Simulcast planning with the TSB-88 Monte Carlo studies in EDX SignalPro

EDX has been a leader in land-mobile radio and DMR planning since it was first founded in 1985. The core SignalPro software includes a collection of studies meant to address the special challenges involved with planning simulcast networks, as well as heterogeneous networks that include both simulcast and multicast regions. These studies can be found both in the ‘basic’ study group and also in the study group called ‘TSB-88’.

The U.S. Telecommunications Industry Association (TIA) publishes and updates a large and complex document known as the TSB-88, which gives a variety of recommendations for mobile radio planning. This document provides information and advice to manufacturers and users of primarily Land Mobile and Public Safety radio systems. It provides guidance on performance criteria for different types of radio technologies and methods for measuring and predicting system performance. Section 5 of the TSB-88.1 document provides several suggested methods for predicting the CPC and the reliability of the signal over a defined area. TSB-88 is similar to the ITU Recommendations in that you are provided with various options for doing an analysis and it is up to each person to decide how they want to proceed. Further, new methods and recommendations are periodically added in updated versions of the document.  

EDX has continually maintained and added to its complement of studies over the years. The EDX TSB-88 studies section is based on the EDX’s interpretation and implementation of the recommendations in the TSB-88 document and on what the majority of our U.S. land mobile radio customers have requested over the decades that we’ve offered these features. 

In 2009 EDX released version 1 of its ‘TSB-88 Monte Carlo’ studies, designed to take fading and simulcast delay-spread into account within the same analysis. These studies were designed in collaboration with the TSB-88 committee and in accordance with the recommendations given in “TSB-88.1-C; Appendix E”, which has also gone through revisions since its first release. 

 The summer 2025 release of SignalPro further includes three new “Version 2” TSB-88 Monte Carlo studies. With all these study options available, it can be confusing to know which studies to choose and how to set them up. In this document we’ll look at the various simulcast-related studies, clarify the differences between them, and suggest some best practices for simulcast planning within EDX SignalPro.

Why specialized studies are required for simulcast networks

Simulcast networks are often used for land-mobile communications in areas where terrain shadowing makes it difficult to reliably cover the entire area with just one site. A simulcast network will have multiple base stations, often including tall macro sites and low power repeaters to fill in difficult spots, with all base stations transmitting the same signal on the same frequency. This effectively creates one large “cell” or simulcast group operating on a single radio frequency. As a mobile unit moves through the area and experiences signal fading from one base station, the signal from other base stations will effectively reinforce the desired signal to provide seamless reliable coverage. 

In order for these co-channel signals from various base stations to reinforce each other at the mobile receiver they must be reasonably aligned in the time domain. The mobile receiver will lock onto the timing of the strongest incoming signal at its location and other incoming signals will arrive with some difference in transit time from the strongest. If this difference in arrival time is too great then the delayed signal no longer reinforces coverage, but instead becomes a source of interference by distorting the desired signal. If the different incoming signals at a receiver have a large difference in arrival time but similar power levels then it can cause time-delay interference (TDI), negatively impacting intelligibility. 

This feature of simulcast networks means that special planning techniques are required to ensure reliable coverage. In general, the goal is to ensure reliable coverage even with substantial signal fading, but also ensure that the delay-spread, or difference in arrival time between the strongest and other incoming simulcast signals, is kept below the maximum allowable value in order to prevent time-delay interference. 

Definition of terms

• CPC, Channel Performance Criteria – This is the power ratio between the strongest incoming simulcast signal (the desired signal) and all sources of interference plus noise. In order to operate at a given digital audio quality (DAQ), a receiver must maintain its required CPC or higher. 
• Delay-Spread – This is the difference in arrival time at the mobile between any two signals in the same simulcast group (or channel). 
• Required Simulcast Delay – This is the time delay window for simulcast signals. Those signals with less delay-spread than this value may be treated as reinforcing the strongest signal. Those with a greater delay-spread may be treated as interference. 
• TDI, Time-Delay Interference – This is an undesirable condition where signals coming into the receiver at similar power levels (within the capture ratio) have significant delay-spread. This distortion of the signal causes a higher Bit Error Rate and the mobile’s performance is degraded. 
• Capture Ratio – The mobile receiver will only be affected by delayed signals if they are within a given power ratio of the strongest (desired) signal. If a signal’s power level is below the capture ratio then it will not cause TDI.
• RMS delay-spread – Once the strongest incoming signal has been established at the remote then the delay-spread values for the other incoming signals are averaged using an RMS weighted-average function so that the delay time of signals at a higher power level will have more impact on the reported delay time. This single delay value, given in microseconds, represents the overall delay conditions at that location.
• Maximum delay-spread – Once the strongest incoming signal is established at the remote then the most delayed signal within the capture ratio range is found and that delay time is reported. 

