Dear all,
I am just stepping into embedded program through my final year project. As my course is a distant education based, I do all this at home. I am very much new to Logic Analyzers and Oscilloscopes. My question is while developing embedded application which is the best option Analyzers or Oscilloscopes? Can't Oscilloscope perform all the functions of Analyzers.
Please guide me through this.
Warm regards
--Himanshu
On 29 May 2006 19:30:54 -0700, "Himanshu Chauhan" wrote in comp.arch.embedded:
This is my opinion only, others may feel differently.
Scopes, even digital storage scopes, are most useful in the domain of analog signals, but that can also include analog problems such as noise, ringing, and so on, on signals that are supposed to be digital.
They have a limited number of input channels.
Logic analyzers are of much less use in troubleshooting analog problems, particularly high speed analog problems. They capture the logical state of digital signals, but they can have a wide number of input channels.
If you need to capture the contents of a 16-bit data bus and an 8-bit data bus plus a few control signals from the processor, you need a logic analyzer with lots of input channels.
-- Jack Klein Home: http://JK-Technology.Com FAQs for comp.lang.c http://c-faq.com/ comp.lang.c++ http://www.parashift.com/c++-faq-lite/alt.comp.lang.learn.c-c++ http://www.contrib.andrew.cmu.edu/~ajo/docs/FAQ-acllc.html
That depends on what sorts of things you're working on.
Most analyzers can capture far more channels than a scope. Some analyzers can group lines and print their combined values in hex or decimal. Some know about protocols like I2C, SPI, CAN, and can decode the lines appropriately. Some know about specific microprocessors and will decode bus cycles and display dissassembled instructions.
OTOH, a scope can let you look at signal rise times, the noise on a DC supply, and other things that a logic analyzer can't.
If I had to pick one or the other, a 4 channel digital storage scope is probably what I'd pick. That said, I picked both. I've got one of these
-- Grant Edwards grante@visi.com
Oscilloscopes let you look at a few signals at a time, typically up to four. Logic analyzers let you look at many signals in parallel, pretty much only bounded by your budget.
O'scopes treat all signals as analog voltages, displayed in time. Logic analyzers treat all signals as digital logic levels also displayed in time or with respect to a clock signal.
O'scopes are good for looking at analog stuff - power supply voltages, real world signals, digital signal integrity issues and such. O'scopes can be used to look at a few parallel signals at a time and can be used to figure out what is happening in a digital system as long as you have the patience. Analog scopes require a signal to repeat, or at least it is MUCH easier to see if it does. Digital storage scopes will capture one-time events and show them to you at your leisure.
Logic Analyzers are good for looking at parallel bus transactions and parallel data transfers of many bits. You can tell exactly what a processor is executing by watching the data, address, and control buses with a logic analyzer. They are useless for analog signals.
If you are working with microcontrollers that do not bring buses out for external access, a logic analyzer is of pretty limited use.
My opinion is that an O'scope should be gotten before a logic analyzer. Later, if you are working on systems with accessible buses and have the money, a logic analyzer will solve some problems that are very very dificult to solve with a scope.
I also think that a good digital multimeter should be gotten before you buy a scope.
Regards, Bob
Put simply, oscilloscopes display analog signals and have fewer inputs. Logic analysers are limited to digital signals but have a lot of inputs. But there a devices that perform both functions.
It depends on how much money you want to spend, more money equals more features.
Dan
Some o'scopes, perhaps, for a limited number of inputs and a limited duration.
If you have external memory then you have the ability to wire up the full address bus, data bus, and control lines and trace program flow and execution on, say, 32 analyzer channels. Tough to do on a scope but a logic analyzer can show you exactly what the processor is really doing.
On the other hand, many current chips with internal memory often handle this kind of inspection with a serial JTAG interface.
For watching general-purpose I/O on typical microcontrollers, a dual channel scope can be sufficient ... although some creativity may be required. It's very helpful if the o'scope can capture and store a single event, as with, for example, "real time sampling" digital scopes.
The logic analyzer really comes in handy when you need to watch a large number of channels simultaneously ("large" being more than two when you're using a dual channel scope). Many have other handy features, such as interpreting selected parallel or serial groupings into things like equivalent hex or assembly mnemonics.
-- Rich Webb Norfolk, VA
Some Oscilloscopes have Digital probe heads, that can capture Digital channels at the same time as they Analog Sample - so you can get the best of both worlds ( for a price.. )
With most embedded design being done on single-chip uC, then Analog matters more than digital; often there is little digital stuff to probe :)
-jg
My favorite in the moderate-budget arena is the HP 54645 family. You get sixteen digital channels and two analog channels, and deep deep memory.
Yes they can but the limitation is the number of channels, a analyzer is just an simplified oscilloscope that gives an on or off indication, whereas an oscilloscope can tell you the magnitude of the signal with 8 bits of resolution (you can think on an analyzer as an oscilloscope with 1 bit of resolution). An oscilloscope typically has 2 channels, the analyzer about 80 channels. You need both and you can get them on ebay for a few hundred dollars, JTAG/NEXUS is replacing the analyzer now a days for processor debugging, but you still need an analyzer for debugging any external digital logic, which seems to never go away no matter how integrated the processor is.
