Shrinking maintenance windows are becoming a key factor to consider when specifying hardware at plants under continuous operation.
A systems integrator recently commissioned a panel PC for a coolroom at a facility under continuous operation. Shortly after going live, the resistive touchscreen developed a calibration issue. It could be recalibrated on site, but the calibration would drift again within hours. Under normal circumstances, the panel PC would come back to us for investigation and potential warranty repair. But with continuous operations, the customer can’t take the system offline for the several days that would take.
This is the kind of situation that was manageable when plants had regular shutdown windows. A weekly or monthly maintenance period allows time to swap hardware, investigate faults and commission replacements, without affecting production. Those windows are disappearing. Plants that once shut down regularly now run closer to 24/7, and the hardware decisions made during specification often haven’t caught up with that reality.
The spare you didn’t budget for
The coolroom situation left the customer with two options: try to diagnose and fix the issue on site with our remote support, or purchase a second panel PC as a hot spare, so the original could be returned for warranty repair without taking the system down.
This is a pattern we see regularly. A customer purchases exactly the number of PCs they need for the project. The system is commissioned and everything works. Then a fault develops, or a unit fails, and they discover there’s no way to replace it without stopping production. At that point, they order a spare, often urgently and at a premium.
The timing of that spare purchase matters more than people realise. RAM and SSD prices have spiked, with some forecasts suggesting further increases over the next 12 months. A spare that would have cost a modest amount at initial procurement now costs significantly more. The customers who specified spares at the quoting stage and procured them alongside the primary units avoided this entirely.
Warranty replacement no safety net
When a unit fails under warranty, a common expectation across many industries is that the supplier should provide a replacement while the faulty unit is repaired. In practice, this is almost never possible for industrial PCs.
Consider the CyberVisuell panel PC range as an example. There are 10 screen sizes, two touchscreen types, and 10 CPU options, plus various RAM and SSD sizing and Windows versions. That creates well over 800 possible configurations. No distributor can hold stock of every variant in Australia. Even the factory does not keep assembled units on hand. Each panel PC is built to order from modular components.
This is not unique to CyberVisuell. It is the reality across the industrial PC market. Standard warranties are return-to-base, and the repair cycle takes days to weeks depending on the fault. If your plant runs continuously and you have no spare, that repair time comes directly off your production.
The decision nobody makes
In our experience, when a plant increases its operating hours to improve efficiency or throughput, the people making that decision rarely stop to consider the downstream implications for equipment like industrial PCs. The focus is on the production process, the staffing, the logistics. Whether the existing hardware is reliable enough to support the loss of maintenance windows, and whether spares are on hand to cover a failure, isn’t part of the conversation.
This is understandable. An industrial PC that has been running reliably for years doesn’t feel like a risk. But reliability is not the same as invulnerability. Every component has a finite life, and the consequences of a failure change dramatically when there’s no scheduled downtime to recover in.
Systems integrators face a similar pressure. They are assembling complex systems with tight budgets, and redundancy planning tends to get left out at the quoting stage because it increases costs and makes the initial quote less competitive. The result is that the end customer receives a system with no built-in resilience, and only discovers this when something fails.
How the spec changes when the window closes
Moving from scheduled maintenance periods to continuous operation doesn’t just increase the consequence of a failure. It changes what you should be specifying in the first place.
- Spares become essential, not optional. When there was a monthly shutdown, you could tolerate a two-week lead time for a replacement PC. Without that window, you need a configured spare on site or at a nearby depot, ready to swap in. The cost of that spare is trivial compared to the cost of lost production while waiting for a replacement to be built, shipped and commissioned.
- Component reliability matters more. A fan failure in a PC that can be swapped out during a shutdown is an inconvenience. The same fan failure in a 24/7 operation, where the PC is mounted inside a sealed enclosure that can’t be opened while the plant is running, is a production event. Fanless industrial PCs eliminate the most common mechanical failure point. They are typically specified for dust and temperature reasons, but the reliability advantage is equally relevant for continuous operations.
- Remote diagnostics save more than time. If a problem develops at 2am on a system that can’t be taken offline, the ability to diagnose remotely before anyone travels to site is valuable. For the coolroom panel PC, our first step was supporting the integrator to run diagnostics on site, checking for mounting issues, Windows configuration problems, or environmental factors like condensation. Remote support cannot fix every problem, but it can often identify whether a site visit is necessary and what the technician should bring.
- Disk images and backups are critical. Replacing a failed PC is only half the recovery. The replacement needs to be loaded with the correct software, configuration and settings. If the only copy of that configuration is on the failed unit, the recovery time extends dramatically. A current disk image, stored in a known location and verified periodically, turns a potential multi-day recovery into a matter of hours.
The real cost equation
The arithmetic is straightforward but rarely done at specification time.
A spare industrial panel PC costs a few thousand dollars. Even at today’s inflated component prices, it is a known, bounded cost. A day of lost production at a plant running 24/7 typically costs many times that. For some operations, the cost per hour of downtime exceeds the cost of the spare PC.
The cheapest time to buy that spare is when you buy the primary units. The same build, the same components, the same prices. Every month you defer, the risk increases that component prices will rise further, that a specific component will be discontinued, or that the configuration will need to be revalidated against newer hardware.
How ESIS can help
We supply industrial PCs and panel PCs into continuous operations across manufacturing, utilities, food production and infrastructure. We understand the practical constraints of specifying hardware for environments where there is no convenient time to intervene.
Whether you are specifying a new system and want to plan spares from the outset, or you have an existing installation that now runs 24/7 and needs a resilience review, we can help you assess the situation and put a practical plan in place.
Contact us to discuss your requirements today.
