← Back to Yu-ting Hsu

Yu-ting Hsu · Dagbok

september 2026

22 entries

Photographs, observations, and the days between.

Calibrating a stepped rotary encoder for volume control without visual checks. A standard twelve-pulse wheel has soft detents that blur together under low finger sensitivity. Adding a tiny notched beryllium copper shim creates a distinct mechanical click at every eighth of a turn. It rattles slightly when spun quickly, but each step delivers an unmistakable jolt through the fingertip. Noting the tooth depth in the bench log before calling it a night.

Yu-ting Hsu · 30 september 2026

Testing an extended paddle lever for a palm-drop switch tonight. The longer lever arm reduces required downforce to fifteen grams, but the weight of the aluminum bar itself almost trips the microswitch if the desk is bumped. Countering it with a small brass slug behind the pivot balances the lever neatly in neutral. The brass bushing is still slightly rough from hand-reaming, but the reset action stays reliable. Recording the balance weight in the ledger before putting the calipers away.

Yu-ting Hsu · 29 september 2026

Calibrating return tension on a foot-actuated rocker switch tonight. A stiff coil spring prevents accidental trigger from a resting heel, but eighty grams of continuous resistance tires the ankle within minutes. Swapping to an offset brass leaf spring drops the break pressure while keeping the snap crisp. The bracket has a rough cut from the bandsaw and the pivot pin needs a dab of graphite, but the reset stroke is instant. Logging the deflection angle in the ledger, then setting the tools aside for the night.

Yu-ting Hsu · 28 september 2026

An adjustable ball-joint clamp has to lock down without requiring two hands or fifty pounds of grip strength. Testing an oversized aluminum lever on the fifth mounting bracket iteration tonight. The cam profile bites into the socket with a quarter-turn, holding twelve ounces of switch housing steady under firm taps. The lever radius needs another five millimeters of clearance from the desk rail so it doesn't pinch knuckles on release, but the locking friction is solid. One adjustment at a time.

Yu-ting Hsu · 27 september 2026

Closing the ledger for the night with one recorded run on a magnetic breakaway connector. When someone catches a controller cord with a foot or wheel, solder joints shouldn't have to absorb the shock. Four neodymium magnets in an acrylic collar release cleanly at two kilograms of lateral pull, though the alignment pins still bind if yanked sideways at an angle. It is an unrefined first layout, but the terminal pads stayed intact. The next iteration just needs beveled pin guides.

Yu-ting Hsu · 26 september 2026

Casting a silicone sleeve over an extended thumbstick tonight to test grip friction under minimal downward pressure. The first pour came out cloudy with a visible air pocket near the base, but the textured ribbing along the crest holds firm against a relaxed palm. Adding three millimeters of diameter reduces the pinching force required to steer, which spares the carpometacarpal joint during long sessions. The mold needs better venting before the next run, but the physical feel is right where it needs to be.

Yu-ting Hsu · 25 september 2026

When working with involuntary hand tremors, spacing auxiliary switches twenty millimeters on center is rarely enough to prevent accidental brush inputs. Filing a recessed stepped collar for a secondary toggle tonight to test clearance. Sinking the toggle head three millimeters below the top plate allows a resting palm to glide past without tripping the contact. The cut looks crude straight off the rasp, but eliminating false actuations gives a player far more confidence than maintaining a slim chassis.

Yu-ting Hsu · 24 september 2026

Testing the actuation depth on a pair of low-profile tactile switches tonight. The spec sheet calls for one millimeter of travel, but under light finger pads, eighty grams of spring weight feels too abrupt when repeated twenty times. Swapping to forty-gram springs softens the initial resistance without losing the break point. It leaves the switch cap sitting slightly uneven in the test collar, but the trigger stroke no longer tires the joint. Finding the right balance on the bench always takes more patience than reading ratings on paper.

Yu-ting Hsu · 23 september 2026

A thumb rest should feel neutral before it feels comfortable. When shaping a prototype grip from scrap pine tonight, the initial contour leaned too far inward, crowding the index knuckle. Shaving down the inner ridge by two millimeters flattened the contact patch. It looks rough from the rasp marks, but testing the reach against an analog stick showed immediate relief in the wrist. An unpolished surface tells the hand the truth about pressure points much faster than a slick finish.

Yu-ting Hsu · 22 september 2026

A neat wiring loom looks satisfying on paper, but a working adaptive controller needs slack where hands actually move. Routing the auxiliary paddle lines tonight, I left extra loop radius around the main chassis hinge instead of bundling them tight with zip ties. Rigid cabling pulls on solder terminals whenever someone adjusts the arm angle. It looks slightly untidy resting against the bracket, but strain relief matters far more than clean symmetry. The joints won't fatigue when repositioned.

Yu-ting Hsu · 21 september 2026

A distinct acoustic click is often as important as tactile resistance when a player has reduced sensation in their fingertips. Listening to the snap of a leaf switch tonight reminded me that audio feedback shouldn't be sharp or harsh to be clear. A dull plastic thud gets lost, but an overt metallic ping becomes tiring after an hour of play. Finding the right enclosure resonance to amplify that clean, low click takes trial and error, but it gives an unmistakable cue every time contact is made.

Yu-ting Hsu · 20 september 2026

An adaptable mounting arm needs to stay rigid under working pressure, but release without forcing someone to reach for tools. Testing the friction hinge on this articulated clamp shows the tension is uneven across the tilt arc—too tight at forty-five degrees, slipping slightly past ninety. It is an awkward first attempt, but the mechanical fault is obvious. Swapping in a slotted belleville washer should smooth out the resistance before I mount any switches.

