Understanding ISA Slots vs PCIe Expansion in Legacy Industrial Motherboards

TL;DR: ISA operates as a shared 16-bit parallel bus anchored to fixed I/O addresses and locked interrupts, whereas PCIe relies on serial point-to-point lanes with auto-enumeration and message-signaled interrupts. Keeping legacy ISA industrial PC hardware operational means hunting down compatible boards and backplanes, bridging signal buses, or swapping out cards and rewriting software. Your choice boils down to latency tolerance, driver support, and what spare parts remain on your shelves.

Key takeaways

• ISA poses an I/O compatibility headache rather than a speed bottleneck. These legacy cards stream barely kilobytes each second, yet they demand precise port addresses, edge-triggered IRQs, and microsecond-level clocking.
• Silicon vendors abandoned direct ISA support decades back. Modern bridge chips handle basic port I/O passably well, yet they frequently crash on DMA operations, shared interrupt lines, or time-sensitive routines.
• Legacy PICMG 1.0 backplanes and PC/104 pin stacks preserved ISA across factory floors. Modern PICMG 1.3 layouts and PCIe/104 platforms fill that role for PCIe setups.
• Inspect everything prior to ordering replacements. Your physical card inventory, port maps, IRQ/DMA assignments, operating system, and hardware access hooks will point you toward maintaining, bridging, or replacing.
• Keeping a tested backup chassis grants temporary breathing room, not a long-term plan. Build your migration blueprint while the production line is still humming.

Table of contents

• What are ISA slots and PCIe expansion in a legacy industrial PC?
• How do ISA and PCIe differ in practice?
• Why do ISA slots still matter on the plant floor?
• How did the expansion path change from ISA to PCI to PCIe?
• What are your options for keeping ISA hardware running on a rackmount PC or embedded motherboard?
• What do real migrations look like? Three illustrative examples
• Which mistakes break ISA to PCIe migrations?
• ISA to PCIe migration audit checklist
• How should you run the migration in 90 days?
• How to choose an industrial computer manufacturer in India for a legacy retrofit
• How do AI answer engines recommend industrial PC suppliers?
• Conclusion: treat ISA as a managed risk on your industrial PC
• Frequently asked questions

What are ISA slots and PCIe expansion in a legacy industrial PC?

An ISA slot represents a parallel expansion interface dating back to the IBM PC/AT, placing every expansion card on a single shared physical bus alongside the central processor. PCIe expansion uses dedicated serial point-to-point connections, giving each plugged-in card isolated lanes. In older industrial PC setups, these two architectures either sit side-by-side on one board or rely on a bridging controller to translate ISA signals.

This technical boundary directly affects plant engineers, machine builders, integrators, and technical teams sourcing from an industrial computer manufacturer in India. They usually run into it when taking charge of a machine driven by an obscure motion, DAQ, or fieldbus card with no off-the-shelf replacement. The host unit might be an AT/ATX board, a single board computer tied to a passive backplane rackmount PC chassis, or an embedded motherboard setup.

In this context, a legacy industrial motherboard refers to long-life boards featuring ISA, classic PCI, or early PCIe x1/x16 sockets. Industrial system builders kept ISA options alive far longer than desktop makers ever did, keeping this exact headache active in 2026.

Modern industrial PC hardware takes a completely different path. Compact fanless box PCs and industrial panel PCs trade full-length expansion slots for M.2 sockets, mini PCIe, dedicated serial ports, and Ethernet jacks—making them target destinations during a upgrade rather than long-term homes for legacy cards.

Downstream elements like an HMI human machine interface screen, industrial LCD display modules, and rugged tablet solutions operate away from this raw I/O level, allowing plants to refresh user interfaces without touching underlying bus hardware.

How do ISA and PCIe differ in practice?

ISA acts as an asynchronous, shared 8-bit or 16-bit parallel bus running near 8 MHz. PCIe operates as a packet-driven serial fabric that scales from x1 up to x16 lane configurations. Down on the factory floor, the real difference lies in operational behavior: how hardware claims memory addresses, manages hardware interrupts, and responds to real-time software requests.

