Pick the wrong filter, and your entire system’s signal quality can go straight down the drain.

In RF systems, filters are easy to overlook – they are not as visible as antennas, nor as “noticeable” as power amplifiers. But this unassuming passive component, when chosen correctly, works quietly in the background; when chosen wrong, it can drag an entire DAS or base station project into endless tweaking and customer complaints.

Over the years, I have seen quite a few projects where the installation crew installed power dividers and couplers properly, but when it came to the filter, they just grabbed whatever was “good enough” and screwed it on. The result? Out‑of‑band interference leaked in, receive sensitivity dropped through the floor, and PIM performance could not be pushed down no matter what. After swapping it out, the problem went away.

This article skips the fluffy theory and gets straight to the engineering selection points – the parameters that actually matter when choosing a filter.

Bandpass, Bandstop, Lowpass, Highpass – Get the Type Right First

Many junior engineers look at a filter selection table and feel overwhelmed. A string of numbers and letters in the model number, and they have no idea where to start. But the first step is simple: figure out what type of filter you actually need.

Bandpass filters are the most common type in RF systems. Their job is straightforward – let signals in a specific frequency band pass through and block everything else. You will find bandpass filters everywhere: in base station transmit and receive paths, DAS front ends, and repeater input stages.

Bandstop filters (notch filters) do the opposite – they block only a narrow frequency band and let everything else pass. These are typically used when you have a specific interference problem, like a strong interfering signal sitting right inside your receive band – a bandstop filter can knock it out.

Lowpass and highpass filters are used less frequently, mainly in DC bias handling, harmonic suppression, and similar scenarios. You often see lowpass filters at the output of power amplifiers to clean up the harmonics generated by the amp.

Insertion Loss – The Signal Disappears Before It Even Reaches the Antenna

Insertion loss is probably the first parameter people look at during selection, but it is also one of the most misunderstood.

Simply put, insertion loss tells you how much signal power is “lost” passing through the filter. If a filter has 0.5 dB insertion loss, feeding 100 W in gives you roughly 89 W out – 11 W of energy dissipated inside the filter. If insertion loss hits 2 dB, 100 W in gives you only 63 W out.

In base station and DAS systems, every extra 0.5 dB of insertion loss means the power amplifier has to work that much harder to compensate. When you push power up, heat dissipation, electricity consumption, and device lifespan all take a hit. So, as long as out‑of‑band rejection requirements are met, insertion loss should always be as low as possible.

Among MS TELERF’s cavity filter products, some models achieve typical insertion loss below 0.5 dB, and certain DAS‑specific bandpass filters can even keep it under 0.5 dB. That is fairly solid performance in the industry – though, of course, the actual number depends on frequency, bandwidth, power level, and other factors. There is no one‑size‑fits‑all answer.

Out‑of‑Band Rejection – Keep Neighboring Band Signals Out

If insertion loss is about “whether the signal that should pass can get through,” then out‑of‑band rejection is about “whether the signal that should NOT pass can be blocked.”

Out‑of‑band rejection is usually expressed in dB – the higher the number, the better the filter’s ability to “keep people out.” A high‑spec base station filter may need 50 dB or even 60 dB+ of out‑of‑band rejection. That means interfering signals outside the passband are attenuated by a factor of 100,000 to 1,000,000.

Why does this matter so much? Today’s wireless spectrum is packed like a rush‑hour subway – the 900 MHz band might sit right next to 1800 MHz, with only a few tens of megahertz of separation between the two systems. If the filter’s out‑of‑band rejection is insufficient, high‑power signals from the neighboring band will leak in. At best, they raise the noise floor; at worst, they can completely swamp the receive channel.

This problem is even more pronounced in DAS systems. A single DAS may need to support multiple frequency bands from multiple operators. If isolation between filters is inadequate, intermodulation products and spurious signals can turn the whole system into a mess.

