TECHNICAL NOTES
- RT1
|
TECHNICAL NOTES
- RT1
|
1) General Description
There are 3 sensitivity ranges supported by the recorder:
Range 0: 10 mV/D <= S < 0.1 V/D
Range 1: 0.1 V/D <= S < 1.0 V/D
Range 2: 1.0 V/D <= S <= 10.0 V/D
where "D" stands for units of distance, and can be either "cm" or "inches".
Ranges 1 and 2 are sufficient for most common applications; range 0 is
intended for high precision A/D conversion of low amplitude analog signals.
There are some aspects regarding the use of this range 0 that require careful
interpretation. This document attempts to clarify these points.
When recording in real time mode, the Analog Input Module (AIM) performs
A/D conversions with a gain of 1 if the channel sensitivity is in either
Range 1 or Range 2. Channel settings with sensitivities in either of these
ranges will be referred to as being in R1/2.
The A/D conversions are performed with a gain of 4 on channels with sensitivity <100 mV/D (Range 0). This improves the precision of the A/D conversions by a factor of 4 for low amplitude signals, in the range from -2.5 to +2.5 V. Channel settings with sensitivities in this range will be referred to as being in R0.
Notice that the recorder will automatically take into consideration the gain of 4 and compensate for it when plotting the traces corresponding to signals in R0. There are however some effects of this difference between R0 and R1/2 that must be kept in mind:
a) Offsets and setpoints.
A direct result of the difference between R0 and R1/2 is that the effective
range of values selected for setpoints and offsets which is normally ±10
V, is reduced to ±2.5 V for channels in R0.
b) Data transmitted via serial port.
The data transmitted in real time via the serial port will reflect
the extra precision for channels in R0. If the data is transmitted in ASCII,
the binary value is converted to a voltage with 1 lsb (least significant
bit) = 5/65536 = 0.00008 V. If binary mode is used, the data is transmitted
directly in complementary 2's complement format. In this last case, in
interpreting the data the receiver must use a scaling such that 1 lsb =
5/65536 instead of the scaling of 1 lsb = 20/65536 which is normally used
for channels in R1/2.
c) Source other than AIM.
This factor also has to be taken into consideration when a channel
is assigned to a source other than the AIM. In particular, for Digital
channels the host sending data to the recorder has to take the sensitivity
range into consideration when encoding the data. For channels in R1/2 the
data has to be encoded in a 16-bit 2's complement format such that 1 lsb
(least significant bit) is equivalent to (20/65536) V (i.e., full scale
from -10 to +10 V). Channels in R0 are encoded similarly but such that
1 lsb is equivalent to (5/65536) V (i.e., full scale from -2.5 to +2.5
V). Notice that the host has access to the information on the sensitivity
range of the channels since it can explicitly select it with the appropriate
command. However, if the host setting is overwritten from the control panel,
going from R0 to R1/2 (or from R1/2 to R0) would have the effect of scaling
the trace on the chart by a factor of 4 (or by a factor of 1/4). (Notice
that the same effect would be seen on the voltage transmitted via the serial
port in ASCII form).
In the case of data coming from the 20-digit BCD module or the ARINC module, the data is converted into the 16-bit 2's complement format used internally, scaled such that 1 lsb corresponds to (20/65536) V (i.e., full scale from -10 to +10 V). In order to have the correct amplitude on the resulting traces, the sensitivity settings have to be in R1/2 for these channels.
A situation similar to the ones described above applies when dealing with digitized channels. Since there are some special cases to consider we analyze them in detail in the next section.
2) Effect on Transient Digitizer Channels
The sensitivity selected for the logical channels used to either plot in real time a signal being digitized (digitizing mode), or to replot a digitized signal from memory (report mode), should be limited to the range from 100 mV/D to 10 V/D (i.e., R1/2).
When recording data in digitizing mode the AIM uses a gain of 1 regardless of the sensitivity selected for the channel, therefore there is no improvement in precision for small amplitude signals when the sensitivity selected is in R0. Furthermore, if the data is plotted by the recorder on a logical channel with such a sensitivity, either in real time or out of memory, the recorder would assume that scaling down by a factor of 4 is necessary and thus the resulting trace would have the wrong amplitude.
Notice that in order to support sensitivities in R0 it would be necessary to make the conversion with the appropriate gain (1 or 4), and also to tag the data somehow to indicate whether the gain used in the A/D conversion was 1 or 4. This complicates matters significantly in that the plotting routines would have to take this tag into consideration instead of the sensitivity of the logical channel used, and this would also have to be somehow taken into consideration by an external instrument receiving the data via the serial interface.
When a second AIM is used, the data plotted in real time originates from this 2nd AIM and therefore A/D conversions for channels with sensitivities in R0 are done with gain of 4. In this case the real time traces will have the correct amplitude, but the same limitations explained above apply with respect to the data plotted or transmitted from memory.
The following summarizes the effects of using a sensitivity in R0 on channels being digitized:
a) If a sensitivity in R0 is used with a single AIM in the system:
|
|
|
|
|
|
|