rbm@afinitc.UUCP (Rick Moll) (02/06/84)
We are presently looking for a logic analyzer for use in our software
engineering department. Units being considered are:
HP 1630
Dolch 64300
Nicolet NPC-764
Tektronix 1240
I would appreciate hearing from anyone who has had experience with any
of these analyzers. Please include information on what processors the unit
is used on. As well as any feedback you may have about the software
analysis features of the unit.
Rick Moll
..!ihnp4!afinitc!rbm
Affinitec
2252 Welsch Ind. Ct.
St. Louis, MO 63146carey@uiuccsb.UUCP (04/19/84)
#N:uiuccsb:4400059:000:1403 uiuccsb!carey Apr 19 13:37:00 1984 Our Computer Science department is considering purchase of a logic analyzer. What do you folks have to say about? What is your experience with particular analyzers? What applications have you used or seen them used for? What are good and bad features of particular units? What features are necessary, trivial, overkill, redundant, really super, convenient? What applications, and necessary capability, could you envision for a large Computer Science Dept., that is other than troubleshooting department equipment? What analyzers allow for expansion, either in existing capability, or more powerful interaction systems. What are people's (and/or users') feelings on value of an analyzer? That is, is this thing really worth 5K, 10K, 15K, 20K, 100K? Are certain analyzers overpriced in some areas, or is what you are getting to start with worth anything withouth investing once or twice that again? Is the modular approach good in some circumstances? And to sum it up about money we are looking for something 10K or under. Our primary usage is not development, but there are or will be people involved in hardware in the department who could use some kind of capability in that area. We are going to act soon on this so please respond quickly if you can. --carey My net address would be the same, I think as other UI dcs people. Who knows it out there?
dont@tekig1.UUCP (Don Taylor) (04/22/84)
X
Well, one thing that I think would be nice to see on a logic analyzer would
be a couple of 16, 20, 24, and 40 pin connectors on the analyzer that could be
attached to a ribbon cable. Just unplug the IC, plug the ribbon cable into the
board under test, plug the IC back in, and tell the la which pins you are really
interested in displaying. This would stop the rats nest of a dozen micro clips,
all attached to the same chip. This same idea could work with a glomper clip,
from the top, and the same method of giving ALL the pins to the la. It would
make 'well, i dunnow, maybe your right, what is the interrupt line doing?,' and
the fun process of trying to get a clip in around that little leg. Just tell
the silly machine to show me that pin! Now building that mux inside the box is
going to take work, but I think the users would love it. (let me hear votes on
that one, if you would)
One other point, watch out for the kinds of clips you get with the thing.
Ive used an HP analyzer, (no offense intended), where the clip was a small tube
with the actual connector extruded from the end. The tube was metal, with paint
to insulate it from adjacent pins. The paint quickly wore off and we kept
shorting pins together.
Don Taylor
tektronix!tekig1!dontcarey@uiucdcs.UUCP (04/23/84)
#R:uiuccsb:4400059:uiucdcs:10400135:000:207 uiucdcs!carey Apr 23 11:45:00 1984 Here is my proper address, thanks DM! John Carey Computer Research Laboratory Department of Computer Science University of Illinois at Urbana-Champaign 1304 West Springfield Avenue Urbana, Illinois 61801
carey@uiucdcs.UUCP (04/23/84)
#R:uiuccsb:4400059:uiucdcs:10400136:000:245
uiucdcs!carey Apr 23 11:50:00 1984
Really! I'm not just trying to make it look like my note is getting
a lot of response! I just forgot the net address stuff:
USENET: ...!{pur-ee,ihnp4}!uiucdcs!carey
CSNET: carey.UIUCyouknowthatfunnylittle"a"withthecirclearounditrand-relaynathan@orstcs.UUCP (04/24/84)
I've looked at most of what's available, and must say I've been impressed with the Hewlett-Packard 1630D, which runs about $10,000 - $15,000. Of course, it will be practically useless if it doesn't support whatever processor you use . . . Incidentally, Tek donated a DAS9000 box to the Oregon State University computer science department; last I heard it was being used as a paper stand. You may be able to talk HP out of a development system or two, which include all the logic analyzer stuff and a computer/compiler etc to go with it. Nathan Myers orstcs!nathan
dmmartindale@watcgl.UUCP (Dave Martindale) (04/26/84)
I think the idea of a 40-pin cable running to a prewired glomper clip (or socket that sits between the processor and the board) is a wonderful idea. Providing a hardware MUX that allows you to select between all those signals under software control would be expensive, but just buffering the signals and then feeding them to a manual crosspoint switch shouldn't cost too much. And sticking pins into a crosspoint switch is a LOT easier and less error-prone than moving tiny clips on an inaccessible PC board. The only thing is, I want more than 40 pins.
