Ice Machine Water Filtration, Explained: How Scale Kills Machines and How to Size a Filter
Ice machine water filtration, explained
An ice machine is a heat exchanger that happens to make a product you serve. Water freezes onto a chilled evaporator plate, the harvest cycle releases it, and the cycle repeats — a few hundred times a day, every day. Everything that was dissolved in the water and did not become ice stays behind on that plate.
That residue is the whole story. It is why ice machines fail early, why they get slower before they fail, and why the manufacturer's warranty has a clause in it about water that most operators never read until the week they need it.
This guide covers what scale actually does, how to size a filter so it protects the machine instead of throttling it, and what the manufacturer documents say — quoted directly, because in this category the documents are more useful than the marketing.
What scale is, and why it costs you before it kills anything
Freezing is a purification process. When water crystallises into ice, it rejects most of what is dissolved in it — calcium, magnesium, silica. Those minerals concentrate in the water that is left, and precipitate out onto the evaporator plate and into the water distribution tubes, the sump and the pump.
The result is a hard mineral crust on the exact surface whose job is to move heat.
That matters because scale is a very poor conductor of heat. The U.S. Department of Energy makes this point in its guidance on boiler heat-transfer surfaces: scale "typically possesses a thermal conductivity, an order of magnitude less than the corresponding value for bare steel," so even thin layers "serve as an effective insulator and retard heat transfer."
The DOE's measured numbers are worth looking at, with one important caveat. They come from firetube boilers, not ice machines — this is data about what scale does to a heat-transfer surface in general, not a measurement of your ice machine:
| Scale thickness | Fuel loss, "normal" scale | Fuel loss, iron + silica scale |
|---|---|---|
| 1/64″ | 1.0% | 3.5% |
| 1/32″ | 2.0% | 7.0% |
| 3/64″ | 3.0% | — |
| 1/16″ | 3.9% | — |
Source: U.S. DOE, Energy Tips — Steam, Tip Sheet #7, extracted from NIST Handbook 115, Supplement 1.
Two things are worth taking from that table.
The first is how thin the layers are. A sixty-fourth of an inch is a film you would struggle to see, and it is already costing measurable energy. The second is the right-hand column: the composition of the scale matters as much as the thickness. Iron in the feed water makes deposits dramatically worse — 3.5% against 1.0% at the same thickness. If your water has iron in it, you have a harder problem than your neighbour with the same machine and the same hardness.
You will find a widely-repeated claim online that 1/8″ of scale costs 25% efficiency. We are not going to repeat it, because we went looking for its source and could not find one — the DOE table it is usually attributed to stops at 1/16″ and 3.9%. The honest version of the claim is the one above, and it is bad enough.
What you actually notice, in order
Scale does not announce itself. It presents as a sequence of small operational annoyances that most operators attribute to something else:
- Longer harvest cycles. The insulating layer slows heat transfer both ways, so the ice takes longer to release. The machine is now running more hours for the same ice.
- Lower daily yield. The spec sheet says 500 lb/day. You are getting 400. Nobody measures this, so it usually goes unnoticed until a hot weekend runs the bin dry.
- Cloudy, soft, fast-melting cubes. Mineral inclusions in the cube. Your ice waters drinks down faster, which your bartenders will notice before you do.
- Distribution problems. Scale narrows the water distribution tubes, so the plate no longer floods evenly and cube formation gets irregular.
- Pump and valve failures. Abrasive particulate circulating through the sump. This is where scale stops being an efficiency problem and becomes a repair bill.
The commercially important thing about that list is that stages one through three cost you money continuously — in electricity, in lost ice, in drink quality — for months or years before anything actually breaks.
What your warranty says about water
This is the part worth reading carefully, because equipment manufacturers have thought about scale considerably more than their customers have, and they have written their conclusions into the warranty.
Here is Hoshizaki, from the current KM Series limited warranty. Under the heading "This warranty does not include parts or labor coverage for component failure or other damage resulting from," the list includes, verbatim:
"External water supply failure or plumbing problems to the Product for any reason."
And:
"Failure to clean and/or maintain Product as set forth in the owner/user manual for the Product."
Read the first one again, particularly the last four words. For any reason. Water coming into the machine is outside the boundary of what Hoshizaki warrants. That is not a hostile position — they cannot control what is in your municipal supply — but it does mean the water side of the machine is unambiguously the operator's problem.
