Logic Analyzer

[DynamoBen]

I'm looking for a reasonably priced logic analyzer. I'm familiar with the Open Logic Sniffe (http://www.sump.org/projects/analyzer/)r, and the Saleae Logic USB analyzer (http://www.saleae.com/logic/), but was wondering if there are others I should consider.

Opinions/thoughts?

[ctmal]

If you have a pickit 2 or 3 there's a logic analyzer in there.

[DynamoBen]

If you have a pickit 2 or 3 there's a logic analyzer in there.

I have a PicKit2 also but was looking for something with a few more channels and higher resolution.

[ags0000]

I clearly don't have the impact on this board that you do, Ben (not that I've earned it, I'm just sayin'). I posted this thread yesterday: http://doityourselfchristmas.com/forums/showthread.php?14922-Which-Logic-Analyzer-would-you-buy
I can still hear the crickets in the background with all the responses it's gotten.

I'd add to the list of those I'm considering the USBee SX, similar in price and features to the Saleae Logic but with a digital signal generator and internal buffer: http://www.usbee.com/sx.html

It seems the Intronix Logicport is the most expensive at about $400, but is the fastest at 500MHz and has 34 channels. Personally, I'm torn as I may never need such capabilities for my projects. On the other hand, I have never (really, never ever) puchased a cheap tool "just for this one use" and not ended up regretting it later on when it failed or wasn't extensible for other uses.

This might be just my own personal bias, but assuming the basic features/functions supported by the hardware meet minimum requirements, I tend to look for the most robust, full-featured, intuitive and powerful software. The ZeroPlus LogicCube 16032 is also in the same price range, but I've read many complaints about the quality of the software (and translation to English). Also, I note that you'll get 4 standard protocol analyzers, plus two of your choice with the hardware purchase, but then any additional would be $70 each or $100 for three.

[LabRat]

If you have a pickit 2 or 3 there's a logic analyzer in there.

Have they updated the PK3 software to support add the Logic Analyzer support?? Last I looked this was only available with the PK2.

Some guys in the office here just picked up some of these (http://www.saleae.com/home/), and seemed happy (so far).

[DynamoBen]

Some guys in the office here just picked up some of these (http://www.saleae.com/home/), and seemed happy (so far).

That's one of the ones listed above...seems interesting.

[Entropy]

I'm looking for a reasonably priced logic analyzer. I'm familiar with the Open Logic Sniffe (http://www.sump.org/projects/analyzer/)r, and the Saleae Logic USB analyzer (http://www.saleae.com/logic/), but was wondering if there are others I should consider.

Opinions/thoughts?

Just one small comment - you link to the Sump project when you mention the OLS. While the OLS is a SUMP derivative, from everything I can see the original SUMP project is kind of dead, so the OLS is starting to deviate significantly from the original SUMP project, especially with the new triggering mechanisms being added by dogsbody.

I guess the biggest disadvantage of the OLS is the rather fragmented nature of its documentation. Its advantage is that its performance/cost ratio is way better than any other option on the market and it is constantly improving.

[DynamoBen]

I clearly don't have the impact on this board that you do, Ben (not that I've earned it, I'm just sayin'). I posted this thread yesterday: http://doityourselfchristmas.com/forums/showthread.php?14922-Which-Logic-Analyzer-would-you-buy
I can still hear the crickets in the background with all the responses it's gotten.

Its not about who posts, its about where. I totally forgot there was a tools section of the forum. ;)


I'd add to the list of those I'm considering the USBee SX, similar in price and features to the Saleae Logic but with a digital signal generator and internal buffer: http://www.usbee.com/sx.html

Looks "cute." I do like devices that do one thing well but I will give this some consideration.


It seems the Intronix Logicport is the most expensive at about $400, but is the fastest at 500MHz and has 34 channels. Personally, I'm torn as I may never need such capabilities for my projects. On the other hand, I have never (really, never ever) puchased a cheap tool "just for this one use" and not ended up regretting it later on when it failed or wasn't extensible for other uses.

This one is easy...too expensive.


The ZeroPlus LogicCube 16032 is also in the same price range, but I've read many complaints about the quality of the software (and translation to English). Also, I note that you'll get 4 standard protocol analyzers, plus two of your choice with the hardware purchase, but then any additional would be $70 each or $100 for three.

Thanks for the heads up on this one.

[budude]

The logic analyzer on the PICKit2 is a neat feature (I had that thing a year before I even knew it could do that!) for sure but it's sample size is very small. The OLS is pretty cool and probably more than enough for what most of us do here. There's talk of adding o-scope functionality to it as well.

