The Sag Formula Is the Easy Part

Two Circulating Calculations, One Off by a Factor of Ten — What the Parabolic Sag Equation Does Not Tell You, and What Line Design Software Actually Computes

1. Executive Summary

Two versions of the same transmission line sag calculation circulate widely on professional networks. Both use a three hundred metre span. Both use the number 1.2 for conductor weight. Both apply the same equation and both do the arithmetic correctly. One arrives at roughly nine metres of sag and the other at three quarters of a metre.



The difference is a unit. One treats 1.2 as kilograms per metre and multiplies by gravitational acceleration to obtain a force per unit length. The other treats 1.2 as already being newtons per metre. The second reading describes a conductor weighing about a hundred and twenty grams per metre — roughly a tenth of any transmission conductor in service — and produces a sag that is physically impossible over that span.

That is a good illustration of the real point, which is that the sag equation is the least difficult part of sag-tension engineering. It is one line of algebra, it is easy to get right, and getting it right tells you almost nothing about whether a line will be safe. The genuinely hard questions sit around it: at what condition, at what tension, with what history of loading, across a whole tension section rather than one span, on supports that are rarely level, and against a clearance requirement measured at the condition that actually governs.

Most of all, the equation takes horizontal tension as an input. In real design, tension is the unknown. It is what the sag-tension calculation solves for, at every temperature and loading condition the line must survive, from a limit the designer sets at one design condition. Presenting tension as a given inverts the problem.

This paper works through the equation and its assumptions, the change-of-state problem that sits behind it, why software is used, what an integrated line design platform genuinely computes, what automated structure spotting can and cannot know, and where survey import decides the quality of everything downstream.

2. Two Posts, Same Number, Two Answers

Set the two calculations side by side, because the comparison makes the point better than any explanation.

Version B is internally consistent and describes a conductor that does not exist. A weight of 1.2 newtons per metre corresponds to about a hundred and twenty grams per metre. A common aluminium conductor steel reinforced transmission conductor weighs on the order of one and a half kilograms per metre — roughly sixteen newtons per metre, more than ten times the figure used.

Correcting only the unit and leaving everything else in Version B unchanged gives a sag of about 7.4 metres rather than 0.75. That is the difference between a line that clears a road and one that does not, and it arrived through a single mislabelled unit in an otherwise flawless calculation.

3. What the Formula Actually Says

The parabolic sag equation states that mid-span sag equals the weight per unit length multiplied by the span squared, divided by eight times the horizontal tension.

D  =  w L² / (8 H)



Three things about it are worth stating precisely, because each is a place where the formula is misused.

• The weight term is a force per unit length, not a mass per unit length. In metric working, a conductor catalogue weight in kilograms per metre must be multiplied by gravitational acceleration first. In North American practice, conductor weight is published in pounds per foot and tension in pounds — both already forces — so no conversion step exists, which is precisely why the error appears when sources from different markets are mixed.
• The tension term is the horizontal component, which is constant along the span. It is not the tension at the support, which is higher because the conductor is inclined there. Conductor strength checks are made against the support tension, so using the horizontal value where the support value is required understates the loading.
• Weight divided into tension gives a length, and that length — the catenary constant — is the single parameter that characterises a span. Sag is simply the span squared divided by eight times that constant. It is not merely algebra; the catenary constant is used directly in vibration assessment, which Section 9 returns to.

For the corrected Version A figures, the catenary constant is about twelve hundred and seventy metres, and the support tension exceeds the horizontal tension by well under one percent. On a long crossing span with deep sag that margin grows and stops being negligible.

4. Parabola or Catenary?

A conductor hanging under its own weight follows a catenary, which is a hyperbolic cosine. The parabola is an approximation to it, valid when sag is small relative to span.

It is worth knowing how good the approximation is, because the answer is reassuring and the limits are real. Working the true catenary for the corrected Version A case gives about 8.84 metres against the parabolic 8.83 — a difference of roughly one hundredth of a metre, or about a tenth of a percent. At that sag-to-span ratio, close to three percent, the parabola is entirely adequate.



The error grows with the sag-to-span ratio, and it grows in one direction: the parabola always understates sag. At a ten percent sag-to-span ratio the understatement is on the order of one percent; at fifteen percent it approaches three percent. For ordinary transmission spans the parabola is fine. For long river and valley crossings, for lightly tensioned distribution construction, and for any case where sag approaches a tenth of the span, the catenary should be used — and design software uses it throughout regardless, because there is no reason not to.

