6288030.pdf

(409 KB) Pobierz
25df1.qxd
BY AUTHOR’S NAME TITLE
abbreviations go by on a screen.
Logic analyzers traditionally fit into
large desktop enclosures, but several man-
ufacturers now offer small, “pocket-sized”
logic analyzers that rely on a host PC for
display, control, and storage capabilities.
We tested seven of these instruments and
report our results for you here. Tests in-
volved the use of a standard microcon-
troller-development kit and simple pro-
grams written in C.
There are several areas of concern to
keep in mind while evaluating these or
similar instruments in your lab.
Connections: Several analyzers come
with small grabber clips. Connecting to
eight signals on a row of pins spaced at
0.10-in. intervals provides a challenge.
Other instruments provide a choice of
female contacts that push onto 0.025-in. 2
pins.
Triggers: All logic analyzers in this
review provide triggering, because that’s
what logic analysis is all about: You want
to see what happens when an event occurs.
Unfortunately, some analyzers provide
triggers that act on only logic levels and,
in one case, could set up a trigger based on
only four signals. These days, analyzers
need level and edge triggering, and some
engineers need sequential triggers, such as
“trigger if Condition A follows Condition
B.” Know your triggering requirements, so
that a purchase doesn’t disappoint you.
Cursors: These movable marks that
locate signal features prove troublesome
in many displays. Software should give
you complete control over cursors and
should show you where the cursors are,
how to control them, and the time mark
they represent in a display. The ability to
calculate the time between sets of cursors
makes them particularly valuable. A dis-
play should not “lose” cursors so that you
have to go and look for them.
And last, you cannot have a good logic
analyzer without solid documentation.
BY JON TITUS CONTRIBUTING EDITOR
HANDS-ON REVIEW:
ANALYZERS REVEAL
DIGITAL SECRETS
L
ogic analyzers capture digital signals and display state
information, timing information, or both. When test-
ing or troubleshooting electronic equipment, engi-
neers use these instruments to extract microcontroller
timing information, monitor sequences of digital
events, watch instructions go to and from a micro-
processor, and so on. Wherever you need to monitor sequences of
bits and bytes, you’ll find that a logic analyzer fills the bill. In
addition to watching the usual flow of signals on parallel lines,
some logic analyzers can decode I 2 C, SPI, CAN (controller-area-
network), and other serial protocols. And, if you plan to debug
an embedded computer, you also can decode instructions so that
you can watch the flow of a program, step by step, as op-code
MORE AT EDN.COM
For details on our testing tools and
procedures, see the sidebar “How we
ran the tests” at www.edn.com/
051205df1. At this link, you’ll also find
the code we used for tests 1 and 2,
which you will read about in the follow-
ing pages, as well as a table summariz-
ing each of the seven logic analyzers.
DECEMBER 5, 2005 | EDN 67
PC-BASED LOGIC ANALYZERS CAN REVEAL MUCH ABOUT DIGITAL
SIGNALS. HERE’S THE LOWDOWN ON SEVEN MODELS WE TESTED.
+
659318725.002.png
HANDS-ON REVIEW
BitScope BS-50 Pocket Analyzer
Price: $345
www.bitscope.com
or for a sequence of events that trigger
acquisition of signals.
We ran the Test 1 code on the Rabbit
board and set triggering for a logic-one on
D0 and a logic-zero on D1. The display
seems upside-down, with D7 at the top
and D0 at the bottom. The software pro-
vides no way to change the color or to
name the traces.
To acquire data, users click “repeat” to
continuously acquire and display data
whenever the analyzer “sees” the trigger
condition and is not busy acquiring data.
A click on a trace, though, initiates a one-
shot capture. The timebase controls let
users select a timebase of 25 nsec to 200
This small module provides logic-analyz-
er and DSO (digital-storage-oscillo-
scope) functions, and it includes a single-
channel AWG (arbitrary waveform gen-
erator). Setup went smoothly, although
engineers may find it easier to work with
a 26-pin header than with individual
wires. The color-coded wires that accom-
pany the analyzer provide a grabber on
one end and a 0.025-in. female connec-
tor on the other. The grabbers crowd the
I/O pins on our test board, but they fit
snuggly and held in place during testing.