There is a second clause that matters even more, and almost nobody quotes it:
"In no event shall the user be entitled to recover incidental or consequential damages, including but not limited to, damages for inconvenience, ice purchase, rental or replacement equipment, loss of profits, water damage, or other commercial loss."
Ice purchase. If the machine goes down on a Friday and you spend the weekend buying bagged ice from a cash-and-carry, that cost is explicitly yours. So is the lost revenue if you run out. The warranty covers the part. It does not cover the weekend.
For completeness, the KM Series coverage itself is three years parts and labour, with a further two years on the evaporator and on the compressor and air-cooled condenser (parts only). Check your own model's warranty statement — terms differ by series and they change.
We go through the equivalent clauses for espresso and other equipment in our warranty roundup.
Ice is food, and the inspector knows it
The other reason this equipment gets attention comes from two definitions in the FDA Food Code, and it is worth walking through them rather than asserting the conclusion.
The Code's definition of "Food" at §1-201.10(B) names ice explicitly. And §4-601.11(A) requires that food-contact surfaces be "clean to sight and touch" — a Priority foundation item, in the Code's own classification.
Put those together and the surfaces that touch ice — the evaporator, the chute, the bin, the scoop — are treated as food-contact surfaces. That last step is our reasoning from the definitions, not a sentence the Code contains, and we would rather show you the join than pretend the Code says it outright.
One important caveat that most articles on this subject skip: the FDA Food Code is a model, not law. Its own preface states it is "neither federal law nor federal regulation and is not preemptive." States adopt it on their own timetable and with their own variations — as of the most recent FDA adoption table only seven states were on the 2022 edition, and Minnesota, where we are, is still on the 2013 Code. Your jurisdiction governs, not the model. Check your state and county rather than assuming the national document applies as written.
Filtration is not a substitute for cleaning — nothing upstream of the machine will stop biofilm growing inside a bin that has not been sanitised on schedule. But the two problems are related in a practical way: sediment and particulate give biofilm something to establish on, and a machine that is being descaled less often is opened and inspected less often.
We cover what inspectors actually look at in the ice machine health inspection guide.
How to size an ice machine filter
Here is where most of the money is wasted, in both directions. Sizing a commercial ice machine filter comes down to three numbers, in this order.
1. Service flow rate — the number that must not be exceeded
This is the constraint people get wrong. A filter cartridge has a maximum service flow rate, and it is a hard limit, not a suggestion. Push more water through it than it is rated for and two things happen: pressure drops across the cartridge, and contact time with the media falls, so the filtration you paid for does not happen.
Manufacturers state this in unusually direct language. From the Everpure Insurice Single-i2000² spec sheet:
"Service flow rate must not exceed 1.67 gpm (6.3 Lpm)."
Your machine's peak draw during a fill cycle — not its average over a day — has to sit under the filter's rating. That peak is in the ice machine's own specification sheet. Here are the Insurice systems, as an illustration of how the range is structured:
| System | Part # | Max service flow rate | Rated capacity |
|---|---|---|---|
| Insurice Single-i2000² | EV9324-01 | 1.67 gpm | 9,000 gal |
| Insurice Twin-i2000² | EV9324-02 | 3.34 gpm | 18,000 gal |
| Insurice Triple-i4000² | EV9325-03 | 5 gpm | 36,000 gal |
| Insurice Quad-i4000² | EV9325-04 | 6.68 gpm | see note |
Source: Pentair Everpure system specification sheets EV9324-01, EV9324-02, EV9325-03, EV9325-04. Note: the Quad's spec sheet states flow rate but does not state a rated capacity; we have not been able to verify the 48,000-gallon figure that appears elsewhere, so we do not publish it.
Notice the pattern. Going from a single to a twin does not buy you finer filtration — it buys you flow and capacity. Manifolding cartridges in parallel splits the flow between them, so each one stays inside its own rating while the system as a whole keeps up with a bigger machine. That is the entire design logic of the range, and it is why "which one is better" is the wrong question. We take that comparison apart in detail in Everpure Insurice: i2000² vs i4000².
If more than one machine shares the line — an ice machine and a fountain, say — add their peak draws together. This is the single most common sizing error we see.
2. Rated capacity — the number that sets your running cost
Capacity is how many gallons the cartridge treats before it is spent. It determines how often you buy a cartridge, which is the real operating cost of filtration.