ignatz@ihuxx.UUCP (Dave Ihnat, Chicago, IL) (05/01/84)
Sitting to my left is a Tetronix DAS 9100 Series Digital Analysis System. (It sez so on the cover.) The straight hardware type I share the office with just got it a couple of weeks ago, and it looks like everything you want. It has room for something like six expansion cards (data acquisition and pattern generator modules), with each module accepting up to 4 probes, and up to 32 channels. These data acquisition probes, or pods, are, themselves, quasi-intelligent, and can be fitted with several types of connectors. Sorry, the ones I've seen are separate clips per channel or logic line. There is the capability for extensive mapping of probe input to logical display ordering and location, and of mapping physical channel or logic line to logical line. There also exists a complex symbolic definition, labeling, and display capability. And they provide a micro-cassette with directory file maintenance to save these complex setup configurations. Overall, a beautiful little machine that I would have given my eye teeth for a couple of years ago when we were stuffing Unix on an 8086 with custom memory managment. The catch? Well, aside from the learning curve for the operator, it's cost (surprise, right?) This critter can run from $16,000 or so on up. If you're interested, however, please contact Tektronix--I've spent enough time on this. (For the legal types....) The description and information provided herein is an unsolicited submission by yours truly, and as such represents only my own opinions and experience. Statements made herein by me are in no way to be construed as reflecting the opinion or endorsement of either my employer or my contract employer. Viel spass, Dave Ihnat ihuxx!ignatz
sct@lanl-a.UUCP (05/01/84)
We use four different analyzers: HP-1615, Biomation K100D, Tektronix 9100 DAS, and the HP-1630D. The HP-1615 is old (about 5 yrs) but was the leading logic analyzer when it came out (my opinion). The K100D was one of the first 100Mhz analyzers but is best used for timing analysis (it only has 16 channels). Biomation does have newer models (k500?) that offer a lot more capability. The two I would like to compare are the 9100 and the 1630D. The 9100 is a large, highly expandable, unit. We have 64 state channels on it and 16 timing channels (100 Mhz). You can get disassemblers for it for most micros but we don't have any. These disassemblers cam be expensive and have to be added to the cost of the analyzer. We consider this unit to be our 'cadillac' analyzer. It is not very easy to set up and use especially after you haven't used it for a month or so. The set up features are somewhat clumsy, the pods use ribbon cable from the unit to the pod (yuk!) and these pods have a spring loaded clip for the ground connection which is awkard to use. The cassette unit works well for saving setup parameters and state listings but you cannot save waveforms. The 1630D has 43 state channels, 16 of which can also be used as timing channels (100 Mhz). Its cost is about $10K (much less than the 9100). It is much easer to use than the 9100 in my opinion. The screen is brighter and easier to read. The probe pods use round cables which are much easier to work with. The cassette is the micro-cassette unit that HP sells with their 41C series of calculators using the HP-IL interface. Even though it is not an integral part of the analyzer it is small enough to not be a hindrence. The cassette stores all setups, state listings and waveforms. We have the Z80 dissassembler and are very pleased with it. The cost of the disassembler hardware was $1200. Another nice aspect of the 1630D is that it is portable (lightweight) which to 9100 is not. l m
holzwort@ittral.UUCP (Paul Holzworth) (05/03/84)
I too have used a Tektronix 9100 and CANNOT recommend that anyone else
suffer through with it. On paper it really does look great but in
practice it isn't such a great machine. We constantly had problems
with it triggering on the wrong state. (You would tell it to trigger on
XYZ and it would trigger on ABC.) It is also extremely slow in writing
the information to the screen so that if you want to take multiple
traces in rapid succession it doesn't keep up.
