TDR Tutorial and Riser Bond TDR Product Review
FILTERING NOISY WAVEFORMS
The Riser-Bond Instruments waveform TDRs come equipped with unique signal filters for cleaning up noisy waveforms.
It is not always possible to completely disconnect the cable under test from all other equipment. A couple of examples of this are a tower cable that goes into an antenna, or a local area network that needs to stay operational during the test.
If the cable is left attached to equipment at the far end, the energy present on the cable may present a danger to the operator o to the TDR. If so, do not proceed with the test until danger is removed. A TDR may damage the equipment on the cable under test. Riser-Bond Instruments TDRs test using a negative voltage signal of less than five volts in amplitude and with varying pulse widths. The amplitude should not damage any equipment on the cable. However, excessive pulse width may cause a digital system to adversely react to the TDR signal.
There are many different types of signals that can be present on the cable under test. They range from 50/60 Hz power, to audio frequency, to data in the 1 to 100 MHz range, to RF. No single filter will eliminate all of these signals. Riser-Bond Instruments waveform TDRs have multiple types and levels of software filters which can eliminate almost any type of problem. In a filtering application, it is necessary for the operator to step through the various filters to see which one will work best in any given application.
Because of the inherent nature of the filters, some operate so fast they do not seem to affect the speed of the instrument display, while others seem to slow the display so much that it makes display manipulation almost impossible. An example of this is filtering power line noise. A 50 Hz power line cycle takes 20 milliseconds to complete and therefore 20 milliseconds for the TDR to create one screen display dot. The update rate of a 256 dot display would then be over two seconds. This is a long time when trying to reposition a waveform. One way around these long delays is to turn on the filter that will do the most good, then simply store the filtered waveform . Waveform storage may take a long time, but the post-storage waveform manipulation will be as fast as having the filter turned off.
When testing data cables that are remaining operational, it is best to keep the TDR in a short pulse width mode, like 2, 10, or, at the very most, 100 nsec. This way, the error detection software of the data system will not see the pulse as "bad data" and continually request for "resend data" which may lock the system.
The TDR noise filters are extremely useful when testing noisy cable but, as with any equipment, experience will enhance success. Practice with different noise sources using different filters.
DETERMINING CABLE ATTENUATION
Any signal will lose some of its energy or signal strength as it propagates down the cable. This loss is frequency sensitive. As the frequency of a signal goes up, the loss becomes greater. With any given set of frequencies, some cables will have more loss, some will have less.
Over a given cable length and within a span of frequencies, the user needs to know the signal attenuation. This specification will be in the cable manufacturer's catalog.
Normally, in order to find this attenuation, the cable is tested with a sine wave signal source (or sweep generator) attached to one end of the cable and a signal receiver (or AC power meter) with terminator attached to the other end. As the signal source is scanned through the frequencies of interest, the cable can be tested for attenuation vs. frequency at a particular cable length.
A TDR transmits signals at different frequencies; that is, the different pulse widths have different fundamental frequencies. Examples of fundamental frequencies for various pulse widths are:
| Pulse Width | Approximate Fundamental Frequency |
|---|---|
| 2 nsec | 250 MHz |
| 10 nsec | 50 MHz |
| 100 nsec | 5 MHz |
| 1 usec | 500 Khz |
| 2 usec | 250 Khz |
It is possible to use a TDR to find an approximate value of cable attenuation. To use a TDR to sweep a cable, simply connect the TDR to the cable under test, set the first cursor to the transmitted pulse, and the second cursor to the reflected pulse. Note the cable length, the pulse width setting, and the return loss reading. Make sure the far end of the cable under test is not connected to a terminator or any other piece of equipment.
Since the TDR has both the signal source and the receiver located at the same end, the signal will have twice the attenuation because the signal has traveled down and back along the cable.
Therefore, if you simply divide the dBRL value by 2, you will have the value of the cable attenuation at that frequency and cable length. It is sometimes helpful to actually make a graph of your most commonly used cable. The graph will show dBRL reading vs. tested cable length with various TDR pulse widths. You can now quickly check against good cable (low loss) and bad cable (high loss).
A number above each pulse width line indicates a good cable (low loss). A number below each pulse width line indicates a bad cable (excessive loss).
Calculating Return Loss
A unique feature of Riser-Bond TDRs is their ability to calculate and display the severity of a fault; the value of dBRL. One general problem with displaying dBRL is that the user would like to know the severity of the fault at the fault. The TDR calculating the severity of the fault is located some distance away at the far end of the cable. Therefore, the dBRL reading has an error equal to the attenuation of the cable between the fault and the instrument.
The TDR could subtract out this error if and only if it knew the type of cable being tested. The Riser-Bond Model 1205C can make this correction. A feature within the instrument allows the user to select the type of cable being tested and the 1205C will give the corrected value of the fault on the instrument display.
The 1205 C can display both the total (uncorrected) dBRL value or, by entering the type of cable being tested, just the value of the fault.
STRUCTURAL RETURN LOSS
STRUCTURAL RETURN LOSS (SRL) in a cable is caused by small imperfections distributed along the length of the cable. These imperfections cause signal distortion and/or micro-reflections. Structural return loss can be caused by manufacturing flaws, installation damage, or by some other means of cable disturbance or degradation.
A TDR is generally used to locate "point" problems rather than "distributed" type problems which cause SRL. Therefore, if you are tempted to use a TDR for checking SRL, it should only be used as a quick, cursory check or evaluation of SRL and not for true or absolute SRL measurements. For true SRL measurements, a sweep generator should be used.
Structural return loss can be viewed on a TDR by looking at the base line of the waveform. A perfectly flat baseline indicates a high quality cable with no damage or structural return loss. A bumpy baseline would indicate a lower quality cable, cable damage, and/or structural return loss.
In order to analyze SRL with a TDR, some basic information must first be determined:
1. What is the cable attenuation vs. distance at different frequencies? You can find this either from the cable manufacturer's catalog or from previous application note.
2. What is the frequency content of the various pulse widths of the TDR being used? You will find this discussed in the previous application note.
3. What is an acceptable/non-acceptable level of dBRL at specific frequencies for the cable being tested? This is a value determined by your particular application.
Remember that when testing cable each pulse width has a specific fundamental frequency, and, as stated above, cable attenuation is frequency sensitive. Therefore, identical readings at two different distances can indicate a different severity of cable problem. Cable attenuation, signal frequency, and distance to the fault must all be taken into consideration when analyzing dBRL.
It is helpful to draw a graph of the cable loss vs. the distance at each frequency (or pulse width). The graph will indicate how the acceptable SRL number will increase with distance.
REMEMBER: When measuring decibels of return loss (dBRL), the larger the dBRL number, the smaller the problem. The smaller the dBRL number, the more severe the problem.
A number below each pulse width line indicates a good cable (low SRL). A number above each pulse width line indicates a bad cable (excessive SRL).
To evaluate the SRL of a cable using Riser-Bond Instruments' Model 1220/1205/1205T, first select the shortest pulse width (2 nsec) and position the first cursor on the leading edge of the transmitted pulse. To scan the cable (on screen) in short lengths, zoom-in and increase the vertical gain. Move the second cursor along the baseline of the waveform, noting the distances at which the dBRL number drops below the acceptable value determined by your graph. When the maximum distance at one pulse width has been reached, switch to the next larger pulse width. You might want to inspect the cable at any point where an excessive dBRL reading is indicated.
|
To be contacted for a confidential consultation please E-mail: jmatk@tscm.com
or send a letter via US Mail to:
or call:
URL: http://www.tscm.com/ |