Work it out from your actual production, not from the machine's nameplate — and take the water-per-ice figure from your machine's own spec sheet, where it is listed as gallons per 100 lb of ice. It varies substantially by machine type, by ambient conditions, and by how much water goes down the purge rather than into the cube, so a general figure is not worth much and we are not going to invent one.
Once you have that number, the arithmetic is one division, and we work it through with a real example in how often to replace a commercial water filter.
Whichever answer you get, compare it against six months — and that brings us to the third number.
3. Six months, whichever comes first
The Everpure Insurice sheets carry the same instruction — this from the Twin (EV9324-02):
"Change cartridges on a regular six (6) month preventative maintenance program."
And separately, change when capacity is reached, or when pressure falls below 10 psi.
Six months is not a marketing cadence, and it is not really about the carbon being exhausted. It is about what a warm, wet cartridge that has been sitting in a foodservice environment becomes over time. Capacity and time are two separate limits and whichever arrives first governs.
In practice this means most operators are on a six-month cycle, and the capacity calculation exists to catch the high-volume sites where six months is too long.
What you can skip
Sizing up beyond your flow requirement buys you capacity, not protection. If you are comfortably inside the flow rating of a single cartridge and you replace it every six months regardless, a twin system mostly buys you a longer interval you are not going to use. Spend the money on water testing instead.
What a filter does and does not fix
Being precise about this is the difference between buying the right thing and being disappointed.
Sediment and particulate — yes. Straightforward mechanical filtration. The Insurice systems filter particles "as small as 0.5 micron in size by mechanical means," per their spec sheets.
Chlorine taste and odour — yes, via activated carbon. This is what makes ice taste like ice rather than like a swimming pool.
Chloramine — usually not, and this catches people out. Many utilities have moved from chlorine to chloramine, which is more stable and considerably harder to remove. It needs catalytic carbon and more contact time. A filter that says "reduces chlorine taste and odor" is not making a chloramine claim — the Insurice sheets do not mention chloramine anywhere. Find out what your utility uses before you buy. Chloramine vs chlorine covers how to tell and what to do about it.
Hardness — this is the important nuance. Scale-inhibiting cartridges do not remove hardness. Everpure describes a "self-contained scale inhibitor feed" on the Insurice sheets: something is dosed into the water rather than taken out of it, so the minerals pass through the machine instead of depositing on the evaporator. The water is still hard; it is just less inclined to build a crust.
We are describing what the manufacturer's sheet says the cartridge does. The underlying chemistry is often explained as sequestration, but we have not sourced that to a manufacturer or independent document, so we are not asserting it. That is usually the right trade for an ice machine, and it is why a filter is not a softener.
Very hard water — at some level of hardness, sequestration alone stops being enough and the answer is reverse osmosis or a blended RO system. That is a bigger install and a bigger budget, and you should know which side of the line you are on before you spend anything.
Existing scale — no. A filter prevents; it does not descale. If your machine is already scaled, it needs cleaning first, or you are protecting a problem you already have.
Biofilm in the bin — no. That is a cleaning schedule, not a filter.
Every Everpure system carries this qualifier, and it is a fair one:
"The contaminants or other substances removed or reduced by this drinking water system are not necessarily in your water."
Which is the argument for testing your water before you buy a filter, rather than after.
Start with your water, not with the catalogue
Two operators with identical machines in different cities need different filters. The variables that matter are hardness, iron, silica, chlorine or chloramine, and pH — and they are knowable. Your municipal utility publishes an annual water quality report, and a test kit will tell you what is actually arriving at your tap after the building's plumbing has had its say.
Spec the filter from a measurement. It is a cheaper way to find out than the alternative, which is finding out from a service call.
- Browse ice machine filters
- Water test kits
- The full picture: the complete bar and lounge water guide
Never miss a filter change.
Join the reminder list — we'll tell you when your cartridge is due, free, whether or not you bought it from us.
Sources: Hoshizaki America, Inc., "Manufacturer's Limited Labor and Parts Warranty — KM Series Ice Machines"; U.S. Department of Energy, "Energy Tips — Steam, Steam Tip Sheet #7," data extracted from NIST Handbook 115, Supplement 1; Pentair Everpure Insurice system specification sheets EV9324-01, EV9324-02, EV9325-03 and EV9325-04. Specifications are the manufacturers' and current at the time of writing; the manufacturer's own current documentation governs.