The practical guidance is simple. Use the parabola for hand checks and order-of-magnitude work. Do not use it to establish a clearance on a long span, and never use it to argue that a marginal clearance is acceptable, because the direction of its error works against you.

5. The Input That Is Really an Output

This is the most consequential thing the circulating posts get wrong, and it is a conceptual error rather than an arithmetic one.



Both calculations treat horizontal tension as given. In design it is not given. Tension is what the sag-tension calculation solves for, and it takes a different value at every temperature and every loading condition the line experiences.

The actual procedure runs the other way round. The designer sets a tension limit at a defined design condition — for example a maximum permitted tension under the governing ice and wind loading, expressed as a percentage of the conductor’s rated breaking strength, together with an everyday tension limit for vibration control. Those limits, with the conductor’s physical and stress-strain properties, define a reference state. From that reference state, the change-of-state equation solves for the tension at every other condition of interest, accounting for thermal expansion of the conductor and for its elastic and permanent elongation. Only then is sag computed, from the tension the calculation produced.

Two consequences follow that matter for anyone reading a sag figure.

• A sag quoted without its condition is meaningless, because the tension it came from belongs to one specific state. The same conductor on the same span has materially different sag at fifteen degrees than at its maximum operating temperature.
• Conductor properties are not optional inputs. Cross-sectional area, elastic modulus, thermal expansion coefficient and the stress-strain characteristic all enter the change-of-state solution. A calculation that needs only weight and tension is not solving the design problem — it is evaluating one already-solved state.
  

6. Which Sag? The Condition Is the Question

A single line has many sags, and the design has to know which one governs which check.

The difference between initial and final sag is not a detail. Conductors elongate permanently through metallurgical creep over years and through settlement under the first heavy load. A line strung to its final sag will end up sagging further than the design allows; a line strung to its initial sag arrives at the design condition years later. Getting the two confused at installation produces a line that is correct in the model and wrong on the ground.

7. Level Supports, and What Happens When They Are Not

The parabolic equation assumes the two supports are at the same elevation. Real alignments almost never are.

With an elevation difference between supports, the lowest point of the conductor moves away from mid-span toward the lower support, and in steep terrain it can move outside the span altogether — the conductor rises continuously from one support to the other, and there is no low point within the span at all. That condition produces uplift at the lower structure, which is a design check in its own right.

The sag figure itself also becomes ambiguous. Sag measured vertically from the straight line joining the two supports is a different number from sag measured below the lower support. Both are used, and a sag value quoted without saying which is being reported is an invitation to error.



The clearance consequence is the one that matters. Ground clearance is not measured at mid-span; it is measured wherever the conductor comes closest to the ground, which on inclined and undulating terrain is somewhere the hand calculation never considered. That is one of the two central reasons line design uses a three-dimensional terrain model rather than a spreadsheet.

8. One Span Is Not a Line

The second central reason is that spans are not independent.



Between two dead-end structures lies a tension section containing many spans, joined at suspension structures whose insulators swing freely. Because they swing, tension equalises along the whole section. Pull one span tighter and the insulators move until the tensions balance again. A calculation that treats a span in isolation, with its own assumed tension, is not describing how the section behaves.

The conventional way to handle this is the ruling span: a single equivalent span that represents the whole section, from which sag and tension for every individual span are then derived. It is a good approximation for uniform terrain with comparable span lengths, and it underlies most routine design.

It becomes unreliable where the assumption of free tension equalisation breaks down — widely varying span lengths, steeply inclined spans, large elevation differences, and short spans adjacent to long ones. And it cannot represent longitudinal load cases at all, because broken conductor, unbalanced ice and stringing conditions all involve unequal tension either side of a structure, which is precisely what the ruling span assumes away. Those cases require a finite element solution that models each span with the structures and insulators as flexible elements.

9. Tension Answers to Three Masters

Choosing the tension a line is strung at is a genuine optimisation, and the three constraints pull in different directions.



• Strength. Tension at the governing loading condition must stay within a defined fraction of the conductor’s rated breaking strength, and it must stay within what the structures, foundations and hardware were designed for. This sets an upper bound.
• Clearance. Higher tension means less sag, which means shorter structures or longer spans for the same clearance. This pushes tension up, because structures and foundations are the expensive part of a line.
• Vibration. Aeolian vibration causes fatigue at attachment points and splices over decades, and susceptibility rises with tension. The everyday tension limit — expressed as a fraction of rated strength, or directly as the catenary constant — is a fatigue control, not a strength requirement. This pushes tension down.