We have recently gotten some Hewlett Packard 1630 analyzers and they are
almost all you could ask for. They are fast in responding to keystrokes.
(The Tektronix almost always fell behind when I tried to rapidly do
a number of commands.) It has never false triggered. It has a number of
nice features such as a histogram mode that shows you what percentage of
time is spent executing what part of the code. (Great for software
optimization) It will also do timing analysis like the Tek. It also
has a cassete drive available for storing setups and the newest version
(1630G) will operate with a microfloppy and contains EEPROM that
allows you to store one setup without the microfloppy present.
The Tek does have a couple of advantages in that you can get some
very high speed interface pods (with a corresponding loss of
logic channels) and it allows you to mix high and low speed
pods. It also has an optional state generator available.
Given all of the above though, I would rather pay twice as much and
get the HP than to have to use the Tek. The beauty of it is though
that they are comparable in price. The HP1630D with cassette drive
and 43 channels costs a little over $11k. I seem to remember the
9100 being similarly priced although I don't have my catalog with me.
I have no axe to grind or affiliation with either Tektronix or Hewlett-
Packard.
Paul Holzworthsct@lanl-a.UUCP (05/03/84)
The new HP1630G logic analyzer is essentially a 1630D with more state channels. I think HP is offering it because the 1630D didn't handle the 16 bit micro disassember job adequately. Thus the 1630G with 65 state channels (if I remember right) will take care of this drawback The 1630G however only has 8 timing (100Mhz) channels. The 1630D however does do a great job with the 8 bit micros. The 1630G is upgradable from the 1630D for about $3K.
rpw3@fortune.UUCP (05/05/84)
#R:uiuccsb:4400059:fortune:28000034:000:831
fortune!rpw3 May 4 21:44:00 1984
I also prefer the H-P analyzers, but the Tek 9100 DAS has one nice feature,
the pattern generator option. It will generate fairly complex streams of
16-bit words up to 25MHz (40ns/word). The patterns can have multi-level
subroutines, jumps, conditional jumps, and even (*yes*) interrupts.
I used it to generate (using only one bit of 16) Ethernet packets (to debug
a PLL), hand-crafting the Manchester coding by running the pattern generator
at 20MHz. It worked fine. (The subroutine feature came in REAL handy!)
But then again, we just happened to have one I could use. I wouldn't buy
it for that reason alone (I would kludge up a RAM and a counter).
Rob Warnock
UUCP: {ihnp4,ucbvax!amd70,hpda,harpo,sri-unix,allegra}!fortune!rpw3
DDD: (415)595-8444
USPS: Fortune Systems Corp, 101 Twin Dolphin Drive, Redwood City, CA 94065rh@glasgow.UUCP (05/24/84)
I have recently surveyed and purchased a logic analyser in the U.K. The H-P 1630D came out as best value for money (about 9000pounds including tape system and a selection of disassemblers). It is very easy to use. Links with a storage scope to capture and display a waveform during a trace. Could be used with a micro for data collection & external control via the 488 bus (we havent tried this). The 'personality' modules for interfacing to specific cpus are very expensive but could be attractive in a commercial environment - we intend to build our own. Hooking up via the spring clips is ok but not wonderful - lack of insulation on the probe body is a mystery. A big plus is the availability this summer (it had better be !) of an internal expansion board to give a total of 63 state channels - 25MHz of which 16 can be used for timing - 100MHz. Glitch detection (on the timing channels by halving the number available) is useful and anythig similar on the opposition cost a lot more. Now someone tell me that I made the wrong choice and should have bought brand X with twice the goodies for half the money. ray herman, Dept Computing Sc., Univ. of Glasgow. - !gl!rh