AVLS Driver/Automotive Arduino Controller

A while ago, I swapped out the engine in my 2001 Subaru Impreza 2.5RS to one from a 2011 model. This new engine had a feature called I-Active Valve Lift System (AVLS) that allowed the engine to run one cam profile for smooth idling and another for a slight increase in power. This system is normally operated by the ECU. Because it's not available for the original ECU in my car, I decided to make my own.

Arduino Uno R3 mounted in a TuxCase lower enclosure and custom AVLS Driver Shield PCB boards.

I used an Arduino Uno R3 microcontroller for this project because of the low cost and ease of use when prototyping. Programming an Arduino is also very easy to learn. Next, I designed a simple circuit on a bread board using a 9V regulator to lower the voltage down from the car, and a Bipolar NPN transistor to switch the AVLS solenoids on. I chose a transistor because it works with a PWM signal, which is used by the solenoids.

AVLS Driver Shield populated.

I then designed a custom Shield to more permanently mount the components using EAGLE PCB by CadSoft. On the custom PCB, I added a second transistor to control future projects. I also added fuse holders, LED indicators, and Molex connectors for power and signal I/O. The boards were manufactured by OSH Park who do small orders of custom PCBs for hobbyists and prototyping.

Spacers.

AVLS Driver Shield attached to lid that also acts as a heat sink.

Lastly, I designed an upper enclosure to fit on top of a TuxCase to house all the components. I designed the AVLS Shield to use the top lid as an heat sink and mount. The components were waterjetted out of aluminium and just stacked to form the the upper enclosure.

Completed enclosure.

While I am not completely satisfied with the enclosure, it will serve its purpose for testing. One day, I would like to design a PCB with both Arduino and my components on a single board. It would have a water tight enclosure to allow it to sit in the engine bay.

This AVLS Driver can also control other devices and has room more inputs to create smarter control algorithms. Once I test this board and make some revisions, I will release the board design as open source so others may use it.

update - individual throttle body manifold for subarus

I explored making the manifold and throttle bodies as a whole system as opposed to an adapter for pre-made ITBs.

The manifold for this system would use a separate throttle plate assembly for each cylinder, with the injector ports on the manifold.

Multiple vacuum ports on the manifold would be plumbed to provide vacuum measurements in addition to IAC Valve use.

Designing a whole system instead of just a manifold would reduce the cost of the system overall. The cost of the manifold itself would not change, as adding ports does not add much in terms of manufacturing. Instead, the savings would come from the throttle plates. They are simpler to manufacture and are cheaper than purchased ITBs. The challenge with this, however, is my lack of knowledge in ITB systems.

individual throttle body manifold subarus

I have thought for the longest time that it would be interesting to put individual throttle bodies (ITB) on my Subaru Impreza. With the introduction of the Subaru BRZ/Scion FRS/Toyota GT86, I felt that there might be a market for a custom ITB intake manifold.

The first design I came up with was for a vertical mounted IDA carburetor style ITB.

IDA vertical ITB intake manifold.

The benefits of the vertical mount include a much straighter path for the air into the intake port of the head as well as a less cluttered appearance in the engine bay.

Top view. The port changes from the round opening of the ITBs to an oval opening on the Subaru heads.

This IDA style intake manifold could also be used in air cooled Volkswagens with Subaru engine swaps to maintain the appearance of a modified Volkswagen engine.

I also designed a horizontal layout manifold for use with SU side draft carburetor style ITBs.

SU side draft style ITB intake manifold.

The side draft style has the benefit of allowing a place for the stock fuel injectors, which are mounted on the angled flats near the bottom flange. The horizontal layout also points the air inlets (or the two heads on a flat 4 engine) towards the center, which would make it possible to have a single plenum and/or use the hood scoop present on Subarus to feed the inlets.

Bottom View.

These manifolds were designed around carburetor style ITBs so that it would also be possible to convert to standard carburetors. This may be desirable for a Volkswagen with a Subaru engine swap to maintain simplicity of the fueling system or to just be different on a Subaru.

I will be taking some measurements of the engine and further refining these designs.

custom snowboard stomp pad

Stomp pad fitted to the board.

This was a quick project requested by my brother for a snowboard stomp pad. He provided me with a vector outline of the word "fidelis" to make into a stomp pad.

A few things considered in the design were: enhancing grip, attaching it to the board, and maintaining its strength. The original request was to make a simple outline of the word. My thought was that this would be too fragile to stand up to being kicked--for example, in instances when it's used to knock the snow off your boot. This design would also be difficult to attach as there would not be enough surface area for adhesive. Additionally, the small pieces like the dot on the "i" and the "d" and "e" would require a template for alignment. My solution was to make it one piece by putting an offset border that would encompass the dots in the "i," then pocketing out the word.