The challenge: Modeling the relationship between coverage fading and TDI

In earlier revisions of both the TIA TSB-88 and EDX SignalPro, the recommended planning practices for simulcast networks treated coverage fading and delay-spread as two different issues and used separate studies to account for them. Coverage studies were typically run using a very conservative propagation model which assumed some excess pathloss margin to represent a “faded” or worst-case received signal power at the remote. Then delay-spread studies would be run to show the delay-spread values for all those areas where the receiver has more than one signal within the capture ratio. If a given location showed too high a delay-spread value then that location would be considered unreliable due to the high potential for TDI at that location. This method has a limitation, though; in the real world incoming signal levels are constantly changing due to Rayleigh fading as well as the slow fading caused by the mobile moving through the environment. By using a conservative propagation model to establish coverage and then treating time-delay separately it’s possible to overlook areas of marginal reliability, where TDI only occurs under certain signal fading conditions. 

The ‘TSB-88’ area studies were developed and added to SignalPro in collaboration with the TIA TSB-88 committee as a more sophisticated method of evaluating the potential for TDI under dynamic signal fading conditions. Version 1 of EDX’s TSB-88 studies were developed in coordination with committee members and the TSB-88 document itself was later revised to include recommendations for a similar Monte Carlo approach. 

The TSB-88 Monte Carlo studies in EDX SignalPro offer a few ways of looking at the potential for TDI and finding the overall predicted reliability of downlink reception in a simulcast network. These studies run a Monte Carlo simulation where the signal levels from all relevant base stations are allowed to fade up and down over a given number of trials. In each trial the time-delay values and relative power levels are evaluated to determine whether the receiver would perform adequately and the percentage overall reliability or a related metric are produced as an area study layer. 

Note that there is also a full selection of uplink studies for land-mobile systems with different voting and diversity schemes. This document covers only the downlink studies related to simulcast. For a full description and best-practices regarding uplink studies make sure to visit the EDX knowledge base or Reference Manual Appendices.

Describing sector and mobile properties

Care should be taken to make sure the link budget and antenna details are correct for all sectors (base stations) and also for the mobile/remote that’s being used for the studies. Each sector has a value for the ‘simulcast delay offset’, which delays the effective launch time of the signal from that sector. The default value is zero and should only be increased if you want to delay that sector’s launch time to better align with other sectors. 

The mobile/remote unit(s) used in the study should also be configured with the correct antenna gain, height, etc. paying careful attention to the ‘Receiver simulcast capture ratio range’. Delayed signals within the capture ratio range can potentially cause TDI, so it’s important to establish the correct value based on the equipment, modulation, and DAQ requirement.

It’s also possible to define the performance of the receiver under different signal-to-noise and delay conditions. The Monte Carlo analysis can be run using a simple static requirement for CPC and delay-spread or it can be run using a ‘simulcast curve file’. This file gives a table of delay-spread values and the CPC ratio needed to achieve the desired performance. 

This file can be created by going to Utilities -> Create/Edit Simulcast Curve File. This file is required by the ‘Hess Simulcast Monte Carlo Reliability’ study. Other simulcast studies use a single CPC and delay-spread criteria. 

Propagation model considerations; average vs faded coverage predictions

SignalPro includes a number of propagation models that are appropriate for any land-mobile frequency band. For land-mobile planning users typically choose a path-based deterministic model such as Anderson2D, Longley-Rice, or Free Space + RMD. Whichever model is chosen, one important consideration is whether the model should be set up to target the average signal level or a “faded” signal level.

Previous versions of the TSB-88 document recommended including additional pathloss margins as part of the propagation model to account for both fast fading and seasonal fading. This was usually done using the ‘% of Location’ and ‘% of Time’ fields, which add more pathloss to the model in order to account for fading. This produces a smaller coverage footprint around the site to represent the signal level in a faded or “worst-case” condition. Since the average real-world signal level at any location should always be higher than the prediction, this method is not ideal for evaluating the potential for interference. 

The TSB-88 Monte Carlo analyses are designed to take fading into account within the study itself, so typically the propagation model setup should target the average signal level that would be measured in the field, rather than a faded signal level. This means that, when running the Monte Carlo analysis,  the ‘% of Location’ and ‘% of Time’ fields should be set to 50, meaning no additional pathloss. 