The BitScope DSO 1.2 software starts
easily, but it takes a moment to realize
that you have to click the power button
to “start” the instrument. The software’s
initial screen looks like a DSO, but a click
on the logic button on the right-hand
menu produces the eight logic signals the
module can acquire. When the software
changes to logic-analyzer mode, the
upper-left trigger area displays each logic
signal and an associated button that lets
you select a logic-zero, logic-one, or don’t-
care condition for triggering. The soft-
ware does not provide for edge triggering
. The
display shows a time-delay value, the
timebase value, and the sample rate. We
chose a 1-
play shows the location of the cursor rel-
ative to your current view of the acquired
data. And you can’t click within a display
of data to retrieve the x cursor from a dis-
tant or forgotten position; you must hunt
for it.
We performed the BIOS interrupt-tim-
ing measurements on the Pocket Ana-
lyzer’s data and calculated an average
period of 488
sec. EDN
zoom
factor, and the data generated by Test 1
appeared in the display area. Although
the display triggers properly, we could not
relate the TD value to anything we had
set or could control.
The display provides an x-axis cursor
and a display of the x cursor’s time rela-
tive to the trigger. But good luck if you
can’t remember where you placed the cur-
sor the last time you used it. No subdis-
sec timebase and a 1
Bottom line: Before you use this
analyzer, the manufacturer should
upgrade its software. The inability to
easily move a timing cursor and the
lack of signal-name assignments
and color codes make the instrument
difficult to use. Because the manual
concentrates on DSO measure-
ments, logic-analyzer operations get
scant coverage.
Cwav USBee ZX Test Pod
Price: $895
www.usbee.com
The display window shows all eight
color-coded signals, which users can
name as they like. Four columns near the
signal names let you select sequential-trig-
ger conditions based on logic-one, logic-
zero, or don’t-care states. If you do not set
a trigger condition, the module starts to
acquire data as soon as you click on the
“acquire” button in the acquisition area.
The trigger levels prove easy to manipu-
late and the module easily acquires data
from the Test 1 and Test 2 program runs.
The display area includes two cursors,
X and O, that you can move freely
through the data. Three windows show
the time position of the cursors, as well
as the time difference between X and O.
You can set X, O, or T (the trigger con-
This module provides eight digital in-
puts, a ground line, an external trigger,
and clock lines. The built-in wires come
with female adapters that slide onto
0.025-in. 2 pins. The unit includes micro-
clips that can grab component leads.
Sampling of digital signals takes place at
nine rates from 1M to 24M samples/sec.
Because the module uses a PC’s internal
memory for storage, acquisitions can run
from 1M to 810M samples, depending
on the PC memory available. Users can
add modules to acquire data from more
than eight signals at a time. Each mod-
ule provides both 5 and 3.3V outputs.
dition) as the reference for time meas-
urements. Measuring the times between
the expected glitch in the data during
Test 1 provides an average time of 485
68 EDN | DECEMBER 5, 2005
sec and a zoom factor of 10 to 20
sec from eight consecutive measure-
ments. The cursors and the display con-
659318725.003.png
trols are easy to use. An additional meas-
urement, Insta-measure, lets you select
width, frequency, period, or byte. Select
one of these measurements, and it appears
next to the cursor as you move it among
the waveforms.
Because the module comes with what
looks like an external trigger input, we
decided to use it to trigger the logic ana-
lyzer. Although the USBee ZX software
includes help files that exist as video seg-
ments, the triggering options that the
video segment describes do not match the
latest version of the logic-analyzer soft-
ware. A printed manual would help.
How hard can it be to configure a sim-
ple trigger input line? We set up condi-
tions to try and clicked on the “acquire”
button. A blue “waiting-for-trigger” indi-
cator flashed for longer than appropriate
for the signal we had selected as the trig-
ger source. So, we clicked on the “stop”
button to abort the run. That action sent
the software into the “not-responding”
condition, and, in this state, the Windows
task manager could not shut down the
program. To double-check the operation,
we tried similar external trigger actions,
and all locked up the program. If you
leave it long enough, the program even-
tually shuts down.
Tim Harvey at USBee at first re-
marked that we had uncovered an un-
known bug in the software, which
should otherwise let users gracefully
“back out” of a condition in which the
software waits for a trigger and a clock
signal. He later explained that we prob-
ably needed to upgrade our USB-port
drivers because early drivers had a prob-
lem that prevented certain timing oper-
ations from working properly. EDN
Bottom line: The logic-analyzer func-
tions work well, and the ability to cas-
cade USBee ZX modules in parallel
makes the logic-analyzer functions
useful. This unit will appeal to design-
ers who develop small embedded sys-
tems and who need basic logic-analyz-
er functions.
