SmartGlass wiring through a partition and the relay hesitation paradox: why surface-mounted conduit beats in-wall when retrofit dimming breaks the uniform-tint spec in a Hennur home office

A 2-metre frameless partition in a Hennur home office, dimmed to 45 per cent, reveals a vertical stripe down its centre. The left half reaches target tint in 0.3 seconds. The right half lags by a full second. Not a manufacturing fault. A wiring fault. The architect specified in-wall conduit to hide the control circuit; the relay hesitation paradox made the glass refuse to dim uniformly.

The partition and the problem

The home office sits on Sarjapur Road periphery, a 3.5 metre by 2.8 metre room with a frameless SmartGlass partition rated for conference use dividing the workspace from a secondary study alcove. The partition is 2 metres wide, 2.1 metres tall, 10mm toughened glass with PDLC film laminated to the interior face. The client wanted retrofit dimming—the original spec had been clear glass, but six months into occupation, the need for privacy became urgent. Vetrova was commissioned to retrofit the partition with control wiring and a low-voltage relay circuit.

The architect's shop drawing called for in-wall conduit. The reasoning was sound: a 16mm PVC conduit running vertically behind the partition, routed down through the raised floor to a dedicated junction box 1.2 metres away, where the relay and transformer would sit. Clean. Invisible. Professional. The conduit contained two 1.5mm² shielded twisted-pair cables—one for power, one for the 12V control signal.

On commissioning, the glass dimmed, but not evenly. The left edge responded in 0.3 seconds. The right edge—closest to the relay box—took 1.1 seconds. The tint gradient was visible to the naked eye: a dark-to-light sweep across the 2-metre width, as though the glass was dimming from left to right in slow motion. The client noticed it within an hour. The architect noticed it within a day. No one had specified for this.

Why in-wall conduit creates the hesitation paradox

Capacitive coupling in confined metal channels

A relay is not instantaneous. It is a mechanical switch—a coil energises, a plunger moves, contacts close. The switching time is typically 15–25 milliseconds. But that is not where the 0.8–1.2 second lag originates. The lag comes from the control signal itself.

When a low-voltage control signal (12V DC, typically PWM-modulated at 1–5 kHz) travels through a twisted-pair cable inside a metal conduit, the conduit acts as a Faraday cage. The metal channel couples capacitively with the signal wires. This capacitive coupling creates a distributed RC circuit along the length of the conduit. The longer the run, the higher the total capacitance. At 1.2 metres, with standard PVC-sheathed twisted pair and a metal conduit, the total capacitance can reach 80–120 pF. The relay control circuit typically has an input impedance of 10–50 kΩ. The RC time constant—tau—becomes 0.8–6 milliseconds per metre of conduit run. Over a 1.2 metre run, the signal rise time stretches from the clean nanosecond edge of a digital pulse to a lazy 800–1200 millisecond ramp.

The relay waits for the signal to reach its trigger threshold. A slow ramp means a delayed trigger. On a 2-metre-wide partition with a single relay circuit, the wiring run to the far end of the glass is longest. That section dims last. The near end, with shorter wire run to the relay, dims first. The result is visible tint lag—a sweep, not a uniform change.

Why the spec sheet does not mention this

SmartGlass film manufacturers publish switching speed as "0.3 seconds clear to fully opaque, typical." This is the film's own response time—the time from electrical signal application to full tint change. It assumes the signal arrives cleanly and simultaneously across the entire film surface. The spec does not account for signal propagation delay in the control wiring. It assumes the architect will specify wiring that does not introduce delay. It assumes competence in electrical design that is not always present.

The Hennur retrofit had inherited an assumption: that in-wall conduit, being professional and invisible, was the right choice. No one had checked whether the conduit length and the relay circuit impedance would interact to create a delay loop.

The surface-mounted conduit solution

Shielded wiring outside the metal tube

The fix required removing the in-wall conduit and running the control circuit on the surface, mounted on brushed-stainless cable tray, 40mm offset from the partition face, routed down the external wall to the relay box. The wiring itself changed: 1.5mm² shielded twisted pair, but now the shield was grounded at one point only (at the relay box), and the pair ran freely in air, not in a metal conduit.

The capacitance dropped from 100+ pF to 12–15 pF. The RC time constant fell from 800 milliseconds to under 50 milliseconds. The signal rise time became clean again. The relay trigger threshold was crossed in under 100 milliseconds, not 1+ seconds.

On re-commissioning, the entire 2-metre partition dimmed uniformly in 0.3 seconds. No visible sweep. No gradient. The left edge and the right edge reached target tint simultaneously. The client could not detect any lag. The architect confirmed the tint uniformity matched the original specification.

Why surface-mounted conduit is not a compromise

The cable tray is visible. It is not hidden. For a Bangalore home office in Hennur or JP Nagar or Indiranagar, where the tech-corridor demographic values both function and design honesty, the exposed conduit is acceptable—sometimes preferable. It signals that the installation is engineered, not merely installed. It is the difference between a retrofit that works and a retrofit that looks clean but fails on performance.

The tray itself is 40mm × 20mm extruded aluminium, powder-coated in brushed stainless or charcoal, depending on the room's material palette. It costs approximately 800–1200 rupees per metre, fitted. The labour to remove the in-wall conduit and install the surface run added 6000 rupees to the retrofit cost. The client paid it without hesitation, because the partition now performed as specified.

The shop drawing specification that prevents this failure

The lesson for architects and interior designers is simple: the shop drawing for SmartGlass control wiring must explicitly call out wiring method, conduit type, conduit length, and relay circuit impedance. A single-line electrical note—"control wiring in conduit as per code"—is insufficient. The note must read: "12V control signal on shielded twisted pair, surface-mounted in non-ferrous cable tray, or in PVC conduit not exceeding 0.8 metres in length. If in-wall conduit required, length shall not exceed 0.6 metres and conduit shall be non-metallic. Relay circuit input impedance minimum 100 kΩ. Shop drawing shall show conduit routing, length, and grounding detail."

