Logic Analyzer Sample Rate Needed?

[beaker353]

I'm looking at purchasing a  Logic or Logic16 from Saleae as a logic analyzed and decoder. One big advantage of the Logic16 over the basic Logic model is increasing the same rate to accurately sample higher by limiting the number of simultaneously monitors channels.  My question to the minds better than my own is what protocols am
I likely to run into in a hobbyist/microcontroller situation that would be above a 6MHz clock rate the basic model can handle. I understand the long term advantage to quality equipment.  I purchased a Fluke 189 shortly after it was released and even though it has been sent in for recal every year, they have never needed to adjust. The Logic goes for $150, and the Logic16 goes for $300, so it's not a huge amount of cash difference, but I'm curious to know if I will ever need that much speed. Opinions???

- EM

[NiHaoMike]

If you like to DIY, get a FPGA board and make your own.

[kripton2035]

I ran into the same questions some times ago
and then i discovered the logic sniffer : http://dangerousprototypes.com/docs/Open_Bench_Logic_Sniffer
I bought it ant it is largely enough for my hobby needs.
and it's so cheap ...

[slateraptor]

If you like to DIY, get a FPGA board and make your own.

Unfortunately, that doesn't compensate for a clean and usable GUI with decoding, which IMO is what makes Saleae's tool worth every penny.

@beaker353: For the 8-channel Logic, you'll alias sampling signals above 12MHz, and resolution isn't the greatest as you approach 12MHz either. That being said, an 8-channel unit is likely all you'll need for the typical hobbyist uC development. If you're prototyping on solderless breadboards, signal speeds will be limited to ~10MHz anyways. I personally prefer not having all the extra wires cluttering my workspace, but to each his own.


EDIT: I think it proper that I mention my personal bias towards Saleae. Why? From an objective standpoint, their design is outstanding, both hardware and especially software: form factor, cleanliness, usability, simplicity. Sure the device necessarily has limitations like sampling rate and not being able to set thresholds, but they've done everything else so well that it's easy to overlook and makes me more willing to design around limitations for the sake of troubleshooting. Secondly, their customer service is outstanding. To date, I've already gotten 3 friends to purchase a unit on their own volition by simply loaning them mine to use in troubleshooting their own projects. I can't vouch for their product enough, and will continue to so long as they maintain the same level of excellence.

[T4P]

You know , there are clones of saleaes out there . They are cheap and very close to the original model , but since it costs so little , if it breaks , WHO CARES ?!

[slateraptor]

You know , there are clones of saleaes out there . They are cheap and very close to the original model , but since it costs so little , if it breaks , WHO CARES ?!

"Who cares" seems to be the prevailing model in Asia...then again, Asia doesn't exactly bleed innovation either: they're too busy copying everyone else's design to come up with their own. =\

[T4P]

You know , there are clones of saleaes out there . They are cheap and very close to the original model , but since it costs so little , if it breaks , WHO CARES ?!

"Who cares" seems to be the prevailing model in Asia...then again, Asia doesn't exactly bleed innovation either: they're too busy copying everyone else's design to come up with their own. =\

You mean , China . It's to cheap to be bothered .
You know , i really have no other choice but to purchase stuff from the country i hate most  well except from korea .

[george graves]

You know , there are clones of saleaes out there . They are cheap and very close to the original model , but since it costs so little , if it breaks , WHO CARES ?!

That's great?  But what software are you going to use it with?

[T4P]

You know , there are clones of saleaes out there . They are cheap and very close to the original model , but since it costs so little , if it breaks , WHO CARES ?!

That's great?  But what software are you going to use it with?

Saleae's software ?
http://www.aliexpress.com/product-fm/522626571-Saleae-24MHz-8Channels-Logic-Analyzer-Fully-Checked-Best-quality-Input-buffered--wholesalers.html

[amyk]

You can also use the Saleae/clones with open-source software: http://sigrok.org/wiki/Main_Page

You know , there are clones of saleaes out there . They are cheap and very close to the original model , but since it costs so little , if it breaks , WHO CARES ?!