The third one is the constraint that gets forgotten, because its consequences appear in year twelve rather than at commissioning. A line strung tight enough to save a structure per kilometre, without damping designed for the resulting exposure, will show broken strands at suspension clamps long after everyone involved has moved on.

Note that the catenary constant from Section 3 is exactly the quantity used in that assessment. The algebra that produces sag also produces the parameter that governs fatigue life, which is a good reason to understand what the terms mean rather than treating the formula as a black box.

10. Units: Where the Errors Come From

The unit error in Version B is not a one-off. It is structural, and it comes from working across two measurement traditions.



• In North American practice, conductor weight is published in pounds per foot, tension in pounds, span in feet and sag in feet. The pound is a force, so the equation applies directly with no conversion. Nothing in the calculation reminds you that a conversion might be needed, because it never is.
• In metric practice, conductor weight is commonly published in kilograms per metre, which is a mass. It must be multiplied by gravitational acceleration to become a force per unit length before the equation applies. Skip that step and every result is too small by a factor of about ten.
• Mixed sources make it worse. A conductor catalogue in one system, a worked example in another, and a spreadsheet inherited from a third is how the error survives review — each individual step looks correct.

The defence is a sense check that takes seconds. Divide tension by weight per unit length to get the catenary constant, then divide span squared by eight times that constant. For ordinary transmission construction the catenary constant lands in the region of one to two thousand metres and sag-to-span sits at a few percent. A catenary constant of fifteen thousand metres, as Version B implies, is an immediate signal that something is wrong with the inputs rather than the arithmetic.

11. What the Software Actually Does

With the preceding sections in view, it becomes clear why line design is done in an integrated three-dimensional model rather than a spreadsheet. It is not because the sag equation is difficult. It is because the equation has to be solved everywhere, at every condition, against terrain, simultaneously.

12. Optimum Spotting: What It Can and Cannot Know

Automated structure spotting is the capability that most impresses people, and it deserves both the credit and the qualification.



What it does is real. Given an alignment, a terrain model, a catalogue of available structures with their capabilities and costs, a set of clearance and structural criteria, and a cost model covering structures, foundations and earthwork, the optimisation evaluates enormous numbers of combinations of structure position, type and height, and returns the arrangement with the lowest modelled cost that satisfies every constraint. No human process can search that space.

What it cannot do is know anything that was not entered.

• It optimises against the cost model it was given. If foundation costs do not vary with ground conditions in the model, it will place structures where the ground is difficult without penalty.
• It respects the criteria it was given. Wrong clearance requirement, wrong loading case, wrong conductor condition — the result is a fully optimised design against the wrong rules.
• It knows nothing about land. Parcel boundaries, unwilling landowners, easement cost and acquisition risk are frequently the dominant economics of a line and appear nowhere in the optimisation.
• It knows nothing about access or constructability. A structure at the optimal electrical location on a slope with no road is not optimal.
• It knows nothing about permitting, environmental constraints, cultural resources, or aviation obstruction assessment.

Used properly, the optimisation is the starting point for engineering judgment rather than a substitute for it. The right workflow is to run it, then walk the result against the constraints the model does not contain, then constrain the model where reality requires it and run again. An engineer who issues the first output has not designed a line — they have accepted a cost-minimisation over an incomplete constraint set.

13. Survey Import: Where the Model Is Won or Lost

Nothing downstream is better than the terrain surface, and most of the expensive errors on line projects are seeded here.

13.1 The Two Import Paths

Airborne lidar is the industry standard for transmission work. Points arrive classified, and the classification is what allows ground to be separated from vegetation, buildings, existing conductors and structures. Filtering large datasets before import, mapping classification codes to the design platform’s feature codes, and generating a triangulated surface from the ground returns is the standard sequence.



Conventional survey arrives as coordinate files, commonly ordered as point number, northing, easting, elevation and description. Column mapping at import, followed by selection of the points that define the alignment, is the equivalent sequence.

13.2 A Detail Worth Getting Right

Classification codes follow a published standard, and the assignments matter. Ground is class two. Conductors are class fourteen, shield wires class thirteen, transmission structures class fifteen, and structure-to-wire connectors class sixteen. Class ten is rail, not wire — a mapping error there puts railway returns into the conductor layer and conductor returns nowhere, which is exactly the kind of silent error that survives to the clearance report.

