Resume
Professional resume and experiences of Sahib Virdee.
Professional resume and experiences of Sahib Virdee.
Information about Sahib Virdee, Mechanical Design Engineer.
[TOC] What is SARIT? SARIT (Safe, Affordable, Reliable, Innovative Transport) is a fleet of enclosed electric three-wheelers developed by Elvy Mobility and deployed across institutional partners in the Toronto area. The vehicles are designed for short-range urban mobility — compact enough for pedestrian zones, enclosed for year-round use, and quiet enough to operate in environments where larger vehicles would be intrusive. Active deployments include the Toronto Zoo, York University's main campus, and several other pilot sites. The research program at York University's BEST Lab sits at the intersection of the hardware and the field. Vehicles are assigned to partners, used daily, repaired when things break, and progressively outfitted with technologies developed by the team — pedestrian detection, telemetry, reverse safety systems, and custom accessories. It is applied engineering work: real constraints, real users, real consequences when something fails in the field. My Role I joined the SARIT project in May 2025 through York University's LURA (Lassonde Undergraduate Research Award) program, supervised by Professor Andrew Maxwell and day-to-day by Victoria, the project coordinator. I worked primarily on the mechanical side of the research fleet — designing, fabricating, and iterating on physical hardware that had to be functional, manufacturable, and field-ready. Over the course of the project I moved through several distinct work streams: camera mounts for the pedestrian detection system, trailer hitches for cargo deployment, a LiDAR sensor mount, and eventually the power delivery system underlying all of the vehicle's added electronics. The work was largely self-directed within each stream — I would identify the problem, research solutions, build and test, and hand off documentation to the team. The fleet lives outside. Vehicles get rained on, driven over curbs, and handed to users who do not read manuals. That context shapes every design decision. The Team The BEST Lab SARIT team is split into mechanical and technology subgroups, with a separate team from Elvy handling manufacturing and fleet logistics. I worked most closely with Usman Ali (project lead, mechanical), the mechanical sub-team, and periodically with the technology team on integration work where the mechanical and electrical systems met. One thing that made the project interesting: because the vehicles were parked in the Bergeron Engineering courtyard and I spent a lot of time working on them in public view, people would stop and ask what I was doing. I ended up giving a lot of informal demonstrations — letting curious students and passersby take a SARIT for a short drive around the courtyard. One of those conversations turned into a recruitment: someone who drove it around asked if there were openings on the team, I saved their contact, and a few months later when the team expanded I reached out and they joined as a software member. The Vehicles The SARIT is powered by a 48V LiFePO4 battery (96 Ah on most research units) driving a 3-phase BLDC motor controlled by an ASI BAC2000 motor controller. The enclosed aluminum frame includes doors, a locking trunk, a covered roof, and a small instrument cluster. Stock vehicles have essentially no accessory power infrastructure — every addition the research team makes has to be engineered from scratch. The fleet at York includes vehicles of varying ages and configurations. Some have the newer head-up display; others are older units without one. Some have the technology team's power distribution hardware installed; others run stock. Part of the mechanical team's job was maintaining continuity across this heterogeneous fleet — making sure parts, mounts, and fasteners were interchangeable wherever possible. ---
