A new 16 channel version of the Saleae Logic is out. In addition to double the channels of the Saleae 8bit analyzer, the new version supports capture up to 100MHz (2 channels only) and has better buffering. We really like the Saleae 8bit analyzer and use it in the lab.
The Saleae is a streaming data logger type logic analyzer. It pushes samples to USB as fast as it can, and can take “unlimited” samples. Too many things sharing USB, or too much system activity, and the maximum capture rates will fail. It has a small buffer, but it heavily depends on USB bandwidth.
This is different than the topology of the Logic Sniffer. It stores samples to internal RAM and then dumps them when the capture is complete. The Logic Sniffer approach is limited to internal memory but can capture the maximum rate for all channels. The Saleae approach has a lot more samples, but the channels decrease with speed, and maximum speed depends on system and USB activity.
Justin writes in the comments:
I think thats the board they use to test the Logic 16. From thier blog post.
Logic16 Testing
We ultimately built a dedicated PCB for testing Logic16 – it uses a CPLD, DAC, and logic level translator to generate a number of variations of conditions which can then be examined in detail for signal integrity. It worked out well because we can use the same board to test each and every Logic16and then check out the photo caption.
The latest Logic16 tester PCB
Our fault! Sorry about the mistake.
We’ve worked on a lot of different logic analyzers, so we’re always interested on someone else’s approach. This picture of the internals of the new Saleae look a lot like the Logic Sniffer. It uses a similar LVC(H)16T245 buffer, and a FPGA. We assume it’s the same 100pin Spartan3 FPGA on the Logic Sniffer because that is the only Xilinx FPGA family that has a 100pin TQFP package.
U3 could be a SC-70 flash ROM chip because it is close to the config pins. This could also be a static configuration Spartan that doesn’t use a ROM chip. U3 could also be a regulator because we only see two power supplies, by the round inductors marked L. This chip usually needs 3: 1.2, 2.5, and 3.3volts.
Why the FPGA? The new version is much faster, the FPGA is doing some heavy lifting. We assume it’s used for triggering and buffering. The tech specs mention compression, maybe the FPGA does some form of RLE.
The 8bit Saleae uses a simple Cypress USB chip to take samples and send them to the computer. Unless there are components on the back of the PCB, this version may implement the USB connection in the FPGA instead of using a stand-alone chip. There are a number of USB cores for FPGAs available, most are expensive and closed source, we haven’t found an open source version yet. It also seems like it would be difficult to upgrade with a FPGA-only design, so maybe there’s still a Cypress USB chip on the back of the board.
Congratulations to Saleae on the new design. You can pick up the 16 channel version for $300, or the 8 channel version for $150. If anyone has one, please drop us a line, we’d like to hear more about the chips on the inside.
Thanks octal! Via the forum.
I think thats the board they use to test the Logic 16. From thier blog post.
Logic16 Testing
We ultimately built a dedicated PCB for testing Logic16 – it uses a CPLD, DAC, and logic level translator to generate a number of variations of conditions which can then be examined in detail for signal integrity. It worked out well because we can use the same board to test each and every Logic16
and then check out the photo caption.
The latest Logic16 tester PCB
You’re absolutely right, my fault for looking at the picture in the forum and ignoring it in the article.
Sorry it’s my fault Ian, I got the image from someone and didn’t checked the original post from Saleae (I was in a move when I posted it on forum).
Sorry :(
Altera has some nice 100TQFP CPLD’s
Could be that they use one of those
Problem with most of this stuff , is that it is overpriced for what it is……
pop over to http://www.pctestinstruments.com/
Get a 32 channel 500Mhz device for a little extra, well worth the money.
The Intronix LA has a very limited buffer of 34 x 2048 samples (http://www.pctestinstruments.com/logicport/specifications.htm). While it supports compression, that is still a very limited buffer. If Saleae Logic really support streaming to PC, I would like to give it a try.
The thing is, you have to look at the total package when you judge whether it is worth $150 (or $300 for the logic 16). I bought one of the 8 bit ones a while back and it has been super useful. I am an embedded developer and most of the stuff I deal with at work is under a megahertz. The sample rate limitations are ok (I was used to one of the high end Agilent models at my previous job). I got the test leads included (price the micro ezhooks if you scoff at this), the software just *works*, including the protocol analyzers, throwing in a case is good (I admit, this isn’t a big deal though). I can store _huge_ captures and export them as vcd files All in all, what I bough was _easily_ worth $150. I have shown it to other embedded developers at work and they promptly ordered their own. Are there minor quibbles, sure – a little less eye candy on the interface, the ability to store more complex triggers, doesn’t like USB ports other than the one it was installed with, etc…
Its just a well though out, nicely made, complete package – to me, thats worth a lot.
I agree that logic is worth the money. It has advantages and disadvantages, and therefore will not be everyone’s choice. As Ian says, it’s good to have more than one kind of logic analyzer.
The reason I am writing, though, is to point out that the Mac version doesn’t seem to have an affinity for the USB port it was ‘installed’ with, because OSX doesn’t have the silly USB model that Windows has. I would not blame those problems on Saleae, but would instead place the blame squarely on Microsoft.
I agree on the e-z hooks. They’re $2+ each even at 1000+. With the machined case, probes, and food and beer for the full-time developer, $150 is a bargain. I’ve used the saleae to debug lots of projects, including the Logic Sniffer logic analyzer :)
Ah, thanks for the clarification, Ian. The new logic description features the increased size of its memory, and I assumed that the USB protocol had changed from real-time to store-and-forward. But now I’m guessing that it’s still attempting real-time transmission, but that the increased memory acts as a temporary FIFO.
In other words, I thought the logic 16 was designed like the OLS, but it’s not?
No I dont think the desing is the same.
OLS stores data and send them to PC when buffer is full. Thus, OLS is limited to the buffer size, but the good thing is that it supports acquisition at full speed on all channels.
Logic8 does only serialize 8 bits of data and send them with a DMA like mechanisme (provided by the Cypress chip) to PC via USB.
Logic16 seems to do the same at 12,5M/sec. At high speed, it seems Logic16 still do the same at 100MHz, it buffers each 8 bits and send them via USB. This explains why you can do 100MHz only on 2 channels (in best cases), you can only serialize 1 byte at high speed USB (24Mhz).
The layout of the board looks like it fits inside the enclosure of this product for the iPad:
http://www.oscium.com/products/imso-104
-d
I really do not see what you’re talking about. The Saleae board is square, has a centered USB mini jack on one side and a centered 18-pin header on the other, plus a 6-pin ICSP header. The Oscium is rectangular, has an iPhone connector on one side and an off-center 5-pin header on the other side along with a BNC. Nothing about the enclosures look anything alike.
rsdio,
for me this post is more an advertisement for oscium than a simple post ;)
No, I saw an ad for the oscium yesterday and was trying to figure out what hardware it was using.
-d
It uses a Cypress PSOC as I recall, a processor with a bit of programmable logic around it.
I talked to Joe Garrison about this design, turns out it does have a FPGA :) A cheap Spartan 3 like the Logic Sniffer.