Why Is the Link Budget Always Off?

The most frustrating part of DAS design isn't drawing the system diagram. It's calculating the antenna port power, thinking it's sufficient, and then measuring it on site to find it's off by 5 or 6 dB. Coverage shrinks, users complain, and when you trace it back you find — component insertion loss was undercalculated, feeder loss was estimated wrong, connector loss was completely ignored.

A link budget is ultimately addition and subtraction. But many people get it wrong. Each stage is off by a fraction of a dB, and cascaded together it becomes a big problem.

This article walks through every step from the RRU output to the handset receiver, breaking down how many dB are lost at each stage, with a fill-in link budget template you can use directly.

1. What Exactly Is a Link Budget Calculating?

The essence of a link budget is answering one question: From the source to the handset, how much power is left after passing through feeders, components, and space?

In a DAS, the signal travels this path:

RRU output → Jumper → Main feeder → Power splitter / Coupler → Branch feeder → Antenna → Free space → Handset

Every step has loss. The link budget adds up the loss at each stage to see how much power remains at the antenna port and at the handset.

Two critical nodes:

• Antenna port power: The power arriving at the antenna input. This determines how strongly the antenna can radiate.
• Handset received power: The power arriving at the handset after free space propagation. This determines whether the user can communicate normally.

If the link budget is inaccurate, everything downstream is guesswork.

2. From RRU to Antenna Port: Component and Feeder Loss

This section is all about "visible" loss — cables, connectors, splitters, couplers — each with a specific loss value.

2.1 Feeder Loss: Diameter and Frequency Decide Everything

Feeder loss depends directly on two factors: the larger the diameter, the lower the loss; the higher the frequency, the higher the loss.

Feeder loss at different bands (dB/100m):

Quick field estimate: 1/2″ feeder loses about 0.65dB per 10m at 1800MHz and about 0.78dB per 10m at 2600MHz. 7/8″ feeder is roughly half of 1/2″.

Real example: RRU to first splitter distance 40m, using 1/2″ standard feeder at 2600MHz.

Feeder loss = 40 × 7.8 / 100 = 3.12dB

If the design estimated 6dB per 100m, it would be off by 0.72dB. That's not much for 40m, but if every stage is a little off, the total adds up.

2.2 Splitter Loss: Theoretical Split Loss + Insertion Loss

Splitter loss has two parts: theoretical split loss (determined by energy conservation) and insertion loss (consumed by the device itself).

Cavity splitters have low insertion loss; microstrip splitters have higher insertion loss. After two or more stages of cascading, the difference becomes obvious.

2.3 Coupler Loss: Mainline Insertion Loss + Coupling Loss

A coupler has two loss concepts that are easily confused:

• Mainline insertion loss: Loss from input to output (the path that continues forward). A 10dB cavity coupler has mainline insertion loss of about 0.1-0.3dB.
• Coupling insertion loss: Loss from input to coupled port, equal to coupling value + extra loss. A 10dB coupler has coupling insertion loss of about 10.5-11dB.

In a DAS link budget, what actually eats the signal is the mainline insertion loss, not the coupling insertion loss. When selecting, make sure you're looking at the right one.

2.4 Jumper and Connector Loss: The Most Overlooked Link

Every RF connector (N-type, 4.3-10, 7-16 DIN) adds insertion loss of 0.05-0.2dB.

A typical DAS may have 20-30 connector pairs. If each pair averages 0.1dB, that's 2-3dB total.

Jumper cable loss: 1/2″ super flexible jumper is about 0.09dB/m at 2600MHz. A 3m jumper is about 0.27dB.

Connector loss is easy to overlook because each one is small, but there are many. When doing a link budget, estimating 0.1dB per connector pair is a reasonable starting point.

3. From Antenna Port to Handset: Free Space Path Loss

The signal radiates from the antenna and travels through space to the handset. This is free space path loss.

3.1 Free Space Path Loss Formula

Lfs = 32.4 + 20×log10(f) + 20×log10(d)

• Lfs: Free space path loss (dB)
• f: Frequency (MHz)
• d: Distance (meters)

Example: 2600MHz, 10 meters

Lfs = 32.4 + 20×log10(2600) + 20×log10(10) = 32.4 + 68.3 + 20 = 120.7dB

3.2 Additional Indoor Loss

The free space formula is ideal. Indoor environments have walls, furniture, and human bodies that cause more loss. In indoor link budgets, an indoor attenuation margin is typically added:

4. Complete Link Budget Template

Below is a fill-in template. Assume a typical scenario: RRU output 33dBm (2W), operating at 2600MHz, RRU to first splitter 40m of 1/2″ feeder, one 2-way splitter, splitter to antenna 15m of 1/2″ feeder, antenna gain 3dBi, coverage distance 10m.

