|
Enhanced
Pulse Measurement and Generation
Using VXI Based Instruments
Abstract
This
paper describes the problems associated with large
rack-and-stack ATE systems in relation to pulse timing
performance. An overview of the limited applications
that can be met with more traditional measurement
approaches will be reviewed together with how merging
instrumentation technologies, such as VXI can be applied
to meet both today's and future needs.
Introduction
Automatic
Test Equipment (ATE) of one form or another has been
around for over twenty years. Early systems utilized
BCD programmable instruments and had to make do with
a great deal of human intervention. The establishment
of the IEEE-STD-488 brought about a revolution in
ATE design. System integrators no longer needed to
be instrumentation specialists. This could now be
left to the dedicated instrumentation companies. Off-the-shelf
instruments were now available with a standard communications
control interface. The systems integrator could at
last get down to the real problems of writing software
and configuring the system.

General Hardware Interfacing
Unit
Under Test (UUT) and signal interfacing became one
of the biggest headaches for the systems integrator.
UUT interfacing is the means of connecting between
the UUT and the ATE system. Signal interfacing is
the routing of signals between instruments (both measurement
and stimulus). The interfacing to the computer controller
was resolved with the advent of IEEE-STD-488.

Hardware Instrument Interfacing
What
are the problems of instrument to instrument interfacing?
The most serious is that of timing between instruments,
we can call this "the orchestration of the instrument
suite".
Consider a large ATE for testing avionics such as
a radar system. A typical configuration is shown
in Figure 1. Notice that the pulse and delay generators
are as close to each other as possible, and close
the UUT interface so that in theory synchronization
pulses have minimum lengths to travel and minimum
differential lengths. On closer examination, the critical
pulse generation and measurement instrument cable
lengths present significant problems. Figure
2 shows how the cable is routed between instruments
in a test rack. As can be seen, it is not just a simple
matter of short interconnection between the two units.
Cable retractors and rack to rack wiring add many
feet of cable.

Unwanted Timing Delays
Why
are the lengths of these cables so important? To synchronize
pulse generators and arm counter/timers, coax cables
are used to connect them to each other. Unfortunately,
it takes a finite time for a pulse to travel down
a cable.
In free space an electromagnetic (E.M.)wave travels
at:
In
free space an electromagnetic wave will travel 12 inches
in 1 nanosecond.
Coaxial cable, however, slows down the E.M. wave,
nominally by a factor of 0.66. (Velocity factor for
solid dielectric coax cable.) Now, instead of taking
1ns to travel 12 inches, it will take approximately
1.5ns.
The time delay then, when connecting two instruments,
even mounted close together, say with 7 feet of cable,
can add up to 10 nanoseconds. This is unacceptable
in modern radar and EW environments. In medium to
long range radars, where the radar pulse is between
1 and 10 microseconds long, and ranges are in excess
of several hundred feet, these instrument to instrument
delays have never been considered a problem.
But today, busy airfield ground movement radars need
to resolve to feet and automatic aircraft carrier
landing systems need control to tens of feet.
Figure 2. Test Rack Instrument to Instrument
Cabling

Application
Figure
3a shows the ideal case for a radar simulator. Pulse
Generator A is triggered and a sync pulse is fed to
Generator B for delay of simulated return pulse. In
the radar system the delay between the transmit pulse
and receive echo is primarily a function of range.
In the ideal case, this range can be programmed into
pulse Generator B. As discussed earlier, there is
now a delay to be considered between the trigger out
from Generator A and the input into Generator B, due
to the propagation delay along the interconnecting
cable. In the case described earlier, this will introduce
an additional delay of 10ns. On a medium to long range
radar, this extra delay would only introduce, at the
most, a 0.1% error. However, for radars requiring
tens of nanosecond delays, the cable delay introduces
a 100% error.
Accurate timing measurements of these pulses is also
a problem as cable lengths to inputs A and B on a
timer/counter must be matched.
All the above errors can be removed through calibration
and software correction, however, this creates more
software overhead and increases the test time.
Figure 3a
Figure
3b-Radar Pulse Simulation

The Solution
An
environment that allows the critical pulse generation
and measurement instruments to sit far closer together
is bound to reduce the time delay problems. Instruments
on a card and the VXIbus concept take care of this
problem in more than one way.
The VXI chassis allows for twelve instruments, plus
a Slot 0 controller, to have inter-module spacing
of only 1.2 inches. Pulse generators and counter/timers
arranged in adjacent slots can now realize transmission
time delays of <0.25ns. Even when mounting two pulse
generators at extremes of the cage and placing the
counter equidistance from both, the worst delay would
be 2ns.

Conclusion
The
emergence of VXI and instruments on a card, with increased
performance to the older rack-and-stack instruments,
means that high performance systems can now be developed
easily.
With this increased performance comes the added bonus
of reduced size test sets. Figure 4 depicts an example
of a VXIbus version of the ATE system in Figure 1.
Figure
4 - VXI Test System
VXI-1
Counter-Timer: 1 slot
DVM: 1 slot
Function Generator: 1 slot
Programmable Digitizer: 2 slots
Calibration Standard: 6 slots
VXI-2
Pulse Generator: 1 slot
Pulse Delay Generator: 1 slot
Display Stimulus Generator : 2 slots
Ratio Transformer: 2 slots
Custom Switching: 2 slots
ICA Panel
VXI-3
Synchro Resolver: 3 cards
Computer and I/O Controller

TOC
1 | 2
| 3 | 4
| 5 | 6
| 7 | 8
| 9 | 10
| 11 | 12
| 13
|