Page 1 of 1
USB based logic analyzers
Posted: Wed Apr 19, 2006 2:32 pm
I am considering buying a usb base logic analyzer. There are quite a few ranging in price from several hundred to several thousand dollars
The ones I am considering are: Bitscope BS-50, Techtools Digiview, Intronix LogicPort and Rockylogic Ant 16. all under $500.
I think I've narrowed it down to 2 products: Digiview or Logicport. The logicport has the better specs except for small trace buffer depth (2048 plus some hardware compression - digiview has 128K). The digiview doesn't analyze serial protocols (logicport does - spi, I2c and async). Logic port has more channels and faster sampling (up to 500 mhz). The logic port is $110 cheaper. The digiview seems like a very well thought out instrument and I have seen several write ups by users that like it. Logicport has no independent reviews.
So, I guess it boils down to whether I can live with the 2K trace buffer depth. The other question is how well does the protocol analysis work on the logicport.
There are several other products that are well above 500 but I'm not willing to spend that much, yet.
Anyone have any thoughts on this?
Posted: Wed Apr 19, 2006 2:43 pm
Strange... I just used for the first time the TechTools Digiview. I only needed to look at 3 signals, so the number of channels was plenty for me. Once I learned how the GUI worked, I really liked the tool. I've never used any of the other ones and this was my first day using this one, so I'm no expert, but I liked it.
Posted: Thu Apr 20, 2006 5:27 am
I have the Bitscope device and one thing I like is the protocol they use allows you to write some neat custom applications communicating directly with their hardware using register commands, etc. Very clever approach. The Bitscope website has some fascinating articles on the design tact taken with all the devices.
Posted: Thu Apr 20, 2006 3:51 pm
thanks guys. I appreciate the feedback.
I ruled the bitscope out due to it only having 8 channels and sample rate up to 40 mhz. It does have nice buffer depth, though.
Dr_when and HighFreq, how long are the typical signals you look at? I know I'll need at least 1K so a 2K buffer size looks a bit cramped.
Posted: Fri Apr 21, 2006 5:52 am
The Bitscope BS50 that I have is definitely not a high end tool but is small, pocketable and a nice tool for looking at things like keyboard data streams, parallel port data, small micro port data, stuff like that. Sometimes the data is a few hundred milliseconds in length. It gets me by when I cannot lug a larger piece of equipment and the price was right but you can do better. I think Jask Ganssle did a review in Embedded Systems Journal which may be archived on their site. He is a pretty good authority on this type of equipment.
Posted: Fri Apr 21, 2006 9:58 pm
Yeah, I read that article. If I was in the market for an 8 channel LA, I'd definitely consider it.
by the way Ganssle has written a number of very good engineering documents. His paper on switch bounce is a classic must-read for any one designing equipment with switches. By far it's the bes ttreatment I've ever seen on that subject.
Posted: Fri May 05, 2006 6:10 pm
I've been using the Logicport for quite a while, and like it very much. The protocol analysis is the best I've found because it shows the interpreted data right in the waveforms, instead of having to relate the waveforms to a separate text list. The Logicport is the only one I've seen that does this.
By the way, you are talking about the 2K buffer, but you seem to have missed the fact that it has sample compression. The compression is particularly effective with protocol analysis. I've seen cases where it captured literally MEGAsamples of data in the 2K buffer.
If you're looking for reviews, take a look at this:
http://empegbbs.com/ubbthreads/showflat ... /0/gonew/1
It discusses all of the analyzers mentioned in this thread, and even shows a photo of the logicport in use.
Posted: Sat May 06, 2006 12:22 am
thanks. that is very useful.
I did note that it has compression but still the 2K seems pretty small. I agree that for protocols, it helps a lot. As well as allowing using a much higher sample rate than the system is using.
Recently, I found out that the digiport has a new beta release of the software that does protocol analysis though I've not been able to determine how good it is.
well, I'd about made my mind up to get the digiport so I'll have to spend the weekend thinking about it.
Posted: Sat May 06, 2006 1:30 pm
I looked at the digiview as well. The thing that put me off was the very limited trigger capability. It has just one level, and extremely limited terms. Basically, the first occurrence of any edge, possibly qualified by a pattern - that's it. Without trigger features, you basically have a data logger.
The logicport has two trigger levels which can be combined logically or sequentially. Each level by itself has much more capability than the digiview. For example it can trigger on the nth edge instead of just the first edge (n = 1 to >1,000,000 range). It can also trigger on a pulse or bus value duration (useful for finding glitches), or even the duration of time when a bus remains within a specified numerical range.
There were other things I liked better about the logicport (better input specs, input voltage range, state mode capability, etc.) but the trigger features were the kicker.
Good luck with your decision.
USB based logic analyzers
Posted: Sat Oct 10, 2009 1:39 pm
Is there anyone out there who has figured out how to use the Bitscope BS100U? Bitscope have an article, from Nuts & Volts, April, 2009, about the BS100U on their download page, but it doesn't go into the more complex features of the unit, which is a shame.
I bought one of these in 2008, thinking I could use it to look at signals at 50MHz or so, despite the 40MHz maximum sampling rate. In fact I discovered that I can do this, despite what naive sampling theory tells us, but how you do this seems to be part of a secret that the Bitscope people don't want to tell you about. If they did, they'd probably sell a lot more of them.
I've at least figured out that, if you know roughly what you're looking for, you can get quite a lot of information about a high-frequency, continuous signal by using the sampling hardware as a mixer. If you want get a close look at, say, the output of a 40MHz oscillator, you won't see anything with a 40MHz sampling rate, but if you reduce the sampling rate to 38MHz, then you see the lower sideband or mixing product, at 2MHz, a frequency Bitscope can display just fine. The input amplifiers appear to be good to above 100MHz, so you can still see some detail in the waveform.
If you have no clue about the signal, you can use the frequency domain display to see what frequency components are present. You can select a subrange of frequencies, as well as something called a "focus frequency". I have't yet figured out what this is, but it seems to be useful for looking at sidebands clustered around a carrier frequency.
An annoying 'feature' of Bitscope is that it appears to correct settings that are not consistent with its processing algorithms. Not understanding what these constraints are, I get the feeling that I'm continually doing something wrong, which I probably am. The manual they sent with the instrument explains none of this; in fact, it appears to have been written for a different Bitscope.
I discovered that the vertical display of my unit is incorrectly calibrated, reading almost 10 percent too high on the ranges that I could easily test. This looks like a silly ommission, one that could be easily corrected by adding an 'adjustments' dialog to the DSO program, or by simply putting calibration coefficients in the "probe" settings file that gets read at start-up. But they seem to have abandoned their earlier policy of publishing the source of the DSO program.
Any help will be most appreciated.
Re: USB based logic analyzers
Posted: Thu Oct 22, 2009 1:24 pm
I know nothing about these things, but I wonder what the major difference is between the units you're looking at the the Saleae:
http://www.sparkfun.com/commerce/produc ... ts_id=8938
This has been a staple on Sparkfun's website for a while. Junk?