A typical propagation model configuration for VHF or UHF simulcast coverage would be the Anderson2D model with the ‘add clutter loss’ option turned on and the % of Time and Location fields set to 50%,  as seen here;

Whatever clutter database is used, it should include a clutter attenuation file which gives default clutter loss values for each clutter type and frequency band. In EDX-provided attenuation files these values are based on the TSB-88 document recommendations for general clutter losses. Any given project area may need to have those clutter loss values adjusted in order to get the most accurate prediction possible. This can be done using field measurement (drive test) data and the ‘Test Against Measurements’ function. Model tuning and validation are an important step to ensure an accurate prediction. While the default values may be acceptable in many regions (especially within the United States, where the recommendations were developed), it’s important to consider the project area carefully and decide whether clutter in the project area necessitates drive-testing and model tuning. For additional information on model tuning and validation, see the EDX Knowledge Base. 

TSB-88 Monte Carlo study setup

Once the equipment parameters and propagation model have been set up then the area studies can be added to the project and run. If using any of the area studies in the TSB-88 study group it’s important to set up the performance criteria and Monte Carlo settings by going to Studies -> TSB-88 Studies Setup.

• Service Area Boundary File: This should point to a polygon boundary around the project’s area of interest. See the EDX Knowledge base for information on creating a boundary file and exclusion areas, if needed. 
• Simulcast Delay vs. Signal file: This file gives the CPC requirement under different delay-spread conditions for the receiver. It’s required if running the ‘Hess Simulcast Monte Carlo Reliability’ study. To create this file, go to Utilities -> Create/Edit Simulcast Curve File. 
• Required Tile Reliability (%): This is the target reliability for each grid cell. The V2 ‘Bounded Area Coverage’ study will only show results for grid cells that meet all the Performance Criteria, including this reliability requirement. 
• Required Simulcast Delay: This is the delay time window used in the Monte Carlo analysis. Signals that arrive within this window relative to the strongest signal are considered to reinforce the simulcast signal, rather than interfere with it. Signals outside this window are treated as sources of interference. 
• Required CPC: This is the required Channel Performance Criteria, which is the ratio of the desired signal power over noise and interference, including simulcast time-delay interference. 
• Random Seed: This is the seed number used for the pseudo-random fading in the Monte Carlo analysis. Use the same number each time to get reproducible results. 
• Standard Deviation (dB): This is the standard deviation of fading used within the Monte Carlo analysis. The default value is 5.6 dB, per the TSB-88 document. In situations where high fading is expected (dense urban, railway scenarios, etc.) this value may be increased.  
• Samples/TX/Location: This is the number of Monte Carlo trials that will be made for each grid cell in the study area. This can be reduced to make the analysis somewhat faster, but the Monte Carlo analysis is most effective with a statistically useful number of trails, so the best practice is to set this to at least a few hundred trials.

Simulcast Study types

Included here is a brief description of each of the simulcast-related downlink area studies in SignalPro. A more complete explanation of the propagation model methods and area study method can be found in the Reference Manual Appendices. This list is meant as a quick reference to clarify some of the differences between them. 

• ‘Simulcast RMS delay spread’ (basic study): This finds the strongest signal and reports a weighted-average of the delayed arrival times of the other incoming simulcast signals within the capture ratio. The delay time is averaged so that signals of greater power contribute more to the reported value. No Monte Carlo analysis is performed within this study.
   
• ‘Maximum simulcast delay spread’ (basic study): This finds the strongest signals and then reports the worst delay time with the capture ratio. No Monte Carlo analysis is performed within this study. 
  
• ‘Number of Servers within capture ratio’ (basic study): This simply finds the number of other signals within the capture ratio of the strongest. No Monte Carlo analysis is performed within this study. 
  
• ‘Aggregate Simulcast Monte Carlo Reliability’: This original version of the study runs a Monte Carlo analysis and returns the probability of a “pass” for each grid cell using the Required CPC criteria and comparing the Required Simulcast Delay value to the RMS time delay value for that cell. The predicted reliability is reported regardless of whether it meets the Required Tile Reliability criteria.
  
• ‘Aggregate Simulcast Monte Carlo Reliability (Ver. 2)’: The updated version of this study uses an improved Monte Carlo analysis which sorts potential simulcast interferers according to whether their delay time falls within the Required Simulcast Delay window. Those with a low-enough delay time are treated as reinforcing the signal, rather than degrading it. The updated fading method in Version 2 was also updated to better align with the current revision of TIA-TSB-88.2-C section 6.10.2.
  
• ‘Hess Simulcast Monte Carlo Reliability’: This original version of the Monte Carlo analysis makes use of the Simulcast Delay vs Signal file. For each trial in the Monte Carlo analysis, the CPC and delay spread values are compared to the Signal vs. Delay file to determine whether that trial is a pass or fail. This study differs from the aggregate reliability in that it uses a full table, rather than a single requirement for CPC and delay-spread.
  
