Field-data from a 150 MW Saudi PV plant (4,200 chemical electrodes, 3.1 Ω average pad resistance) and ME/Africa/SEA export shipments: why standard driven rods fail in sandy, rocky and laterite soil — and what the procurement document has to say instead.

The conversation I want to anchor this article on happened in August 2020, when a Madrid-based EPC sent the inquiry that turned into the 150 MW Saudi PV project this article’s field data comes from. The inquiry email asked about chemical earth rods for a utility-scale PV site where the soil-resistivity survey had come back in the 1,500 to 2,200 Ω·cm band. The geotech report was clear, the EPC was competent, and the site sat in a caliche belt the lead engineer had seen on a smaller project in Chile in 2018 where a watered-rod inventory had drifted past the touch-voltage margin in year three because two scheduled watering visits were skipped during COVID travel restrictions. He did not want to repeat that. The article you are reading is the engineering argument that came out of that email — the soil-type framework the EPC eventually wrote into the procurement document, and the field data from the plant that produced it.

This guide is for procurement teams on the receiving end of a similar inquiry: ME, Africa and SEA project owners whose geotech report is on the desk, whose design team is asking for a chemical electrode rather than a driven rod, and whose procurement document has to be written in the next two weeks. The three soil types below — sand, rock and laterite — are the ones that show up most often in those reports, and the failure modes are specific to each. What does not change across the three is the procurement-side answer.

1. Sandy Soil — the Wire-in-Air Failure

Dry sand is the most counterintuitive of the three failure modes. On paper it looks like a perfectly fine earthing medium: it is loose, it drains well, and it is easy to auger. The Wenner four-pin reading, however, tells a different story. Dry desert sand at 1 m depth typically reads between 3,000 and 10,000 Ω·cm; dune sand at 1.5 m depth can read above 20,000 Ω·cm in the worst case. The reason is ionic, not mechanical: dry sand has almost no clay content and almost no dissolved mineral content to conduct current away from the rod surface.

What changes the calculation is moisture at the rod surface. A chemical earth rod is a copper-clad core inside a 100 mm augered annulus, with the annulus filled by a hygroscopic backfill compound (typically bentonite compounded with magnesium sulfate). The compound absorbs ambient moisture from the soil and from seasonal humidity, holds it against the rod surface, and slowly releases ionic content into the immediate soil volume. The result is that the rod-to-soil interface behaves like a soil of perhaps 200 Ω·cm even when the surrounding bulk soil reads 8,000 Ω·cm. The NEMA GR-1 rod geometry is unchanged; what changes is the soil contact.

On our shipments into the Middle East between 2020 and 2025, the median pad resistance on chemical-earth-rod installations in the 2,000 to 8,000 Ω·cm sand band landed at 3.6 Ω against a 5 Ω design target, without a watering cycle. The driven-rod-only installations on adjacent sites in the same band read between 18 and 42 Ω at commissioning, and were re-augered within the first year. The rod was the same diameter. The difference was the backfill.

2. Rocky Soil — the Caliche-Layer Bottleneck

The Saudi site that produced the field data in this article had a caliche layer at 3.5 m on the southwest array corner. The original spec called for 2.4 m driven rods — which worked on 94% of the array. The remaining 6% sat on caliche too shallow to drive through and too resistive to ignore. The six pads in that corner read above 4.7 Ω at commissioning even with the rod extended to 3.0 m. We did two things: switched to delta-pattern doubled electrodes (two 2.4 m rods in a triangle around each pad) and added column graphite ground modules at the MV transformer pads in parallel. The year-five retest came back at 4.7 Ω worst — within the design margin — because the modules extended the equipotential surface laterally where the vertical path was blocked.

Rocky sites fall into two practical categories for the procurement team. The first is fractured rock with soil-filled gaps — augerable, but the rod only makes contact with the soil fraction, not the rock matrix. The second is intact rock or caliche (calcium carbonate cemented soil) at or near the planned electrode depth, where driving becomes either impossible or uneconomic. The Middle East and the western U.S. see a lot of the second category; large parts of Saudi Arabia, Qatar and the UAE have a caliche layer between 2 and 4 m, and the ISRIC World Soil Information catalogue puts the same profile across the Sahel and the Horn of Africa in seasonally dry zones.

