S2's Build Thread


Lifetime Supporter

Bellcrank bushing (left) and fore chassis billet upright weld joint (right)

There has been some speculation regarding two potential weak points in the rear chassis; the aluminum bellcrank bushings and the weld joints between the 2” x 2” tube that supports the bushing and fore chassis billet upright. Endurance cars provide useful data points because they are subjected to more abuse than most cars, certainly mine, will ever see. After 180+ hours of intense racing (big slicks, large downforce, 1,200+ pound springs, being rubbed by many cars, off-track excursions, etc.) a routine inspection of one such car noted stress cracks at the aforementioned weld joints. The fix requires grinding and rewelding the joints and adding gussets. The good news is that the rocker bushings had no apparent issues.

There are two straightforward ways to mitigate potential issues in this area; brackets and gussets. Several cars, including the endurance one mentioned above, feature brackets that place the bellcranks in double shear. Whether the rocker bushings would have experienced wear without the brackets isn’t clear, but I fabricated a set in a previous post. I had been planning on adding gussets for some time and now that I’m about to modify a bunch of chassis tubes to fit my exhaust it made sense to do it now.


Temporary spacer replacing bellcrank, gusset tacked into place and preheating with torch

Welding a thick aluminum gusset to the 2” x 2” billet upright generates a lot heat which will melt nylocs, cook bearings and has the potential to warp the chassis. The first step was to remove the engine and everything that’s heat sensitive from the engine compartment. Once that was done, I reinstalled the rear chassis brace and rear roll cage legs with plain nuts (nylocs would have melted). Because the bellcranks have bearings I couldn’t leave them in place so temporary spacers were fabricated to allow the bellcrank support brackets to be torqued. The large hunks of aluminum also act as a heat sink.

Once everything was torqued, a temporary frame was fabricated to further lock everything in place. Four 1/4” thick steel plates were fabricated and bolted to the suspension suspension mounting points and 1” square steel tubing was used to connect them.


Temporary steel brace tying the fore chassis billet uprights together

I fabricated the gussets from 1/2” 6061. The hole reduces weight and the amount of heat required to weld it in place. I considered making the gussets larger, but Abe didn’t want to put that much heat that close to the bushings. The billet uprights and gussets were preheated with a propane torch. The amount of heat dumped into the chassis was impressive… even the aft billet upright got hot!



Gusset welded


Lifetime Supporter

Given my power level and the length of the primaries, I decided to step the headers from 1-7/8” (green blocks) to 2” (yellow blocks). This required the headers to be redesigned and the new version rubs four of the 2” x 2” chassis tubes. D'OH!


Top and slanted tubes have been scalloped

The solution was to scallop the top and slanted chassis tubes. As can be seen in the picture above, the cut in the top tube is larger than the slanted tube. While 90-degree cuts would have been easier, 45-degree cuts were used to prevent stress risers. Cutting the tubes was tedious and after some experimentation, this is what worked best:
  • Holes were drilled in the 45-degree corners.
  • An angle grinder with a 4-1/2” abrasive cutoff wheel was used to plunge cut the vertical face.
  • A Dremel with an abrasive cutoff wheel was used to plunge cut the 45-degree angle from the previous cut up to the holes.
  • A reciprocating air saw with a Bosch T227D blade was used on the longitudinal cuts. 8 TPI seems too aggressive for aluminum, but that’s what the blade was designed for and it works well.
  • The Dremel and a small right-angle grinder with a 2” abrasive wheel were used in areas where the other tools wouldn’t fit.
  • A combination of a 6” orbital sander, a right-angle grinder with a 2” sanding disk and a deburring tool were used to clean up the edges.
The tube has 1/8” walls and since I was removing material I decided to oversize the inserts. Fortunately, McMaster stocks 3/8” x 1-3/4” tall 6061 bar, so I didn’t need to trim the height. Scrap pieces of round and right-angle steel were used to fabricate male and female dies. They had no provision to ensure alignment so the male and female die had to be carefully aligned with a small machinist’s square for each bend. Given that only eight bends (nine if you include the test bend) were needed, they worked well enough. I added duct tape to the three locations that come into contact with dies to reduce marring.


