How to read a coolant datasheet without misreading it

A table that answers a different question

Open the LM-4A series data and a single column of figures looks like this:

Read left to right and the density barely moves — about 2 % across the whole range. Read the viscosity row and it climbs from 3.26 to 243 cP, a factor of roughly 74.

Now read the last row. The viscosity figures are not all measured at one temperature. Each concentration is paired with the temperature at which that concentration would plausibly be used, so the row spans 5 °C down to −30 °C. The steep rise reflects both effects at once: more concentrated fluid and colder fluid.

This is the single most useful habit in reading a coolant datasheet. Ask what conditions the number describes before asking whether it is good.

What "reference values" is telling you

Most property tables carry the heading "reference values", and the column headers name the conditions — for example "Density g/cm³ (20℃)" or "Viscosity cP (20℃)". Some tables add the concentration; others define the figures as stock solution, meaning the fluid as supplied, before any dilution. The LM-1 and LM-3 series entries are printed this way — a single set of stock-solution figures rather than a curve across concentrations.

That combination does three things at once. It tells you:

1. The temperature. A measurement at 20 °C is a laboratory reference point, not an operating condition.
2. The concentration — or that there isn't one, because the figure describes the undiluted product.
3. That this is a supplied value for selection, not a warranty for your circuit.

None of that makes the number unreliable. It makes it specific. The point is made plainly in the data itself: these are reference values, so request the current technical data sheet and safety data sheet for the exact model before use.

Temperature moves viscosity further than you expect

Most engineers have an intuition that viscosity rises as fluid gets colder. Fewer have a calibrated sense of how much.

The LM-4A row above is a good calibration. Between −15 °C and −30 °C — a 15-degree window — the viscosity goes from 28.60 to 243 cP. That is not a gentle slope; it is a near eight-fold change over a span smaller than many winter ambient swings.

Why it matters: viscosity drives pressure drop, and pressure drop drives pumping power. At a fixed flow rate, a fluid that is eight times more viscous demands far more from the circulator, and the relationship between pumping power and viscosity is more than linear once you include the effect on flow regime.

This is the same mechanical reality behind the argument that pumping energy, not heat capacity, decides a low-temperature coolant: the coolants in a given family differ only modestly in volumetric heat capacity, but their pumping demand can differ by an order of magnitude.

The practical consequence for reading a datasheet is direct. If the figure you were given is at 20 °C and your loop runs at −25 °C, you have not been told the number you need. The same problem runs in the other direction: the high-temperature families, such as the LM-16 series, are tabulated against an upper operating window that a room-temperature figure does not describe.

Concentration changes several properties at once

Dilution is not a neutral act. Moving a glycol-type fluid from 50 wt% to 30 wt% to save cost changes freezing point, boiling point, density and viscosity together, and it changes inhibitor concentration with them.

The concentration tables make the trade-off visible. In the LM-4A data, going from 20 wt% to 55 wt% buys freezing protection from −7 °C down to −41 °C, while viscosity at the relevant operating temperature rises by two orders of magnitude. There is no free direction to move in. Not every family is designed to be diluted at all — the salt-based LM-8A and several anhydrous families are supplied for use as received, with no dilution step in the specification.

That is why "what concentration will you run" belongs in the specification before properties are compared, and why a corrosion or performance result at one dilution does not transfer to another — the same reasoning set out in what corrosion protection actually means.

It is also why interpolating between two tabulated columns is orientation, not specification. If your target sits between 30 wt% and 40 wt%, ask for 35 wt%.

Three unit traps in real datasheets

Specific heat in cal/g·°C. Specific heat is printed here in cal/g·°C. Most modern datasheets use kJ/kg·K, and the two are not close in magnitude. One cal/g·°C is about 4.184 kJ/kg·K, so a fluid printed as 0.585 cal/g·°C is roughly 2.45 kJ/kg·K. Compare the raw numbers without converting and the SI datasheet looks about four times better than it is — or worse, the other way round, depending on which way you are converting. Always convert before comparing.

Stock solution versus operating concentration. A stock-solution viscosity figure describes the product as shipped. If you dilute it — and many of these families are designed to be diluted — that figure does not describe what your pump will see.

Freezing point versus pour point versus burst point. These are three different measurements and datasheets do not always print all three. A freezing point tells you when crystals begin to form; a pour point describes flow behaviour; and neither is the same as the concentration needed to prevent a pipe bursting. The last of the three is the one worth understanding before you buy protection you think you have — see freeze protection versus burst protection.

Where misreading actually costs you

Unit errors are the easy mistake to catch. The expensive one is selecting the wrong model because you read a family figure as a model figure.

Within a single family the variants are not interchangeable. In the glycol-type range alone, the LM-4A series and LM-5 sit in the same general category but are tabulated against different operating windows, and the mixed-metal variants — LM-1L among them — exist precisely because a standard formulation does not cover every metal set. Choosing the family name from a summary table and then ordering a model from a different part of the range is how a system ends up with a fluid that is chemically correct but dimensionally wrong for its circuit.

Family ranges and model ranges are separate figures, and the product pages keep them apart. Treat the family line as a shortlisting aid and the model line as the specification.

What to ask for instead

Reference values are the right tool for shortlisting. They are the wrong tool for specifying. When you have narrowed the field, request the following for the exact model:

• The property at your operating concentration and temperature — or the curve, if one is available, so you can read the value where you will run.
• The measurement basis for any figure you intend to rely on: temperature, concentration, method.
• The current TDS and SDS for the exact model designation, not the family.
• Confirmation of the wetted-material list, including seals, gaskets and any aluminium grade.
• The monitoring parameters and action thresholds for the fluid in service.

Bring your operating window, concentration and material list with the request. A question that names those three can be answered with data; a question that asks whether the fluid "performs well" cannot.

The selection guide sets out the six application questions worth preparing before you approach a supplier, and you can request documentation for a specific model rather than a family. Where a reference value and a model-specific datasheet disagree, the datasheet governs — the glycol-type versus organic-salt comparison explains why older website material can differ from the supplied product data.