[TOC] Background The SARIT fleet was identified early as a potential last-mile delivery platform. Getting a trailer behind a three-wheeled enclosed electric vehicle introduces genuine engineering challenges — hitch height, tongue weight, multi-axis articulation, and the constraint that you cannot significantly modify the SARIT's frame. A previous hitch design had been developed by an earlier team and used on some vehicles, but it had a known nose-down tilt issue and had never been documented in a way that made it easy to build or reproduce. My task starting in late May 2025 was to design a new one from scratch. Wike Trailer — Design Process The cargo trailer selected for the York University use case was the Wike cargo trailer — a lightweight, open-frame unit rated for approximately 150 lbs (70 kg). Before designing the hitch I built a full 3D model of the Wike in SolidWorks from physical measurements, since no manufacturer CAD existed. This model became the reference for all subsequent hitch geometry. v1 was a first pass based on research into hitch types — pintle, lunette, and ball-mount. I selected ball-mount based on the project's need for multi-axis rotation (the SARIT turns sharply and the trailer needs to follow without binding). The v1 design went straight to the machine shop for a Design for Manufacturability review, which produced the most useful feedback of the entire process: sharp internal corners that looked fine in CAD would have required expensive secondary operations to produce. I incorporated that feedback into v2 the same week. v2 added FEA simulations in SolidWorks and a first Bill of Materials. The simulations flagged thin cross-sections near the frame attachment points that were not obvious by inspection. v3 was the version I would have been comfortable fabricating. I completed a comprehensive BOM at this stage — part numbers, quantities, material specifications, estimated masses, and supplier sources. The BOM discipline here was new to me; it is not just a parts list, it is a procurement and assembly document, and a small error propagates into the wrong parts arriving or budget overruns. v4 resolved remaining geometry issues and transitioned the SARIT-side bracket material from steel to 6061 aluminum. The weight savings were meaningful given the trailer's light load rating, and switching to aluminum also eliminated the galvanic corrosion risk between the steel tube and the SARIT's aluminum frame. The machine shop confirmed 6061 availability and cost through a local metal supplier (Mr. Metal, Etobicoke), and the final BOM was built around their pricing. Fabrication The fabrication week (July 14–18, 2025) was the most satisfying stretch of the whole project. I machined the square tube components myself on the mill at the Bergeron machine shop, and cut the steel bracket plates on the water jet. The machinist on duty (Gurjit) belt-sanded the water jet edges and chamfered them — a finishing step I hadn't asked for but that made the assembly look considerably more professional. Drilling the mounting holes into the SARIT frame required a 3D-printed jig to locate the holes accurately. I also printed nut carriers to hold fasteners deep inside the hollow aluminum frame sections where a wrench couldn't reach — a constraint that only became clear during physical test-fitting, not in CAD. The finished hitch was assembled, mounted, and load-tested on the orange SARIT. It cleared all frame members with the trailer attached and tracked correctly through turns. Impact Trailer — Second Hitch Shortly after the Wike hitch was complete, the team acquired a MotoAlliance Impact trailer — a heavier-duty unit rated at 1,500 lbs, roughly ten times the Wike's capacity. The design process for the Impact hitch was significantly faster because the workflow was already established: physical measurement, SolidWorks model, FEA, BOM, machine shop review, fabrication. The key difference was scale. The Impact trailer is substantially heavier empty, which required thicker tube walls in the SARIT-side bracket and larger fastener sizing (3/8" rather than 1/4"). The tow ball specification also changed — a 2" receiver versus the Wike's smaller coupling. Both hitches were completed and deployed. As of early 2026 the orange SARIT had a functional hitch and five assembled Wike trailers were staged in the Bergeron courtyard. Build Quality Notes Working closely with these vehicles over several months meant encountering their imperfections firsthand. The panel gaps on some units are substantial enough that small objects can fall through — a fact I confirmed personally when my phone slipped out of my pocket, slid through a gap in a door panel while I was doing alignment work in the parking lot, and was run over by approximately ten cars before I found it. The Bluetooth headphones saved it — I wouldn't have noticed otherwise. The case was scuffed. The phone was fine. The gap is still there. This kind of observation is actually useful engineering information. It informed how I thought about weatherproofing for the electrical enclosures and why certain mounting solutions that looked adequate in a controlled environment needed to be more robustly constrained for field use.