Male and female bending dies


The dies where placed in a hydraulic press and a digital angle finder was used to determine when 23.5 degrees was reached. Since I wanted a 45-degree bend and the aluminum is being pressed into a right angle die each side should be 22.5 degrees plus one degree for spring back.


The dents made by the dies were sanded out of the front and back sides to reduce the potential for stress risers. In addition, the bending process causes the sides to deform and these bulges must be sanded flat for the piece to slide into the chassis tube. One that was done, the ends were rough cut on the bandsaw.


The edges were then trued up with a 3/4” end mill. Fixturing was as simple as dropping the piece in the vice and tightening it. This ensured a perfect fit with the inside of the chassis tube.


We anticipated that once the piece was inserted into the opening we wouldn’t be able to move it into place or remove it to trim it so we drilled and tapped two 1/4”-20 holes which will be used later to mount a heat shield. However, once we had inserted the piece we had a hard time using the bolts to move it. Abe then had the idea of using a slide hammer so he welded a 1/4"-20 screw to a 5/8” nut to adapt it to our tapped holes. It worked perfectly.


Pre-weld fit up is excellent


Fully welded


Lifetime Supporter

One of the challenges with high-boost supercharged engines is belt slip. Harrop offers multiple pulley diameters, but they only have 8-ribs which doesn’t provide enough surface area. Upgrading to a 10-rib belt requires a bunch of changes including a new supercharger pulley. Harrop’s pulley has a large offset and multiple recessed mounting holes, not something you’d want to reverse engineer.

Fortunately ZPE manufacturers pullies for a broad range of superchargers. They offer 8-rib Harrop pullies in nine different diameters ranging from 60 to 85 mm. ZPE was willing to do some non-recurring engineering (NRE) to make me a custom 10-rib pulley with a GripTec Micro finish. GripTec Micro is a patented micro-ablation machining process that creates multi-directional ridges and valleys that increase gripping force and provide escape ducts for trapped air and debris. As can been seen in the picture below, the grooves are micro.


Note that despite increasing the number of ribs by 25%, the mounting screws are still proud of the belt.

I need to pull the engine to install the pulley, so hopefully there aren’t any surprises!


Lifetime Supporter
When I dropped the engine back in the automatic tensioner for the accessory belt hit the 2” x 6” chassis tube. This happened because I increased the supercharger belt from 8 to 10 ribs and I moved the A/C compressor and alternator to a 6-rib belt running off of a pulley mounted to the front of the super damper. I had two options: (1) modify my elegant design (if I say so myself LOL) which would likely require me to swap the automatic tensioner for a manual one (ugh) or (2) I could scallop the chassis. I decided to do the latter.


I needed to cut a semicircular hole and the easiest way to do that was with a hole saw. However, the geometry of the cut located the pilot bit 80% into the edge of the 1/4” thick tube and 20% hanging out in free space, probably the worst possible location. The solution was to fabricate a drilling jig from scrap 1/4” steel. Two screws mount it to holes tapped into the 2” x 6” and the bottom hole retains the pilot bit. The jig stays in place because the hole saw was only plunged 5/8” and the bottom edge is stiff enough to keep it in place. Once the semicircle was cut a cutoff wheel was used to cut the underside of the tube.


An insert was fabricated from two pieces of 1/8” 6061 and welded into place. Plenty of clearance was provided to mitigate heat soak and to prevent binding during engine installation and removal.



OK, I think that now, short of some other unforeseen issue, work can begin on the exhaust!


Lifetime Supporter
I mounted the transaxle cooler to the custom tail sub frame (more about that later). I used the same vibration isolators that I used on the intercoolers and mounted them with four tabs that I designed and laser cut from 0.104” mild steel. The upper two were bent and tacked to the tail sub frame and the bottom two were tacked to a bracket that I fabricated from 1/2” mild steel tube — I’m glad that I didn’t use 4130 because it was difficult to shape with a manual bender!


