XIAC Museum — Swiftpack counting sensors (Swiftpack • DT Swiftpack • Kalish • IMA)

The story

This is not one single “model” of sensor — it’s a lineage. Across decades, Swiftpack counting systems changed mechanically and cosmetically, and the branding shifted between Swiftpack, DT Swiftpack, Kalish, and IMA Swiftpack. Yet the counting head sensors kept the same core approach: conservative optical geometry, repeated modular channels, and design margin aimed at long service life.

What makes this Museum piece special is the continuity: good industrial design that didn’t need to shout. It simply stayed in service.

Museum takeaway: good industrial design doesn’t shout — it persists.

What these sensors are

Robust optical detection mounted beneath the discharge path, built for vibration, dust, and long duty cycles.

• Through-beam optical detection aligned with product discharge
• Modular channel construction repeated across multi-channel heads
• Conservative margins focused on stability and serviceability

These sensors were seen everywhere — and rarely noticed — precisely because they just worked.

Machines & eras

One platform, many badges. Swiftpack / DT Swiftpack / Kalish / IMA eras — the sensor concept remained consistent.

Two generations: without AGC vs with AGC

Over time, two distinct electronics generations emerged: early fixed-gain heads and later heads with AGC.

Fixed gain → AGC. A small but significant circuitry addition to compensate ageing and support wider spacing.

• Older type: fixed-gain analogue front-end (stable, but ageing reduces margin)
• Later type: AGC added to prolong sensor life and tolerate larger emitter–detector distance
• Outcome: suitability for later “Swiftpack wide” models with greater spacing

Repair first, then redesign

Designing replacement sensors was tricky because a complete matched reference set (emitter + detector) wasn’t available. The starting point was repairing original boards — both to understand how they behaved in operation and to obtain accurate dimensional measurements from working hardware.

• Functional understanding: observe real margins, tolerances, and failure modes
• Dimensional truth: confirm board outline, connector heights, and alignment offsets as-built

Principle: you can’t responsibly replace what you don’t first understand.

Repair & measurement references

Modular channel banks. Repetition is reliability.

Original boards. Repairs provided operating insight and reference behaviour.

Bare board reference. Used for accurate dimensional capture.

XIAC replacement sensor design (from 2010)

The replacement design was intentionally based on proven architectures used in the King TC‑8 and Cremer CF‑830 upgrades: keep the analogue path clean and fast, then add digital control for stability, margin, and adaptability.

• Receiver: pre‑amplifier (typically I‑V conversion) followed by an analogue multiplier
• Control: gain control via PIC16F73
• Emitter: 8 DAC channels over I²C for adjustable LED brightness
• Result: exceptionally wide usable dynamic range (adjustable LED + adjustable gain)

In practical terms: LEDs could be tuned for brightness, and the receiver could be tuned for gain — without resorting to fragile high‑gain analogue designs.

Board lineage

Lineage. Original Swiftpack boards and later XIAC replacements (heritage from TC‑8 + CF‑830).

The progression from individual single-channel sensors to integrated multi-channel boards was a direct consequence of experience in the field. Once synchronisation and precise optical alignment became essential for reliable counting, these characteristics could no longer be treated as independent per-channel adjustments. Timing, geometry, and signal integrity had become system-level concerns — best solved by fixing the relative alignment and synchronisation directly in hardware.

The current XIAC infrared sensor bar platform for PSR35 and PSR25 heads represents the structured continuation of this development path — integrating fixed mechanical pitch, synchronised channel timing, and self-calibrating optical control into a single architecture.

View current PSR35 / PSR25 sensor platform →

Practicality — single-channel replacements

While the fully integrated, synchronised multi-channel system represents the best technical solution, it is not always the most appropriate one for every machine in the field.

Left: integrated emitter & detector boards (synchronised, aligned by design)  •  Right: OEM HeadBox with individual channel sensor boards

In many cases, only a single channel within a multi-channel head has degraded or failed, while the remaining channels remain mechanically sound and stable. For these situations, replacing an entire head assembly is neither necessary nor economical.

• Targeted replacement of one failed channel
• Minimal disruption to existing alignment and wiring
• Lower upfront cost and faster return to production

Supplying both integrated multi-channel systems and single-channel replacements reflects a deliberate service philosophy: offer the right solution for the machine, not just the most advanced one.

Museum note: the “best” design didn’t replace the earlier ones — it put them in context.

From sensors to systems (from 2017)

The first sensors were sold in 2010. As they proved themselves in the field, the next limitation became clearer: legacy control electronics were increasingly the weakest link. From 2017, XIAC also began manufacturing replacement controllers using a Mitsubishi PLC and a colour touchscreen HMI.

• Modern diagnostics and clearer fault reporting
• Long-term serviceability with readily available control components
• Modern operator interface without abandoning proven mechanics

Controller replacement. Mitsubishi PLC + colour touchscreen HMI (introduced from 2017).

Gallery

DT Swiftpack era. Dual counting heads with later-style HMI.

Earlier control style. The sensor concept persists through changes.

In-line installation. The quiet workhorse lives under the hood.

Brand variety. Swiftpack / Kalish / DT / IMA across markets.