Feed water pre-treatment to meet the electrolyser specification
We source and supply the water treatment train that fits your electrolyser. We set the arrangement from your source water, unit count and daily demand, and carry the overseas purchase and customs ourselves.
WaterTreatmentCirculationCell / stackDryingAnalysisCompressionWhat this part does and what decides it
The electrolyte circulates and keeps coming back. Whatever the feed water carries in has nowhere to go, so it builds up in the loop. Ions end up on the membrane and the catalyst surface, organics and microbes build a layer on the flow field and the membrane, and particles such as sand or rust block the circulation pump and the narrow channels. Feed water quality is not something you set once at commissioning; it is a condition you hold throughout operation. This is why "our water is clean" does not answer the question here. It usually means the water is fit to drink, and drinking water quality and electrolyser feed water quality are two different specifications. The minerals that make water taste good are what form scale on our side. We work with two feed water figures. The unit used here, µS/cm, measures how many ions the water carries, and a smaller number means fewer ions. The HX recommended basis is ASTM D1193 Type II, at or below 1.0 µS/cm at 25 °C. It is deliberately conservative, and it is the figure printed on the datasheet. The HX in-service basis is 30 µS/cm or below, which is closer to our actual operating record and what we normally work to. The in-service figure is a line we set from operating experience. Crossing it does not mean the unit fails, and it does not mean the warranty is gone. But where a contract or specification states 1.0 µS/cm, that is the figure that applies.
On HXB-V1 you connect your own water tank to the feed inlet and the built-in gear pump refills as needed, so no external feed pump is required. What matters to us is therefore not the pump but what goes into that tank. We set the feed water basis the electrolyser runs on. So when a treatment package is being chosen, we check first which of the two figures above the water leaving it actually reaches. Bought separately, that check is yours to make.
The figures are the conductivity of the water leaving each stage, in µS/cm. They are a reference range for mains water as the source, not guaranteed values. Korean mains water runs roughly 100 to 300 µS/cm and varies by region and season. Reverse osmosis is the only stage that actually brings conductivity down; the three ahead of it are there to protect the RO membrane.
The sections and tables below use water treatment terms. Ten of them cover the rest.
Each species attacks a different place. Read it against your own water analysis.
Water carrying a lot of calcium and magnesium, that is, hard water, gets softening or scale control ahead of the RO. Fed straight to the membrane, that water crusts the surface and the membrane is the first thing to go. What goes at the front is not set from hardness alone; we calculate it together with alkalinity and recovery. All three terms are defined in "The water-side vocabulary first" above. The HX in-service basis is reachable on reverse osmosis alone. But it leaves no margin if the source water moves, so we look at an ion exchange stage alongside it. To reach the HX recommended basis, a demineralisation stage after the RO is mandatory.
The Treatment column carries more than reverse osmosis and ion exchange. Knowing what each stage is there to remove makes it clear why a given row reaches the quality it does. The Remote monitoring column describes how far the unit watches itself: whether someone has to write a reading down every day, or whether the unit measures and tells you.
More ions in the water means it conducts better. Conductivity reads that total ion load immediately, which is why it is the daily monitoring value: the instrument is inexpensive, it is easy to install and logging is easy to add. What conductivity does not tell you is which ions are present. Water can read well and still be high in chloride or hardness. So conductivity is for day-to-day monitoring, and a full analysis is run at initial approval, once a year, and whenever something looks wrong. The easier things to miss are the ones that are not ions at all. Organics, microbes and particles do not register as conductivity, so water reading 1.0 µS/cm can still carry microbes. Those are not watched with a number; they are handled by putting sediment filtration, activated carbon and, where needed, a disinfection stage at the front. That is what the first three stages in the diagram above are for.
Purifier catalogues quote resistivity while we quote conductivity. At 25 °C the two are reciprocals, so 0.25 MΩ·cm is about 4 µS/cm. Because 0.25 also appears as a number in the Type III row of the grade table, that specification looks like Type III when it is closer to Type IV. More importantly, an ASTM type is not a conductivity-only criterion. Each type carries TOC, sodium, chloride and silica conditions alongside it, so naming a type from conductivity alone produces a wrong label. We therefore quote a conductivity figure rather than a type. That works in your favour as well: asking for a type invites the assumption that laboratory-grade equipment is needed, whereas a conductivity figure makes it plain that ordinary reverse osmosis and ion exchange will do.
The options and the models we handle
Start here and we adjust to your conditions. We can usually source what is not listed.
Start here before buying anything. The permeate figure on the analysis is all we need to decide. If it meets the specification, no additional equipment is needed and you can skip the table below.
Ask us to review the analysis→With mains water as the source and no 1.0 µS/cm figure in the contract, one RO unit is the whole train. Nothing is added behind it. The three rows in the group For the 30 µS/cm basis below are the ones.
Quote this option→A stage that takes out the ions left behind goes behind the RO: ion exchange, or EDI (electrodeionisation). That is normally one unit, not two. Even with no 1.0 µS/cm figure in the contract, a site whose source water swings with the season is looked at here for the margin. The three rows in the group Where 1.0 µS/cm is required below are the ones.
Quote this option→With groundwater or high hardness source water, softening or filtration goes ahead of the RO whichever of the above you pick. That is a separate axis from the target quality, so it does not change the choice above; it is added once the main unit is set. The pre-treatment units are not in the table below, and how many stages you need can only be set from a source water analysis.