• ‘Hess Simulcast Monte Carlo Reliability (Ver. 2)’: Version 2 of this study still uses the Signal vs Delay file to define dynamic performance criteria. The updated version also uses an improved Monte Carlo analysis which sorts potential simulcast interferers according to whether their delay time falls within the Required Simulcast Delay window. Those with a low-enough delay time are treated as reinforcing the signal, rather than degrading it. The updated fading method in Version 2 was also updated to better align with the current revision of TIA-TSB-88.2-C section 6.10.2.
  
• ‘TSB-88 Monte Carlo Bounded Area Coverage’: This original version of the study runs a Monte Carlo analysis and returns the probability of a “pass” for each grid cell on the basis of required CPC. This also produced some metrics on overall area coverage; this functionality has been replaced by the Study Queries tool. Due to inconsistencies in how delay-spread is treated relative to other study methods, EDX no longer recommends using this version of the Monte Carlo analysis. The Version 2 Bounded Coverage study is recommended instead.
  
• ‘TSB-88 Monte Carlo Bounded Area Coverage (Ver. 2)’:  This updated study will calculate the reliability percentage at each grid cell and show results for only those grid cells which pass the defined Performance Criteria for percentage reliability, CPC, and delay-spread. Like the Version 2 Aggregate Reliability study, this updated version of this study uses an improved Monte Carlo analysis which sorts potential simulcast interferers according to whether their delay time falls within the Required Simulcast Delay window. Those with a low-enough delay time are treated as reinforcing the signal, rather than degrading it. The updated fading method in Version 2 is also used to better align with the current revision of TIA-TSB-88.2-C section 6.10.2. The updated fading method in Version 2 is used to better align with the current revision of TIA-TSB-88.2-C section 6.10.2. It will also subject each grid cell to the Required Tile Reliability performance criteria; those cells which fail the reliability requirement will show no result. 
  
• ‘Simulcast delay spread (Hess)’: This study reports the delay-spread in microseconds, using the “Hess” calculation method as recommended in TIA-TSB-88.2-C. This method will typically report twice the value that the basic studies ‘Simulcast RMS delay spread’ does. This is because the Hess method doubles the basic difference in arrival time to account for potential multipath. 
  
• ‘TSB-88 radius-of-operation overlap Reliability’: This is a specialized study which runs the Version 1 Monte Carlo analysis, but for only a single focus sector and single interferer. Additionally, the “bounded area” in which reliability is calculated is determined by the overlap of the service contours of the two transmitters.  The program requires the user to produce two polygon files that define the radius of operation area of the serving and interfering transmitters. These files should be placed in the \CTR folder of the Project and are named using the Transmitter/Sector ID with a “.bna” file extension.

Which Monte Carlo studies should be used?

Choosing and interpreting studies are always engineering decisions which require a complete understanding of the network’s real-world requirements and the study types available. EDX has updated the analysis, but left the previous versions of the TSB-88 Monte Carlo studies intact so that users can choose to continue with the original method for established projects or go back and reproduce previous results. Our engineering team feels that the improvements to the Monte Carlo analysis should represent more realistic and accurate simulcast performance in most situations when the proper inputs are used. 

If the information is available to populate the Simulcast Curve File with signal vs. delay values then the most complete version of the reliability prediction is the ‘Hess Simulcast Monte Carlo Reliability (Ver. 2)’. If you’d prefer to use only a single requirement for CPC and delay-spread then the ‘Aggregate Simulcast Monte Carlo Reliability (Ver. 2)’ is recommended. 

You may also wish to produce a layer showing the generalized delay spread in microsends for each location. In this case the ‘Simulcast Delay Spread (Hess)’ is recommended, particularly in environments which may exhibit high multipath, such as urban canyons or rugged terrain. 

Including interference from outside sources (neighboring networks)

An additional feature of only the Version 2 Bounded Area Coverage and Aggregate Reliability studies is that that can be configured to consider not only simulcast time-delay interference from transmitters in the same simulcast group, but also interference from other non-coordinated sources in a second group. For example, if two adjacent network operators are using the same frequency then in-network sites may generate TDI, but nearby out-of-network sites will cause traditional cochannel interference (regardless of any timing considerations). To represent this in a SignalPro project, the user can put all the coordinated simulcast sectors into one Tx Group, which is set as the Primary Tx group in the study details. Then put sectors from the non-coordinated interfering network into another Tx group and set it as the secondary Tx group in the study details. The reliability percentage calculated by the study will take that second group interference into account when calculating the CPC. This is an important feature for modelling the potential for inter-network interference and how it might degrade an otherwise reliable simulcast system.