[TOC] Overview Every technology added to the SARIT — the reverse buzzer, cameras, display, telemetry, pedestrian detection system — requires power. The vehicle's stock electrical architecture provides nothing for accessories. All of it had to be designed and built by the research team, layered onto a 48V traction system that was never intended to support secondary loads. I worked on the electrical system starting in May 2025, initially as part of the trailer hitch and camera mount work (both of which required understanding what power was available and where), and increasingly as a primary focus through the fall as we moved toward a formal power delivery architecture and parasitic drain investigation. System Architecture The SARIT's power distribution is organized around three voltage rails: Rail A — 48V Traction. The battery, main solenoid, and ASI BAC2000 motor controller. This rail sees large inrush currents (10–30A) when the battery is first connected — normal behavior from the controller's 1.64 mF input capacitance, not a fault. The solenoid acts as a master switch for the accessory system: it connects to the ignition so that added accessories only receive power when the vehicle is on. Rail B — 12V Accessories. Headlights, horn, relay coils, and most user-facing accessories. Converted from 48V via a DC-DC buck converter. One important architectural note: in earlier configurations this converter was wired before the solenoid, meaning it remained live even with the key off. This was identified as a primary suspect for the parasitic drain issue. Rail C — 5V Logic. The ESP32 telemetry module, hall sensors, GPS, and small signal electronics. Must remain stable at exactly 5.0V to prevent sensor drift. Converted from 48V via a dedicated step-down. Reverse Detection Circuit The SARIT has no built-in way to signal that it is in reverse — no dedicated output, no indicator on the harness. The team needed a reverse signal to trigger the buzzer and switch the display to the rear camera view. The solution: we tap a wire on the signal path between the handlebar's forward/reverse switch and the ASI motor controller. The voltage on this wire changes slightly depending on direction. That change is too small to directly drive a relay, so a custom amplifier circuit boosts it to a level that can switch a relay. When the relay fires in reverse, it simultaneously activates the buzzer and sends a signal to the display to switch camera feeds. This is cleaner than the previous approach, which involved soldering directly onto the handlebar switch — a fragile modification that failed under vibration and caused the kind of intermittent wiring problems that are difficult to diagnose. The ASI BAC2000 and Access Constraints The ASI BAC2000 is the motor controller at the heart of the SARIT's drivetrain. It is a capable unit, but getting any visibility into its operating state is not straightforward. The BACDoor app — ASI's tool for reading and adjusting controller parameters — requires an OEM or authorized dealer account. Accounts are explicitly not available to end users or research teams. This matters for development work. Without low-level access you cannot read the controller's internal state, adjust regenerative braking parameters, or confirm whether a behavior you're observing (high current spike on startup, for example) is normal or a fault condition. I eventually found a read-only monitoring app from ASI that provides enough visibility for diagnostic purposes without requiring the full OEM credentials, but the limitation is real and shapes what kinds of electrical investigations are practical. The reason for the restriction makes sense: at 48V with the inrush currents this controller handles, incorrect parameter adjustments can seriously damage the motor. The protection exists for good reason. It is just inconvenient. Mid-2024 Context The project went through an uncertain period in mid-2024 following a legal development involving the company's primary financial backer. For several weeks the team's focus shifted to documenting all active projects and partnerships for York University, and some technical work was paused while the organizational situation was assessed. Contracts were eventually renewed and the project continued, but the disruption is visible in the timeline — a gap in active development that had nothing to do with the engineering. --- See also: [[SARIT — Power Delivery System]] for the parasitic drain investigation and proposed monitoring hardware. See also: [[SARIT — Parasitic Power Drain Investigation]] for the full experimental report.
[TOC] Background The heated accessories research established what could be done with off-the-shelf 12V products. The Climate Caddy project was a different question: could you mount a compact space heater inside the SARIT's cabin and connect it to an intake duct, creating something closer to an actual climate system? The answer was technically yes, practically marginal. This was an exploratory prototype — not a commissioned product, not something that went to any partner site. It ran in the orange SARIT (SARIT 8) and stayed there. The Heater The heater unit was a compact 12V ceramic element designed to mount inside a cup holder — roughly the size of a large travel mug, with adjustable louvers on the output end. The model was identified from Amazon and the CAD was built from the product photos and listed dimensions, since no manufacturer files existed. This is a normal part of working with off-the-shelf components at this scale: the product page gives you enough to model around, even if the tolerances require a few print iterations to get right. The mount positioned the unit low on the dashboard, near knee level. This was a compromise driven by the SARIT's interior geometry — there isn't much flat, accessible surface in the cabin, and the area near the handlebar cluster was already occupied. Low mounting meant the heat output pointed roughly toward the occupant's legs rather than their chest or face, which limited its effectiveness. The louvers could be adjusted upward, but at that range the warm air mostly dissipated before reaching you. The Hood Scoop The hood scoop was a 3D-printed intake duct designed to route outside air into the heater's intake port. The geometry was derived from a scaled screenshot of the SARIT's front hood panel — not ideal, but workable given the tolerances involved. The duct clipped onto the hood panel edge and connected to a flexible hose running to the heater intake inside the cabin. The purpose was to ensure the heater was drawing fresh outside air rather than recirculating cabin air — a basic requirement for any active ventilation system. In practice, the volume of air moved by this heater was small enough that the distinction between fresh and recirculated air was probably not meaningful, but it was the correct engineering decision and it made the system feel more complete. Performance It worked, in a limited sense. On a cold winter morning you could feel the heat within a few minutes of driving. The cabin warmed noticeably compared to no heater. Whether it warranted a winter jacket depended on how cold it was — at -5°C it was a genuine improvement; at -15°C it was still a winter jacket situation. In summer it was just a fan blowing warm air, which was worse than useless, so the unit was simply left unplugged when not needed. The positioning was the main limitation. A heater mounted at knee level in a small cabin is always going to underperform versus one mounted at chest height with directed airflow. The interior geometry of the SARIT didn't offer a better option without a more significant installation. What Came of It The prototype demonstrated that a 12V cabin heater is viable in the SARIT platform and that the hood scoop intake concept is printable and installable without permanent modification. It also established the failure modes: low mounting position, insufficient airflow volume for the cabin size, and no value in warm weather. The project informed the heated accessories research specification — particularly the conclusion that seat and handlebar heating are a more effective investment per dollar than a convective air heater at this power level. The Climate Caddy itself was not recommended for production.
