Passive intermodulation, or PIM, is interference a radio system makes for itself. When two or more strong transmit signals pass through a passive component that is not perfectly linear, the component mixes them and creates new signals at other frequencies. Some of those new signals fall in the receive band, where they raise the noise floor and reduce the receiver’s sensitivity.
A passive component, a connector, a cable or an antenna, is meant to carry signals without changing them. PIM appears when a tiny non-linearity, such as a loose or corroded joint or a magnetic metal, behaves like a weak mixer. Two transmit carriers at frequencies f₁ and f₂ then create products at:
PIM product frequency = m · f₁ ± n · f₂
The order of a product is m + n. The third-order products, 2·f₁ − f₂ and 2·f₂ − f₁, are usually the troublemakers, because they fall closest to the original carriers and often land in a nearby receive band.
PIM behaves unlike most interference in two awkward ways. More transmit power makes it worse, not better, and once a product lands inside the wanted receive band it cannot be removed without also removing the signals the receiver is trying to hear. So it has to be stopped at the source.
Two transmit carriers at f₁ and f₂ create products at every combination of m · f₁ ± n · f₂. The order is the sum of the multipliers, m + n. Even-order products exist as well, but the odd-order ones, particularly third, fifth and seventh order, are usually the most troublesome, because they tend to fall closest to the original carriers and so are the most likely to land in a nearby receive band. The third order is the strongest and the most common offender:
2·f₁ − f₂ and 2·f₂ − f₁
Here is a real example. A GSM900 base station transmits two downlink carriers at f₁ of 935 MHz and f₂ of 960 MHz. The third-order products are:
2 × 935 − 960 = 910 MHz 2 × 960 − 935 = 985 MHz
That same base station receives the uplink band of 890 to 915 MHz. The product at 910 MHz lands squarely inside it. The site is now interfering with its own receiver, on a frequency it cannot filter away because it is the band it needs to hear.
Warning: do not ignore the higher orders
Each step up in order, from third to fifth to seventh, produces a weaker product, so it is tempting to design around the third order alone. On a high power site with many carriers, fifth and seventh order products can still sit well above the noise floor, and more carriers mean far more combinations that can reach a receive band. Check the orders your actual band plan allows, not just the third.
Most sites transmit and receive at the same time on different frequencies. A PIM product that lands in the receive band behaves exactly like interference from outside: it raises the noise floor the receiver has to work against. The receiver cannot tell the difference between a faint handset at the cell edge and a PIM product sitting on the same frequency.
The result is desensitisation, usually shortened to desense. As the noise floor rises, the weakest signal the receiver can still decode has to be stronger, so the cell shrinks. Calls drop at the edge, uplink data rates fall, and the coverage suffers on the uplink while the downlink still looks healthy.
Third-order PIM also rises about three times faster than carrier power: raising the transmit power by 1 dB lifts a third-order product by roughly 3 dB. That non-linear scaling is one reason PIM can stay hidden on a low power test and then turn severe once the site runs at full power.
Because the cause sits inside the site, driving around with a test handset never finds it. The symptom is a coverage problem, but the fault is a junction on the tower.
PIM is quoted in two ways. In absolute terms it is the power of the product in dBm. In relative terms it is in dBc, which is how many decibels the product sits below the carriers that created it. The more negative the dBc figure, the better the component.
The standard test, set out in IEC 62037, drives the component with two carriers of 20 W each, which is +43 dBm, and measures the third-order product. A good connector might be rated −160 dBc, while a poor or damaged one can be −120 dBc or worse.
Putting numbers in shows why the rating matters. The dBc figure is referenced to each +43 dBm test carrier, not the two combined, so:
A −160 dBc connector makes a product at 43 − 160 = −117 dBm. A −150 dBc connector makes a product at 43 − 150 = −107 dBm.
A typical LTE or GSM receiver floor may sit near −118 dBm in a narrow receive channel, though the exact value depends heavily on channel bandwidth and receiver design. The −117 dBm product from the good connector is right at that floor and barely moves it. The −107 dBm product from the poorer connector is about 11 dB above the floor, which is severe receiver desensitisation. A 10 dB difference on the datasheet is the difference between a healthy site and a crippled one.
