- What is the cause of messages Attenuator Error or Maximum Drive Exceeded?
- How should the sigma clipping level be set?
- What do the different status panel colors indicate?
- What does the Specialized Channel Processing Option apply to?
The attenuator is used to adjust the random spectrum power level (or the random background level in Sine on Random). When the random control methods call for more power, such as in increasing the test level, the attenuator is often used to allow more voltage (range) to be output from the Dac. Once the attenuator has reached the 0 dB setting (either at max drive voltage or at the maximum output available) then the user will see the "Maximum Drive Exceeded" message. In the Sine on Random application, the user may alternately see an "Attenuator Error" message. While these alarms are generally not terminal to the test, too many of these messages may indicate that the test control accuracy is being compromised. To increase the available power level, the user may try lowering the sigma clipping level value (if the test specification allows). The available power is approximately the max drive volts (from the safety limits parameters, normally set to 12 volts) divided by the sigma clipping level. Too low of a clipping value ( i.e. less than about 2 or so) will generally cause severe random output signal distortion which causes problems in maintaining control of the random spectrum during a test. Too high a value (i.e. greater than 5-6) can cause low power output availability and "spiking" of the output random signal. Finally, note that an Attenuator Error may also be caused by a true hardware error condition. This case is normally followed by an immediate test shutdown with a related "System Error" message as well.
In the Sine on Random application, there are two competing utilizations for the Dac output voltage levels. The first is the random background spectrum, the second is the added sweeping sine tones. Both of these components are summed together and then output through the Dac. Ordinarily, the sigma clipping level is set to maximize the drive power level. With the addition of (possibly) multiple sine tones, some additional "headroom" is required to allow for the sine tones to be added to the output buffer without being clipped due to the sigma scaling operation of the background random signal. Generally, this "headroom" is provided by using a higher sigma clip level than the random-only test case. In order to assist the user in setting the sigma clip level, the additional parameters for Tone Clip Level (%) and Tone Abort Level (%) should be utilized. The tone clip levels are defined as the number of output buffer data points that have been truncated to the maximum allowable output level (this number is normalized to the number of total points available). For example, a clip level of 10% indicates that for a 400 line test, over 102 data points (out of 1024) had to be clipped during the output process. Severely clipped output buffers generally lead to problems in control of the background random spectrum due to generation of distortion from the clipped sine tones. Normally, the user should try to set the clipping alarm to about 1% and observe the frequency of the clipping level messages. If the messages come too often (with each or every other background loop update) or if the background spectrum has problems with control due to a large dynamic range requirement, then the user should increase the sigma clipping level. Increasing the sigma clip level allows for more dynamic headroom which reduces the tone clipping probability. One should also note that the higher frequency sine tones (i.e. more peaks per buffer) generally also increase the chance of tone clipping so the user should be aware of the clipping requirements during the entire test. A situation to be aware of is when all sine tone sweeps tend to end together at the higher frequencies. The user should also be aware that setting the sigma clipping level too high may bring on the problem of "Maximum Drive" as mentioned in an earlier FAQ therefore the setting of this parameter is a balance between the sine tone test requirements as well as the background control power level. See Sigma Clipping in a related FAQ for more information.
The color of the status panels indicate if a channel is on/off in the random, sine on random, and random on random applications.
|Grey||Channel is inactive or not available.|
|Green||Channel is active.|
Located in Puma under > Setup > Controls > Control Setup tab > Tone Extraction option
The Sine on Random vibration control program augments the usual broadband random control spectrum by adding a set of user defined tones to the broadband test specification that are allowed to sweep within defined limits. The tonal amplitudes are controlled independent of the broadband spectrum. The channel processing has the capability to extract individual tone data from the measured composite broadband spectra. The channel processing is always applied to the control channels, but it may be optionally applied to ALL active channels as described below.
Tone Extraction Options
1) The default (Control Chan) tone extraction processing is applied to the control channel in order to separate the 2 control components of tone and broadband. The algorithm relies on knowledge of the control spectrum, reference spectrum and tone frequency and sweep rate in order to perform this task. The control channel is special in that the reference spectrum provides information as to what level of broadband is to be found at the tone frequency of interest. For this option, measurement channels are not processed for tone data. Measurement channel data will appear to have a composite spectrum, as one would expect when using a standard analyzer type of measurement. There are no tone amplitude displays available for measurement channels in this option because tone data has not been extracted from the measurement spectra.
2) The other (All Channels) tone extraction method applies to the measurement channels. Measurement channel tone extraction is enabled with this option. Control channels are processed the same as with the default tone extraction method. Measurement channels, however, differ from control channels in that there is no reference spectrum available to differentiate the residual broadband component from the extracted tone level component. The only option is to extract all tone information from the broadband measurement. As an example, for a fixed tone on broadband, the broadband spectrum will appear to have a drop out (i.e. very little energy) at the tone frequency being extracted. The removed energy can be displayed in the tone amplitude display. Because the tone is a discrete frequency (if fixed), the value may not line up exactly on a broadband spectral line. The resulting tone removal in the broadband display may appear to be spread across several spectral lines (analogous to spectral leakage effects).