Magnetic water treatment has a mixed reputation, and a buyer is right to ask why. This page starts with the man who designed these units and the patents he filed, then sets out how they work and what the published record says, including the studies that found nothing. Primary sources are linked throughout. Nothing here asks you to take our word for it.
Not a plumbing catalogue. A research lab, and an accident.
In 1955 John Meckling was hired by Brunswick Corporation as a research engineer. He was told to build a pump with no moving parts. Five years later Martin Marietta, a predecessor of Lockheed Martin, was awarded the contract and he moved with it, working on magnetohydrodynamic propulsion: driving a conductive fluid with a magnetic field rather than a mechanical impeller.
The discovery was incidental, and it came from a cleaner.
A cleanup man remarked of how easy it was to clean my lab. Minerals did not stick the same as other labs. At first I thought not the same water. In reality, it was the same, only I was influencing a change in crystal bond angles.
John H. Meckling
He described that moment as the birth of the scale control system. When he saw what the field was doing to the minerals, he understood that the future of scale control was not in barrels of chemicals and annual acid washes. The principle Faraday demonstrated in the 1830s, that a conducting fluid moving through a magnetic field becomes the seat of an induced current, would do the work instead.
Meckling spent the mid-1970s in permanent magnet design and manufacture, and the filings are public.
The 1974 filing matters most. Focusing flux by laminating magnets with spacers is the technique the water conditioner depends on, and it was on file eighteen months before the first commercial sale of a scale control unit. When Amoco described the device they tested in 1984, they described magnet arrays in a spool piece with opposite poles across from each other, strongly focusing the field across a slot.
The Louisiana trademark record dates first commercial use of the Hydrodynamics mark to 15 September 1975, from Bogalusa. It was registered for self-contained magnetic units to control scale formation in pipes and boilers.
The Meckling family entity was registered in Bogalusa in December 1984 as Hydro America, Inc., took the Hydrodynamics Corporation name and mark in October 1987, absorbed Hydro Industrial, Inc. by merger in December 1988, and became Meckling MHD Systems, Inc. in January 1998. Helga Meckling served as President, John H. Meckling as Vice-President and Secretary/Treasurer. The business has been Canadian-owned since 2009.
The name moved because the business did. John Meckling sold the original company; the sale went wrong, and recovering the business took years. The Hydrodynamics name and trademark came to his own entity in October 1987, and after a further period of dispute over the name he stopped contesting it and put his own on the company instead, in January 1998. Much of the field record, including the Coast Guard article and the test at the Department of Energy geothermal facility, was made under the earlier name, which is why it appears throughout the documents below.
Two charters carry the Hydrodynamics Corporation name at different times. The earlier, registered in Baton Rouge in May 1976, was sold and renamed HDC, Inc. in August 1987. The Meckling entity took the name and trademark that October. The officer records for the earlier charter between 1976 and 1987 are not yet in hand, so the documentary chain across the sale is incomplete even though the trading history is continuous.
The product line runs from half an inch to sixty inches, flanged to ASA specification, and has been installed on plant loops in 42 countries.
Magnetohydrodynamics is the behaviour of an electrically conducting fluid moving through a magnetic field. An MHD unit applies that principle to water carrying dissolved hardness.
The unit is a flanged stainless steel body containing high-powered permanent magnets and venturi tubes. Water is routed through the tubes, which accelerate it across a focused field oriented perpendicular to the direction of flow.
What is observed is a change in which crystal form calcium carbonate takes as it leaves solution:
Rhombohedral crystals that nucleate on the hot metal surface itself and grow into a continuous adherent layer. This is the hard scale that insulates tubes and narrows passages.
Needle-like orthorhombic crystals that nucleate in the body of the water rather than on the wall. They stay in suspension and are carried through the system to blowdown.
John Meckling put it more plainly: the same chemical substance precipitates as a snowflake rather than an ice cube, and a snowflake will not support mass or resist flow impact.
John Meckling's explanation was electrical. A conductive fluid moving through a magnetic field generates an electropotential, the same reaction as a DC generator. He held that this ionising condition alters the outer valence charges of the dissolved minerals and changes crystal bond angles. On that account the effect scales with total dissolved solids, because conductivity is what drives it.
The current leading proposal in the literature is different. Coey argues the effect turns on the gradient of the applied field rather than its absolute strength, acting on nanometre-scale prenucleation clusters in calcium carbonate solutions, by Maxwell-like stress or singlet/triplet mixing of proton dimers. That account addresses the long-standing objection that there was no plausible route for water to hold a magnetic memory. It does not depend on flow velocity or conductivity.
The two are not reconciled, and we do not claim to have settled it. What both agree on is the outcome: a magnetic field applied to hard water shifts the deposit toward aragonite.
Field evidence bears on the question. The 1984 refinery trial concluded that two focused fields of about 1,700 gauss, with water moving through them at 6 m/s or faster, gave adequate scale control, and that low flow led to scaling while 6 m/s did not. On two separate occasions a blocked strainer dropped velocity through the field, scaling began, and it stopped once flow was restored. That is consistent with a velocity-dependent effect, though not decisive on its own, since reduced flow through a heat exchanger increases fouling regardless of any treatment.
Prevention and removal are different claims, and the removal one is harder. We can describe what has been observed rather than assert a mechanism for it.
At the Coast Guard's New Orleans radio station the pattern was gradual and documented week by week. Units were fitted in January 1978. By the nineteenth of that month the galley heat exchanger was running 5°F warmer and granular calcium carbonate was appearing in the blowdown. Another 5°F by early February. Another 10°F by the twenty-third, with enough loosened deposit in the system to interfere with faucet closures. By March and April large quantities of carbonate were being blown from the boiler and lines. When the boiler was opened on 19 April the tubes that had been 40% closed were clean, and four thirty-gallon cans of loose carbonate came out of the hot water storage tank.