"Who cares" seems to be the prevailing model in Asia...then again, Asia doesn't exactly bleed innovation either: they're too busy copying everyone else's design to come up with their own. =\

The Saleae isn't exactly innovative either... it's essentially the Cypress reference design with all the GPIOs brought out and the firmware does nothing more than read them and send the data over USB. I've heard it referred to as the Mac of logic analyzers, and I think that's a rather good comparison. They should be selling logic analyzer software, not hardware.

Also you might want to look up the "shanzhai" phenomenon, that's where the Chinese innovation is. They do have different views on doing it, more about taking an existing design and improving it than trying to make something completely new.

Now, I think the hardware in something like this is more worth the cost: https://www.eevblog.com/forum/product-reviews-photos-and-discussion/intronix-logicport-34-channel-logic-analyzer-teardown/

...and on the other hand you can get much more value with e.g. something like this:
http://www.seeedstudio.com/depot/preorder-open-workbench-logic-sniffer-p-612.html?cPath=75

[alm]

How's the Sigrok software coming along? Last time I checked it was not even close to usable, but that was a while ago.

The Saleae isn't exactly innovative either... it's essentially the Cypress reference design with all the GPIOs brought out and the firmware does nothing more than read them and send the data over USB. I've heard it referred to as the Mac of logic analyzers, and I think that's a rather good comparison. They should be selling logic analyzer software, not hardware.

Weren't they the first with the Cypress dumb logic analyzer? I believe they were there before Usbee, but I'm not sure about any others. It's quite a neat hack in my opinion. Yes, most of the cost is probably in software, although the nice case and probes also add to the cost. Just selling the software would probably be a support nightmare, however, it would effectively require them to troubleshoot issues with any random board with a Cypress USB SoC on it.

[amyk]

I think it's usable now, but like most free software doesn't have the "polish" of commercial offerings.

USBee and Saleae both came out in 2005, the Cypress chip came out ~2001. There's evidence on the Internet that the Chinese had designed a logic analyzer around it in 2002, and 8051-based (using Atmel's chips) USB-hosted ones seemed the norm before then.

[0xdeadbeef]

To get back to the topic, I think you sometimes need 100MHz to measure some delays between clock/chip selects signals etc. For protocol decoding of 1MBaud, usually 10MHz are more than enough.

Anyway, I'd like to mention the Zeroplus "LogicCube" family.
I have both, an Open Logic Sniffer and a Zeroplus LAP-C(16032) which I patched to 128k (just ID in serial EEPROM).
Both are worth their money, but of course you have to live with some restrictions this way or the other.

The OLS is really neat for the price and I use it at work indeed sometimes when I need to analyze SPI buses. It's very portable as you don't have to install drivers and software. Then again, it has somewhat limited memory, the run length encoding leaves a bit to be desired, the GUI is neat to use, but still very basic and partly buggy: E.g. it still doesn't support bus grouping, the advanced (demon core) trigger options are still not accessible, the times displayed in 200MHz mode are still wrong, the pre-trigger time setup is more or less completely ignored. Also the lack of a case, the missing pull devices and input protection are a bit problematic for industrial use. Last but not least there are only a few protocol decoder available (the most important ones are there though).

The  LAP-C(16032) is much more sturdy and the software is pretty good, although some things could be improved. I can even display analog data from my Owon SDS in parallel (though this is currently limited to 1k sampling depth). The ability to patch the low end 32k version to the middle range 128k version and the availability of >30 free protocols are the most interesting points. The major drawbacks are the very limited trigger possibilities and the download/installation process. You need to download every protocol separately though some are already included in the normal installations. Download takes forever and each protocol wastes a lot of disk space due to very large Chinese PDFs.
Also the design of the protocol decoders is a bit clumsy. E.g. to decode SDI and SDO, you have to double the CS and CLK signals, then group CS/CLK/SDI and CS/CLK/SDO to separate buses and apply the protocol decoder to both buses separately as this is possible with other LAs.