IndieRF Coax Loss Calculator

Coax cable loss calculator

Pick up to three cables, a run length, and a frequency. You get total loss, power at the load, and delay for RG-58, RG-316, RG-142, LMR-400, Ecoflex, semi-rigid, and lab cables, fitted to each manufacturer's printed attenuation table. The plot shows where the table ends and extrapolation starts.

Loading the calculator…

Above the printed frequency

Loss keeps climbing on the same curve. Conductor loss rises as √f and dielectric loss as f, so a fit that matches the table keeps working past its last row. Fit Belden 83284 (RG-316) on its rows up to 1 GHz and it predicts the printed 3 GHz row within 3%. Fit LMR-400 the same way and it lands within 1% of the printed 8 GHz row. That is why each curve continues as a dashed line past its table.

A datasheet stops at a frequency for other reasons, and each one arrives at a different place:

  • A second mode. Above the TE11 cutoff, f ≈ 2c·VF / (π(D + d)), the cable carries a second mode, and loss and impedance stop being one number. With D the dielectric diameter and d the center conductor, LMR-400 cuts off at 16.3 GHz, RG-213 at 13.3 GHz, and LMR-600 at 10.4 GHz. RG-316 is near 66 GHz. Bigger cable, lower ceiling. Curves stop there.
  • The connector. PL-259 (UHF) is a VHF connector. BNC is good to about 4 GHz, Type N to 11 GHz, SMA to 18 GHz, and 2.92 mm to 40 GHz. Read the connector's own sheet: a cheap SMA on RG-316 can be the worst return loss in the run long before the cable is.
  • Shielding and impedance ripple. A single braid leaks more as the wavelength shrinks toward the weave. Foil plus braid (LMR) or a solid tube (semi-rigid) holds up. Small periodic flaws from manufacturing add return-loss spikes, which is why Belden specifies RG-316 VSWR only from 1 to 3000 MHz.
  • Power. The loss is heat in the center conductor, so the power rating falls as loss rises. Times rates LMR-400 at 3.33 kW average at 30 MHz, 0.58 kW at 900 MHz, and 0.17 kW at 8 GHz.

Each curve ends at the lowest of the sheet's rated maximum, the TE11 cutoff from its printed diameters, and 4× its last printed row. A table that ends at 1 GHz says little about the dielectric term at 10 GHz. Below the first row the fit extends down to 1 MHz at most. Under that the skin depth approaches the conductor size and the √f term stops holding.

Picking a cable: loss against size, weight, and bend

Thicker cable loses less and fights you more. LMR-195 (4.95 mm) prints 16.9 dB/100 ft at 2 GHz, weighs 0.021 lb/ft, and allows a 2.0 in repeated bend. LMR-400 (10.29 mm) prints 6.0 dB/100 ft at the same row, at 0.068 lb/ft and a 4.0 in repeated bend. About twice the diameter buys 2.8× less loss for 3× the weight. Ten common cables ranked at 2.4 GHz and 10 m, with the SWR a 2:1 antenna shows at the radio: coax loss at 2.4 GHz.

Read both bend figures. The static (installation) radius is a one-time bend you clamp and leave. The repeated radius is for a cable that moves, and on LMR it is three to four times larger. Bend tighter than that and the dielectric deforms, which shows up as a return-loss bump at that spot.

Foam polyethylene (LMR-195 and up, Ecoflex, Aircell, H155, VF 0.80 to 0.87) is the lowest loss per millimeter here. Solid polyethylene (RG-58, RG-213, VF 0.66) is cheaper and tougher. PTFE (RG-316, RG-178, semi-rigid) costs more per dB and survives a soldering iron. Semi-rigid is a solid copper tube: 100% shield and stable phase, bent once with a tool. A lab assembly like SUCOFLEX 104 keeps its phase while it flexes, which is what a VNA test port needs, at 72 g/m and a 16 mm static bend.

How the number is made

Each curve is a least-squares fit, a·√f + b·f, to the attenuation table on one manufacturer datasheet. a is the conductor term and b is the dielectric term. The fit is reported against those printed points under each cable's details. Matched loss is that attenuation times length, at 25°C and VSWR 1, which is how the sheets specify it. A load VSWR then adds the extra line loss from the ARRL expression. Power at the load is the power you type in, minus that total.

