1. Why Do You Need to Measure Both BER and MER on Digital Signals
Presented by
UNAOHM
is a trademark of START S.p.a

2. Introduction
Most digital analysers measure Bit Error Ratio (BER) and Modulation Error Ratio (MER).
BER and MER each have their limitations.
This seminar explains why it is important to measure both BER and MER and tells you what types of impairments will be missed if you only measure one or the other.
Viewing of the Constellation Made Easy Seminar and the Modulation Error Ratio Made Easy Seminar is recommended prior to viewing this seminar.

3. Modulation Error Ratio
RMS error magnitude
average symbol magnitude
MER is defined as follows:
MER is expressed in dB.
10 log
RMS error magnitude
Ideal symbol
Average symbol magnitude

4. Constellation with “good” MER Constellation with “poor” MER
MER effectively assigns a value to the fuzziness of the symbol cluster.
The larger or fuzzier the cluster, the poorer the MER.
The further from the ideal locations, the poorer the MER.
Constellation with “good” MER
Constellation with “poor” MER

5. Correct locations fall Locations in error fall outside
How Errors Occur 1 3 5 7 -1 -3 -5 -7
Each symbol on the constellation is
framed by decision boundaries.
When the signal falls inside the
decision boundaries the information
is transmitted error-
free.
Correct locations fall
within decision
boundaries
Locations in error fall outside
decision boundaries

6. Comparison Between Three Error Free Constellations
All constellations below have a perfect BER with no errors, because the cluster always falls within the decision boundaries.
The constellations to the right, however, have significantly better MER with less noise.
When the cluster falls within the decision boundaries, BER is not an effective measurement of quality because the BER is perfect. 1 3 5 7 -1 -3 -5 -7 1 3 5 7 -1 -3 -5 -7 1 3 5 7 -1 -3 -5 -7
Poor MER Perfect BER
Good MER Perfect BER
Best MER Perfect BER

7. Note there are no errors in this range of MER.
Theoretical MER vs BER With Only Gaussian Noise Impairing a 64 QAM Signal
Note there are no errors in this range of MER.
BER
No errors
MER
In practice errors will tend to occur at higher MERs due to other forms of impairments besides Gaussian Noise.

8. Why Measure BER?
Since MER can quantify signal quality when no errors exist the question can be raised, why measure BER at all if MER will do?
The major limitation of MER is the inability of the measurement to capture fast intermittent transients.
A signal can have a very good MER, but poor BER due to intermittent interference.

9. Examples of Intermittent Interference That Cause Poor BER But Good MER
Loose Connections
Corroded or loose connections.
Sweep System Interference
Sweep pulses from a sweep system set up to sweep empty spectrum.
Laser Clipping
Occasional overload of the laser due to analogue sync pulses lining up.
Microphonics
Vibration of digital origination equipment can cause intermittent errors.

10. Troubleshooting By Measuring Both MER and BER
One way to determine if you have intermittent problems is to measure both MER and BER.
If the MER is high, but you still see errors, then the errors are probably caused by an intermittent problem.

11. Intermittent Errors on a Constellation Display
Intermittent errors will show up on a constellation display as lone dots away from the main cluster.

12. BER vs MER vs C/R Q&A Q. What is MER?
A. MER is a measurement of the modulation impairment that affects the ability of a digital receiver to recover data bits.
It accounts for modulation problems such as non linearity, group-delay and flatness variations, filter mismatching, and ingress.
Q. When is MER useful considering?
A. When considering digital video system margin or SNR issues.
Q. What is the difference between MER and C/N?
A. A good C/N accounts for the effect of noise only while MER accounts for modulation problems.

13. BER vs MER vs C/R Q&A
Q. Can we calculate MER by measuring C/N (for instance using a spectrum analyser) and then applying the relationship between C/N and MER(SNR)?
A. No, because C/N accounts for noise ONLY while MER accounts for ALL kinds of signal impairments —noise, signal leakage, IQ level and quadrature imbalance etc —.
Q. What minimum value of MER is measurable?
A. To all practical purposes, down to 18 dB for DVB-C systems, and down to 5dB for DVB-S systems. This is due to the fact that a test equipment must be able to demodulate the signal before MER measurement. If the signal is so poor that this process fails, the measurement result cannot be trusted.

14. BER vs MER vs C/R Q&A
Q. What are we going to use for faultfinding when MER is below 18 dB (DVB-C) or 5 dB (DVB-S)?
A. At this point C/N (measured as channel power/noise power) can be used instead of no (MER) measurement. Noise only will be accounted for this way, but it is better than nothing.

15. BER vs MER vs C/R Q&A Q. What is the difference between BER and MER?
A. BER reflects noise and modulation impairments severe enough to cause bit errors, remaining insensitive to subtle trends in the digital modulation. MER allows the installer to detect modulation impairments being concealed by the system’s equalisation and error correction (before they become errors). In other words BER starts counting when problems surface as errors. MER is useful to detect impairments underneath the surface that is, while impairments are still within the reach of the correcting capabilities of the system.
Q. What does a good BER mean?
A. It indicated proper service delivery.
Q. What does a bad BER mean?
It highlights impaired service, but does not identify the cause of the problem.

16. BER vs MER vs C/R Comparison Table

17. Digital signal analyzer with COFDM/QAM/QPSK testing
Conclusions
In order to see the effects of all types of impairments on a QAM signal you need to measure both BER and MER.
MER can quantify the quality of a digital signal that does not have any errors.
MER has the limitation of not being able to see intermittent errors so a signal can have a good MER but a poor BER (Bit Error Ratio).
The most common type of intermittent errors are caused by laser clipping.
UNAOHM EP3000
Digital signal analyzer with COFDM/QAM/QPSK testing