- Robotics & Automation News) Every channel on a cold chain tracker exists to answer a question somebody will eventually put to you. Was it too warm. Was it dropped. Was it opened. Where was it when that happened.
The useful way to specify one of these devices is to work out which of those questions you actually get asked, and buy the channels that help answer them.
Most specification conversations run the other way – the supplier lists options, the buyer picks a configuration that sounds thorough – and that ordering is how a fleet that never leaves the ground ends up carrying pressure sensors.
Which channels those are depends on which argument you expect to have, and that varies more between shippers than most option lists suggest.
A cold chain tracker is an evidence-collection device. Working backwards from the evidence you need is the only specification method that reliably produces the right configuration.
So start with what has actually gone wrong on your lanes in the last two years. Not what could theoretically go wrong – a rejected load, or an argument with a customer.
Those are the disputes the hardware exists to support. A channel that would have helped in one of them earns its place. A channel that would not is a line item.
Two things make this decision heavier than it first looks.
The first is that a channel cannot be added retroactively to a shipment that has already moved. Devices in this category are specified once, deployed in volume, and live in the field for years.
Some channels can be switched on later by configuration, if the sensor was fitted at build time. A sensor that was never fitted is a different matter, and that gap runs through every record the device produces until the fleet is replaced.
The second is that channels need owners. Temperature usually belongs to quality. Location usually belongs to operations. Shock data often belongs to nobody in particular, because the function that would use it – claims – is frequently not in the room when hardware is specified.
Before adding a channel, it is worth naming the person who will look at it. If that name does not exist, the channel produces data nobody opens, which resembles not having bought it in every respect except the invoice.
Channels are not interchangeable. They answer different questions, with different degrees of certainty.
2-300x262.png 300w" sizes="auto, (max-width: 662px) 100vw, 662px" />
Four of those deserve more than a row.
External probe versus ambient temperature. An ambient sensor reports air. A probe reports whatever it is in contact with. During pull-down on a loaded reefer these diverge substantially and for some time, so an ambient-only record can show an excursion on a load whose product never left range – or, less often and more expensively, the reverse.
The caveat is that a single probe describes a single location: it is evidence about the load, not proof about all of it, and how representative it is depends on where it sits and what your acceptance criteria say.
If you ship pharmaceuticals, USP General Chapter <1079> on good storage and distribution practices is worth reading on this point. It is an informational chapter rather than a binding standard, but it is the vocabulary your customer's quality team will be using.
Shock data speaks to a different argument. Temperature claims and damage claims are separate disputes, often with separate adjusters, and a temperature-only device leaves you unable to say anything at all about handling. Shock readings are also only interpretable against a reference.
The ISTA test procedures your packaging was qualified against describe what it was built to survive – but comparing a field reading to a qualification threshold is not a simple number-to-number match, because pulse duration, axis and how the device is mounted all change what a given peak means. Treat a shock reading as the start of an investigation, not as a verdict.
Humidity is narrower than it is often sold. Its real job is condensation and packaging integrity: fibreboard loses strength as it takes on moisture, some produce respires and drives humidity up in a sealed space, and condensation on pull-down can damage labels and cartons even when temperature never leaves range.
Those are genuine failure modes worth instrumenting for. What humidity is not is a general-purpose early-warning channel – in a sealed, loaded trailer it broadly follows the air condition rather than leading it, and how it behaves depends on the load, the outside air and where the sensor sits. Specify it for the packaging and product reasons, not as a proxy for something else.
Light and a door sensor are not the same channel. Light is inexpensive and works well in containers and cartons, where opening admits daylight. In an enclosed trailer backed tight against a dock, a door can open, a pallet can leave, and a light sensor may register very little: the dock is dim, the opening is brief, and the device may be deep in the load.
A dedicated open-close sensor answers the question directly. If your custody disputes happen at docks rather than on the road, that distinction decides whether the channel is useful or decorative.
How to decide which channels you need
Work from the dispute backwards rather than from the datasheet forwards.
If your recurring problem is temperature rejection at the receiving dock, you need ambient temperature, an external probe, and location. The probe speaks to whether product temperature left range.
Location speaks to where it happened – in transit, or while the trailer sat in the receiver's own yard – which is often the fact the commercial argument turns on, though what the contract says determines who carries the cost, not the sensor.
If your recurring problem is physical damage, shock and location do the work and temperature becomes secondary. Shock without location tells you the load was struck but not where, which is the half of the answer that does not help you fix the lane.
If your recurring problem is shrinkage or tampering, light or a door sensor plus location is the set. Temperature is close to irrelevant to that argument.
If you move moisture-sensitive product or ship in fibreboard, humidity earns its place for the reasons above.
If you ship by air, pressure. If you never ship by air, it is usually skippable – with one caveat: a pressure trace can raise a flag on a shipment that was moved to air without anyone telling you.