Dynon Instruments Elab-080
Price: $495
www.dynoninstruments.com
logic-analyzer functions, we ran Test 1 so
that we had some data to capture. Click-
ing a green arrow button in the control
window started an acquisition, but we
could see no changes in the data, and, no
matter how we adjusted the sample rate,
the display timing remained fixed.
Off to one corner, the control window
displays a group of two buttons: (DSO/
LA Zoom). Yes, the label is in parenthe-
ses. The buttons carry labels S and nS.
Clicking the buttons zooms the display
out and in, respectively, so we could see
more or less of the data. After zooming
appropriately, we could see the eight
logic-zero signals.
Clicking on the logic-analyzer tab in
the control window opens a subwindow
that lets you quickly label signals, desig-
nate a signal’s source, and change a sig-
nal’s displayed color. Changing a name
does not change the signal’s channel
number—a nice feature that makes it easy
to name signals and still know what
points they come from on a circuit. A
“buses” tab lets users group signals and dis-
play bus information in octal, binary, dec-
imal, or hex format.
A “trigger” tab lets you select an op-
tional trigger, but trigger choices appear
limited. You can select to trigger only on
patterns that involved channels 00 to
03. The trigger-setting window accepts
any binary pattern except for 00002.
The unit’s help file explains that a hard-
scope probes. We confined
investigations to the logic-analyzer por-
tion of the unit and its software. The
package includes the main instrument
box, a large plug-in power supply, and
connection leads. The leads connect to
a header that furnishes eight signal leads
and a ground for each. Labels CH1, CH2,
and so on mark the leads so that you need
not rely solely on the color-coded wires
for identification at your system under
test.
Unfortunately, the package lacks a CD-
ROM, manual, and quick-start guide.
Instead, a slip of paper asks users to visit
the Dynon Web site to download the lat-
est software. We downloaded the software
onto a USB memory stick and moved it
to an isolated lab PC. The setup went
well, and the installed software displayed
windows for the logic-analyzer wave-
forms, oscilloscope signals, and spectrum
display.
Controls make it easy to set a sample
rate (1 kHz to 100 MHz in one, two, or
five steps) and a sample size (1k sample
to 32k samples in 2n steps). To test the
ware limitation causes that problem.
The analyzer may accept an external
trigger signal on a separate pin, but
nothing in the trigger window provides
a clue about using this input line. Sev-
eral trigger combinations work as you
would expect. When you remove the
check mark from the enable-trigger
option, the analyzer simply acquires data
as soon as you click on the green arrow
button.
The display provides two timing cur-
sors, A and B, that have little value.
The Elab-080 software displays numer-
ic-timing information in a separate win-
dow so that you can view state changes
as they relate to sample times (timing
analysis). Unfortunately, the display
simultaneously shows only 19 time-
stamped samples, and you cannot ex-
pand the size of this window.
The lack of a manual and a quick-start
DECEMBER 5, 2005 | EDN 69
The Dynon Instruments Elab-080 pro-
vides a 16-channel logic analyzer, a two-
channel digital oscilloscope, and a one-
channel arbitrary- waveform generator.
The package includes two 60-MHz
1
/10
659318725.004.png
HANDS-ON REVIEW
guide troubles us. A user new to logic ana-
lyzers and other PC-based instruments
needs some hand-holding that this prod-
uct doesn’t provide. The software comes
with a help file that includes some expla-
nations of controls and displays, but the
information about the timing cursors, for
example, simply explains that they exist
and that a user can move them. EDN
gering options and cursor-timing in-
formation, however, limits the use of
this instrument as a logic analyzer.
When compared with the other in-
struments, this one fails to deliver the
performance you should expect from
an instrument that costs almost $500.
Bottom line: The instrument provides
more functions than you would ex-
pect, which may make it attractive in
some lab situations. The lack of trig-
Link Instruments LA-2124-128K
Price: $800
www.linkinstruments.com
offers 14 1-2-5 steps from 5k samples/sec
to 100M samples/sec.) A right click on
the signal-name column lets you change
the designation for each signal and its
color. Trigger setup involves clicking on
“trigger” and then on “trigger word.” You
can set up a pattern of logic levels and
don’t-care conditions (logic zero or logic
one) for a group of eight signals. You can
select a trigger or trigger-false condition,
and the software lets you establish triggers
in binary, hexadecimal, ASCII, or deci-
mal formats. The trigger menu also pro-
vides a threshold setting that ranges from
Link Instruments provides a family of
logic analyzers that operate with a PC
over either a parallel-port or a USB-port
connection. We chose the least-expen-
sive model, which uses the parallel port.