The word was etched repeatedly to a depth of about 1/16" and then cut out with a laser. I used some leftover 1/4" rubber from a previous project. It was attached using double-sided mounting tape.

The stomp pad worked perfectly with zero slips throughout a day of snowboarding and my brother was very happy with the design.

redbull flugtag sf 2012

Entertaining the crowd in the "hangar" area .

This November I participated in Red Bull Flugtag with a team called the Slammin Tri Lambs. We were trying to build a plane that would fly as opposed to one for comic relief. The design was supposed to resemble a double-necked guitar with wings and picks for a tail.

Quick CAD work for the event application.

The plane was a short wing, long cord, canard aircraft designed to use ground effects to gain extra lift. For this project I did most of the initial design work, with input from the team, and all of the CAD work.

Version 1.

Version 2.

The plane was designed around easily available and cheap materials from the local hardware store. The spars and fuselage were made from a plywood and foam sandwich for strength. These were surprisingly strong and could easily hold up a person when cantilevered out. The wings on each side could be pulled off for transportation. All the complex shapes were cut with a water-jet.

Water-jetted foam pieces.

Water-jetted plywood pieces.

Nearing completion. Elevator moved to the back.

After some model testing, the team decided to move the elevator to the rear for ease of calculating the center of gravity (CG) of the plane. If we had more time, I believe we could have made the canard work. However, the plane was built in only about two weeks.

Painted and ready for some final touches.

McCovey Cove with the 30 ft launch platform.

End result.

Unfortunately, the plane did not fly. After the event, we suspected two mistakes that may have caused the plane to nosedive off the platform. First, we had the plane sitting on the cart too far forward, which nosed us down as the cart went off the platform. Ideally, we should have had the CG of the plane closer to the back of the cart, allowing the plane to separate from the cart more easily. The second mistake was using paracord for the the elevator controls. After watching video from the event, we noticed that the elevator moved about a third of what it should have been capable of, meaning that the paracord was probably stretching.

A video of the live stream can be seen here. We come on at the 1:36:00 mark.

"Z" for zoom

Original design concept.

This is a project I did for some friends who are about to have a baby. They wanted to decorate the baby's room with the alphabet, where each letter was made by a different friend. I chose the letter "Z" because I thought it was cool to choose the end of the alphabet. The letters were supposed to represent the maker, so I went with a car theme using some engineering to get the wheels to do a burnout.

Car assembly.

 The car was made from six layers of plywood. Each was laser cut and glued together.

Car Interior.

The two center layers were just in the roof area so that car would straddle over the bottom of the "Z". I put in a sub frame piece to support the body of the car.

The "Z" was made from four layers of plywood, also laser cut and marked. The bottom of the "Z" from the "OOM" was one layer thick and the slot allowed the space for the sub frame and axles to slide through. In the back, there were also cutouts to allow it to hang onto nails.

Painted and in the starting position.

 The car was painted with spray paint while the wheels were painted with Plasti-Dip to give it a rubbery feel.

ZOOM.

The axles are a 1/4" hardwood dowel riding on nylon bushings that were pressed in the car. PET spacers sit between the car and the "Z" to take up space and to allow the car to slide smoothly. When the care slides across, the axles ride against the bottom of the slot causing the wheel to spin--giving the appearance of a four wheel burnout.

Here is a quick video of it in action.

home still: complete assembly, cooler, and temperature control

I haven't posted anything about the still in a while, but the work has been steady. In fact, the construction is complete.

Completed Assembly.

With some preliminary testing with water, I realized that there was a slight flaw with the condenser design. It condenses the vapors well, but since the condensed drops just fall off the coil, the liquid is still very hot. I decided to add a cooler in order to bring the temperature down.

The cooler is the smaller tubes in the foreground.

The cooler is a simple tube and shell cooler. The cooling water is plumbed from the outlet of the condenser so that I can operate both with a single pump. With this configuration I'm able to get the condensate to almost room temperature.

Another component completed is the temperature control for the pot heater.

Temperature control unit.

The control unit is the same as some DIY Sous Vide controllers that utilize a PID heater controller and solid state relay. The control unit powers a water heater element which is connected to the pot with a sanitary fitting.

Heater element with detachable cord.

I did notice some inconsistencies in the temperature readings as compared to other thermometers, but this heater unit can hold consistent temps. I believe this may be due to a slow reaction time of the RTD probe.

Other experiments have also shown that it will be better to use the dephlegmator to control vapor temperature, so I may use the temperature controller to power the dephlegmator pump instead.

The experiments will just continue from here...