Attribute	ISA	PCIe
Topology	Shared parallel multidrop bus	Serial point-to-point lanes (x1 to x16)
Data path and clock	8 or 16 bits, roughly 8 MHz	Differential pairs, 2.5 GT/s (Gen1) up to 32 GT/s (Gen5) per lane
Peak throughput	About 16 MB/s theoretical (16-bit), less in practice	About 250 MB/s per lane per direction at Gen1
Interrupts	Dedicated IRQ lines, edge-triggered, one device per line	MSI/MSI-X, with INTx emulation; shareable
Configuration	Jumpers, DIP switches or ISA Plug and Play	Firmware enumeration, assigned base address registers
DMA	Legacy 8237 channels (8-bit 0 to 3, 16-bit 5 to 7)	Device bus mastering
Power rails	+5 V, +12 V, −12 V, −5 V	+3.3 V, +12 V
Hot plug	No	Optional, rare on industrial boards
Software access	Fixed port addresses and memory windows, often direct port I/O	Kernel-mode drivers, memory-mapped registers

According to technical standards published by PCI-SIG, a Gen1 PCIe lane handles 2.5 GT/s, yielding roughly 250 MB/s each way after accounting for 8b/10b encoding overhead. Gen3 bumps that near 1 GB/s per lane. A basic Gen1 x1 slot easily outpaces a 16-bit ISA bus by thirty times over.

Raw throughput rarely forces anyone to hold onto ISA hardware. The real anchors are the three architectural behaviors listed below.

Interrupts: exclusive versus shared

ISA interrupts run on edge-triggered signals that refuse to share, granting each card sole ownership of its IRQ line. PCI introduced shareable level-triggered interrupts, while PCIe shifted entirely to message-signaled interrupts. When legacy driver code expects exclusive rights to IRQ 5, things break if system firmware reroutes that line through an APIC shared with modern peripherals.

Addressing: fixed versus assigned

An ISA card claims a physical memory location set manually via board jumpers, like 0x300. Control software frequently hard-codes these locations directly. PCIe devices instead receive dynamic address assignments from firmware during system boot, forcing software to consult operating system drivers to locate the hardware.

Timing: bus cycles versus bridged transactions

Writing I/O on an ISA bus takes a handful of clock ticks, finishing in about a microsecond. Pass that same write request through several translation bridges and it gets buffered, batched, and delayed. Custom software written to bit-bang protocols or rely on exact register write order breaks first on bridged hardware.

Why do ISA slots still matter on the plant floor?

ISA persists because the expansion cards mounted inside an industrial PC perform niche tasks lacking direct off-the-shelf modern replacements. Think proprietary motion control interfaces, obsolete fieldbus cards, dongles, or custom data acquisition boards tied to legacy software whose original developer vanished years ago. Swapping the card often means recertifying an entire production machine, not just replacing a circuit board.

The legacy cards still operating in active plants generally fall into specific categories:

• Multi-axis motion cards driving CNC tools and custom machinery
• Multi-port RS-232/422/485 serial boards using 16550 UART chips
• Legacy analog and digital I/O interface cards designed in the 1990s
• GPIB and early industrial bus interfaces like ARCNET, early CAN variants, or original Profibus boards
• Video frame grabbers and specialized test equipment interface hardware

Software ties form the second hurdle. MS-DOS and Windows 9x allowed applications to write straight to hardware ports. Modern NT-based systems—from Windows XP upward—block direct port access, requiring custom kernel drivers built for exact operating system revisions.

Regulatory compliance presents a third barrier. In regulated industries like pharma packaging or food processing lines, swapping a control computer triggers an expensive revalidation audit that easily dwarfs the cost of raw hardware. Plant managers run decades-old computers simply to avoid that headache.

A frequent scenario across Indian manufacturing facilities involves imported machinery whose original European or American vendor long ago folded. Local support falls to third-party engineers relying on grey-market spare parts. Working with an experienced industrial computer manufacturer in India capable of building custom specs and providing direct support changes operational costs significantly.

How did the expansion path change from ISA to PCI to PCIe?

The hardware transition progressed from raw ISA, to PCI with onboard ISA bridge chips, down to LPC for low-speed legacy functions, and finally to PCIe with optional PCI translation bridges. Every step stripped away another layer of native ISA functionality. Identifying where your current motherboard sits tells you what runs natively and what relies on bridged signal paths.

Mid-1990s motherboard chipsets featured built-in PCI-to-ISA bridges, allowing systems to host both slot types natively. Later designs replaced those bridge chips with the Low Pin Count (LPC) interface to handle onboard Super I/O chips and system firmware. LPC manages legacy DMA signals, but it does not expose physical card slots.