The part number is the part. “RG-58” on this page is Belden 8262, not an average of every cable that has ever used that name. Size, weight, and bend radius come from the same sheet as the attenuation table. Each cable's details give the sheet and the date it was read. A blank means the sheet does not print it.

Checked against scikit-rf: the cases, the tolerance and what the check does not cover.

FAQ

How do I calculate coax cable loss for a length?

Pick the manufacturer part, the length, and the frequency. The loss is the datasheet fit, a·√f + b·f, times the length. Belden 83284 (RG-316) prints 47.2 dB/100 ft at 2400 MHz, so 1 m is 1.55 dB. Times Microwave LMR-400 prints 6.8 dB/100 ft at 2500 MHz, so 1 m is 0.22 dB. A frequency past the printed table is extrapolated and labeled that way.

Can I use a coax cable above its datasheet frequency?

Usually, up to its TE11 cutoff and the connector's rating. Loss keeps rising on the same √f and f curve: fitting RG-316 on its rows up to 1 GHz predicts its printed 3 GHz row within 3%. What stops a cable is a second propagating mode (16.3 GHz for LMR-400, about 66 GHz for RG-316), the connector, braid leakage, and power rating, which falls as loss rises.

Why does coax loss go up with the square root of frequency?

The center and outer conductors carry the RF current in a skin a few micrometers deep. That depth shrinks as 1/√f, so the resistance, and the conductor loss, rise as √f. The dielectric adds a second term that rises in proportion to frequency. The fit on this page is those two terms. At a few hundred megahertz the √f term is almost the whole loss. By several gigahertz the straight-line term is visible, especially on PTFE.

Why do thin cables like RG-174 and RG-316 lose so much at 2.4 and 5.8 GHz?

A thin center conductor is a small resistor, and skin effect makes it worse as frequency rises. Belden 83284 (RG-316, 2.5 mm jacket) prints 47.2 dB/100 ft at 2400 MHz, 1.55 dB for a 1 m jumper. Belden 8216 (RG-174) stops printing at 1000 MHz (34 dB/100 ft), and its fit extrapolates to 1.98 dB/m at 2.4 GHz. Times Microwave LMR-100A, a 2.79 mm cable in the same size class, prints 64.1 dB/100 ft at 5800 MHz, 2.10 dB per meter. LMR-400 at that same 5800 MHz row is 10.8 dB/100 ft, 0.35 dB/m.

When should I use LMR-400?

Use LMR-400 for an outdoor or long feed where the run, not the connector, sets the loss. It is 0.405 in (10.29 mm) foam-polyethylene cable. Ten meters is 3.79 dB on Belden 8262 (RG-58) at its 400 MHz row, 2.17 dB on Belden 9258 (RG-8X) at 400 MHz, and 0.89 dB on LMR-400 at its 450 MHz row. It is a poor bench jumper: Times rates it for a 4.0 in repeated bend radius and 0.068 lb/ft, where a short RG-316 or SUCOFLEX assembly bends in an inch.

How does VSWR add loss on a coax cable?

A mismatched load reflects power. That reflection travels back through the lossy cable, so the cable dissipates a bit more than its matched loss. The extra term here is the ARRL expression: with matched loss ratio A = 10^(α/10) and ρ = (S−1)/(S+1), total ratio = (A² − ρ²) / (A (1 − ρ²)). On 10 m of LMR-400 at 450 MHz and load VSWR 2, the matched loss is 0.89 dB and the extra is 0.18 dB. The input return loss is better than the load return loss because the cable attenuates the reflection on the way back. This is not a full source-and-load mismatch cascade.

What are phase and delay on this page?

Delay is length divided by velocity factor times the speed of light. Electrical length is that delay times frequency, shown in wavelengths and degrees. It uses the velocity factor printed on the datasheet. The Taoglas RG-316 sheet does not print one, so delay stays blank for that part. The model does not add dispersion.

Question or a correction?

Loss by cable

Each page has the maker's attenuation table and the loss at every band from 1.8 MHz to 5.8 GHz. Compare all 24 cables at one frequency and length.