The M6 screws were swimming in the bushings so I replaced them with M6 shoulder screws with an 8 mm shoulder diameter. The shoulders were too tight so I opened the bushings with a drill bit.



The next step is to mount the thermostat and fabricate the lines. When I get around to fabricating the new tail, I’ll mold a fiberglass duct to the underside of the tail which will transition to an aluminum duct around the cooler.
Last edited:


Lifetime Supporter
Chris, thanks for the compliment. It's a lot of work!

Fabrication of the 180 cross-under headers has begun. The first step when routing anything, especially hard tubing, is to establish the start and end targets. Setting up the starting points was easy. I simply bolted cast stainless steel exhaust flanges from Ultimate Headers to the heads. The merge collectors are the end targets. Since mine are located as low as possible I clamped a piece of 3/4” plywood to the bottom of chassis, 3D-printed spacers and clamped the backsides of the merge collectors to the plywood.


3/4” plywood clamped to underside of chassis and rear of merge collector affixed via 3D-printed spacer and clamp

The fronts of the merge collectors were held in place by a long 8-32 screw passing through the gap between the primaries — I didn’t have one that was long enough so Abe welded to 3” long screws end to end.


Note how far the bellhousing flange extends below the Daily dry-sump oil pan. This is what provides the space for the primaries to cross under the engine.


Upper half of left side. Several of tubes have been tacked, the horizontal one is clamped


This picture illustrates why it was necessary to scallop the chassis tubes. A tack-welding clamp in middle of outer primary.

The icengineworks tack-welding clamps are very useful. They allow the orientation of the tubes to be clocked while ensuring that they are concentric. When everything is where you want it you tighten the nuts and nothing will move. They have large openings that allow the tubes to be tacked. I have two sets of clamps, 1-7/8” and 2”, to support the stepped primaries. We were not able to get either to work at the transition joints so we took one of each apart and created two clamps with a different CLR on each side. It worked well.

Removing all of the marks made by the bending dies is a lot of work. I used the tube polisher discussed in an earlier post, starting with 120-grit sandpaper and finishing with a Scotch-Brite surface conditioning belt. The trickiest parts are the outside radiuses because any lateral pressure causes the belt to pop off of the tracks and fall on the floor.

There is a lot more fabrication required to finish these up.


Lifetime Supporter
I just finished a neat part that would have been very difficult to machine without a rotary table. I thought the following accelerated fabrication video might be of interest (there are basic comments in the lower right corner). I sure wish I could fabricate parts that fast!

The part is a spacer that sits between the custom front accessory mounting plate and the automatic tensioner. It accomplishes the following:
  • Aligns the tensioner/pulley with the accessory serpentine belt.
  • Provides a boss that indexes the inside of the tensioner’s mounting hole to ensure no lateral movement.
  • Prevents the tensioner from rotating by capturing its anti-rotation post.
  • Enables the tensioner to be clocked.
Since the entire serpentine system is custom, I wasn’t sure of the belt length so it was important that I could easily clock the orientation of the tensioner. After some thinking I came up with the following solution which enables it to be clocked 360 degrees in five-degree increments:
  • The mounting plate has six sets of mounting holes located 30 degrees apart
  • The spacer has three sets of tapped mounting holes located 20 degrees apart
  • The spacer has two anti-rotation holes located five degrees apart when the spacer is rotated 180 degrees.
When I need to measure something accurately, I often 3D print mini tests to validate dimensions. As can be seen below, it usually takes multiple attempts to get it near perfect. In this case I was dialing in the ID of the mounting hole, the OD of the anti-rotation post, the center-to-center distance of the anti-rotation post and the mounting hole, the angle of the anti-rotation post and the arc of the tensioner.