13.3 The Errors That Cost Real Money

• Datum and geoid. Confusing ellipsoidal with orthometric height, or applying the wrong geoid model, offsets the entire corridor vertically. Every clearance is then wrong by the same amount, consistently, with nothing looking anomalous.
• Projection and scale. Over a long route, a projection chosen for convenience introduces distance distortion that accumulates into span lengths and therefore tensions.
• Misclassification. Vegetation classed as ground raises the surface and hides clearance problems. Ground classed as vegetation does the reverse and produces phantom violations.
• Interpolation. Dense canopy, water and steep slopes produce sparse or absent ground returns, and the surface is interpolated across them. Interpolated ground under a critical span is a place for a check shot, not a place to trust the model.
• Missing features. Crossings, buildings and utilities that were never captured or coded do not exist as far as automated checking is concerned, and the clean clearance report will not mention them.
  

14. New Line Versus Upgrade

The two workflows share a platform and very little else.

The as-built recovery capability deserves emphasis because it is undersold in most descriptions. Lidar of an energised line captures the conductors in their real position at a known temperature and loading, which allows the actual catenary and tension state to be recovered. That is a far better starting point than the original design drawings, which record what was intended rather than what was built, and which do not reflect decades of maintenance, replacement and structure settlement.



One terminology note for upgrade work: in this context, the constraint is a clearance-limited location rather than a thermal one. The conductor is capable of running hotter; the ground clearance at the worst location is what caps the rating. Calling those locations thermal constraints confuses the diagnosis, and the mitigation — re-tension, raise a structure, shorten a span, or reconductor — addresses clearance, not temperature.

15. Reading the Source Material Correctly

Check the units before checking the arithmetic

Both circulating calculations divide correctly. Only one describes a real conductor. Divide tension by weight per unit length, look at the catenary constant, and ask whether a span behaves that way. It takes seconds and it catches the error the arithmetic cannot.

Tension is not an input

The formula takes horizontal tension as given. Design does not. Tension is solved for at every condition from a limit set at one design condition, using the conductor’s stress-strain properties. A calculation that needs only weight and tension is evaluating a state that somebody else already solved.

The note about temperature, wind, ice and creep is right and understated

Both posts end by saying that real design should account for those things. That is correct, and it understates the position: those factors do not refine the answer, they determine it. Sag at maximum operating temperature after creep is a different number from sag at stringing temperature when new, and only one of them is the clearance case.

Software descriptions use marketing language for engineering functions

Terms such as digital twin and real-time physics engine describe genuine capability in promotional register. The underlying functions — an integrated three-dimensional model, a change-of-state sag-tension solution, a finite element structural interaction, and automated clearance checking against terrain — are what an engineer should evaluate and specify.

Optimisation is bounded by its inputs

Automated spotting finds the lowest modelled cost against the criteria and cost profile it was given. Land, access, constructability, permitting and environmental constraints sit outside the model, and on most projects at least one of them governs.

Classification codes are not interchangeable



Class ten is rail. Conductors are class fourteen. A mapping error there is invisible in the workflow and visible only in a clearance result that was computed against the wrong points.

16. Keentel Transmission Line Design Services

Transmission line design is core practice at Keentel Engineering. We work across the full delivery chain — survey through detailed design and construction support — and carry the line model forward into the power system studies that depend on it.

16.1 Modelling and Analysis

• Three-dimensional line model development including terrain, features, crossings and existing utilities, with independent verification of datum, geoid, projection and point classification before the surface is used.
• Sag-tension analysis by ruling span and by finite element methods, using manufacturer stress-strain data, with initial and final states distinguished and every governing condition evaluated.
• Longitudinal and security case analysis — broken conductor, broken shield wire, unbalanced ice — and uplift assessment, which the ruling span method cannot address.
• Clearance analysis in three dimensions against ground, features, crossings, other circuits and between phases, at the governing conditions including blowout and insulator swing.
• Structure spotting and optimisation, run against a cost model that reflects the actual project and then reconciled against the constraints the model does not contain.

16.2 Structures, Foundations, and Electrical Design

• Structure family development and capability envelope derivation with the load case basis documented, and analytical modelling of lattice, tubular pole, H-frame and guyed structures under actual loads.
• Insulator assembly selection and swing analysis, arm geometry, and right-of-way width determination from blowout and clearance.
• Foundation reactions issued by load case with definitions attached, foundation type evaluation, and geotechnical coordination scoped early enough to influence spotting.
• Conductor selection and thermal rating analysis including high-temperature low-sag conductors, insulation coordination, shielding and lightning performance, corona and field effects, and grounding and footing resistance.
• Vibration and galloping assessment, damping and spacer design, and everyday tension selection as a fatigue control.

