DXLook Blog

If you've used DXLook's VOACAP view, you've seen colored arcs stretching across the map, showing where your signal should propagate based on ionospheric predictions. But what's actually happening behind those arcs?

This post goes deep into the technical choices we made to bring VOACAP — a serious propagation prediction system — to a simple web interface. It's written for hams who want to understand the math, the model, and the trade-offs.

What Is VOACAP?

VOACAP stands for Voice of America Coverage Analysis Program. It was developed by the U.S. government in the 1980s (with roots going back even further) to predict HF radio propagation for international broadcasting. The underlying model uses decades of ionospheric measurements — millions of ionosonde soundings — compiled into mathematical coefficients that describe how the ionosphere behaves under different solar conditions.

The engine itself is a Fortran program. Yes, Fortran. It reads fixed-width input files with very specific formatting (character position matters!), runs complex ionospheric calculations, and outputs predictions for signal strength, noise levels, and reliability.

Our job was to wrap this venerable engine in something practical for amateur radio operators.

The Origins and Targets Concept

VOACAP predicts propagation between two points: a transmitter (origin) and a receiver (target). To show predictions for the whole world, we need to define both.

Origins: Where Your Signal Comes From

We've defined 239 transmitter locations across the globe, concentrated in areas where amateur operators actually live and operate. These include major population centers in North America, Europe, Asia, South America, and Oceania — places like Seattle, Chicago, London, Tokyo, São Paulo, and Sydney.

Each origin has a latitude, longitude, and a regional grouping (like "NA_West" or "EU_Central"). When you enter your grid square in DXLook, we find the closest origin to your location and use that as your transmitting point.

Targets: Where We're Checking Reception

The receiver side is a grid of 328 points spread across the globe, from 75°S to 75°N latitude, with points every 10-15 degrees of longitude. Think of it as a mesh covering the Earth's surface where we sample "can this signal be received here?"

For every origin-target pair, VOACAP calculates whether propagation exists and how good it is. That's 239 × 328 = 78,392 predictions per hour.

The Ionospheric Model

VOACAP doesn't guess. It uses the CCIR coefficients — a mathematical model of the ionosphere developed by the International Radio Consultative Committee (now ITU-R) based on decades of ionosonde measurements from stations worldwide.

These coefficients encode how the ionosphere's electron density varies with:

From these coefficients, VOACAP calculates the Maximum Usable Frequency (MUF) — the highest frequency that will be reflected back to Earth for a given path. This is the same MUF concept we discussed in an earlier post, but here it's calculated mathematically rather than inferred from real-time spots.

Antenna Assumptions: We Kept It Simple

This is where we made deliberate simplifications.

Transmit Antenna: Isotropic at 100W

For the transmitter side, we use an isotropic radiator — a theoretical antenna that radiates equally in all directions. It's not a real antenna (nothing radiates perfectly uniformly), but it's a neutral baseline that doesn't favor any direction.

We set the power to 100 watts, which is typical for an amateur station running barefoot (no amplifier). If you're running more power, your actual results will be better than what we show. If you're QRP at 5 watts, they'll be worse.

Receive Antenna: Short-Wave Whip

On the receiving end, we model a vertical whip antenna — again, omnidirectional and with modest gain. This represents a "generic" receiving station without directional antennas.

Why These Choices?

We could have modeled directional antennas like Yagis or dipoles oriented specific ways. But that would require knowing how your antenna is oriented, what height it's at, and where it's pointed. That's information we don't have.

By using omnidirectional antennas on both ends, we show you what propagation exists in principle. If you have a beam pointed at Europe, your actual performance will be better than our predictions. The predictions are a floor, not a ceiling.

Frequency Bands: All 10 Amateur HF Bands

We run VOACAP predictions for 10 frequencies corresponding to the amateur HF bands:

Band Frequency
160m 1.8 MHz
80m 3.5 MHz
40m 7.0 MHz
30m 10.1 MHz
20m 14.0 MHz
17m 18.068 MHz
15m 21.0 MHz
12m 24.89 MHz
10m 28.0 MHz
6m 50.0 MHz

For each origin-target pair, VOACAP calculates whether each frequency can propagate. We then determine the optimal band based on which frequency gives the best combination of signal strength and reliability.

What VOACAP Actually Calculates

For each prediction path, VOACAP outputs several metrics. Here's what they mean:

MUF (Maximum Usable Frequency)

The highest frequency that will be reflected by the ionosphere for this path. If you're transmitting on a frequency above the MUF, your signal goes into space instead of back to Earth.