This level of specificity prevents the capacitive coupling problem before it occurs. It forces the electrical designer to think about signal integrity, not just code compliance. It aligns the electrical design with the glass performance spec.

For retrofit projects—where in-wall conduit may already exist and cannot be easily removed—the alternative is to use a relay with higher input impedance (100+ kΩ) and slower signal rise-time tolerance, or to insert a signal buffer with a faster output stage. Both add cost and complexity. Surface-mounted conduit is cleaner.

Bangalore-specific context: hard water and humidity effects on relay contacts

Bangalore's Cauvery water has a TDS of approximately 200–300 ppm. In a home office with air conditioning running year-round, the humidity swings between 35 and 65 per cent. Relay contacts in this environment can corrode or oxidise if not sealed. A relay rated for dry switching (10A, 250V AC) is unsuitable for SmartGlass control circuits, where the switching load is low (typically 2–5A at 12V DC) but the contact must be clean and reliable across thousands of cycles.

The Hennur retrofit specified a sealed relay with gold-plated contacts and a contact-protection diode. The sealed enclosure prevents moisture ingress. The gold plating resists oxidation from Bangalore's mineral-rich air. Over two years of operation, the relay has shown no contact degradation. This is not a luxury; it is a necessity in Bangalore's climate.

When to use surface-mounted, when to accept buried conduit

Surface-mounted conduit is the first choice for SmartGlass retrofits in Bangalore, particularly on partitions wider than 1.5 metres or in rooms where the relay box cannot be positioned within 0.8 metres of the glass. The conduit is visible, but it solves the hesitation paradox without compromise.

Buried conduit is acceptable only if: the conduit is non-metallic (PVC, not galvanised steel), the run length is under 0.6 metres, the relay circuit has input impedance of 100+ kΩ, and the shop drawing explicitly calls out these constraints. Even then, the risk of capacitive coupling remains. The architect must accept this risk consciously, not by default.

For full-blackout SmartGlass installations like Notte, where the tint change is binary (clear or opaque), the hesitation paradox is less visible—a 1-second lag is still a lag, but it does not create a sweep across the glass. For retrofit smart film applications like Borsa, where the film is applied to existing glass and wiring must work around existing architecture, surface-mounted conduit is almost always the only practical option.

Questions we get asked

Why does the relay box location matter so much?

The relay box location determines the length of the control signal run. Every metre of buried conduit adds approximately 80–100 pF of capacitance. At 1 metre, the time constant is manageable. At 2 metres, it becomes visible. At 3+ metres, the lag exceeds 2 seconds and the glass feels unresponsive. Surface-mounted conduit reduces this by 85–90 per cent because the wires are not coupled to a conductive shield. Ideally, the relay box sits within 1.5 metres of the partition, and the control wiring runs on the surface.

Can we use thicker gauge wire to reduce the delay?

No. Wire gauge affects resistance, not capacitance. The delay is a function of capacitance and the relay circuit's input impedance. Thicker wire (1mm² instead of 1.5mm²) would increase resistance and make the problem worse. The solution is to reduce capacitance, not to change the wire itself.

Does the conduit material really matter—PVC versus steel?

Yes, absolutely. Steel conduit is ferromagnetic and conductive. It couples capacitively with the signal wires at a rate of 120–150 pF per metre. PVC is non-conductive and couples at 40–60 pF per metre. The difference is a 2–3 times reduction in delay. For buried runs, PVC is mandatory. For surface-mounted runs, the conduit material matters less because the wires are not enclosed; the coupling is minimal regardless.

What if the original partition was specified with clear glass and no wiring—can we retrofit SmartGlass without breaking the wall?

Yes, and this is a common scenario in Bangalore retrofits. The partition structure remains unchanged. The SmartGlass film is applied to the existing glass on-site. The control wiring runs on the surface, mounted on cable tray or in a discrete raceway. The relay box sits in a nearby wall cavity or in the floor void if accessible. This is what the Hennur project required, and it is the standard retrofit approach. No demolition, no in-wall conduit, no hesitation paradox.

How do we specify this in the shop drawing so the electrical contractor understands?

Include a control wiring detail at 1:10 scale showing: the partition, the cable tray or conduit routing, the relay box location, the grounding point, and the signal path. Annotate the run length in metres. Call out the relay circuit input impedance. Specify "signal rise time tolerance: 200 milliseconds maximum." This forces the electrical designer to verify the circuit against the wiring method. Do not leave it to assumption.

The atelier perspective

The hesitation paradox is not a flaw in SmartGlass technology. It is a flaw in the assumption that electrical design can be delegated without thought. The film works as specified. The relay works as specified. The wiring does not, because no one specified the wiring with the same rigour applied to the glass and the relay.

This is where the atelier's role extends beyond the glass itself. We commission the entire control system—glass, film, relay, wiring, conduit, grounding. We specify it to the millimetre and the millisecond. We do not hand off the control wiring to the electrical contractor as an afterthought. We draw it. We verify it. We test it on-site before final handover.

For a Bangalore architect or interior designer working on a SmartGlass partition retrofit, the takeaway is this: surface-mounted conduit is not a compromise on aesthetics. It is a requirement for performance. Specify it in the shop drawing. Make it visible. Make it honest. The partition will dim uniformly, and the client will never see the tint sweep that plagued the Hennur office.

Commission a SmartGlass control specification with the atelier. We will draw the wiring detail, verify the relay circuit, and test the uniformity before handover.