Candidates we have already reviewed. Tell us your source water and target quality and we will pick from these and quote. Reverse osmosis alone reaches the 30 µS/cm HX in-service figure. If you need the 1.0 µS/cm HX recommended figure, it has to be one with ion exchange or EDI. A demineralisation stage does not on its own guarantee 1.0 µS/cm, so read the output column alongside it. For BWT that column is not a guaranteed figure but a target the user selects, either 5 or 15 µS/cm, and the choice changes cartridge life. Capacity is rarely the constraint: one HXB-V1 uses about 13 L a day, so even a 15 L/h unit covers that in under an hour.
| Model | Treatment | Output | Capacity | Feed method | Remote monitoring | Origin | Quote this |
|---|---|---|---|---|---|---|---|
HumanNew Pure RO 260↗ | Reverse osmosis | – | 25 L/h | Storage tank required | None | Korea | Quote this→ |
HumanNew Pure RO 130↗ | Reverse osmosis | – | 15 L/h | Storage tank required | None | Korea | Quote this→ |
HumanGreen RO 350↗ | Reverse osmosis | 1 – 30 µS/cm | 35 L/h | Storage tank required | None | Korea | Quote this→ |
| RO + mixed bed DI | Target setting of 5 or 15 µS/cm | 100 L/h | Pump built in; 8 mm port for an external tank | Built-in display plus Bluetooth app; permeate conductivity measured continuously | Austria | Quote this→ |
| Model | Treatment | Output | Capacity | Feed method | Remote monitoring | Origin | Quote this |
|---|---|---|---|---|---|---|---|
ELGA / VeoliaPURENERGY 30↗ | RO + EDI | Below 1 µS/cm | 30 L/h | External 25 L reservoir (optional) · recirculation built in | Monitoring and alarms built in; the manufacturer does not publish a remote interface | United Kingdom | Quote this→ |
ThermoSmart2Pure Pro 16↗ | RO + ion exchange polisher + UV + ultrafiltration | 0.055 – 1 µS/cm | 16 L/h | 30 L tank included | None; local display and alerts only | Germany | Quote this→ |
You can buy direct from the manufacturer. Coming through us means we carry the international payment, the customs clearance and the minimum order, and you deal with a domestic invoice.
What to settle before a quotation
Only what has come up more than once in real enquiries.
Its conductivity is well above the basis, and it varies by region and season, so not as it comes. Against the 30 µS/cm in-service figure, reverse osmosis alone can get there. If you need the 1.0 µS/cm recommended figure, an ion removal stage has to follow the RO.
A softener is not a purifier. It exchanges calcium and magnesium for sodium, so it lowers hardness while the total ion load stays where it was. Water out of a softener reads almost the same conductivity as the water going in. A softener is not a demineralisation step; it protects the reverse osmosis membrane from scale.
The output quality stays the same and only the front end changes with the source water. Municipal supply in Korea and Japan is generally on the easier side. Singapore supply makes carbon for chloramine removal important. Parts of Germany and the US southwest are hard enough that softening or scale control is added. For groundwater in India and the Middle East the country name is not the input; the analysis certificate for that particular well is. All of this is for a first look only — the design input is the water quality report for your address, or the well analysis. And note that a location needing a larger front end is often perfectly fine as drinking water. Drinking water quality and electrolyser feed water quality are two different specifications.
We have not configured a system for rainwater. Conductivity alone can come out lower than tap water, but catchment contamination, organics and microbes are the variables, and it swings from one rainfall to the next. Send us a water analysis and we will work out how many stages you need. It will likely put storage and disinfection at the front.
Seawater without desalination is outside our scope. With chloride present, chlorine evolution competes with oxygen evolution, and we hold no performance data for that condition. If a desalination stage in front brings the feed within our water spec, everything downstream is the same as any other source.
If your contract or specification states 1.0 µS/cm, yes. If it does not, we work to the HX in-service figure of 30 µS/cm or below. Why we keep two figures is set out under Why it matters above. Which one applies is settled together against your contract terms and target lifetime.
No. An analytical ultrapure unit is more than electrolyser feed water needs. On normal municipal water a sediment filter, carbon, reverse osmosis and ion exchange are usually enough. Groundwater or process water needs a larger front end, so we do not fix a cost before seeing the source water analysis.
Yes. We set the arrangement and we procure it. We carry the dealing with the overseas manufacturer and the customs clearance, and hand it over as a domestic transaction with a tax invoice. The manufacturer warranty passes through to you.
We check up front whether the water leaving that package reaches our feed water basis. Buy the two separately and that check is yours to make, and if they do not match, responsibility is split.
We would not. A single HXB-V1 consumes about 13 L a day, so even a 15 L/h unit runs under an hour. More throughput only raises the equipment price and the consumables bill. If you plan to add units, we look at it again then.
You judge it by output conductivity. On the domestic ion exchange filters we handle, reaching 20 to 30 µS/cm is the sign that replacement is close. The only item with a manufacturer-stated interval is the feed pre-filter; the rest go by conductivity. It is worth putting a conductivity meter on the feed line — though some units measure permeate conductivity continuously and report it to an app, in which case you do not need a separate one.
Start with temperature compensation and sensor calibration. More often than not it is the instrument rather than the water, and skipping this step means replacing perfectly good ion exchange resin. Then measure conductivity right after the reverse osmosis rather than only the source and final water. If the rejection rate has dropped, the problem is on the RO side; if rejection is normal but the final water is poor, it is the ion exchange side.
Record the final water conductivity and the storage tank level every operating day. A rising trend is the signal, not any single reading. Once a month, compare the conductivity of the source water, the reverse osmosis permeate and the final water together, and you will see which stage is degrading before it becomes a problem.
Once these are set, selection and quotation happen in one step. If they are not, we start there.
Fill in only what you know. We ask about the rest.
Request a quote for this item