Variable length velocity stacks

[gropula]

Hi, all! I'm in the process of building a set of variable length velocity stacks. Inspired by mohawk's yoshi style velocity stacks and variable systems built by MV Agusta and Yamaha I've decided to do something similar on my '02 VTEC.

 

I bought a whole other airbox, with velocity stacks included. 6th gen comes with two long and two short stacks, so I have 4 of both. The long ones are repurposed to be fixed at the bottom, this way the bore diameter is increased and the length of the stack is reduced. The short ones will be kept stock and made to move up and down. When down they will insert into the bore of the bottom ones. The length and bore diameter in lowered position will equal the original long stack. When in upper position the upper stack will move above the lower stack. In picture below is the initial idea. I have much longer bolts now that will enable the upper stacks to move up to fully clear the bore of the lower ones. The bolts have threads only on the bottom. The upper stacks have these hollow cylindrical pieces inserted in bolt holes. This enables the upper stacks to slide up and down on the bolt.

 

 

I cut off the radius part from the longer stacks and made them flush with the bottom of the airbox. I used dual compound epoxy to glue them. After curing I used a file and sandpaper to match the holes of the bottom of the airbox to throttle body bores. That is the bottom part done. 

 

 

For the movement I decided to repurpose the flapper assembly. I glued the vacuum diaphragm actuator upside down in the middle of the airbox. It clears the fuel rails and connectors beneath with no problems. It has about 12mm of travel, but I need around 25-30mm of travel so that the upper velocity stacks fully clear the bottom ones. Because of that I decided to use a lever system. This way the movement is reversed and the range is extended. I drew a schematic of how it will be done, only one side is drawn for simplicity, but it will be simetrical. The lever system will be mounted on a metal U shaped profile (red), bolted to the bottom of the airbox. A bolt (cyan) will go horizontally through the U profile and the lever arm (green) will rotate around the bolt. The lever has rotational movement, while the actuator and the movable stacks have linear movement (yellow arrows). To solve that problem I will simply drill the holes in the lever arm that are wide (the ellipsoidal shape of the holes in the schematic). There will be a bracket to connect the front movable stacks together (cyan). Same in the back. The bracket will slide on long bolts. I intend to intercept the VTEC signal and use it to activate a relay, which will activate the flapper solenoid. The original signal for the flapper activates at 5500rpm which is too early. 

 

 

 

I will update the thread with more details when I take more pictures. I'd like to know your thoughts on this mod, if you have any ideas to improve the design feel free to share them. When I finish the build I intend to test the stacks. I can assemble a stockish setup with 4 short original stacks. That will be the baseline. Then I can compare it to the modified setup. I will try to bring it to a dyno. If I can't get it dynoed I have an idea to do 3rd gear pulls from 3000rpm to limiter. Time every 1000rpm segment from video or from Rapid bike logger and compare gains/losses. 

[Mohawk]

I've thought about doing this & on the 5th gen using the flapper valve vacuum which opens at 5krpm IIRC 👍

 

Also thought about standoff fi using the same trigger to switch feed from in TB injectors to above velocity stack ones. Not sure how much it would gain but an idea. 

[V4 Rosso]

11 hours ago, gropula said:

The lever has rotational movement, while the actuator and the movable stacks have linear movement (yellow arrows).

The stacks don't need to have linear movement, do they? Or can you otherwise not get enough clearance?

[gropula]

The movable stacks slide on long bolts straight up and down, so they can't follow the arc alongh which the lever arm moves. By having stacks slide on bolts they are always in the right position, and the lever and pivot point can be positioned by trial and error. The U shaped profile will also have holes bigger than bolts so it can be moved to the right position without having to drill holes at the exactly the right place. Since this is all made by hand I need big tolerances on the moving parts. Also, by having it slide on bolts more parts are held in securely, I can't have stuff falling apart in the airbox and getting sucked into the engine. It could be done so that the lever holds the bracket "in the air" without the support of the long bolts, but then the bracket and the lever should be one part, or solidly connected. That would require much more precision in positioning of the actuator and pivot points, also the length of the lever should be exact. 

[gropula]

Got an update on the build. The weather improved and I've been able to test the various positions of the stacks to determine the best performing lengths and corresponding rpm ranges. Unfortunately no dyno time, but a good test was performed on a flat abandoned road. This was the setup: 

Phone video recording the tachometer at 30 fps

Straight, flat road, 2 runs forward and 2 runs backwards with each stack length for good averaging

3rd gear WOT pull from 3000 to 12000 rpm, from about 40 to 150 km/h with my gearing (15/44)

Stacks at 4 different heights, supported at various heights by a piece of vacuum hose

d=0 - upper stack fully inserted - length equaling long factory stack

d=10 - upper stack lifted 10 mm from the base, creating a longer stack with a very small gap in between the upper and lower stack

d=20 - upper stack lifted 20 mm from the base, creating a dual stack with a sizable gap in between the upper and lower stack

d=30 - upper stack lifted 30 mm from the base, as high as it can go without interfering with the air filter, basically disabling the upper stack and breathing from the short lower stack

 

I've used a video editing software and excel, determining the frame number for each time the needle went over the x1000 rpm notch. Each frame lasts 0.033s at 30fps. For example, the difference in frame number for 5000 and 4000 multiplied 0.033s gives time elapsed to accelerate from 4k to 5k rpm.

 

 

Averaging and comparing the times gave results. The results for 3-4 rpm are hard to read because of big vibration that occurs when WOT at 3k so I've decided to discard them. Also going WOT at 3k is not something I do. The d=20 stack doesn't excel at any rpm range so they're discarded as an option. d=10 has slight advantage at 5-7k rpm, but not enough to be worth anything. Also, as expected, they totally choke the engine from 9 to 11k rpm, losing 5% over d=0 and even more compared to d=30. 

 

Stock long stack length works best for 7-9k rpm range, also expected as that's where maximum torque occurs, 2% better over others. Honda obviously knows what they're doing so that's why I had one configuration in stock long length. d=30 works better in the range of 10-11k rpm where maximum power occurs, 3% better than d=0. Nice to see, that means maximum power has increased with this mod. 

 

I've decide that d=0 and d=30 will be the two positions and they will move at 9k rpm. That way I have the maximum torque from 7-9k and maximum power from 10-11k rpm. I ordered an RPM switch from ebay to do the electrical switching at 9k.

 

Interesting to see that the engine totally chokes from 11-12k and that there is very little difference between stacks. It's the cams limiting the performance at extreme rpm, not stacks. Shift at around 11k.

 

All in all I'm very satisfied with results and measurements. Even though some of the differences are small, barely measurable, they are what is to be expected by theory of the stacks so I trust that even 1% differences are accurate, while 2-5% differences are significant. Averaging 4 runs with close results makes results trustworthy.

 

 

 

 

[gropula]

Made a bit of progress. Video of stacks

Just need to fix this securely and wire up the RPM activated switch. 

 

I made a more understable graph from the measurements. The first graph is the force that accelerates the bike at the measured rate. That is force remaining after all friction and resistances are overcome. It resembles the torque graph. The other graph is power. All calculated from gear ratios, calculated wheel diameter and assuming bike+rider weight is 350kg. Don't mind the absolute values, the differences are the point. Only showing the two best options, stacks fully up and fully down. The final moving assembly will take the best of both.