The deposit does not disappear. It stops adhering, and then it has to be carried out of the system or collected. That has two practical consequences. A blowdown or flushing schedule matters during the first months on a scaled system, and velocity matters, because where flow is slack the loosened material will settle in sumps and low points and need removing by hand.
Timescales vary with the water. A visible difference in deposit typically appears within about 90 days. Scale is a composite of whatever minerals that particular source carries, so no two behave alike; higher calcium content degrades faster and other minerals slow it down.
Whichever mechanism is operating, repeatable results require controlled conditions: a known field, a known velocity, a unit sized to actual flow, and a stated water chemistry range. That is what an inline unit provides and what a device with no specifications cannot. The engineering case does not rest on the mechanism argument being settled.
Units are sized to flow rate, not to pipe diameter. The acceleration happens inside the unit, so what matters is the volume passing through it.
Treated water retains the effect after leaving the unit. Doctoral work at Baylor University documented 30 hours in a stirring environment (paper not yet located); Coey and Cass measured persistence beyond 200 hours in still samples. We use the 30-hour figure as the conservative design basis, which has always proven sufficient for the distance flow covers in that time.
Coey and Cass also report an incubation period of several hours before the effect becomes apparent. In recirculating loops this is immaterial. On once-through systems where water reaches the exchanger within seconds, it is worth discussing before specifying.
There are no moving parts, no electrical connection and no consumable element. Field strength is effectively permanent. A unit inspected at 25 years of service showed 5% loss.
Where we can, the primary sources are hosted here in full; otherwise we link to the publisher. Read them rather than taking our word for it.
Two of these are not general literature about magnetic treatment. The Coast Guard article and the geothermal test at the Department of Energy facility were both conducted on MHD units supplied by Hydrodynamic Corporation. That is why that name, rather than the current one, appears throughout the documents below. The naming history is set out above. Both spellings, Hydrodynamic and Hydrodynamics, appear in the original documents.
Most of what is sold as magnetic water treatment is a magnet clamped to the outside of a pipe. What those devices lack is control: no specified field strength or gradient, no defined flow path, no sizing to actual flow, and no stated water chemistry envelope. Two installations of the same product on different systems are not the same treatment.
Results in this field have been inconsistent because the category lumps engineered inline equipment together with devices that specify nothing, and because the mechanism itself is still being worked out. Reproducibility follows from controlled conditions, and controlled conditions are an engineering matter rather than a theoretical one.
Grutsch and McClintock addressed this directly in 1984. They observed that the international literature on non-chemical water treatment is almost entirely split, the American literature negative and the rest of the world positive, and that the American papers leaned on ridicule, not data from technically defensible experimental protocols. On the Clemson doctoral dissertation most often cited as proof of no effect, they wrote that their review found its experimental protocol unequivocally inadequate to measure one.
The null results are worth reading directly. One of the most cited is a Lawrence Livermore internal report from December 1996. Read it and the scope is narrower than the reputation: it evaluated one commercial device, a Descal-A-Matic unit from Norfolk, Virginia, at a single LLNL groundwater facility, against no treatment and against a polyphosphate additive. No beneficial effect was found for that device in that application.
That device is not a clamp-on. Descal-A-Matic units are inline, with a housing, a fixed orifice and a magnetic core, and the company positions its product against externally clamped magnets. The report should be read on its own terms rather than dismissed as a test of a no-specification gadget. Two differences are worth noting: Descal-A-Matic specifies AlNiCo magnets, a much weaker material than modern grades, and the scale at that facility formed because air stripping raised the pH rather than because of heating, a different precipitation route from the heat-driven scaling these units are designed for.
Installations, operating histories and inspection records.
The fleet fouled every 15 days. A single car was fitted with an MHD unit and put back on the line with instructions to run until it failed.
It did not fail. After 140 days, more than nine times the fleet interval, the car was taken off the line and the vessel opened for inspection. No scale. The company president did not accept the finding and ordered the unit cut in half. Still no scale.
Fifty units were ordered, followed by a hundred the next year to cover the remainder of the fleet.
Phase 3 of the Grutsch and McClintock trial describes two 36-inch units on a cooling circuit recirculating about 25,000 gpm, with magnet arrays of opposing poles focused across a venturi slot and a pressure drop near 4 psi. The description matches this equipment closely. The paper does not name the manufacturer. On our side, John Meckling's correspondence of October 2001 describes units at Amoco's largest refinery. We state both records and leave the connection to the reader.
Units were sold and installed through agents in Mexico, Korea, Kenya, Jordan, Germany, India and Greece, among other countries, with warranty files maintained by territory.
A unit inspected after 25 years of continuous service showed 5% loss in magnetic field strength. Equipment was still operating in its 26th year as of 2001. Heavy industrial units carry a ten year warranty.
Scale has a thermal conductivity roughly an order of magnitude below bare steel, so even thin layers act as insulation and retard heat transfer. The result is overheated tube metal, tube failures and lost efficiency.
The U.S. Department of Energy puts fuel waste from boiler scale at around 2% for water-tube boilers and up to 5% for fire-tube boilers, with roughly 2% lost to a thirty-second of an inch of normal-composition scale. On a plant burning 450,000 MMBtu a year, DOE's own worked example puts a 2% loss at $72,000 in additional annual fuel cost at $8.00/MMBtu.
Fuel is only part of it. Scale also drives chemical treatment spend, blowdown volume, cleaning intervals, downtime and condenser approach temperature. Those are the numbers a facility already tracks.
Send the flow rate, the total dissolved solids and what the system feeds, and we will tell you which model fits, or that the application is not a good candidate.