The result shows a link margin of only 1.9dB, almost no margin. If the design missed connector loss (-0.4dB) or underestimated feeder loss by 0.7dB, the link margin becomes negative — users at the cell edge drop calls directly.

5. Common Mistakes: Why Is the Link Budget Always Off?

Mistake 1: Estimating feeder loss using "typical values".

Many designers habitually estimate "1/2″ feeder at 7dB per 100m", but actual values vary greatly with frequency. At 2600MHz it's 7.8dB, at 3500MHz it's 10.2dB. The higher the frequency, the greater the estimation error.

Correction: Check the actual attenuation curve in the feeder datasheet. Don't use "rough values".

Mistake 2: Forgetting connector loss.

A DAS has dozens of connectors. Each 0.1dB doesn't seem like much, but added up it's 2-3dB. Many people only count feeders and components, ignoring connector loss entirely.

Correction: Include 0.1dB per connector pair, or add a 5% margin to total loss.

Mistake 3: Confusing mainline insertion loss with coupling insertion loss.

The "10dB" of a coupler refers to coupling value, not insertion loss. Mainline insertion loss is only 0.1-0.3dB. If you mistakenly use coupling insertion loss (10dB+) as mainline insertion loss, the link budget will be severely underestimated.

Correction: For couplers, only look at mainline insertion loss. Coupling insertion loss does not affect the main path signal.

Mistake 4: Ignoring PIM's impact on the uplink.

The link budget only calculates downlink power. But PIM raises the uplink noise floor, causing uplink coverage shrinkage. No matter how accurate the downlink budget is, if the uplink is suppressed by PIM, users still can't make calls.

Correction: After completing the downlink budget, separately check the uplink noise floor. If PIM is not up to standard, solve the PIM problem first, then talk about power budget.

6. FAQ

Q1: In a link budget, should I count both the "theoretical loss" and "insertion loss" of a splitter?

Yes. Total splitter loss = theoretical split loss + insertion loss. A 2-way splitter has 3dB theoretical loss plus 0.1-0.5dB insertion loss, for a total of 3.1-3.5dB.

Q2: What indoor attenuation margin should I use?

Depends on the environment. Open office: 5-8dB; cubicle office: 8-12dB; concrete walls: 15-20dB. Too small a margin causes edge coverage dropouts; too large a margin requires excessive antenna port power, possibly requiring a higher-power source.

Q3: How much link margin is safe?

At least 3dB is recommended. System aging, environmental changes, and component performance drift all consume margin. 3dB means about a 2× signal strength margin.

Q4: How much less loss does 7/8″ feeder have compared to 1/2″?

About 40-50% less. At 2600MHz, 1/2″ standard feeder is 7.8dB/100m and 7/8″ is 5.4dB/100m, a difference of about 31%. But 7/8″ is stiffer and has a larger bend radius, so it can't be used at equipment ports — only for main runs.

Q5: Why is the measured received power lower than calculated in the link budget?

Common reasons: connectors not tightened (adds loss), actual feeder length longer than drawing, actual antenna gain lower than nominal, environmental attenuation greater than estimated. It's recommended to measure power at each node with a spectrum analyzer during commissioning and compare with the budget table to find the source of deviation.

7. Practical Advice for Designers

First, use a table, not mental math. List each component, each feeder segment, each connector on its own row. Mental math is prone to omissions.

Second, check actual feeder loss data. Different manufacturers and batches have different feeder attenuation. Maniron's feeder assemblies come with measured insertion loss data, ready to use.

Third, leave enough margin. Link margin should be ≥3dB. Don't design right at sensitivity; the field environment is far more complex than the lab.

Fourth, calculate the uplink after the downlink. PIM's impact on the uplink can be greater than link loss. Low PIM components (≤-155dBc) are the foundation for uplink coverage.

Fifth, create a complete link budget template and reuse it for every project. Pre-fill component parameters, feeder loss, and connector loss. Change distance and frequency to calculate.