Lifetime Supporter
Dog-box transmissions are about rapid shifts with minimal pause in power production— the exact opposite of how what you’d expect from a synchronized box. That’s fun on the track, but it’s typically going to necessitate the use of a clutch that will be a nightmare on the street. The combination of a dog box and a cerametallic clutch is fairly binary — it’s either engaged or disengaged with very little slipping between the two states. This is of particular concern for me because the driver, as opposed to GCU, is responsible for shifting from neutral into first or reverse. That doesn’t happen much on the track, but it’s constant on the street.


Carbon clutches are more slippable and I looked at one from Tilton, but it wouldn’t fit the bellhousing. After talking with Agile Automotive we decided to purchase a triple carbon/carbon clutch from RPS. Agile worked directly with RPS to develop their proven carbon/carbon clutch to work in this application. This involved tweaking the geometry to fit the ST6-M’s bellhousing and adding a shim between the piston on the slave cylinder and the throw-out bearing. Many carbon clutch packages utilize metal-on-metal friction surfaces on the flywheel, floater disc and pressure plate. The RPS package utilizes carbon on all of the friction surfaces (a F1 trickle-down technology), hence the carbon/carbon moniker. The package has many benefits for the street and track with the only downside being cost:
  • Greater than 1,200 ft/lbs torque capability
  • Smooth slippable engagement
  • Minimal pedal effort compared to cerametallic clutches of the same capacity
  • Massively improved lifespan over cerametallic clutches with the same number of discs
  • Improved reliability due to all friction surfaces being carbon eliminating glazing, warpage, and cracking common with other clutch types
  • Very light weight
  • Fully rebuildable

On pump gas my engine produces 860 ft/lbs at 2,000 RPMs and over a thousand at 4k RPMs, so when running E85 I’ll push the torque threshold.

The complete unit including the flywheel, pressure plate, clutch and mounting hardware weighs 26 lbs. All of the weight is in the flywheel which will make driving on the street more enjoyable. In comparison the flywheel for my Ricardo is 14.3 lbs and the pressure plate and clutch are 31.6 lbs so it’s >46 lbs when the bolts are added.

Going forward Agile will recommend this package for their SL-C and Aero builds for street, road racing and endurance racing. Plans for future development include a longer piston for the slave cylinder to eliminate the shim and have an ultra-light flywheel for pure race applications.
Last edited:


Lifetime Supporter
Crossing four of the exhaust primaries under the engine is only possible due to the low-profile Daily Engineering dry-sump pan. It includes a -10 AN adapter which replaces the stock oil filter. While it’s massively smaller than the stock unit, it was too close to the exhaust. In the picture below, the blue tape is protecting the oil pan (it’s billet and I don’t want to scratch it) and the oil adapter is the threaded AN fitting. The hole in the bell housing (white arrow) mounts to the stock oil pan which illustrates how much smaller the the Daily pan is vs. the OEM pan.


The solution was a shorter adapter from Kurt Urban. I might need to modify a socket to tighten it, but I won’t know that until I finish tacking the exhaust and remove it for final welding.


Steven Lobel

Crossing four of the exhaust primaries under the engine is only possible due to the low-profile Daily Engineering dry-sump pan. It includes a -10 AN adapter which replaces the stock oil filter. While it’s massively smaller than the stock unit, it was too close to the exhaust. In the picture below, the blue tape is protecting the oil pan (it’s billet and I don’t want to scratch it) and the oil adapter is the threaded AN fitting. The hole in the bell housing (white arrow) mounts to the stock oil pan which illustrates how much smaller the the Daily pan is vs. the OEM pan.

View attachment 114919

The solution was a shorter adapter from Kurt Urban. I might need to modify a socket to tighten it, but I won’t know that until I finish tacking the exhaust and remove it for final welding.

View attachment 114920

Mine got longer when I added the AN adapter for the larger line.
2018-03-13 18.36.56.jpg


Lifetime Supporter
To stiffen the chassis I’m going to use four indexable links to make the bellhousing a stressed member that triangulates the fore billet chassis uprights. This is the same approach that I took with the aft billet chassis uprights as shown below.


