. The photo below shows a typical vintage tree. The mount gussets are welded on the centerline of the tubing which we now know is not a good way to this.



This one has been drilled for conventional handlebars. The offset looks to be around 1-inch. Even though this tree has 'old chrome' and a lot of rust I'm not so sure it's from a Harman fork set and might be somebody's early 'modification. I'm pretty sure this belongs to our old friend 'Jbfrmca' who is a frequent contributor to several of the various discussion boards.

On a 'modern' Spirder you need to be using 1.125x.25 wall DOM tubing for the bolt sleeves. This permits the use of 5/8" shoulder bolts which have a shear strength of around 28,000 psi compared to only half this for the older 1/2" fasteners originally used.

You'll need a good rigid design layout/welding table for the upcoming work. It doesn't have to be steel but it does need to absolutely flat. The dimensions of the table depends on what size forks you'll be building. A piece of 1.25-inch plywood floor underlayment ripped in half to 24"x96" work fairly well.

I prepare all of my Girder and Spirder control drawings to handle forks with up to 42-inches of extension over a stock Harley Springer leg which is typically 19.5 to 20-inches measured from the bottom edge of the lower bearing cup (top of the lower tree). This allows us to do design development on forks intended for up to 60-degrees of rake and up to 8-inches of upstretch. These long forks are usually used on Trikes but I've built a few for two-wheelers.

Here's a diagram of a typical layout table we might be using for a project like this.


To start the sketch for the various control points we layout a line running in the 'long' direction of the work table, about 9" or 10" down from the top edge, that is the edge away from the side you'll normally be working on. starting about 9-inches away from the right hand side of the table strike a line perpendicular to the baseline laid down previously. This establishes a line for the forks upper mounting bolt sleeve. Then you can draw another perpendicular line 70-inches to the left and this sets the control line for the rocker pivot point on a set of forks extended 42-inches over stock. With this system we can design and build almost any Girder or Spirder. The reason the handlebars are shown hanging off the end of the table is because they'll be having bends that are not planar to the main body of the forks.

A lot of the 'extra' space on the table will be used to mount various fixtures for tube coping, welding, bend tweaking and beer can storage among other things.

The sketch below illustrates where the fork mounting bolt sleeves are located on the front tubes. The upper sleeve will always be located as shown. The center point of this bolt is the master control point for all subsequent dimensioning.

The one place I've noticed a difference in the forks I've measured is in the length of what I call a 'riser section' in the sketch shown above. I have it dimensioned as being 4.5" from the upper mounting bolt center to the line of tangency of the first handlebar bend. I've seen this as short as 3-inches and as long as nine-inches. I think the reason for the difference is of course the vintage of the forks and perhaps it also may have been open to a buyers specifications where tank clearance was needed for instance. Also keep in mind that many of the Spirders I've dealt with are not authentic Harman forks or at least they were not part of his 'production' run and may have been built by some of his friends in the early days. It really doesn't make any difference as you can set that measurement to pretty much anything you need.

Notice that I show the lower mounting bolt sleeve welded hard up to the front tube, in alignment with the upper sleeve. You'll see that on most Harmans's this lower mount point is simply a bolt hole in the gusset plate that's welding between the front and rear tube legs. Changing the location of this point changes the trail of the forks. I personally prefer to run a Spirder with about 2.5-inches of trail so having the lower sleeve positioned as shown works for my design no matter what the extension. The gusset is still used but I didn't bother to show it in this sketch. We'll talk more about changing this location later in the article.

Unfortunately this sketch is just a 2-D representation so you can't see the side bends in the handlebars. I won't go into any details about the actual bending of the 'bar' part of the forks as there are just to many possible variations. Do some serious in-depth web searches and you'll find several different ways of bending the Spirder handlebars from a simple drag-bar to pretty complex 6-bend setups and even some straight pullbacks. I don't build Spirders with a traditional top tree and bolt-on handlebars. If you like that style then it's easier and cheaper to just build a regular Girder. For me part of the appeal to Johns style is in the integrated bars.

I started another page where I can start posting general Harman pictures. The first few show some good views of the handlebar end of the forks. Click Here to see the pics.


More to follow ..................................

Follow the links below to other pages:

Go To Page 1

Go To Page 2

Go To Page 3

Go To Page 4



| Main Page | Copyrights | Terms of Use | Warranty Disclaimer | Security and Privacy | Contact |

Copyright 2003-11, All Rights Reserved