Question

I am measuring transient events and need to have a cable length of 500 feet (152 meters) from my signal conditioner to my data acquisition system. Is it possible to obtain meaningful data under the above conditions?

Answer

We often get this question and the answer is a qualified yes. A test was conducted to determine the effects on the output of the Model 136 when driving long cables. The test was performed using the following equipment:

• Function Generator


• Two Channel Oscilloscope


• True RMS Voltmeter


• Capacitance Decade Box


• 50 feet (15 meters) of RG-58 Coaxial cable


• Endevco Model 136 Bridge Amplifier

RG-58 type cable has a capacitance of 30 pF per foot (92 pF/meter), so a length of RG-58 cable was used for the 50 foot length. For lengths greater than 50 feet, a capacitor decade box was used to simulate the additional lengths. Measurements were made to a simulated length of 500 feet (152 meters). Figure 1 shows the relationship between cable length and total capacitance:



Figure 1: Showing the relationship between cable capacitance and length based on 30 pF per foot (92 pF/meter)

In order to establish a baseline and also determine the frequency response of the 136 with a 50 foot and a 500 foot cable, a test was conducted using a sine wave input. While the model 136 has a maximum specified flat frequency response to 50 kHz (with low-pass filter disabled). A flat frequency response to 125 kHz was realized, on the unit tested, with 50 feet of cable attached to the 136. The cable was un-terminated.



Figure 2: Showing frequency vs. amplitude for a 50 foot cable (blue trace) and a 500 foot cable (red trace). X axis is in Hz and Y axis is in mV.

The frequency response information also provides information as to distortion when analyzing transient events. Transient events usually contain high frequency information thus the limited bandwidth can be seen when observing a square wave as shown in figure 3.



Figure 3: Showing the effect that a 200 foot cable (6000 pF) has on an input signal of 15 kHz. Note the rounding of the waveform.

The user can expect to see some overshoot often followed by ringing if the rise time is < 2.5 µ sec. See figure 4.



Figure 4: Illustrating overshoot encountered when the rise time is < 2.5 microseconds. The overshoot will be present regardless of frequency.

Repeating the same test with a 147 Ω resistor in parallel with the output mitigates the overshoot/ringing (see figure 5 below) with no deformation of the waveform. The resistor should be installed between the center conductor and shield at the input of the data acquisition system.



Figure 5: Same as figure 4 with a 147 Ω terminating resistor.

This information is provided only as a guide and does not represent equipment specifications. These data were the result of a test of one Endevco model 136 chosen at random. It is expected, while not confirmed, that similar results can be expected from the Endevco model 133. While a typical test will probably be a transient event, the square wave is an effective test of rise time and the effects of high frequency bandwidth limitations. (repetition rate).