[DynamoBen]

Just one small comment - you link to the Sump project when you mention the OLS. While the OLS is a SUMP derivative, from everything I can see the original SUMP project is kind of dead, so the OLS is starting to deviate significantly from the original SUMP project, especially with the new triggering mechanisms being added by dogsbody.

I guess the biggest disadvantage of the OLS is the rather fragmented nature of its documentation. Its advantage is that its performance/cost ratio is way better than any other option on the market and it is constantly improving.

The only thing that makes me nervous is it seem to be focused more on the *nix community and less on Windows.

[Entropy]

Odd, my biggest issue with it was how long it took them to release a Linux-based loader. Admittedly Ian does not like Windows, but is committed to Windows support - "I only stick with Windows because the majority of our users are on it and support is a must. I’m ready to jump ship any time, and I do make extensive use of VirtualBox VMs when I can :)"

Similarly, until Jawi took over client development, a bunch of hardcoded fugliness in the serial handling routines of the client made using it on non-Windows machines a massive pain.

In general, Mac/Linux support for the OLS has improved drastically without Windows being negatively impacted in terms of support.

[Tabor]

I own the OLS board, but also own a Zeroplus LA.
http://www.zeroplus.com.tw/logic-analyzer_en/

I tend to always grab the ZeroPLus unit when i need one.

[ags0000]

I own the OLS board, but also own a Zeroplus LA.
http://www.zeroplus.com.tw/logic-analyzer_en/

I tend to always grab the ZeroPLus unit when i need one.

Would you elaborate on why? I've heard that the software is rough. How would you compare it to others?
Have you paid the $100 for additional protocol definitions?

I'm not arguing, I'm just hoping to build on and take full advantage of the experience of others that have used any of these offerings.

Thanks.

[ags0000]

I'm still processing this topic in the background. (I typically have a dozen things I'm "processing" at once.) I've spent some time looking at the ZeroPlus LAP-C. It's available here: http://www.nkcelectronics.com/zeroplus-lap16032u-logic-ana16032.html for $120. It seems (by specs) to be superior to the Saleae Logic ($150, 8 channels, 24MHz).

You can go to the ZeroPlus site and download the software here: http://www.zeroplus.com.tw/logic-analyzer_en/technical_support_search.php?model=LAP-C%2816032%29&class1=1 which will run in demo mode (and also you can read the documentation). The docs are pretty rough IMO, but the s/w doesn't look too bad. I haven't played with any other LA s/w, so I can't offer a first-hand comparison. It looks pretty full-featured. I can't comment on how intuitive it is as I've never owned a LA so don't know what a reasonably-knowledgeable user would be looking for.

I also just saw that a Propeller-based LA is being released: http://forums.parallax.com/showthread.php?129310-Preview-of-the-Propeller-Platform-Logic-Analyzer There's been no pricing information provided yet, but someone has already commented on the OLS board being available for $45 so that's at least a valid reference point. The Prop LA is not in the same league as the LAP-C (or Saleae Logic) as it has only a sample depth of 500 w/8 channels. However, it's designed to piggyback directly on a Prop Dev board so it would be convenient for that use.

Finally, I commented ealier in this thread about the cost of protocol analyzers for the ZeroPlus. That was a big concern for me. It turns out that there is a special going on whereby if you purchase the LAP-C, you'll get 30 protocol analyzers included in the price. I'm about the pull the trigger and just get one. Seems like a pretty good deal, although not a full-featured standalone unit (which would be >$500 I suppose).

I hope this helps.

[DynamoBen]

I ended up going with the Saleae Logic, overall I like the look and feel of the software and the hardware. They are also a start up based in SFO and they are very focused on making the best product they can (read their blog). I suspect over the next year they software is going to get better by leaps and bounds, and I don't have to make any additional purchases. Finally going with the Saleae Logic I was able to apply my SparkFun free day money to the purchase, so it was $40 off the regular price. :)

[jeffl]

The Saleae looks like a very nice setup, great GUI and support but what about the speed? It looks like it will run up to 24Mhz which from what I understand will suffice most projects. With some of the newer prjojects around isn't it possible for the chips to start running at faster speeds than 24Mhz?

If the speed works the Saleae product looks like a very good value.

[DynamoBen]

The Saleae looks like a very nice setup, great GUI and support but what about the speed? It looks like it will run up to 24Mhz which from what I understand will suffice most projects. With some of the newer prjojects around isn't it possible for the chips to start running at faster speeds than 24Mhz?

The speed barrier isn't the device or the software, it's the USB buss. They are doing bulk transfers which speeds things up and increases the amount of data that can be sent, but does have its limits. Bulk transfers also have a lower priority than other data so you have to be sensitive to what is already on the USB buss.

Their response on speed:
"Logic samples each channel at up to 24M times per second. A large fraction of practical, real world applications run at less than 10MHz, and Logic is ideal for these."