Users who have another device con-
nected through the parallel port may opt
for a USB-compatible model with a faster
sampling rate and more channels but at
higher cost. The software setup goes
quickly until Windows asks you to install
a device driver. The company’s technical-
support person claims that the driver
exists within the installed software direc-
tory (C:/la2124), not on the supplied CD-
ROM. He also suggests downloading Ver-
sion 1.34 of the software.
Hardware setup goes easily, too, al-
though we dislike pushing individual wires
onto small pins. We’d prefer a cable con-
nected to a single header. The analyzer
comes with 30 color-coded E-Z-Hook
(www.e-z-hook.com) clips, though, so you
can attach the clips to the signal pins and
the clips to the analyzer’s wires. This
model accommodates as many as 24 logic
inputs, and it provides for an external
clock and trigger signal.
After connecting eight wires plus
ground, we ran the LA-2124 software and
set up Test 1 on the Rabbit test board. A
click on the displayed “go” button caus-
es the unit to acquire data in single, nor-
mal, or automatic mode. In single or nor-
mal mode, the analyzer waits for a trigger
and then acquires data. Normal operation
acquires a new buffer’s worth of data after
each trigger. Automatic mode acquires
data regardless of the trigger settings.
After adjusting the sample rate with a
plus and a minus button, we could see use-
ful test results. (The sample-rate control
2.80V. The default is 1.4V.
Set the relative size of the pretrigger and
post-trigger record lengths by dragging the
red trigger cursor along a time line that
represents the data-buffer’s depth. A small
box, or timing window, on this line rep-
resents the amount of information in the
display window below it. Cursors A
through D let you mark positions in the
data, and you can use cursors A and B to
measure the time between two points or
between each cursor and the trigger point.
The BIOS interrupt period is 488
a processor so you can examine instruc-
tions and data as they flow over buses in
a system. By adding extra columns of data
in different formats, you can examine
state information as binary and hexadec-
imal codes.
The software neither “collapses” a
group of signals into a bus nor displays a
hex, decimal, or other code for bused sig-
nals. But the state window provides some
of those capabilities. The state window
and the timing window do not link to
each other, although you can coordinate
cursors between them. You can insert a
“spacer”—essentially, a colored bar—
between groups of signals to separate
them from other signals. EDN
sec.
The cursors are easy to use and helpful.
You can track a cursor by its color, and a
check mark in a small display also lets you
know which cursor you have under con-
trol of the mouse. Cursor movements
involve “grabbing” one with the mouse,
or you can select one in the upper-left cur-
sor box and use six buttons to move it left
or right in time increments. The cursors
also appear along the time line, so you
always know their locations. Menu
choices let you quickly move the cursors
into the display area. The display pro-
duces the state data as hexadecimal, bina-
ry, decimal, or ASCII characters. You can
assign your own values to the display and
can include a translation table. The trans-
lation table lets you establish op-codes for
Bottom line: There’s a lot to like
about this instrument. The built-in
ability to handle 24 inputs as well as
an external clock and trigger makes
this instrument useful for embed-
ded-system designers. The ability to
add modules makes it easy to ex-
pand digital inputs. The menus are
generally intuitive, although we had
to ask tech support how to remove a
spacer. If the analyzer lacks any-
thing, it’s the ability to set up edges
as triggers and to establish sequen-
tial trigger conditions. We give this
unit high marks, and, in this case,
the lack of a printed manual causes
no problems.
70 EDN | DECEMBER 5, 2005
1.15V to
659318725.005.png
HANDS-ON REVIEW
NCI GoLogic-72 System
Price: from $4300
www.nci-usa.com
more than basic triggering, plan to spend
time learning about the various triggering
modes by experimenting with them. To
start acquisition of data for Test 1, we
wanted to trigger on a falling edge on sig-
nal A0. Clicking on the setup tab lets you
select a simple edge trigger, but we could-
n’t figure out how to select the A0 signal.
After we gave up and switched back to
the waveforms display, we noticed small,
gray squares associated with each signal.
Clicking on these squares lets you select
without going back to the setup menu.