[TOC] The Problem The SARIT fleet had a range problem that was not what it appeared to be. Vehicles were losing significant charge while parked — not from driving, not from leaving accessories on, but simply from sitting. A vehicle with a full charge after a 1-hour charge cycle would show dramatically reduced range after sitting for 72 hours with the charger physically connected to the battery but unplugged from the wall. This was confirmed through a controlled experiment: | Condition | Charge time | Idle time | Resulting range | |---|---|---|---| | Baseline | 1 hour | 0 hours (immediate drive) | ~15 km | | Drain test | 1 hour | 72 hours (charger connected, wall unplugged) | ~6 km | A 60% range loss from three days of parking. The drain was estimated at approximately 100–200 mA continuous — silent, persistent, and significant at 48V. Investigation Diagnosing low-level parasitic draw on a 48V system is harder than it sounds. A handheld clamp meter (Klein Tools CL800) was the first tool tried — it produced wildly fluctuating readings between -0.1A and 0.2A and was useless for detecting milliamp-level leaks. This is a documented limitation of clamp meters for DC leakage below 1A; it is not a fault of the tool, but it is not the right tool for this task. The large inrush current spikes (10–30A) observed when connecting the battery manually were the second red herring. These look alarming but are normal — the ASI BAC2000 controller's input capacitance charges up rapidly on connection. The spikes settle to zero within milliseconds and are not the drain. The actual finding: when the on-board charger was disconnected from the battery terminals, the drain dropped to zero on subsequent measurements. The hypothesis is charger back-feed — the charger's internal circuitry lacks a blocking diode, allowing the battery to slowly discharge back through the charger's capacitors or status LEDs when wall power is absent. The full investigation is documented in [[SARIT — Parasitic Power Drain Investigation]]. Proposed Solution The architectural fix has two components: Immediate: Install an AEV250-G contactor (SPST-NO, 48–72VDC coil) in series with the charger connection. This industrial-grade component is rated for 500A to handle the BAC2000's capacitive inrush, and its coil economizer draws only 0.03A at 48V — it does not introduce a new parasitic load. This isolates the charger from the battery when the vehicle is parked. Longer term: A Digital Fuse Box — a set of current and voltage monitors across each power rail, communicating via I2C to the telemetry system — would allow the drain to be measured precisely and tracked over time. The selected components: | Component | Part | Purpose | |---|---|---| | I2C Power Monitor × 4 | Adafruit 5832 (INA228) | 85V-rated, 20-bit ADC; monitors Battery, 12V, 5V, and 48V rails | | Traction Current Sensor | Allegro ACS758-200B | Hall-effect isolation for the high-voltage, high-noise motor rail | | DC Shunt | FL-2-100A / 75mV | External shunt for main battery negative terminal (INA228 onboard resistor is only rated 10A) | | STEMMA QT Cables × 4 | Adafruit 4401 | Plug-and-play I2C daisy chain | Total estimated BOM cost: $124.02 CAD The I2C addresses were assigned as: 0x40 (main battery), 0x41 (12V rail), 0x44 (5V rail), 0x45 (48V rail). The software team was flagged to verify these do not conflict with existing modules before ordering. KiCad Handoff In January 2026, the project was handed off to an incoming team member for formalization in KiCad. The deliverable was a hierarchical schematic with separate sheets for Traction (48V), Accessories (12V), and Logic (5V), plus footprint verification for the ACS758 and external shunt to confirm they fit in the battery compartment. The I2C addressing scheme, component datasheets, and full BOM were documented and archived. The next steps are hardware procurement and a second controlled drain test with the contactor installed to confirm the hypothesis before committing to the full monitoring build.