There are mixed signal instruments that combined both an oscilloscope and analyzer (HP 1631's for instance)
What an interesting scope. And the price seems pretty reasonable too.
Dan
My favorite as well. The very deep memory lets you capture a complete event (like an SPI transfer) and still allows you to examine it in detail. I always feel limited when using another scope: After just a few clicks, the zoom function hits the end of it's capabilities.
-- Stef (remove caps, dashes and .invalid from e-mail address to reply by mail)
Other folks pretty well outlined some of the differences between the two, so I will not add to that.
If you putter with digital stuff very much you will probably want a logic analyzer at some point. If you have the money to purchase a commercial unit, fine. If not, consider building one after you've learned more about micros. I don't necessarily mean clone a proper, CRT-&-pods logic analyzer, but rather build something that gives you what one of those would: use a micro, some RAM or SRAM, and appropriate glue logic to capture and save to memory the relative times of a signal's up/down transitions. After an event, upload the data to a work station for analysis.
I did this 3-4 years ago to aid in reverse-engineering the "chip" on Epson's inkjet ink cartridges, and I built another "logic analyzer" of much simpler design just last week to capture the data streams from several infrared remote controls (TV, DVD/VCR, CD player). Total devices required: 1 PIC uC (16F877A),
1 8k-byte SRAM (Toshiba TMM2064P-1-), 2 open-collector TTL buffer chips (SN7407), an IR receiver cannibalized from an old Panasonic VCR, a 25-pin D-shell connector (PC parallel port), and some pull-up resistors. I also had to write the PIC firmware (assembly language, using Microchip's MPLAB) as well as the upload code for my PC (MS BASIC PDS). Everything except the PIC and the SRAM came from junk boxes; the PIC was new and the SRAM chip was pulled out of my EPROM burner (temporarily). Got everything running over the long weekend and have already figured out one of my remote controls.Note: you might not even need RAM or SRAM, depending on what uC you decide to use. I understand (not through my own experience though) that some Atmel uC's have tons of user RAM. Since I use PICs, and PIC's don't have nearly as much RAM as I thought I'd need, I used external SRAM.
-- Michael
Thank you all! But I found Logic Analyzers and O'scopes are pretty expensive beasts. I will have to find other way round. May be using digital multi-meter. I am not sure. What do you think about this:
Well Michael, Ain't it a problem of Chicken or Egg. To build a logic analyzer using a uC at home, don't we need LA? Can you please throw a light on this?
Warm regards
--Himanshu
Michael wrote: [ ... ]
I once had an idle thought that I might build a logic analyzer around a FIFO RAM. Should finish that sometime, see if it works.
Mel.
Not really.
You can get a lot done with a $20 logic probe and maybe an old used scope.
If you use an proto/eval board or a well integrated SoC, there isn't that much logic analyzing to be done.
--
Grant Edwards grante Yow! ... I'm IMAGINING a
at sensuous GIRAFFE, CAVORTING
visi.com in the BACK ROOMof a KOSHER
DELI --
(16F877A),
There are some excellent and inexpensive PC-based logic analyzers available today. A do-it-yourself project is always a good learning experience (and valuable in and of itself) but harder to justify in a straight build-or-buy decision anymore.
My current favorite is the Logicport USB-interfaced analyzer from
#disclaimer: just a happy customer.
-- Rich Webb Norfolk, VA
I've heard good things about it. The SPI, async, and I2C decoding look particularly cool.
But you've got to run Windows, which is particularly uncool.
--
Grant Edwards grante Yow! I am a jelly donut. I
at am a jelly donut.
visi.com
Ya, there is that. I dual-boot the laptop that has my development and analysis "stuff" on it but haven't been able to eschew MS Windows entirely, yet.
The serial interpreters are seriously cool if you've spent much time like I have going clock edge to clock edge and hand translating SPI conversations. Some of them are loooong (like initializing ADI's DDS chips) and used to be a real PITA to read, bitwise. Now it's just read the hex right off the screen; major timesaver.
-- Rich Webb Norfolk, VA
Himanshu Chauhan wrote: (snip)
I IIt could be. Might not be. Depends on several things, e.g. how extensive/expansive your design is, how *good* your design and proto are when you power up the proto for the first time. Also how many wiring mistakes you make and what kind of mistakes they are. Stuff like that.
The "analyzer" I described last time is very simple: just 1 channel. The uC simply timestamps each high/low transition of that channel and saves to SRAM. I made three wiring errors, and finding them using an ohmmeter cost me about 4 hours. Having a proper LA would not have cut any time off that. What *would* have made things a whole lot easier is using several colors of wire when building, instead of just the YELLOW I had!
-- Michael
Haven't tried it for several years, but from what I've read, I don't think it really provides applications with access to USB peripherals. I think VMWare might be able to do that, though.
--
Grant Edwards grante Yow! If I felt any more
at SOPHISTICATED I would DIE
visi.com of EMBARRASSMENT!
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.