I suspect this is the fastest I will be able to go with a USB attached device w/o any fancy logic in between (FPGA or other more intelligent device). Most analyzers I've seen that are faster tend to be a lot more expensive or are a bit too rough around the edges for my tastes. This seems pretty solid, made in the USA, so to me it's worth a try.

[Entropy]

That's the big difference between the Saleae and the OLS - The Logic can do continuous sampling, but only at 24 MHz.

The OLS can't do continuous sampling, only triggered sampling, but at 100 MHz in all cases and 200 MHz in certain cases (I believe the difference was internal vs. external clock) Sample depth depends on channel count - I think it's 24k in 8-channel mode. There's an RLE compression function that can greatly extend this - it was broken for quite a while but is fixed in some betas (not in official bitstream releases yet).

Jack Gassett's state analysis in his tutorial at http://www.gadgetfactory.net/blog/?p=337 is a great example of where RLE encoding would have been beneficial - long times in many states followed by brief "flurries" of activity. (Unfortunately it wasn't ready at the time he made that tutorial)

[jeffl]

How does this unit compare? I know it's not just a logic analyzer but a scope as well. I think there also more expensive so you could probably purchase a good scope and logic analyzer for the price of this unit.

http://www.bitscope.com/product/BS120/

[Entropy]

How does this unit compare? I know it's not just a logic analyzer but a scope as well. I think there also more expensive so you could probably purchase a good scope and logic analyzer for the price of this unit.

http://www.bitscope.com/product/BS120/

Yup. A Rigol DS1052E is $400 and has either 1 GS/s or 2 GS/s sampling rate (I forget which) for one-shot signals.

That only does 40 MS/sec for one-shot signals. (The 1 GS spec is when it repeatedly samples a periodic signal with slightly different clock offsets) - So its sample rate is actually LOWER than its specced analog bandwidth, which simply doesn't work. (General rule of thumb for scopes is that sample rate should be 5-10x the analog bandwidth - the Nyquist limit isn't good enough for seeing good waveform detail.)

Similarly only 40 MS/sec for the logic part - the OLS does 100.

The benefit is that the BitScope could behave as an MSO (both analog and digital at once), but it may be possible to do that by tying the trigger output of an OLS to the trigger input of a scope. Or you spend the extra cash to get the MSO option for the Rigol scope, much more capability for not much more than the price of the BitScope.

[jeffl]

I know it's more money but what about this one?

http://www.pctestinstruments.com/index.htm

[ags0000]

I ended up going with the Saleae Logic, overall I like the look and feel of the software and the hardware. They are also a start up based in SFO and they are very focused on making the best product they can (read their blog). I suspect over the next year they software is going to get better by leaps and bounds, and I don't have to make any additional purchases. Finally going with the Saleae Logic I was able to apply my SparkFun free day money to the purchase, so it was $40 off the regular price. :)

Ben, have you received this yet and had time to play with it? I'm curious as to the ease-of-use of the software (I loaded the ZeroPlus LAP-C s/w and coudn't do much just through intuition. Then I (tried to) read the manual and it got worse. Note that I've not used an LA like this before).

I'm wondering how hard it would be to hack (write) a custom protocol analyzer for the Saleae that understood the proprietary GE ColorEffects single-wire protocol. It sure would be nicer to look at Start & Stop, and Bulb/Intensity/Color values than counting pulses. Of course, if the bit-banging code is bad, chances are the edges are in the wrong place anyway, and the protocol analyzer wouldn't be able to help much. Just thinking, as the GE lights are my personal favorite and the target of my current obsession (as you well know)...

[Entropy]

I know it's more money but what about this one?

http://www.pctestinstruments.com/index.htm

Someone here does happen to use it - I think jstjohnz? That analyzer was mentioned in one of the recent logic analyzer threads (maybe even this one?)

It depends on if you need the significantly higher samplerate or not.

[DynamoBen]

Sorry for the delay my email notifications from this forum are a bit sporadic.


Ben, have you received this yet and had time to play with it? I'm curious as to the ease-of-use of the software (I loaded the ZeroPlus LAP-C s/w and coudn't do much just through intuition. Then I (tried to) read the manual and it got worse. Note that I've not used an LA like this before).

Still on back-order, should see it by end of month. I can fill you in more then.


I'm wondering how hard it would be to hack (write) a custom protocol analyzer for the Saleae that understood the proprietary GE ColorEffects single-wire protocol. It sure would be nicer to look at Start & Stop, and Bulb/Intensity/Color values than counting pulses. Of course, if the bit-banging code is bad, chances are the edges are in the wrong place anyway, and the protocol analyzer wouldn't be able to help much. Just thinking, as the GE lights are my personal favorite and the target of my current obsession (as you well know)...