Yu-ting Hsu · 19 september 2026

Documented tonight’s small experiment before cutting the bench supply: measuring contact bounce on a batch of low-force lever switches. Under very light finger pressure, electrical chatter averages three milliseconds before settling. A five-millisecond firmware window will filter it cleanly without adding noticeable input lag. It is tempting to mechanically stiffen the leaf spring to eliminate bounce at the contact, but that only adds unwanted resistance for the player. Better to let the code handle the chatter and keep the physical actuation effortless.

Yu-ting Hsu · 18 september 2026

A design schematic only holds value if someone else can build it with standard hand tools. When drafting assembly notes, I try to avoid assuming a specialized jig or rare tooling. If an adaptive mounting bracket requires an impossible fold or proprietary hardware, it fails the accessibility test before anyone even plays. Sticking to off-the-shelf fasteners and simple bend radii keeps repair in the player's hands, where it belongs.

Yu-ting Hsu · 17 september 2026

Lining up the hex keys and layout fluid on the corner shelf before turning in. The temptation is always to start rough-cutting the next mounting plate immediately, but beginning a manual cut past eleven rarely goes well. Clean edges come from steady forearms, not impatience. The dimensions are penciled onto the aluminum stock, and the drill press is cleared. That is enough preparation for one evening; the first holes can wait for tomorrow’s daylight.

Yu-ting Hsu · 16 september 2026

The foot pedal's return spring recorded 120 grams of actuation force on the gauge just now. The stroke feels slightly stiff through the final two millimeters, though the internal wiring harness cleared the pivot channel without catching. There is always a temptation to swap in a lighter coil immediately, but modifying assemblies late at night usually produces sloppy solder joints and stripped threading. The measurement is logged in the notebook. Morning will show whether that stiffness is mechanical resistance or simply tired hands.

Yu-ting Hsu · 15 september 2026

Jag klämde fast prototypens handflatsfäste vid kanten av arbetsbänken innan jag släckte taklampan i kväll. Det första grova snittet i aluminiumplåten har en tydlig grad längs innerkurvan, men vridvinkeln känns rätt under handflatan. Det finns ingen anledning att fila det slätt innan jag har verifierat räckvidden till den nedre paddelbrytaren. En ren finish kan vänta; först måste jag kontrollera frigången med hänsyn till belastningen på leden.

Yu-ting Hsu · 14 september 2026

Det tog längre tid än väntat att packa upp komponentlådorna i kväll. När behållarna flyttades mellan tidszoner hamnade några avdelare fel, så mikrobrytare med låg aktiveringskraft blandades med vanliga taktila knappar. Att sortera tillbaka dem efter aktiveringskraft med digitalt skjutmått är tyst, repetitivt arbete, men det måste göras innan någon montering börjar på morgonen. En blandad låda leder bara till tvekan mitt under lödningen. Morgondagens testchassi kräver förutsägbart motstånd, och precision börjar med att veta exakt vilken brytare jag håller i handen.

Yu-ting Hsu · 13 september 2026

Londons äldre byggnadsbestånd har en annan akustisk tyngd än det jag är van vid. Tunga trädörrar och infällda lås stängs med en dämpad, lågfrekvent duns i stället för ett skarpt metalliskt klick. Den mekaniska resonansen är värd att notera. Vid utformning av handkontroller beror den upplevda taktila återkopplingen nästan lika mycket på det invändiga höljets hålrum som på fjäderspänningen. En brytare känns fastare när höljet inte absorberar aktiveringsklicket, något jag måste ta hänsyn till när jag testar sekundära paddelfästen här.

Yu-ting Hsu · 12 september 2026

Att landa i London innebär att skapa en rutin innan utrymmet är helt organiserat. Den första impulsen är att ordna varje hörn direkt, men en pålitlig arbetsbänk måste komma först. En plan yta, bra belysning och ett öppet anteckningsblock räcker som förankring. Att flytta känns ungefär som att börja med en obekant prototyp: säkra arbetstoleranserna först och låt sedan resten anpassa sig genom den dagliga användningen.

Yu-ting Hsu · 11 september 2026

Det finns en ihållande impuls att lösa tre mekaniska problem på en gång innan jag lämnar arbetsbänken. Min bättre vana är mer avgränsad: testa en variabel, anteckna resultatet och gå därifrån innan tröttheten börjar utge sig för att vara tålamod. En fjäder som kärvar eller en brytarvinkel som nyper leden brukar se tydligare ut efter en paus än efter tre framtvingade justeringar. Anteckningsboken håller räkningen ärlig.

Yu-ting Hsu · 10 september 2026

Att börja skriva den här loggen känns som att fastställa baslinjeparametrar. De första anteckningarna blir förmodligen grova, men ett felfritt första försök är mycket mindre viktigt än att ha en ärlig jämförelsepunkt. Jag vill ha en stadig liggare för att följa vad som håller, vad som glider ur läge och vad som kräver tålmodig justering när saker tar form.

Yu-ting Hsu · 9 september 2026

End of september 2026Back to Yu-ting Hsu →

Att avstå från valfria verktyg begränsar inte din åtkomst till Lyfera. Google Tag Manager läses in först efter valfritt samtycke.

Ditt val sparas i sex kalendermånader. Nya ändamål eller väsentliga förändringar kräver ett nytt val. Läs mer.