Power Handling – Do Not Let Your Filter Burn Out

Power handling looks straightforward – how many watts can it take? – but this is where selection often goes wrong.

A lot of people only look at average power and ignore peak power. In base station scenarios, signals are not constant. Voice and data traffic fluctuate between busy and idle periods, and power swings widely. If the filter’s power handling is only “just enough” for the average power, a peak can cause problems – from degraded PIM all the way to internal arcing or dielectric breakdown.

Another easily overlooked point: power handling and insertion loss are tied together. A filter with higher insertion loss generates more internal heat at the same input power. Heat causes cavity expansion, frequency response drifts, and PIM gets worse. It is a vicious cycle.

MS TELERF’s base‑station‑grade filters typically handle between 150 W and 200 W, and some outdoor models come with IP67 weatherproofing. If your project involves outdoor tower‑top installation, make sure you check the ingress protection rating – an indoor filter taken outside can be ruined by a single rainstorm.

PIM – The Invisible Killer

Passive intermodulation (PIM) was covered in detail in a previous article, so here we will only touch on filter‑specific aspects.

Filters are among the highest‑risk components for PIM in an RF chain. The reason is simple: a filter contains multiple resonant cavities connected by coupling structures, and every contact surface and every tuning screw can become a PIM source.

In DAS and multi‑operator shared sites, a filter’s PIM performance directly determines how low the system’s noise floor can be pushed. A filter with poor PIM, when excited by high power, can produce intermodulation products that land right inside the receive band – it is like the system generating its own interference.

The industry’s PIM requirements for base‑station‑grade filters are getting tighter. MS TELERF’s high‑performance filter series can achieve PIM down to -153 dBc or even -155 dBc, which meets the stringent requirements of most 5G base stations. But one thing to keep in mind – PIM specs are measured under specific test conditions (typically 2×43 dBm carriers). In actual use, lower power gives better PIM, higher power makes PIM worse. You need to leave enough margin.

Why Cavity Filters Dominate Base Station and DAS Applications

After covering the parameters, let us address a common selection question: cavity filter vs. ceramic filter – which one to choose?

Ceramic filters are smaller and cheaper, giving them an edge in consumer electronics and certain size‑sensitive applications. But in professional communication systems like base stations and DAS, cavity filters are the absolute mainstream.

The reason is not complicated. Cavity filters have high Q factor, low insertion loss, and large power handling. For the same out‑of‑band rejection requirement, a cavity filter can achieve it with fewer stages and even lower insertion loss. Moreover, cavity filters are structurally stable – temperature changes and mechanical vibration affect their performance much less than ceramic filters.

Of course, cavity filters have downsides too – larger size, heavier weight, higher cost. But in base station and DAS scenarios, performance takes priority over size, so cavity filters remain the first choice. MS TELERF’s product line is also cavity‑filter‑centric, covering multiple bands from 700 MHz to 2700 MHz.

Selection in Practice: A Checklist to Guide Your Decision

With all that theory out of the way, here is a practical selection checklist. When you get a filter datasheet, go through these items one by one:

Is the filter type correct? (Bandpass / bandstop / lowpass / highpass)
Does the passband frequency fully cover your operating band?
Is insertion loss within the link budget allowance of your system?
Does out‑of‑band rejection adequately suppress adjacent‑band interference?
Does power handling cover peak power (not just average power)?
Does PIM performance meet the system’s noise floor requirements?
Does the connector type match the rest of the system? (N‑type, DIN, 4.3‑10)
Is the installation environment indoor or outdoor? Is the IP rating sufficient?
Does the operating temperature range cover the climate conditions at the project site?

Run through these nine items, and your filter selection is unlikely to have major issues.

Filters – choose the right one and they work silently in the background; choose the wrong one and they can turn your entire system into a nightmare. Hopefully, this article helps you avoid a few pitfalls in your next project.

If you have filter selection questions, or need specific product parameters and test reports, feel free to reach out to the MS TELERF technical team.