16.3 Uprating and As-Built Assessment

• As-built model development from lidar capture of energised lines, with the actual catenary and tension state recovered rather than assumed from original drawings.
• Uprating and reconductoring studies including identification of the clearance-limiting locations, option comparison across re-tensioning, structure modification, span shortening and reconductoring, and structural capacity verification under revised loading.
• Facility rating determination traced to the clearance-limiting analysis, and vegetation clearance basis derived from the model.

16.4 Deliverables, Studies, and Owner’s Engineer Support

• Plan and profile sets, staking reports, bills of material and cost estimates from the model of record; stringing charts and sag tables issued at the correct condition, with initial and final states clearly distinguished.
• Line constants calculation — sequence impedances, susceptances and zero-sequence mutual coupling — taken directly from the designed geometry, and the short-circuit, coordination, relay setting, stability and transient studies that use them.
• Design review of third-party line design packages, QA/QC of models and deliverables, and independent verification where a design has been produced elsewhere.

Keentel Engineering holds a Florida Certificate of Authorization and maintains offices in Tampa, Austin, Sacramento, and Baltimore, supporting projects across the interconnections.

17. References and Further Reading

The following are referenced by subject in the body of this document. The edition adopted by the authority having jurisdiction governs code requirements, and the current published edition of each standard governs its own content.

Code and Loading



• ANSI C2, National Electrical Safety Code — clearance requirements, district and extreme loading cases, overload factors and grades of construction  —  Institute of Electrical and Electronics Engineers
  https://standards.ieee.org/products-programs/nesc/
• ASCE Manual of Practice No. 74, Guidelines for Electrical Transmission Line Structural Loading  —  American Society of Civil Engineers
  https://www.asce.org/
• IEC 60826, Design criteria of overhead transmission lines, for projects following the international framework  —  International Electrotechnical Commission
  https://webstore.iec.ch/

Conductor, Sag-Tension, and Vibration

• IEEE Std 738, Standard for Calculating the Current-Temperature Relationship of Bare Overhead Conductors — the thermal rating basis that sets the maximum operating temperature  —  IEEE Standards Association
  https://standards.ieee.org/
• Aluminum Association conductor data and stress-strain characteristics, and the manufacturer’s published physical, thermal and stress-strain properties for the specific conductor selected  —  The Aluminum Association
  https://www.aluminum.org/
• IEEE Std 524, Guide to the Installation of Overhead Transmission Line Conductors, covering stringing practice, sag tables and construction loading  —  IEEE Standards Association
  https://standards.ieee.org/
• CIGRE technical brochures on aeolian vibration, conductor self-damping and fatigue endurance, on galloping, and on the application of high-temperature low-sag conductors  —  CIGRE
  https://www.cigre.org/

Survey, Structures, and Ratings

• ASPRS LAS specification and point classification definitions, and the ASPRS Positional Accuracy Standards for Digital Geospatial Data  —  American Society for Photogrammetry and Remote Sensing
  https://www.asprs.org/
• ASCE 10 for latticed steel transmission structures, ASCE 48 for steel transmission pole structures, ASCE Manual of Practice No. 91 for guyed structures, and IEEE Std 691 for transmission structure foundation design and testing  —  ASCE and IEEE Standards Association
  https://www.asce.org/
• NERC Reliability Standards — including the facility ratings and vegetation management standards whose inputs originate in the line design model  —  North American Electric Reliability Corporation
  https://www.nerc.com/pa/Stand/Pages/ReliabilityStandards.aspx
  

17. Frequently Asked Questions

Notice and Disclaimer

This document is original technical content prepared by Keentel Engineering LLC for general professional information. It is not project-specific engineering advice and does not constitute a design, a study, or a calculation for any line. Sag, tension, clearance and structure design must be established by project-specific analysis using verified survey data, the manufacturer’s published conductor properties, and the code edition and utility standards applicable to the project.

Worked figures in this document are calculated from values published in third-party material for the purpose of demonstrating method and identifying an error in that material. They are illustrative and are not design values for any conductor, span or installation. Representative ranges quoted for catenary constants, sag-to-span ratios and approximation errors are general engineering discussion and vary with conductor, terrain, loading and design practice.

Software capabilities described here are discussed generically. Software platforms referred to are the products of their respective owners, and all product names and trademarks are the property of those owners. Keentel Engineering LLC is an independent engineering consultancy and is not affiliated with, endorsed by, sponsored by, or a partner or reseller of any software vendor, and no such relationship is implied. Capability or performance claims made in third-party promotional material are those of the vendor and are not adopted or endorsed here.

Reference to any standard, code, industry organisation, regulator, or equipment category in this document does not imply affiliation with, endorsement by, or sponsorship from any such organisation.