To accomplish this, four of the bellhousing-to-engine bolts will be replaced with longer bolts to accommodate rod ends and misalignment washers. The first step was to design a plate (orangish) from 1/8” 4130 to place the rod ends in double shear.


I added three M10 1.5mm bolts and aluminum spacers (green) to each side to stiffen the double shear plate. This required me to remove the bellhousing and drill and tap it. It’s a beautiful piece of billet so I was careful not to f’ it up. I tried to fixture it via the flanges, but that obstructed the areas that I needed to drill. I found the best way to accomplish it was using long bolts through the center as shown below.


Having access to a large mill with a DRO made this a straight-forward process.


Everything worked as planned. The next step is to fabricate the spacers, double shear plates, indexable links and the brackets that mount to the billet chassis uprights.


Lifetime Supporter

In the last post I removed the bellhousing to machine it so the next step was to reattach it to the transaxle. Should be easy right?

Well I couldn’t get the Loctite residue off of the studs. I tried using acetone with a wire brush and a metal pick, but that didn’t do much. I tried brake cleaner to little avail. Apparently chlorinated brake cleaner works better, but the good stuff is illegal in Massachusetts. Part of the issue is that acetone, and I assume brake cleaner, are less viscous than water so it was difficult to keep the solvent in contact with the residue. I considered removing the studs to let them soak in the solvent, but removing seven studs from cast aluminum didn’t seem like the right approach.

After some research I discovered Loctite Chisel which is designed to remove Loctite residue, gaskets, paint, etc. It’s hazardous stuff and I couldn’t find it locally, at McMaster or on Amazon. It’s available from Pegasus Auto Racing, but UPS applies a hazardous shipping charge of $54.50 on top of overnight shipping charges.

Back to research. The primary active ingredient is methylene chloride. I noted that a most paint strippers indicate “non-methylene chloride” so I pulled an old can of aircraft stripper out of the cabinet and in big bold letters was “WARNING: Contains Methylene Chloride.” It’s nasty stuff which can blind you, give you cancer, kill you, etc.

Preemptive statement to prevent verbal barrage… Mom, wife and daughter, I wore gloves, goggles and did everything in front of an open garage door.

Aircraft stripper is viscous and clings to vertical surfaces, so I applied a small amount with a Q-Tip, let it sit for 15 minutes and then removed the residue with a combination of the pick and wire brush. It worked extremely well. The studs have a shiny black coating and broken Loctite is white, so it was easy to determine that all of the residue was removed.

I then used acetone to ensure that all of the stripper was removed. Albins recommends Loctite 7649 primer before applying Loctite thread locker. I had never used primer before and I wonder if it made the residue more difficult to remove.

Howard Jones

Loctite can be softened with heat and then removed with a wire brush. I use a propane bottle with a standard torch. A few hundred degrees does it, no need to get it red hot on anywhere near that.


Lifetime Supporter
Howard, a torch will work. However, standard Loctite needs at least 325F and the high temp stuff needs at least 500F. At 500 you’re almost half the melting point of aluminum and it’s hard to be precise with a torch. While I’m not worried about melting anything, I didn’t want to degrade the shiny coating, the temper of the stud or tapped aluminum hole or the anti seize or whatever might be on the other end of the stud and the aluminum. For this reason, I’ve used a heat gun and piece of scrap aluminum with a hole in it as a heat shield to not heat the surrounding area. My heat gun broke and IMO the stripper was best solution for this situation... now that I know what to use:)


Lifetime Supporter
I’m going to use the door actuators pioneered by Peter and implemented many times by Allan and others. Allan has two videos (here and here) that describe what to purchase and how to do the installation. I made a couple of changes:

The actuators are powerful and exert a lot of tension on the door when closed. The original design pivots on the threads of a 1/4”-20 flathead screw tapped into a 1/4” thick piece of angle aluminum. I replaced that screw with the largest flathead shoulder screw that would fit in the base plate. I then added a high-load oil-embedded bronze sleeve bearing. As can be seen below, the bearing projects 3/16” below the bracket which isn’t an issue because the self-lubricating plate is 1/4” thick.