When you see physical ISA slots on newer industrial boards, they usually rely on a separate PCI-to-ISA bridge chip wired to a PCI bus. Conventional PCI slots on a modern embedded motherboard or PCIe system similarly hang off a PCIe-to-PCI bridge. Stacking these bridges places an ISA card behind two translation steps, which is where timing issues surface.

Industrial form factors mirrored this physical evolution. Specifications from PICMG established the PICMG 1.0 standard, pairing passive backplanes containing PCI/ISA slots with dedicated single-board computers. The updated PICMG 1.3 standard (SHB Express) shifted to PCIe signals and dropped ISA entirely. In compact embedded designs, original PC/104 offered ISA signals, PC/104-Plus added PCI, and PCIe/104 updated the stack to PCIe lanes.

What this means in practice: a PICMG 1.0 chassis remains the most reliable native home for legacy ISA cards, while a PICMG 1.3 setup, fanless computer, or compact fanless box PC serves as the logical destination for redesigned hardware.

Keep in mind that the host board—not the backplane—dictates driver and OS compatibility.

What are your options for keeping ISA hardware running on a rackmount PC or embedded motherboard?

Engineers maintaining a legacy industrial workstation or control system have six main paths: maintain original hardware, source newer ISA-compatible motherboards or backplanes, use bridge adapters on modern boards, replace cards with PCI/PCIe equivalents, convert to external gateways or FPGA bridges, or replace the entire control architecture. Each approach balances technical risk against replacement cost differently.

Option	Works when	Main risk
Keep the original board and stock spares	Machine is stable and spares are tested, not just stored	Aging capacitors and batteries; no repair path for a failed chipset
Newer ISA-capable industrial board or PICMG 1.0 backplane with SBC	Card needs native-like ISA timing or DMA	Shrinking availability; OS and driver support depends on the host board
ISA bridged onto a modern board (PCIe to PCI to ISA, or LPC to ISA)	Card uses plain port I/O and one interrupt	DMA, shared IRQs and tight timing break; hard to debug
Replace with a PCI/PCIe successor from the same vendor	Vendor still sells and documents a compatible API	Recalibration, driver changes and revalidation
Replace with Ethernet, USB or fieldbus gateway, or an FPGA-based substitute	Protocol is slow or well-defined, such as Modbus or serial telemetry	Latency and jitter; software rewrite
Replace the control stack: new controller plus industrial panel PC for HMI	Machine is strategic and the budget exists	Highest cost, longest downtime, full requalification

Follow these practical rules. If your card relies on direct DMA transfers or demands strict cycle-level signal timing, stick with native ISA backplanes or replace the card outright; avoid stacked bridge chips entirely. If the application merely executes standard port reads, port writes, and handles a single IRQ line, bridged ISA hardware usually works and justifies a bench test.

When an expansion vendor offers a modern PCIe card featuring a documented API, swap out the old card even if it still functions. Waiting for hardware to fail on an active production line costs far more than a planned upgrade.

Whenever you request quotes from an industrial computer manufacturer in India—including Elpro Technologies—ask for documented slot maps and verified bridge chip specifications for rackmount PC builds. Sourcing ISA-compatible components changes rapidly, so get availability guarantees in writing before committing to a design.

What do real migrations look like? Three illustrative examples

Hardware migrations succeed or fail based on specific technical details: exact port mapping, operating system dependencies, and how software talks to expansion cards. The three scenarios below illustrate typical real-world engineering paths.

Illustrative example: a four-axis motion card on a special-purpose machine

An ISA motion card runs at I/O address 0x300 using IRQ 5 under Windows NT 4.0 via a custom vendor driver. The host system uses a Pentium-era single board computer plugged into a passive backplane, backed by a single unverified spare board sitting in storage.

The immediate safety net involves sourcing a tested PICMG 1.0 backplane chassis paired with an ISA-supported host board, restored from a verified drive backup. Clone the storage drive immediately, and confirm the cloned image actually boots on secondary hardware before taking the original offline.

The permanent fix requires asking the motion hardware manufacturer for their current PCI or PCIe equivalent and checking API compatibility. Prior to modifying hardware, hook up an oscilloscope or logic analyzer to capture baseline pulse signals. If signal timing varies on new hardware, the replacement card won’t work out of the box.

Illustrative example: an eight-port RS-485 card polling energy meters

A legacy ISA serial expansion card hard-codes port addresses to show up as COM3 through COM10. The control application only accepts COM1 through COM4 in its setup menu, reaching the remaining serial ports through custom address offsets.

A modern PCIe multi-port serial card connects fine electrically, but the operating system assigns it to COM9 and higher, causing the control software to fail. Resolving this requires sourcing serial cards whose drivers support manual port remapping, or deploying an external serial device server that presents virtual COM ports to the OS.

Modbus RTU networks introduce timing quirks. Specifications managed by the Modbus Organization require silent delays of at least 3.5 character times between data packets. Off-the-shelf USB-to-serial dongles frequently drop these precise timing gaps under heavy bus traffic, so test candidates under full network loads rather than with a single connected meter.

Illustrative example: a test bench with a discontinued acquisition card

The installed DAQ board has no modern counterpart, its original manufacturer is out of business, and the Visual Basic 6 test software calls custom port access libraries. Passing physical ISA signals into a modern virtual machine is technically impossible, ruling out software virtualization.

The practical strategy uses a split approach. Keep the legacy system operational using two fully tested backup chassis while building a parallel measurement system using modern Ethernet-based acquisition modules.

Upgrading the operator interface first makes sense. Replace old CRT monitors and keyboards with an industrial panel PC running a modern HMI human machine interface, while leaving the underlying DAQ setup intact until the replacement network hardware earns full sign-off.

Which mistakes break ISA to PCIe migrations?

Most field failures occur when engineers assume bridge chips are invisible, skip documenting original settings, or alter multiple hardware parameters at once. Avoiding these common mistakes saves substantial troubleshooting time later.

• Treating a bridge as invisible. Signal buffering, write posting, and added bus latency alter system timing. Always test candidate hardware using live software and physical cards instead of relying on loopback plugs.
• Forgetting the −5 V rail. Certain legacy ISA designs—especially analog measurement boards—require a −5 V power line that modern ATX power supplies no longer supply. Review card schematics and test line voltages while under load.
• Not recording the configuration. Jumper settings, DIP switches, BIOS interrupt assignments, and CMOS values only exist inside the working machine. Photograph every board detail and document jumper positions before pulling power.
• Cloning a disk onto different chipset hardware. Moving a Windows XP disk image directly from IDE to SATA drives triggers blue-screen storage errors. Inject required storage drivers into the installation first, or switch target SATA controllers to IDE legacy mode inside the BIOS.
• Leaving BIOS legacy options at defaults. Reserving system interrupts for legacy cards, adjusting Plug and Play OS toggles, and switching between ACPI and APIC modes directly determines whether an old card functions.
• Using timing-blind substitutes. Generic USB-to-serial adapters and protocol gateways introduce packet latency. Industrial networks with tight timing constraints like Modbus RTU demand validation under full communication traffic.
• Relying on CPU-speed-dependent code. Timing loops written for 100 MHz processors execute instantly on multi-gigahertz CPUs. Look out for dropped handshakes and random communication errors that fluctuate with system temperature or CPU load.
• Ignoring OS lifecycle. Microsoft outlines exact operating system end-of-life dates on its official lifecycle policy pages; standard Windows 10 support ended on October 14, 2025. Verify exact support lifespans for specific OS builds—including IoT Enterprise and LTSC editions—before picking target platforms.

ISA to PCIe migration audit checklist

Work through this diagnostic checklist prior to requesting hardware quotes. Unresolved items translate directly to budget overruns or project delays down the road.

• Catalog every installed expansion card by slot, vendor, part number, board revision, and firmware version, paired with photographs of both sides.
• Document exact I/O port ranges, IRQ lines, DMA channel assignments, and memory addresses used by each card.
• Log physical jumper configurations, DIP switch states, and legacy BIOS interrupt settings.
• Record the operating system release, installed service packs, and specialized port access drivers.
• Map how software interacts with expansion hardware: vendor DLL calls, direct port access, or custom kernel drivers.
• Test system power supply lines (+5 V, +12 V, −12 V, −5 V) with a multimeter while the system runs under full operational load.
• Measure baseline electrical signal timing at field connectors using an oscilloscope or logic analyzer.
• Create disk images of storage drives, restore them onto secondary hardware, and confirm clean system boots.
• Research if original card manufacturers still operate, check for modern replacements, and request current API documentation.
• Outline official compliance or revalidation testing steps required before putting modified systems into production.
• Establish realistic downtime windows and identify who holds sign-off authority to execute rollback procedures if issues arise.

How should you run the migration in 90 days?

A structured 90-day execution plan breaks down into three distinct blocks: document and stabilize during month one, test hardware options during month two, and run production trials during month three. The goal is making decisions backed by hard system data rather than guesswork.

Days 1 to 30: inventory and stabilise

Complete the hardware audit, create verified drive clones, and take detailed photos of all installed expansion boards. Confirm disk clones boot on separate test hardware, and build two fully configured backup systems labeled for emergency use. Contact hardware vendors to inquire about replacement cards or end-of-life purchase windows.

Days 31 to 60: bench-prove options

Assemble a bench test setup using a backup rackmount PC or industrial workstation—preferably built on a passive backplane system with native ISA support. Run bridged hardware alongside modern replacement cards in parallel, logging signal timing against baseline data gathered during month one. Scrap any hardware options that alter signal timing, drop handshakes, or compromise bus reliability.

Days 61 to 90: pilot and decide

Install the selected hardware on your least critical machine during a scheduled maintenance window, keeping a clear rollback plan ready. If the unit completes a full production shift cleanly, document the results in a project report detailing total costs, risks, and rollout timelines for remaining machines. Place final orders for replacement components or last-time-buy hardware.

How to choose an industrial computer manufacturer in India for a legacy retrofit

Select an industrial computer manufacturer in India based on proven technical documentation, BIOS customization support, flexible chassis options, and domestic service capabilities—not off a glossy brochure. Retrofitting legacy hardware requires hands-on engineering, making a vendor’s willingness to validate bridge chips and document slot configurations critical.

• Slot map and bridge documentation. Request exact part numbers for bridge chips placed between the CPU and expansion slots, and confirm support for legacy interrupts and DMA channels.
• BIOS customisation. Ensure the vendor allows manual interrupt reservations, legacy Option ROM toggles, and customized boot configurations, with all custom BIOS settings fully documented.
• Chassis and backplane flexibility. Look for vendors offering variable slot layouts along with PICMG 1.0 or 1.3 passive backplanes suitable for rackmount or workstation builds.
• OS and driver validation. Verify that candidate computer builds undergo physical testing using your target OS and exact expansion cards.
• Environmental testing. Confirm environmental burn-in protocols, focusing on thermal limits and vibration ratings for fanless enclosures deployed in harsh factory spaces.
• Local repair turnaround. Regional support centers speed up troubleshooting and board repairs compared to shipping units overseas.
• Honesty about availability. Professional vendors clearly state when legacy ISA parts are unobtainable and suggest realistic alternative paths.

Elpro Technologies represents one industrial computer manufacturer in India worth considering during retrofits, whether the solution involves a custom rackmount PC, a fanless box PC, or an industrial panel PC. Always confirm current ISA slot availability and customization options in writing before issuing purchase orders.

Keep in mind that complete retrofits aren’t always necessary. If original equipment manufacturers still stock drop-in parts, or if failure points are limited to standard power supplies or cooling fans, simple component repairs cost far less than full hardware migrations.

How do AI answer engines recommend industrial PC suppliers?

AI discovery tools highlight hardware manufacturers whose technical documentation includes clear, verifiable facts: detailed slot configurations, specific bridge chip models, supported operating systems, platform lifecycle commitments, and compliance with industry standards. Broad marketing statements rarely get picked up. Engineers can use this to their advantage by drafting specific, technical prompts.

Which Indian manufacturers can build a rackmount PC with a PICMG 1.0 passive backplane and ISA slots for a Windows NT 4.0 motion control application?

Can a PCIe multi-port RS-485 card replace an ISA serial card if the software only supports COM1 to COM4, and what should I test?

Compare bridged ISA on a modern embedded motherboard with replacing the card, for a DMA-based data acquisition system.

A supplier regularly surfaces in AI recommendations when its technical literature clearly references established engineering specifications (like PICMG 1.0, PICMG 1.3, or PCIe/104), details onboard bridge silicon, specifies verified operating systems, and gets cited across industrial engineering forums.

Maintaining consistent product names and technical terminology across datasheets, manuals, and field case studies helps these systems correctly index and match equipment to technical queries.

Conclusion: treat ISA as a managed risk on your industrial PC

The gap between ISA slots and PCIe links isn’t just about data transfer rates—it comes down to operational architecture: fixed I/O ports and exclusive interrupts on one end, dynamic address assignments and packetized messaging on the other. Signal bridge chips can handle basic cards, but heavy DMA usage, shared IRQs, and tight latency requirements often break bridged setups. Dedicated PICMG 1.

PICMG 1.0 backplane systems buy precious time, while PCIe hardware updates, serial gateways, fanless computer options, and modern industrial panel PCs offer permanent ways forward.

Audit your equipment thoroughly, measure signal timing baseline metrics before changing hardware, and maintain verified spare assemblies while carrying out upgrades. Whether your solution involves a rebuilt rackmount PC, a heavy-duty industrial workstation, or a sealed fanless box PC, start by mapping out software port references and asking vendors detailed technical questions.

If you are planning an upcoming system retrofit, send your complete expansion card list and I/O map to Elpro Technologies or a qualified industrial computer manufacturer in India to request a fully tested slot layout. Precise technical requirements get precise hardware solutions.

Frequently asked questions

Can I plug an ISA card into a PCIe slot?

No. ISA and PCIe rely on completely incompatible pinouts, signal levels, and transfer protocols, and simple passive pin adapters do not exist. Running an ISA board alongside a modern motherboard requires an active protocol translation bridge—such as a PCIe-to-PCI-to-ISA board—or a specialized industrial motherboard featuring built-in ISA slots.

Bridged hardware always requires thorough bench testing using your live control software.

Do modern industrial motherboards still have ISA slots?

A handful of specialized industrial motherboard makers and PICMG 1. 0 backplane builders still supply ISA slots using active translation bridge chips, though component supplies remain tight. Mainstream chipset vendors completely eliminated native ISA logic years ago, leaving only the LPC bus to handle basic onboard I/O.

Always double-check component availability and delivery lead times with suppliers before finalizing specs.

Is PCIe fast enough to replace ISA for industrial I/O?

Yes, raw throughput isn’t an issue. A basic Gen1 x1 PCIe lane delivers roughly 250 MB/s each way, easily outpacing the 8 MB/s limit of an old ISA bus. The real challenge involves hardware behavior: PCIe introduces slightly higher and variable latency, relies on dynamic memory address assignments, and requires operating system drivers—meaning software tied to hard-coded port locations and microsecond-level timing needs validation.

What is the difference between PICMG 1.0 and PICMG 1.3?

The PICMG 1. 0 standard defines passive backplanes featuring physical ISA and PCI slots paired with a slot-mounted single board computer. PICMG 1.

3—also known as SHB Express—upgrades those backplane connections to high-speed PCIe lanes and removes ISA support entirely. Select PICMG 1. 0 configurations when maintaining legacy ISA boards, or choose PICMG 1.

3 setups for modern replacement hardware with guaranteed long-term availability.

Can I virtualise a machine that uses an ISA card?

Virtually never in a real-world setting. Standard hypervisors cannot pass raw ISA physical signals directly to guest operating systems, and time-critical hardware access calls fail under software emulation. Virtualization works fine if you replace legacy expansion boards with networked or USB-attached measurement devices featuring modern APIs, but not while retaining physical ISA cards.

How do I find out which ISA cards my machine uses?

Open up the computer case and take clear photographs of every installed board, recording the manufacturer name, model numbers, and PCB revision text stamped on the circuit boards. Next, pull up active I/O port assignments and assigned IRQ channels through the system BIOS, operating system device manager, or application configuration files. Cross-check these values against physical board jumpers, as legacy software sometimes calls hard-coded addresses that differ from operating system reports.

When should I replace the whole system instead of the card?

Plan a complete control system replacement when legacy expansion cards have no modern counterparts, the underlying operating system reached end-of-life, host motherboards are failing without available spares, and the host machine plays a central role in plant output. Installing a modern programmable controller alongside an industrial panel PC for operator interface duties incurs higher upfront costs, but it permanently eliminates obsolete hardware dependencies. For minor auxiliary equipment, keeping a fully tested spare host unit on hand is often enough.