[TOC] Context The pedestrian detection system uses an OAK-D stereo camera (Luxonis) to track people and obstacles around the vehicle. Getting that camera mounted cleanly and rigidly to the SARIT's frame — without drilling into anything structural, without it shifting in the field — was the mechanical problem I was handed when I joined the project in May 2025. By the time I started, there had already been two previous mount iterations, both designed in Fusion 360. My work began with v3 and continued through v4. Version 3 — Finalizing a Known Design The first week was a review of the existing v3 prototype. I checked fit, form, and function against the physical camera and the mounting point on the SARIT's interior frame. The changes from v2 to v3 had addressed most of the structural concerns, but there were small CAD details to clean up before the design could be considered final — unnecessary fillets that added print time without adding strength, and tolerances that needed to be tightened slightly for a snug fit without requiring force to assemble. I completed the final documentation package for v3 — detailed drawings and assembly instructions — and archived it. This was my first project in SolidWorks after the previous two versions had been done in Fusion 360, which required me to rebuild some of my modeling habits around SolidWorks' feature tree and assembly workflow. Version 4 — Adjustable Mount The need for adjustability came from the detection team. A static mount works once the system is calibrated, but during development the team needed to physically reposition the camera to adjust its angle — both tilt (up/down) and swivel (left/right) — without reprinting or refastening. This drove the v4 specification. I spent a week researching multi-axis mounting mechanisms before committing to a design direction: mechanical fasteners with friction-fit adjustment joints, no servos. The goal was something a technician could loosen with a hex key, reposition by hand, and retighten — stable under vibration, simple to reproduce. The v4 CAD work centered on the adjustment joint. Designing a functional joint for 3D-printed parts is not straightforward: too little clamping force and the camera drifts; too much and the plastic deforms or the joint seizes. I went through several iterations on the geometry of the locking mechanism before settling on a design that balanced adjustment range against rigidity. Mounting holes for the OAK-D Pro's VESA M4 pattern were included as a fallback in case the primary retention method needed reinforcement. Print orientation was optimized to minimize required supports — all flat faces oriented to the print bed. Print spec: 20% infill, additional perimeter walls rather than higher infill density for bolt-clamping resistance. Version 5 — Return to Static, OAK-D Pro Once the detection team had dialled in their calibration numbers, the requirement flipped back: a simpler, lighter, static mount optimized specifically for the OAK-D Pro (the previous static version had been designed for the OAK-D Lite). The v4 adjustable mechanism was no longer needed in daily use. The v5 mount was a clean-sheet design in SolidWorks, built around the Pro's physical envelope. Overhang was minimized in the print direction, unnecessary fillets were removed, and the assembly was designed to go together with standard ¼-20 hardware already stocked in the lab. The camera's VESA holes were retained as secondary attachment points. One detail that shaped the design: the SARIT frame has a mounting hole positioned very close to the roof — seemingly off-center and not part of any obvious pattern. It turned out to be exactly where the camera mount needed to go. It almost felt like it had been left there specifically. Lessons Designing for 3D printing is not the same as designing for machining. The constraints — layer orientation, overhang limits, wall count vs. infill — are different enough that a feature that looks fine in CAD can fail in print or behave unexpectedly under load. Going from Fusion 360 to SolidWorks mid-project also surfaced some differences in how the two tools handle assembly files and version history; I ended up using Google Drive's file versioning as a lightweight workaround for the latter.