The difficulty of writing an analyzer really depends on your experience. They provide a C++ SDK and this is what they say about how long it would take:

"...implementation of a custom protocol will probably take an experienced c++ developer at least a full day, and possibly up to a week, to complete. You will probably want to have had considerable programming experience, probably significant C++ experience – before taking it on.

That said, you can get away with implementing only a small part of the full capability of the analyzer. For example, you could have only one setting (the input channel(s) to use), you could skip creating simulated data to test against; and not bother providing for data export or tabular display."

[ags0000]

The difficulty of writing an analyzer really depends on your experience. They provide a C++ SDK and this is what they say about how long it would take:

"...implementation of a custom protocol will probably take an experienced c++ developer at least a full day, and possibly up to a week, to complete. You will probably want to have had considerable programming experience, probably significant C++ experience – before taking it on..."

OK, so I'll wait for you to have the Logic in hand for a day or two - and then you can share your opinions on the unit as well as the new protocol(s) you will have developed by then :-)

[DynamoBen]

OK, so I'll wait for you to have the Logic in hand for a day or two - and then you can share your opinions on the unit as well as the new protocol(s) you will have developed by then :-)

Funny! Too bad I don't write in C++, I will however request that they create a wizard for this sort of stuff.

[griffixdc]

Sorry for the delay my email notifications from this forum are a bit sporadic.



Still on back-order, should see it by end of month. I can fill you in more then.



The difficulty of writing an analyzer really depends on your experience. They provide a C++ SDK and this is what they say about how long it would take:

"...implementation of a custom protocol will probably take an experienced c++ developer at least a full day, and possibly up to a week, to complete. You will probably want to have had considerable programming experience, probably significant C++ experience – before taking it on.

That said, you can get away with implementing only a small part of the full capability of the analyzer. For example, you could have only one setting (the input channel(s) to use), you could skip creating simulated data to test against; and not bother providing for data export or tabular display."

I would have to agree to the post 2 posts below me....makes me curious as to your occupation, your very knowledgeable when it comes to programming, i swear you can write a complete program/firmware before i understand a program :)

[DynamoBen]

I would have to agree to the post 2 posts below me....makes me curious as to your occupation, your very knowledgeable when it comes to programming, i swear you can write a complete program/firmware before i understand a program :)

It's off topic but: I'm a software tester during the day, and I do electronics as a hobby at night. When I first started I wrote bad slow code and learned by altering other peoples code. I've since gotten better, and now write most of my code from scratch...I guess practice makes perfect (or better in this case). ;) My coding skills started in high school with QBASIC, then moved to the basic stamp, then PIC (in BASIC), then Parallax SX, and now Propeller. Beyond firmware and electronics I do occasionally do scripting at work in VBScript and Python. While I don't love programming, I do like the end result.

[ags0000]

Funny! Too bad I don't write in C++, I will however request that they create a wizard for this sort of stuff.

See if the wizard can figure out what you want and just write the entire thing automatically...

I can see an opportunity for some bartering/cooperation here: I'll gladly write 10k lines of C++ if you'll write all that SPIN & PASM stuff.

Spare cog, anyone? :-)

[DynamoBen]

So I received my Saleae Logic analyzer a week ago and I have to say that I love it. It's easy to use and is technically very capable. In a matter of minutes I was able to sort out a timing issue I was having on a pixel string and this weekend I was able "see" the data I was sending to my shift registers which made development a breeze. At $140 it was worth it and over the coming months I'm sure it will pay for itself.

Thanks for all the feedback and suggestions.

[ags0000]

That's great news, congratulations. Have you started on the GE G35 protocol analyzer yet? Wait, I was supposed to do that in C++. Ok, if you'll just send the Logic out to me I'll get right on it... :-)

Seriously, I'm glad you're happy with your decision. I haven't ordered anything yet, waiting on better times.

[ags0000]

I own the OLS board, but also own a Zeroplus LA.
http://www.zeroplus.com.tw/logic-analyzer_en/

I tend to always grab the ZeroPLus unit when i need one.

I'm still thinking of going with the ZeroPlus (in the next few months) - they currently have a "30 free protocols" special going on. I think they might be releasing some new hardware (or their sales have slumped and while giving away the protocols eats away at margins their cost (CoGS) is almost nothing, even amortizing R&D).

I've downloaded their software and played with it a bit. I did not find it intuitive - but I've not used a PC/USB-based LA before. The most often voiced negative opinions I've heard regarding the LogicCube were that the documentation was a terrible translation into English (yes, it is pretty bad) and the software was not very polished or user-friendly. It may just be that I need to dedicate some more time to decoding the documentation and playing with the interface. Do you have anythng to add from your experiences?

Thanks,

Al