It is easy to assume that a connection with a good impedance match, meaning a low VSWR and a high return loss, must also be low in PIM. That is not how it works.
Return loss measures how much signal a junction reflects. PIM measures how much new signal a junction creates through non-linearity. The two come from different physical causes, so a connector can reflect almost nothing and still generate damaging PIM. A return loss and VSWR sweep will happily pass a connection that a PIM test fails, which is why careful sites test for both.
PIM comes from any junction where current does not flow cleanly and linearly. The usual sources are:
PIM is far cheaper to prevent than to chase. In rough order of return:
PIM has grown into a bigger problem as sites have become busier. Carrier aggregation, multiband antennas, massive MIMO and active antenna systems, and shared towers carrying several operators all push more high power carriers through the same passive hardware, and every extra carrier adds more pairs that can mix into a receive band.
Higher powers raise the product levels, and denser spectrum leaves fewer empty frequencies for the products to fall into harmlessly. It is one strand of the wider trend covered in why RF interference is getting worse.
What is passive intermodulation in simple terms? Passive intermodulation, or PIM, is interference a transmitter creates for itself. When two or more strong signals pass through a passive component that is slightly non-linear, such as a loose or corroded connector, the component mixes them into new frequencies. Some of those new signals fall in the receive band and reduce the receiver’s sensitivity.
What causes PIM? PIM is caused by non-linear junctions in the passive RF path: loose or under-torqued connectors, contamination, corrosion, ferromagnetic metals such as nickel and steel, cracked solder, and damaged cable. It can also come from corroded or loose metal in front of the antenna, known as the rusty bolt effect.
How is PIM measured? PIM is measured by driving the system with two carriers, usually 20 W each under the IEC 62037 standard, and reading the power of the resulting intermodulation product. It is quoted in dBm as an absolute power, or in dBc as the number of decibels below the test carriers. A more negative dBc figure is better.
What is a good PIM value? For a connector or component, a rating in the region of −150 to −160 dBc with two 20 W carriers is considered good, and lower is better. The figure that matters in practice is whether the resulting product sits below the receiver noise floor for that site.
Is PIM the same as VSWR? No. VSWR and return loss measure how much signal a junction reflects, while PIM measures how much new signal it creates through non-linearity. A connection can have an excellent impedance match and still generate damaging PIM, so the two are tested separately.
How do you fix PIM? Use PIM-rated components, reduce the number of connectors, torque every connection to specification, keep contacts clean and dry, avoid ferromagnetic metals, and test with a PIM analyser after installation. For external PIM, remove or bond loose and corroded metal near the antenna.
Build it in noIM₃
The PIM calculator works out where the intermodulation products from your transmit carriers fall and flags the ones that land in a receive band, and the intermodulation calculator extends that to multi-carrier sites where the combinations multiply. When the question is the match rather than the linearity, the VSWR and return loss calculator covers the reflection side of the same connection.
Passive intermodulation is the interference a site makes for itself when strong transmit carriers mix in a junction that should be linear but is not. Its products can fall straight into the receive band, where they raise the noise floor and shrink the cell, and no amount of transmit power or downlink tuning will clear them. The cure is in the hardware and the workmanship: PIM-rated components, fewer and tighter connections, clean and dry contacts, no magnetic metals, and a PIM test before the site goes live. Stop it at the junction and it never reaches the receiver.
Related Insights
EIRP is the power a perfect isotropic antenna would have to radiate to match what your real antenna sends along its main beam. You get it by adding the transmitter power and the antenna gain, then subtracting the feeder loss. Here is the formula, how EIRP differs from ERP, a worked example, and the licence limits and safety distances that all depend on it.
Jun 27, 2026
A good VSWR is usually 1.5:1 or better, and 2:1 is the common acceptance limit. Here is what VSWR actually measures, how it converts to return loss, reflection coefficient and reflected power, a conversion table, target values by application, and the real reasons a high VSWR matters.
Jun 18, 2026
A coaxial cable is not a single part number, it is a specification: impedance, loss at frequency, shielding, power, environment, and a connector chosen independently for each end. Here is how to get it right.
Jun 5, 2026