[TOC] Background The SARIT's stock side mirrors are small, fixed, and positioned low on the door panels. For a vehicle operating in pedestrian zones — around students, cyclists, and people on narrow paths — the sight lines they provide are marginal. The situation is worse for taller drivers, and worse again when the vehicle is being operated in reverse. The side mirror mount project started as a quality-of-life improvement and ended up as a production recommendation. The mirrors themselves are a specific part: a KEMIMOTO magnetic tractor mirror — a rubber-coated, 114 lb-hold magnetic mount originally designed for Kubota tractors and forklifts. They're cheap, broadly available on Amazon with same-day shipping, and mount without tools. This is relevant because the mirrors get stolen. I found out firsthand one afternoon when I came out to work on the trailer and someone had walked off with the maintenance SARIT's mirror. I ordered a replacement the same day. The Design Problem The stock mirror position is the real issue. Mounted at door height, the mirrors give you a view of the ground about two meters behind the vehicle — useful for spotting dropped objects, less useful for seeing a cyclist coming up on your left. The goal of the mount was to raise and extend the mirrors outward, creating better rear coverage without requiring any permanent modification to the SARIT's body panels. The constraint: the SARIT's door frame is thin-walled aluminum. Any mount that clamps onto it needs to distribute load across a wide area or the clamping force deforms the panel. This ruled out a number of simpler bracket designs early. Five Versions v1 established the basic geometry: an extendable arm that clamps to the door frame and positions the mirror head at approximately shoulder height on the outside of the vehicle. Printed in PLA as a proof of concept. v2 added a hinge point so the arm could fold flat when not in use — important because the mirrors extend past the vehicle's nominal width and would catch on tight gaps. The hinge introduced slop that made the extended position feel unstable. v3 replaced the hinge with a detent mechanism — a ball-and-socket that locked at 90° (deployed) and 0° (stowed). Cleaner, but the printed socket wore faster than expected under repeated cycling. v4 addressed the wear problem by replacing the 3D-printed socket with a metal insert. This required a redesign of the housing geometry to accommodate the insert and maintain printability. The locking force was adjusted by changing the ball's diameter relative to the socket — a tolerance problem that took a few prints to get right. v5 was the clean-up iteration: reduced material, refined clamping geometry, all faces oriented for clean printing without supports. This is the version that went into the recommendation to Elvy as a potential production accessory. The Instagram Reel At some point during the v3–v4 transition, someone filmed the mount getting stuck in the deployed position on the orange SARIT and posted it to the team Instagram. The ball detent had seized — the ball had deformed slightly under temperature cycling and wouldn't release. This is the kind of failure that doesn't show up in a static load test and only surfaces with field use over time. It accelerated the decision to switch to the metal insert for v4. Outcome The v5 mount is in use on the orange SARIT and has been recommended to the company for fleet-wide adoption. The mirrors themselves are already stocked and reordered as needed — their availability on Amazon same-day is genuinely part of why the design works at scale. A custom mirror that requires a two-week lead time is a problem. A mirror that can be replaced same-day is just a consumable.
[TOC] Background In July 2025, after the trailer hitch project wrapped, I transitioned to designing a mount for the SARIT's LiDAR sensor — a Garmin LIDAR-Lite v3 (40m range) being integrated by the technology team for object detection and distance measurement. The mounting challenge was different from the camera work: LiDAR sensors are sensitive to vibration in a way cameras are not. Data quality degrades if the sensor platform is not rigid, and if the sensor needs to rotate (to get 360° coverage), the mechanism needs to be precise enough that the rotational position is repeatable. BOM and Research Phase The first week was documentation and research — establishing a Bill of Materials for the mount and surveying how existing LiDAR installations handle the vibration and rotation problems. The key findings: vibration damping is typically handled either through compliant mounting materials (TPU pads, rubber grommets) or by isolating the sensor platform from the vehicle chassis mechanically. For rotation, a slip ring is required to pass signal and power through a continuously rotating joint without tangling cables. The sensor in question had six wires coming out — manageable for a slip ring design. v1 — First Physical Prototype The v1 CAD model was completed in early August 2025 and printed immediately as a physical prototype. The print is not the final material — it is PLA, not the eventual PETG or nylon — but having something to hold, test-fit, and hand to the sensor operator surfaced problems that were not visible in the CAD model. This is the value of rapid prototyping: the iteration cycle between digital and physical is fast enough that printing a v1 is almost always worth doing before committing to final materials. v2 — Bearing Preload and Rigidity Feedback from the v1 prototype drove the v2 specification. The main change: the rotation joint needed to incorporate bearing preload — a small axial load applied to the bearing to eliminate play. Without preload, even a well-fitted bearing will have measurable wobble, which in a rotating LiDAR mount becomes angular error in the distance data. Designing a bearing preload mechanism in a 3D-printed assembly is non-trivial. The forces involved are small but the tolerances need to be tight, and plastic parts creep under sustained load in a way metal parts do not. I spent the better part of three weeks iterating on the v2 geometry — making incremental improvements to rigidity, simplifying the assembly sequence, and refining the preload mechanism. The work was interrupted once by a CAD file corruption issue (the assembly had accumulated too many suppressed features and interdependent sketches to edit cleanly) which I resolved by rebuilding the assembly from scratch. Having a clear mental model of what the final design should look like made the rebuild faster than the original construction. Discontinuation In early October 2025, the project was discontinued. Team priorities had shifted — the Power Delivery System work was more urgent — and the LiDAR integration timeline had slipped far enough that the mount was no longer on the critical path. All design files were archived. This was a useful lesson in the realities of research project management. A project being stopped is not the same as a project failing. The v1 prototype exists, the v2 CAD is documented, and the work can be resumed when the LiDAR integration becomes a priority again. The bearing preload problem is solved. The slip ring specification is identified. Whoever picks this up next has a substantially shorter path than I did.