The logic-analyzer software also pro-
vides a sample-by-sample numeric display
of information. So, the software lets you
monitor data value on a bus rather than
waveforms. You can still use the cursors to
mark values and determine the time
between them. If you have a waveform
and a numeric display open simultane-
ously, moving around in one display also
moves your view in the other. Thus, you
can easily track logic levels and numeric
data. The logic analyzer detects the anom-
aly that the Rabbit board’s BIOS causes
and measures the time over 10 occur-
rences over a period of 489
The GoLogic-72 System provides for as
many as 72 digital inputs and eight clock
inputs, which makes it a powerful tool for
testing and debugging embedded systems
that require monitoring of several buses,
I/O ports, and control signals. The unit
comes with all 72 connections, split into
two groups. Each group of eight labeled
connections passes through a small harness
that makes it easy to handle them all
simultaneously. This feature simplifies
setup when attaching connections to a
board. The wires each furnish a female
connector that pushes onto 0.025-in. 2
pins, which is how we used the logic ana-
lyzer. NCI also provides micro-grippers ($4
each) and nano-grippers ($7 each) to con-
nect with small and tiny component leads.
The GoLogic system comes with a CD-
ROM, and it takes little time to install the
software and get it running. The module
connects to a supplied power cube and to
a PC through a USB cable. The compre-
hensive manual provides basic informa-
tion on how the analyzer works and then
goes into details about sampling, clock-
ing, state analysis, timing analysis, and
triggering. Unfortunately, the manual
looks formidable, and no quick-start tuto-
rial exists to take you on a tour through
the many options and menus.
Lacking such a tutorial, we plunged in
and connected the lines labeled A0 to A7
plus two ground lines to the test board. A
setup menu lets you select a sample rate,
memory depth, and the type of sampling.
Nine options for sampling include I 2 C
and SPI modes for troubleshooting seri-
al communications in embedded systems.
Next, you select the channel groups,
which show the wire color codes and indi-
cate signal activity on lines A0 to A7. The
threshold for the signals is 1.58V, but you
can reset it for ECL signals or adjust it from
sec. The dis-
play also shows two “glitches” on the A7
and A6 lines where the other signals
change state. The period of these glitch-
es always equals the sampling period: 8
nsec at 125M samples/sec and 20 nsec at
50 MHz. So, it seems unlikely that the
glitches existed; they do not expand and
shrink in sync with the sampling clock.
But glitches might arise from simultane-
ously switching so many loads (the LEDs).
Do the glitches you observe with the
logic analyzer but not with the scope
exist, or does the logic analyzer “see”
something that doesn’t exist? You’d need
a storage scope to tell. EDN
a rising, falling, rising-or-falling, or
unused triggering attribute. You can scan
through the trigger settings you used pre-
viously, which makes it easy to keep sev-
eral settings at hand for complex tests.
Displays that include at least one chan-
nel (signal) include three pairs of cursors
you can move with the left or the right
mouse button to measure times between
events. Luckily, you need not drag a cur-
sor from place to place. Clicking on an
area of the waveform display moves the
selected cursor to the clicked-on position.
The cursor notations need some enhance-
ments, though. It is difficult to tell which
set of cursors is active and which mouse
button controls a given cursor.
Moving on to the binary-count test
(Test 2) lets you set up a pattern trigger for
FF (hexadecimal) and set the duration of
the trigger for at least 8
sec, which put it
about in the middle of a 1-LSB level. That
trigger setup works, and the display pro-
vides the expected data. In this test, the
gray trigger squares indicate logic levels
rather than transitions. You can click on
these squares to change trigger conditions
Bottom line: Although priced much
higher than any other analyzer, the
wealth of triggering options provides
a strong justification for buying this
unit. The display options and the cur-
sors make it a solid analytical tool.
Plan to spend time learning about the
triggering modes and options.
Unfortunately, lack of time prevented
us from exercising many capabilities,
such as searching data for patterns
and triggering a storage scope and
simultaneously displaying analog and
digital data. The analyzer links with
disassemblers and displays C/C
source code so you can observe op-
code-by-op-code operation of a pro-
gram. Software add-ins that provide
these functions cost extra.
5.27V for other logic families.
We downloaded Test 1 to the Rabbit
test board to make information available
to the analyzer. If you need to perform
72 EDN | DECEMBER 5, 2005
4.90 to
659318725.001.png
Zgłoś jeśli naruszono regulamin