Right-angle brackets, bottom plates and flathead shoulder screw

As can be seen below, the shoulder extends past the bearing so it’s important to use a 3/8” (shoulder diameter) rather than a 5/16” (thread diameter) washer to ensure that the nyloc binds on the bearing rather then the shoulder. Washers have one side which is nicer than the other and I usually face the nice side towards the nut because it looks better. However, since there may be some movement between the washer and the bearing I oriented the nice towards the bearing.


The shoulder screw and nyloc collided with the bracket so I machined a hole in the bottom of the bracket. Since I had the end mill chucked up I decided to machine 12 additional holes to lighten the bracket.


Bracket with 12 weight reduction holes (left). Bottom of bracket with slot to accommodate the shoulder screw and nyloc (right). Two weight reduction holes where subsequently added to the bottom.

The part numbers are as follows:
  • High-Load Oil-Embedded 863 Bronze Sleeve Bearing, Flanged, for 3/8" Shaft Diameter and 1/2" Housing ID, 1/2" Long (McMaster #2938T7)
  • 18-8 Stainless Steel Shoulder Screw; 3/8" Shoulder Diameter, 5/8" Shoulder Length, 5/16"-18 Thread (McMaster #92944A132)
Allan recommends a 12” long piece 6” x 6” x 1/4” right-angle aluminum and cutting 3” off of one of the legs to use as a base plate. Instead I purchased a 12” piece of 6” x 3” x 1/4” right angle aluminum and a 12” long piece of 3” x 1/4” flat aluminum. This reduces the amount of cutting and results in perfect edges. I also drilled a 2"-5/8” hole to lighten the bracket.

Allan drills and taps the actuators’ rods to mount the 3/8”-24 rod ends. This is a bit tricky because the cross section is small (see picture below) and since the hole is going into the tip it’s hard to fixture in a mill or drill press. Instead, I welded a Grade 8 hex nut to the tip. I used a “high” (also know as a “tall”) nut because I wanted more thread engagement than what a standard nut provides (McMaster #90565A360). To accomplish this I removed the aluminum C-channel cover to provide access to the weld joint and to facilitate removal of the grease.


Tall 3/8”-24 nut welded to the tip of the rod (left) and 3/8”-24 rod end on top of rod (right)

Peter has designed an emergency release which I will test once the actuators are installed. I am also looking into a more sophisticated motor controller.
Kurt, in a previous post I had purchased Ferrari 458 and Ferrari California headlights to see if I could make them fit -- nope. I've tried a bunch of OEM lights and nothing seems to fit. My conclusion is to 3D print a smaller custom bucket, cover it with carbon fiber, trim the stock polycarbonate lens on all sides and add a dual beam projector and a switchback DRL. Since the lens is stock, I don't need to change any of the curves on the body. Rather, just fill in the gap.
Have you tried the jaguar f type headlights? , I got some and will give it a shot , will post pictures when I will get to that point


The actuators is next in my list of things to do, might just have to steal your mods because it’s about the only thing I can imitate that you have done!


Lifetime Supporter
might just have to steal your mods
Kyle, Peter did all of the hard work. Hopefully I'll be able to get the motor controller I bought to work.

Have you tried the jaguar f type headlights?
Hector, funny that you ask. I've had one sitting on my bench for months. Your post prompted me to cover it with packing tape and wax it (a little sloppy, but good enough for a basic check). I'll make a fiberglass test piece this weekend and post the results.

I bought a very cheap set on Ebay with minor tab damage , I also found a place that got me new pig tails and was able to get a wiring diagram. fingers crossed, they will be beautiful if we can make them fit. Thanks for doing the leg work, I am going to try super hard to make them fit . Sorry to distract you form other projects :cool: