A radio cache looks like a solved problem when it is sitting in the closet with everything charged and the lids closed. It stops looking solved somewhere around the second operational period, when the alkaline clamshells are gone, the sign-out sheet has forty illegible lines on it, half the issued radios are in the pockets of people who went home, and two mutual aid companies on the same street are pressing the button on a channel named TAC 3 that is not the same TAC 3. This piece covers cache programming, charging at scale, issue and recovery accountability, and the failure that puts a radio in every hand and no traffic on the air.

What changes on day three

The first operational period of a large incident runs on inventory. Somebody opens the cases, the radios come out at full charge with fresh batteries, the people receiving them are mostly from the home agency and mostly know the fleetmap, and the person handing them out remembers who got what because there were only eighteen of them. Nothing in that picture stresses the system, which is why tabletop exercises and the annual inventory almost never surface the problems that matter.

By the second operational period the initial batteries are flat, the second set is in chargers that are competing with the light plant and the coffee pot for a generator that was sized for neither, and the people arriving are from agencies whose radios do not share a single common channel with yours. By the third period the original cache manager has gone home for crew rest and handed the clipboard to somebody who was not there when the first sixty radios went out. Recovery from that point is very hard, because the record of who has what was never written in a form a second person could read and continue.

The reason this gets called a logistics problem rather than a radio problem is that everything failing at that point is consumable, replaceable or countable rather than technical. You need cells, charger pockets, electrical capacity, a table with a person behind it, a legible log, and a supply chain that can bring more of all of it before you run out. The National Wildfire Coordinating Group has organized the interagency Communications Unit around exactly that distinction for decades, which is why the unit has a Communications Unit Leader for planning, a Communications Technician for the equipment, an Incident Communications Center Manager for the cache and the incident communications center, and radio operators for the traffic. Small agencies collapse all four into one person, and day three is when that person runs out.

What is in a cache and who owns which tier

Caches exist in tiers, and knowing which tier you are actually going to get is most of the planning. At the local level, a county fire or police agency typically keeps some number of spare portables, spare batteries, a few mobile radios in ready boxes, and possibly a portable repeater or a gateway. That cache is sized for a working structure fire with heavy mutual aid or for covering a multi-day search, and it is the only tier that will be on scene in the first two hours.

At the state level, many states maintain a cache through the state emergency management agency or the state forestry agency, frequently purchased with homeland security grant money after 2003 and frequently sitting in a trailer with a mobile communications vehicle. These vary enormously in size, condition, programming and how fast they can be requested, so the useful step is to find out now, in writing, what your state holds, what the request path is, and how long it took the last time somebody used it.

At the national level, the interagency wildland fire community operates the largest and best documented public safety radio cache in the country, run by the National Interagency Incident Communications Division at the National Interagency Fire Center in Boise. It ships preassembled kits rather than loose radios, which is the single most important design decision in the whole model, because a starter system, a command repeater kit, a tactical kit and a logistics kit each arrive as a defined package with radios, programming, antennas, cables and documentation, so the receiving incident does not have to assemble a working system out of parts under pressure. Ordering runs through the established resource ordering process rather than through a phone call to Boise, and the current kit contents, ordering procedure and return requirements are published by NIFC, so pull them from the source rather than from the memory of whoever went to the class in 2014.

FEMA’s mobile emergency response support detachments and several urban search and rescue and metropolitan task force programs hold their own caches on similar logic. None of the outside tiers arrive quickly enough to matter in the first operational period, which means the local tier and the neighbor’s local tier are the ones that determine whether the first day works.

Template fleetmaps and the clone nobody verified

A cache fleetmap is a different object from a daily-use fleetmap, because the person holding a cache radio has never seen it before and will be reading it in the dark with gloves on. The design goal is that a stranger can find the channel the Incident Radio Communications Plan tells them to find, which means the zone and channel numbering has to be stable across every radio in the cache and has to match the words on the ICS-205 exactly.

What works is a fixed zone order that never changes between codeplug revisions. Zone one carries the host agency’s operational channels, zone two carries incident tactical channels in numeric order, and a dedicated zone carries the national interoperability channels with the standard names published in the National Interoperability Field Operations Guide, so VCALL10 and VTAC11 through VTAC14 on VHF, UCALL40 and UTAC41 through UTAC43 on UHF, and the 700 MHz and 800 MHz interoperability sets on systems that have them. Use the NIFOG names and nothing else in that zone, because the whole point of those channels is that an arriving unit from four counties away already knows the name. Verify the current NIFOG edition for channel names, bandwidths and the correct tone or NAC values rather than trusting a codeplug somebody built years ago, since the narrowband conversion changed names and parameters and not every fleet caught up.

The clone process is where this comes apart. A cache is only as consistent as the last time somebody read every radio back and compared it, and I have opened cache cases where three radios out of twelve had an older codeplug because they were out for repair on the day the fleet was updated. Keep a written codeplug version number, put it on a label on the radio and in the radio’s alias or a dedicated text field where it displays on power-up if the model supports it, and record the version in the inventory sheet. When you update the fleet, the radios in the shop and the radios in the chief’s trunk are the ones that will be missed, so the inventory list is the checklist rather than the pile in front of you.

One more thing belongs in the template, which is a printed one-page channel list in every case, listing zone, channel position, the name displayed, the plain description of what it is for, and whether it is repeated or simplex. People will read paper when they will not scroll a radio, and the paper is what lets a division supervisor tell a stranger where to go by position number.

Cache radios on a trunked system: IDs, affiliation and keys

Everything above assumes conventional channels, and a conventional cache is genuinely simple. A trunked cache carries three administrative dependencies that have nothing to do with the radio in your hand and everything to do with a database you may not control.

The first is the subscriber ID. Every radio on a P25 trunked system needs a unique individual ID provisioned in the system’s subscriber database, and a cache of sixty radios needs sixty valid IDs that are enabled, assigned to the correct talkgroups, and not duplicated against anything else on the system. Duplicated IDs produce behavior that is maddening to troubleshoot in the field, and an ID that was disabled two years ago after an inventory scare will simply refuse service with a message the user cannot interpret. The fix is administrative and slow, so it has to happen before the incident, and it has to be rechecked, because system administrators clean up unused IDs and cache radios look unused by definition.

The second is capacity. Sixty additional subscribers affiliating on one site and keying on a handful of talkgroups is real loading on that site, and on a busy Phase 1 system in a rural area it can be enough to produce busies during the exact hour you needed the cache. If your system administrator can pull channel usage and busy statistics for the site covering your likely incident areas, ask for them and look at what a large deployment would do to the numbers before you find out live.

The third is encryption. A cache radio that is not keyed cannot hear an encrypted talkgroup, and a cache radio that is keyed is a loaded key in a case that gets handed to strangers. Decide which way you are going deliberately, write down who holds the key loader and who is authorized to key a cache radio, and make sure the arriving mutual aid company understands before the first assignment whether their talkgroup is in the clear. The safer default for most agencies is an unencrypted incident talkgroup or conventional channel set for anything the cache touches, with encrypted traffic kept on home-agency radios.

Batteries and charging at scale, with the arithmetic

Manufacturer battery life figures are published against a standard duty cycle, conventionally five percent transmit, five percent receive and ninety percent standby, and that is not what an incident does to a radio. A division supervisor or a safety officer on a busy channel transmits far more than five percent of the time, scan and encryption and display backlighting all draw more, and cold weather takes capacity off the top. My working planning number is that a rated twelve-hour battery gives you something closer to six to eight hours of hard use, which matters because it turns a twelve-hour shift into a two-battery shift and a twenty-four-hour assignment into three.

The rule of thumb that has served me is three batteries per radio in the cache, so that one is in service, one is charging and one is staged, and I would rather buy the third battery than the fourth radio. Batteries also age in storage. Lithium-ion packs lose usable capacity over time whether or not they are cycled, and a cache full of eight-year-old batteries that all read full on the fuel gauge will hand you a set of radios that quit in ninety minutes. Date-code every battery when it enters service and retire on age, not on whether it still appears to charge.

On the nickel-cadmium memory effect, it is worth naming the myth as a myth: the claim that routine partial discharge gives a radio battery a permanent memory is largely false, since the true memory effect was documented in tightly controlled cyclic applications and almost nothing in field radio use resembles those conditions. What actually kills packs in the firehouse is heat, age, and sitting in a rapid charger that keeps them warm for months. It matters less now than it did, because most fleets have moved to lithium-ion, and lithium-ion should not be deliberately deep-discharged at all.

The charging arithmetic is the part that surprises people. Read the input power nameplate on your own multi-unit charger, multiply by the number of chargers you intend to run, and compare that to what the generator, the inverter or the building circuit can actually deliver, because ten six-bay chargers on one twenty-amp circuit in a school cafeteria will trip the breaker and nobody will notice for two hours. Count charger pockets against radios in service and remember that a pocket occupied by a fully charged battery nobody removed is a pocket you do not have. This is precisely why the interagency wildland cache leaned for so long on disposable alkaline cells in battery clamshells, since a case of AA cells moved by the Supply Unit is a logistics problem the supply system already knows how to solve, while three hundred charger pockets and the power to run them is one it does not. If you buy clamshells, check your model’s manual first, because some portables reduce transmit power output when running on an alkaline clamshell.

Issue, track, recover: the paperwork that actually holds

Accountability for issued radios fails in a specific and predictable way. Somebody puts a legal pad on the table, people write a name and a badge number, the handwriting deteriorates as the line gets longer, nobody records the radio’s serial or ID, and at demobilization there is no way to tell whether the radio numbered 27 came back because nothing on the sheet identifies radios by number. The fix costs nothing and has to be built before the incident.

Number every radio in the cache with a large, durable external number that is readable across a room, and put the same number on the battery, the charger pocket if it is assigned, and the case foam cutout, so an empty cutout tells you at a glance which unit is out. Preprint the issue log with the radio numbers already in the left column, one row per radio, so the person at the table fills in the name, agency, unit assignment and time rather than inventing the record. The rows that stay blank are the radios still in the box, and the rows that are filled in are your outstanding list, which any relief can pick up and continue.

Issue to a person and not to a crew, because a radio issued to Engine 12 goes home with whichever of the four people happened to be holding it. Record the accessory too, since remote speaker microphones and surveillance kits disappear at a much higher rate than radios and cost enough to notice. If you have barcode or asset-tag scanning in your supply system already, use it, though I would still keep the paper log as the primary record on an incident, because paper does not need a network and the relief cache manager can read it without a login.

The recovery mechanism that actually works is tying the radio return to the demobilization check-out. The ICS demobilization check-out form carries unit sign-offs, including one for Logistics and Communications, and if the Demobilization Unit enforces that a person does not get released until the Communications Unit has signed, your return rate climbs sharply. Where that discipline does not exist, plan for a nonzero loss rate, price it, and put the replacement cost in the budget request so it is a known number rather than a surprise.

Everyone had a radio and nobody could talk

The Oakland Hills firestorm of October 20, 1991, remains the cleanest documented example in the American fire service. The fire killed 25 people and destroyed close to 3,000 dwellings, and the precise structure count differs between the U.S. Fire Administration’s technical report on the East Bay Hills fire and later published accounts, so give the range rather than a tidy figure if you cite it. Mutual aid arrived from across the region on a scale the Bay Area had rarely seen, and the federal technical report and the subsequent state and local reviews documented that arriving companies could not communicate with Oakland units because their radios were not on compatible frequencies, alongside the better remembered finding that outside engines could not connect to Oakland hydrants because of incompatible hose thread. Every one of those companies had radios, and every one of those radios worked exactly as designed on a system nobody else was on.

The wildland side has its own record. The interagency investigation of the 1994 South Canyon fire on Storm King Mountain in Colorado, which killed fourteen firefighters, documented communications problems among them, including crews from different organizations operating on frequencies that did not give everyone a common path. Read the investigation report itself rather than a summary, because the communications findings sit inside a much larger set of factors and a slide deck version of that fire will mislead you about what the investigators actually concluded.

The modern version of this failure is subtler and, in my experience, more common. Everyone arrives with a programmed radio, the Incident Radio Communications Plan assigns divisions to TAC 2 and TAC 3, and three of the responding agencies each have a channel labeled TAC 3 in their own home fleetmap on a completely different frequency, so crews select it, hear nothing, assume the radio is broken, and switch back to their home dispatch channel where they can at least hear somebody. The cache exists to break that cycle by putting the same codeplug in every hand, but only if the cache radio is the one actually used, only if the ICS-205 names channels by the exact alias displayed on the cache radio, and only if somebody at the staging area tells arriving crews to take one instead of leaving the cases stacked by the command post.

Demobilization and the reset that makes the cache real

A cache comes back from an incident in worse condition than anyone admits in the after-action meeting. Batteries are flat and mixed up, some radios have been reprogrammed or had channels changed by users on models where that is possible, antennas are bent, belt clips are missing, and one radio in ten has water or ash in the speaker grille. If the cases go back into the closet in that condition, the next incident starts with a cache that looks full and performs like a scrapyard.

Build the reset into the demobilization assignment rather than treating it as something to do later. One person, named in advance, inventories against the numbered list, reads each radio back against the current codeplug version and reflashes any that do not match, cycles every battery through a charger with an analyze function if you have one, replaces missing accessories from stock, and signs a single sheet saying the cache is restored to baseline with a date. That sheet is what tells you next month whether the cache is actually ready, since the closed lid tells you nothing.

Restocking the consumables is the part that gets skipped because the consumables are boring. Alkaline cells, earpieces, zip ties, printed channel lists, blank issue logs, adapters, spare fuses for the mobile kits and a few feet of coax all get used up on a long incident, and none of them are on the equipment inventory. Keep a consumables list inside the lid, check it during the reset, and route the replacements through your normal supply requisition so the cost shows up in the budget where it belongs.

If you drew from a state or national cache, the return requirements are contractual and they are specific about condition, packaging and timelines, with charges for missing or damaged items. Assign that return to a named person with the paperwork in hand before the demobilization rush starts, because a kit returned incomplete three weeks late is a bill and a reputation problem the next time your region orders one.

What to do at your agency

• Have your communications officer open every cache case this month, read each radio’s codeplug back against the current version, and write the codeplug version number on a label on each radio and in the inventory sheet so the next person can tell at a glance which ones are stale.
• Print a numbered issue log with one preprinted row per cache radio, laminate a copy of the channel list showing zone, channel position and displayed alias, and put both inside the lid of every case before it goes back on the shelf.
• Read the input wattage from the nameplate of one of your multi-unit chargers, multiply by the number you would deploy, write that total and the pocket count on the outside of the case, and give the number to whoever sizes your generator or mobile command vehicle power.
• Ask your radio system administrator, in writing, for a list of the subscriber IDs assigned to your cache radios with their enabled status and talkgroup assignments, and keep the reply in the communications file so you can prove the cache will register before you need it.
• Have the training officer put one agenda item on the officers’ meeting that already occurs this month: every ICS-205 written by this agency will name channels using the exact alias displayed on the cache radio, with zone and channel position included.
• Pull the date codes on every battery in the cache, retire anything past the age your manufacturer specifies regardless of how it reads on the charger, and put the reorder in this budget cycle rather than next.
• Call your state emergency management agency or state forestry communications contact and find out what state-level cache exists, what the request path is, and how long the last activation took from request to arrival, then write those three answers into your communications annex.

Takeaways

• A cache that performs well on the first operational period tells you nothing about day three, because the failures are consumables, electrical capacity, staffing of the issue point and continuity of the record rather than radio engineering.
• The interagency cache model at the National Interagency Fire Center ships complete kits with programming, antennas, cables and documentation packaged together, and copying that packaging discipline at local scale is more valuable than owning more radios.
• Cache fleetmaps need a fixed zone and channel order that never moves between revisions, with national interoperability channels carrying their published NIFOG names so an arriving crew from another region already knows what they are looking for.
• Trunked cache radios depend on valid, enabled, non-duplicated subscriber IDs in a database you may not control, on site capacity that a large deployment can exhaust, and on a deliberate decision about whether cache radios carry encryption keys at all.
• Published battery life figures assume roughly a five percent transmit, five percent receive and ninety percent standby duty cycle, which is not what an incident does, so plan around three batteries per radio and retire batteries on date code rather than on whether they still appear to charge.
• The nickel-cadmium memory effect as commonly described is a myth for field radio use, and what actually destroys packs is heat, age and long residence in a warm rapid charger.
• Number every radio externally, preprint the issue log with those numbers, issue to a named person rather than to an apparatus, and tie release from the incident to a Communications Unit signature on the demobilization check-out.
• Oakland Hills in 1991 killed 25 people and destroyed close to 3,000 dwellings, and the federal technical report documented that arriving mutual aid companies could not communicate with Oakland units despite every company carrying working radios, which is the same failure that happens today when three agencies each have a different channel named TAC 3.
• A cache is only ready if somebody named in advance inventoried it, reflashed the mismatched radios, cycled the batteries, restocked the consumables and signed a dated sheet saying so.