It is suggestive rather than conclusive, since ground elevation and weather also move the reading, so treat it as something to check against the location record and the carrier's paperwork.
The logic is easier to see applied than described, so here is an illustrative case – a composite, not a specific customer.
A produce shipper runs refrigerated LTL into regional distribution centres and keeps absorbing rejected pallets. The receiving sites cite temperature.
The carrier insists the reefer ran in spec throughout. Both can be telling the truth, because the shipper's devices sit in the trailer and report air.
The channel set that addresses this is ambient temperature, an external probe in the load, and location. The probe produces the measurement the receiver's own specification is written against.
Location establishes where any excursion occurred – and on multi-drop LTL, distinguishing an earlier stop with repeated door openings from something happening at line-haul speed is often the whole question. Shock, light and pressure contribute nothing to that particular argument and would be specified out.
Three channels, not seven, and the three-channel device is the correct one.
Selection criteria summary
Once the channel set is settled, the questions that separate suppliers are narrow:
Are all channels timestamped against a single clock, or does each carry its own timebase? This determines how much reconstruction work a cross-channel question costs you later.
What is the sampling interval per channel, and can channels be configured independently? Shock generally needs to be event-driven; temperature generally does not.
What happens to the record during a connectivity gap – does the device buffer and backfill, or is the window simply absent?
Can the external probe be replaced in the field, and does replacing it affect the calibration record?
Is the calibration document traceable to the specific unit you receive, or to a production sample? Ask to see one.
What does an exported record actually look like? Ask for one from a real shipment rather than a sample file.
The first question on that list is the one that gets asked least and matters more than it sounds.
What the extra channels cost you
Channels carry costs that do not appear on the quotation.
Power is the first, though the size of it depends on the design. Additional sensor reads and larger payloads draw on the same cell, but a channel configured as event-driven costs very little when nothing happens, and devices that batch transmissions behave differently from devices that do not.
Ask what your specific configuration does to the projected service life rather than assuming every added channel is equally expensive.
Data volume is the second, and it compounds across a fleet on plans priced by volume.
Interpretation is the third and the most underestimated. Every channel produces a trace somebody has to read, and a report with seven of them is not seven times more useful than one with two – it can be less useful, because the signal that matters is harder to find.
A configuration nobody has the time to read is a configuration that defaults back to looking at temperature alone.
Interpretation cost is also where the timestamping question from the previous section stops being a technical detail: how much work a cross-channel question takes is determined by how the channels were stamped in the first place.
Suppliers differ in how they handle that, and describe it differently. Eelink, for instance, documents a single internal timebase across channels for its multi-sensor cold chain tracker.
Others take different approaches, and a shared clock on its own does not resolve everything – sampling intervals and what exactly a timestamp refers to still differ between channels.
The point is to ask the question and get the answer in writing before deployment, whoever you buy from, because it is not a detail anyone wants to discover during a claim.
Common mistakes
Specifying the maximum configuration by default. It feels prudent and it is expensive. The configuration you need follows from your disputes, not from the supplier's option list.
Buying temperature-only to save money when the argument turns out to be about handling. This is the opposite error and it is harder to recover from, because the missing channel cannot be added to a shipment that has already failed. If damage claims appear anywhere in your recent history, treat shock as a requirement rather than an option.
Reading ambient temperature as product temperature. Covered above, and it is a recurring source of rejections that neither party can resolve from the data they hold.
Not asking how channels are timestamped. It sounds academic until the first time you need to establish whether a temperature rise came before or after an impact and the two records do not line up.
Assuming the calibration certificate describes your unit. Ask whether it is traceable to a serial number, and ask before the purchase order rather than after.
Configuring shock on a fixed sampling interval. Shock is an event, not a trend. A device reading acceleration every fifteen minutes will miss most impacts worth knowing about.
Letting the pilot configuration become the production configuration. Pilots are rightly run with everything switched on – the point is to find out what the lanes do.
The mistake is rolling that same configuration into the fleet without revisiting it. The pilot was an investigation; production is a standing cost, and the channel set should narrow once the investigation has an answer.
Not asking whether the channel set is a configuration or a hardware decision. On some devices channels can be enabled, disabled and re-intervalled remotely after deployment. On others the sensor is fitted at build time or not at all.
That is the difference between a decision you can revise next quarter and one you own for the life of the fleet, and it is often mixed within a single product. Ask which channels fall on which side.
Finding the right fit
No configuration is correct for everyone, and a supplier who says otherwise is reading from a stock list rather than listening.
The process is short. List the disputes you have actually had. Map each one to the channel or channels that would have helped. Specify those, plus location, and leave the rest off unless somebody can name the argument it would win.
Then work through the six questions above and pay attention to how quickly the answers come back and whether they arrive in writing.
The suppliers worth working with will answer directly, including when the honest answer is that a channel you asked about will not help you. That answer is worth more than a longer option list, and it arrives less often than it should.
Comments
No comment