[TOC] Context The SARIT is enclosed but not insulated. In Toronto winters, "enclosed" means you're not being rained on, but you are sitting in what is effectively an aluminum box at ambient temperature. For users operating the vehicle outdoors in December — maintenance staff, security, zoo workers — this is genuinely uncomfortable. The company had a requirement to explore climate control options for fleet vehicles, and the mechanical team's answer was to start with the simplest viable intervention: off-the-shelf heated seat pads and handlebar grip heaters, powered from the vehicle's 12V accessory rail. What We Evaluated The research covered two categories: Seat heating: USB-powered (5V/2A) heated cushions with adjustable temperature settings. Several options were evaluated — flat seat-only pads, full back-and-seat combos, and low-profile designs intended to be concealed under an existing cushion. The key constraint was the SARIT's sliding seat backrest: any heating element attached to the back panel had to allow the backrest to move without binding or pulling at wires. Most of the back-and-seat options failed this constraint. The flat seat-only pads that could be installed under the existing cushion cover were the best fit. Handlebar grip heaters: 12V DC units designed for motorcycle handlebars. The SARIT uses a 3-wheeler handlebar steering configuration, and standard motorcycle grip heaters fit the handlebar diameter with minor adapter considerations. These are available from $13 to $64 CAD with broadly similar performance. Voltage selection: | Voltage | Assessment | |---|---| | 5V DC | Available via USB, very safe, but insufficient power for meaningful heat output in a small cabin | | 12V DC | Standard for vehicle accessories, many products available, recommended | | 48V DC | Efficient at high power but almost no COTS products designed for it; requires conversion | 12V was the clear recommendation. The 12V accessory rail is already present on the research vehicle fleet, and the product ecosystem is vast. Legislation Running 12V heated accessories on a SARIT in Ontario requires compliance with federal safety standards under the Canada Consumer Product Safety Act (CCPSA) and the Highway Traffic Act. The SARIT is treated equivalently to an eBike or low-speed motor vehicle (LSM) for regulatory purposes. Products with CSA/UL certification and built-in safety features (overheat shutoff, adjustable settings) satisfy these requirements without requiring modifications that would change the vehicle's regulatory classification. The full legislative analysis covers: CCPSA, Highway Traffic Act Ontario, Ontario Regulation 141/21 (eBike Pilot Project), Ontario Regulation 587 (Equipment Standards). What Actually Happened The research covered all of this. The recommendation was 12V, scalable, removable, low-profile, CSA/UL certified, with adjustable heat levels. A shortlist of specific products was compiled with pricing (seat pads: $39–$130 CAD, handlebar grips: $14–$64 CAD). Nothing shipped. The project stayed in the documentation phase — no procurement, no installation, no field test. It remains a ready-to-execute specification if the priority comes back around. The practical reality of a heated seat in a SARIT cabin is probably modest anyway. The cabin volume is small, the vehicle is aluminum, and the seat heater is warming the occupant's lower body rather than the air. For short trips in moderate cold it would help noticeably. For a two-hour shift at -15°C, you'd still want a winter jacket.
FRONTIER_CubeSat_v1.4.1.glb-scale125
[Previous Versions |
[2.59 km, 17:02, 6:35/km] [View on Strava |
Current Rotation NY-based duo Take A Daytrip, known for their "Daytrip took it to ten" tag recently released the experimental oscilloscope track "HEART EYE REFERENCE.001." Recently Played Live feed. Updated on every site {reload}. {Recently Played} Fresh Favorites Most played in the last 500 songs. {Fresh Favorites} Top Songs Most played songs, all time. {Top Songs} Top Artists Most played artists, all time, with their respective most played songs. {Top Artists}