For a deeper procurement-side question on rocky sites, the FAO soil classification data hub lets a procurement team pull the regional soil map before the geotech contractor mobilizes — the cheapest place to learn whether the site sits on caliche or weathered-granite substrate. The FAO World Soils reference document is the field reference for soil-class names that end up on a BES (Bid Evaluation Sheet).

3. Laterite Soil — the Seasonal Swing Problem

Laterite is the soil type where a standard driven rod passes commissioning and then quietly fails the touch-voltage window by the dry season. Laterite is an iron- and aluminum-rich residual soil common across tropical Africa (Nigeria, Ghana, Kenya, Mozambique), South Asia (much of peninsular India, Sri Lanka, parts of Bangladesh) and parts of South America (the Brazilian northeast, Guyana). The soil profile is well-drained and has high iron oxide content, which is what gives it the characteristic red-brown color.

The electrical problem is the seasonal swing. Published field surveys on West African laterite typically report wet-season resistivity in the 200 to 800 Ω·cm range — perfectly workable for a standard driven rod — and dry-season resistivity above 5,000 Ω·cm — outside the workable range for the same rod. The driven rod does not change; the soil around it does. A 6-month watering cycle can re-stabilize the dry-season reading, but the same watering discipline problem reappears: a watering crew that does not show up at the right time leaves the pad inside the touch-voltage window’s failure band for the rest of the dry season.

For SEA laterite sites specifically, the same seasonal logic applies with slightly different seasonal timing: the dry season in Indonesia and the Philippines is shorter but more intense than the West African dry season, which means the resistance swing is compressed into a tighter window. The chemical-earth-rod operating model — moisture-retaining backfill, no watering cycle — works in both cases.

4. What the Procurement Document Has to Say

Across the three soil types above, the engineering decision is the same: the standard driven rod is not the right product. The procurement document has to call out what is.

Three procurement-side controls make the difference between a spec that solves the problem and a spec that creates the problem.

1. Type-test report attached to the PO. A chemical earth rod is not just the copper tube — it is the chemistry inside it. The supplier has to ship a four-property type-test report: resistivity reduction factor at reference moisture, long-term ionic stability at 60 °C / 30 days, moisture retention at 90-day dry-down, and ESIA-stage environmental compliance. We lost a 2019 batch to a sodium-chloride blend (chloride attacks copper-clad cores inside the sheath), and our 2022 supplier batch failed the ionic-stability test at 65% retention — below our internal 80% threshold — which is why we qualified an in-house compounding line. The supplier who cannot ship that report is the supplier who turns up at the end of year two with a “we have a new improved compound” email.
2. Connector standard locked at IEEE 837. Mechanical clamps fail the impulse-current test in IEC 62561-1 and should not be used on a utility-scale project. The connector standard is the part of the spec that protects the asset across the full 25-year design life.
3. Rod geometry per NEMA GR-1. Diameter 17.2 mm to 25.4 mm, copper-cladding thickness ≥ 254 µm. The cladding thickness matters most on impulse-current loads (ESE terminations, MV pads) — a thin cladding fails first.

5. Closing — What the Procurement Side Has to Ask First

The 150 MW Saudi plant that opened this article came back at 3.1 Ω average pad resistance against a 5 Ω design target at the year-five retest, with the worst pad at 4.7 Ω in the southwest corner where caliche stopped vertical driving. None of those numbers existed as a procurement argument when the Madrid EPC sent the August 2020 inquiry email — they are the result of writing the backfill chemistry, the connector standard, the rod geometry and the soil-target resequencing into the same document, and then letting the install crew do the work.

If you are a project owner in the Middle East, Africa or Southeast Asia and the procurement document is open, the question that determines whether the project stays inside its design window for the next 25 years is not “which rod do I buy” — it is “what is the soil resistivity at the planned electrode depth, and which of the three failure modes above applies.” Get the soil answer first; the product answer follows. The procurement question is then which of the two product families — chemical earth rod or column graphite ground module (or both, on the toughest sites) — fits the soil band and the depth limit. The full category page at /chemical-grounding-electrode/ has the geometry options, and the Top 5 Electrolytic Ion Ground Rod Manufacturers for 2025 blog has the supplier shortlist.

For a sample pack and a four-property type-test report on a specific project, the fastest path is the contact-us page. Sample shipments for ME, Africa and SEA projects typically leave the factory within the same week, with technical replies inside one working day.