Water quality for solar panel cleaning: Hardness, TDS and what 23 module makers allow

When a rinse dries, the water leaves and whatever was dissolved in it stays on the glass. Clean, soft water evaporates almost without trace. Hard or salty water leaves spots and haze, and those deposits build up wash after wash.

Water quality is also a warranty matter. Module makers set limits for cleaning water in their manuals, and their letters of conformity for cleaning products come with conditions. See will cleaning chemicals void my panel warranty?

And in South Africa, water is scarce. The country averages about 490 mm of rain a year, about half the world average (WWF-SA), and the Department of Water and Sanitation projects a water deficit of about 17% of supply by 2030. A full wet wash of a 10 MWp plant uses roughly 50–150 m³. Wasting a wash on bad water costs twice.

What does cleaning with bad water cost? A wash with bad water can leave the glass worse than no wash at all. Read the Malaysia case. The lesson for South African plants: test the water before you judge a cleaning method.

• TDS (total dissolved solids) is everything dissolved in the water: calcium, magnesium, sodium, chloride, sulphate and more. It is reported in mg/L, which for water is the same as ppm. A TDS meter estimates it from electrical conductivity.
• Hardness is the calcium and magnesium part. It is what forms limescale. Labs usually report it as mg/L (ppm) of calcium carbonate, CaCO₃, but not every module maker does (see below).

All hardness is part of TDS, but not all TDS is hardness. You need both numbers, because module makers limit both and because the treatments differ. Two traps catch people:

• A TDS meter can’t tell you hardness. Once Water Softening Agent is dosed, Chemitek warns that a TDS meter reads the agent itself and says nothing about the remaining hardness. Use hardness test strips.
• Say which number you mean. “250 ppm” of TDS and 250 ppm of hardness as CaCO₃ describe very different waters.

USGS classes water above 180 mg/L as CaCO₃ as very hard. The bands are those in RenewChem’s ChemCalc dosing tool.

We read the current cleaning, O&M or installation manual of 23 module makers, from the maker’s own site where it could be downloaded and otherwise from a distributor’s copy of the maker’s own document. Nine give a numeric hardness limit for cleaning water:

The other 14 publish no numeric hardness limit. Where their manuals give other figures, they are listed here:

“Not stated” means the manual doesn’t give it. TW Solar is Tongwei’s own brand, so it shares Tongwei’s row. Where a limit is given, chloride is capped at 1,000 mg/L by LONGi and Jinko, and at 3,000 mg/L by Risen, DAS Solar, Huasun and Znshine. Pressure and temperature limits are in warranty-safe methods. Limits change between manual versions, so always read the current one for your module series.

Hardness figures are not all measured on the same basis, and mixing them gives the wrong answer.

• As CaCO₃. Trina and Aiko state their limits as calcium carbonate, the way most water labs report hardness.
• As ions. Huasun and Znshine state theirs as the mass of the calcium and magnesium ions themselves. The same water reads lower that way. To convert calcium to CaCO₃, multiply by about 2.5 (the molar mass of CaCO₃ is about 100, calcium about 40); magnesium converts at a higher factor, so the exact figure depends on the mix.
• Not stated. LONGi, Jinko, First Solar, Risen and DAS Solar give a number in mg/L without saying which basis it uses. Risen’s and DAS Solar’s figures and wording closely match Huasun’s and Znshine’s, but their manuals don’t confirm the basis.

Side by side, 0–40 mg/L and 600 mg/L look fifteen times apart. They aren’t: 40 mg/L of calcium ions is about 100 mg/L as CaCO₃. So check that your water report and the maker’s limit use the same basis before you decide the water passes or fails.

First Solar is often quoted as a flat 75 mg/L. Its 2017 cleaning guidelines are more specific:

• reverse-osmosis water gives the best results; otherwise tap water with total hardness under 75 mg/L, or deionised water;
• water of 75–180 mg/L may be used on uncoated modules, but only if it is squeegeed off to prevent scale;
• the same 75–180 mg/L water is prohibited on modules with an anti-reflective coating (ARC);
• water above 180 mg/L isn’t covered at all;
• TDS must be under 1,500 mg/L, and the pH 6.5–8.5 window applies to the final water-and-detergent mix at 25 °C.

The guidelines don’t say whether the hardness figures are as CaCO₃.

It depends on how hard, and whose modules. Take borehole water that tests at 250 mg/L as CaCO₃, which is very hard on the USGS scale. Against the tables above, that water:

• sits inside Trina’s and Aiko’s 600 mg/L as CaCO₃;
• sits inside LONGi’s and Jinko’s 450 mg/L, if those are CaCO₃ figures too, which their manuals don’t say;
• is above anything First Solar’s guidelines allow;
• is well above Huasun’s and Znshine’s 0–40 mg/L as ions, even after converting: if all 250 mg/L were calcium, it would be about 100 mg/L of calcium ions.

Even inside a maker’s limit, hard water leaves scale as it dries, so we pair routine washes on hard-water sites with softened water. A working rule: if your water is near or above your maker’s limit, or you see spots after a wash, treat the water before the next one.

All three reduce the minerals that leave scale. They differ in what they remove, what they need and where they fit.

The choice depends on the site. A plant with a treatment building and a steady supply may already run RO. A mobile crew working from a bowser, or a site where the water source changes, is often better served by dosing in the tank. Chemitek compares real-time WSA dosing with purification systems in a post listed in the sources. The dose table for each hardness is on the Water Softening Agent page.

Sometimes the right amount of water is less. Dry cleaning versus distilled water at Kalkbult is in soiling losses in South Africa. That doesn’t hold for cemented dust or scale, which need the right water and chemistry. See low-water cleaning.

1. Sample the water the crew will use, from the tank or nozzle, as well as at the source.
2. Measure hardness with test strips and record it as mg/L of CaCO₃. Chemitek makes water hardness test strips for this.
3. Measure TDS or conductivity with a meter, before any softening agent goes in.
4. Check pH against your maker’s window, for example 6–8 for LONGi and Trina, or 5–7 for Risen, DAS Solar and Huasun.
5. Send a lab sample for chloride and turbidity if your maker lists them, as LONGi does.
6. Note the water temperature against the module temperature at washing time. Temperature limits are in warranty-safe methods.
7. Compare every result with the current manual for your module series, on the same basis, then choose the treatment.
8. Re-test after treatment, using strips for hardness, not a TDS meter.
9. Re-test when the source changes, for example a new borehole or supplier.

We run these tests on site and check them against your module maker’s limits: see cleaning-water testing.

• Clean when it pays. Measure the soiling rate and set the interval by cost, not the calendar. See how often to clean.
• Use chemistry that saves water. Chemitek claims up to 50% less water with Solar Wash Protect and Antistatic Solar Armor 2.0; see Solar Wash Protect. At a Scatec-operated plant in Honduras, the O&M team used much less water with Solar Wash Protect, because water was only needed to dilute the product and rinse.
• Budget the water honestly. A wet clean uses about 1 L per m² of module in US literature, and a South African dissertation cites about 3 L/m² (NREL 2018; Naicker 2018). Parts of the Northern Cape get less than 70 mm of rain a year (SAEON), so every litre counts.