MUF varies dramatically — from as low as 5 MHz during solar minimum nights to over 40 MHz during solar maximum days. It's calculated independently of your chosen frequency using the ionospheric physics model.

SNR (Signal-to-Noise Ratio)

How many decibels your signal exceeds the noise floor at the receiver. A positive SNR means your signal is audible; negative means it's buried in noise.

For digital modes like FT8, you can decode signals down to about -20 dB SNR. For SSB voice, you typically need +10 dB or more. We calculate SNR for each band and show you which one has the best performance.

Reliability

The probability (0-100%) that the circuit will work at the required SNR threshold. This accounts for the statistical variability in ionospheric conditions — even with the same solar activity, propagation varies from day to day.

A 70% reliability means that on approximately 21 out of 30 days in a given month, under similar solar conditions, the circuit should achieve the required SNR threshold. We filter out paths with less than 3% reliability since they're essentially unusable — a path that works only one day per month isn't worth showing.

The Scoring Formula

To turn SNR and reliability into a single "how good is this path?" number, we combine them:

Score = (Reliability × 0.7) + (Normalized SNR × 0.3)

Reliability gets 70% weight because it's the primary indicator of whether the path works at all. SNR gets 30% weight to distinguish between "barely works" and "works great."

This score drives the visual display: higher scores mean thicker arcs and brighter colors.

Noise Environment

We assume a residential noise environment — not quiet rural, not noisy urban. If you're operating from a mountaintop with no power lines for miles, your actual noise floor will be lower and results will be better. If you're in an apartment with LED lights and switching power supplies everywhere, it might be worse.

This is another simplification. VOACAP can model different noise environments, but we'd need to know your specific QTH. Using a middle-ground assumption gives reasonable results for most operators.

Pre-Calculation: How We Make It Fast

Running 78,392 VOACAP predictions takes about 2-3 minutes even with parallel processing. That's way too slow for a real-time web interface.

So we pre-calculate everything.

We generate predictions for:

That's 28 million predictions in a batch run, stored as CSV files. When you load the VOACAP view:

  1. We check the current UTC hour
  2. We get the current Sunspot Number
  3. We round SSN to the nearest 5 and load the matching pre-calculated file
  4. We find the origin closest to your grid square
  5. We display the predictions for that origin

The entire process takes milliseconds because all the heavy computation was done in advance.

What the View Shows You

When you enter your grid and select VOACAP:

Hover over any arc to see the detailed metrics: SNR, reliability, MUF, and score.

Limitations: What This Isn't

Let's be clear about what VOACAP predictions can and cannot tell you.

It's a Model, Not Reality

These are theoretical predictions based on average ionospheric behavior. The actual ionosphere doesn't read the CCIR coefficients. Some days will be better than predicted, some will be worse.

No Sporadic E or Aurora

VOACAP models F-layer propagation (the high, stable layer). It doesn't predict sporadic E openings on 6m or 10m, nor does it model auroral propagation. Those are real but unpredictable phenomena.

Fixed Geometry

We model short-path (great circle < 180°) propagation only. Long-path propagation exists but is less common and harder to predict reliably.

Omnidirectional Antennas

If you have a directional antenna, your actual performance will vary significantly from our predictions depending on where the antenna is pointed.

Noise Assumptions

Your local noise environment might be quieter or noisier than our assumed residential level.

How to Use It

The VOACAP view is best for:

  1. Planning: Before an activation, check which bands might work to different parts of the world
  2. Understanding MUF: See how the maximum usable frequency varies by direction and time
  3. Comparing bands: Understand why 20m might work to Europe but not Japan at a given time
  4. Learning propagation: Watch how predictions change through the day as the ionosphere evolves

It's not a replacement for real-time spots (use our Summary or Realtime views for that), but it gives you a theoretical foundation for understanding why certain paths work at certain times.

Final Thoughts

VOACAP is a powerful tool with a steep learning curve. We've done the work to flatten that curve — choosing sensible defaults, pre-calculating millions of predictions, and presenting the results visually.

Is it perfect? No. The ionosphere doesn't follow equations exactly, and your antenna isn't isotropic. But it gives you a physics-based understanding of propagation that complements the empirical "what's working right now" data from spots and reports.

When you see an arc from your grid to Japan with a 75% score on 20m, you're seeing decades of ionospheric science distilled into a single prediction. The math says it should work. Whether it actually does? That's what makes ham radio interesting.

73 de AK6FP Rodrigo https://dxlook.com