Since I do not have any of the hardware, I must make it clear that I am only reviewing the software that is meant to be used with their analyzers. Good logic analyzer products will usually make their software available free, with either some demo files or a means to generate random or test waveforms.

Saleae Logic 1.1.15

http://www.saleae.com/logic/
Written in C++, Qt. Runs on Windows, Mac and Linux. Proprietary software.

Zooming in and out is achieved by clicking or right-clicking in the waveform window, respectively. The zooming is animated, allowing your eyes to keep track of which portion of the waveform is being zoomed.

Random data generation (simulation). It generates random data if you click the “Start Simulation” button on the top, which has no meaningful data. However if you add a I2C analyzer, it generates some random I2C bus data, allowing you to see how a I2C capture might look like.

Analyzers. You can add analyzers for the captured signals. The analyzers require that you either select one or more channels, depending on the protocol you are trying to analyze. Once activated, the active analyzers sit in the right pane, allowing you to change channels later on, or remove the analyzer altogether, which removes the annotations. You can also export analyzer data from the right pane.

Annotations. The annotations are smart. When zooming out of the waveforms, the annotations will definitely get squashed as more data is being displayed. Logic compensates for this by displaying lesser information in the annotations.

Cursors. It has support for two cursors but the UI is a little unintuitive. In the Measurements pane on the right, you will notice T1 and T2, which are underlined, hinting that they are clickable like hyperlinks. After clicking on T1 or T2 and moving your mouse over the waveform, the clicked cursor will appear and follow your mouse. Depending on which channel your mouse is over, the cursor also snaps to the signal transitions in that channel.

Extensibility. Although Logic is not open-source software, Saleae has community support with API documentation and a user forum where you can develop analyzers for it. You probably can’t add support for alternative input devices. Access to the forum and API documentation requires registration, which I didn’t do.

OpenBench Logic Sniffer (OLS) 0.9.5

http://ols.lxtreme.nl/
Written in Java. Runs on Windows, Mac and Linux. Open-source software, based on the SUMP client written for the SUMP logic analyzer.

Zooming in and out is achieved by the Ctrl+mousewheel action. The zooming action is centered around the mouse cursor, so you won’t lose sight of what you are zooming into. There’s also a minor nit, which prevents zoom from working when you mouse-over the wrong area such as the time “ruler” at the top, or the blank part of the waveform window.

I do not know if there are bugs in the software or if it is intended behaviour, but I cannot change the waveform colours and background as the tabs are disabled.

Random data comes from a “test device”. Since each device displays a settings dialog, you can select the type of test data, the number of channels and samples to be generated from that dialog. It includes different types of test data, such as I2C data.

Cursors. To perform measurements between two points, you will need to use cursors. OLS supports up to 10 cursors, labelled T1 to T10, which you can also use as bookmarks. Under the Tools menu there is an option called “Measure”, which launches a dialog where you can view the time at which each cursor is placed, as well as measure the time difference and frequency between any 2 cursors. You can also measure the frequency and duty cycle of a channel as a whole. The problem with the cursors is that they do not snap to the signal transitions, therefore to precisely measure the time difference, you need to zoom in all the way to set each cursor, which becomes a chore.

Analyzers can be applied from the Tools menu, which also display a settings window to select the channels and other settings. After analysis, data is generated in the same window, which can be exported as a HTML file, as well as annotated in the waveform over the respective sections.

Annotations are added by analyzers. When zooming out, the annotations will shrink and when they can no longer accommodate the contained text, it disappears altogether, showing empty bubbles in the waveforms. It could be made better if it displayed a shorter version of the text, much like Saleae Logic does.

Extensibility. The whole project is licensed under GPLv2, so you can pretty much extend or change anything. There is ongoing work to define an API for writing new device “drivers”, tools for analysis, exporters and more.

32leaves LogicAnalzyer (LA)

http://logicanalyzer.sourceforge.net/
Written in Java using Eclipse Rich Client Platform (RCP). Runs on Windows, Linux and Mac. Open-sourced under GPL.

While I applaud the authors for trying to develop an open-source logic analyzer, I have to respectfully say that it is crap.

Zooming cannot be done with the mouse. You have to select a toolbar icon called “Set Scale” and manually enter a scale value in ns/pixel. Worse still, zoomed out waveforms have scaling artifacts, making them look very unsightly.

It is also slow, especially when redrawing waveforms while you’re making a selection. I have a pretty top-of-the-line workstation, but I can still see the selection being redrawn, especially when parallel selection has been enabled. This doesn’t contribute to a good user experience.

LogicAnalyzer has the concept of “sinks”, which is an unnecessary concept to communicate to the end-user and it doesn’t help at all. I selected “Random data” to generate some data I could play with, after which I was prompted for sinks. Among the sinks available was the “I2C analyzer” and the “Editor sink”, so I added both and it “oops” on me. If I had selected the I2C analyzer it just shows a table full of the decoded values and I couldn’t see the waveform at all, and likewise for the editor, it shows the waveform which I can’t analyze later.

At this point, I gave up playing with the software any further.

ChronoVu 3.06

http://www.chronovu.com/
Written in C++, Qt. Runs on Windows, Linux and Mac. Proprietary, closed-source software.

While all 3 platforms were supported, the installers for the “bus decoders” were separate downloads. This means if I were to install the software on a Windows PC, I would download the main installer, as well as the 4 additional installers, one for each decoder.

ChronoVu lacked the ability to generate random data or simulate a capture for me to play with. It had sample files but they were split into “setup files” and “data files”. The data files seemed like just regular square waveforms of different frequencies and duty cycles, so I couldn’t analyze them.

Zooming in and out had to be done by toolbar buttons. Again, this irritates me, but not as much as LogicAnalyzer.

Cursors can be set by left or right clicking at points in the waveform, which are marked with L and R markers, but I do not know what you can do with this. I am also unsure how you can measure the duration and frequency between the markers.

Conclusion

It should be obvious after reading the reviews that I favour Saleae Logic (I do not work for them). It is the best logic analyzer software I have seen so far, and I like the intuitive layout a lot. I didn’t have to read any manuals to figure out how to use the software at all. However, their software works only with their logic analyzers, so I would recommend buying this if you can afford it.

If you wish to use the SUMP logic analyzer or the Open Workbench Logic Analyzer, it is clear that you should use the OLS client. If the issues I raised can be addressed, it would be even better.

If you are looking to buy a logic analyzer, I suggest you look for one with extensibility to their software. The software is just as important as the hardware. I have seen those expensive machines (or hardware) which works with their proprietary software that runs only on Windows, looks like shit and is a total chore to use.

If you are going to spend a lot of money on a hardware product, make sure their software is worth your money too.

A zener regulator is set up like this:

The concept is simple: zener diodes are specially designed to have a specific breakdown voltage (in my case 2V) when put in reverse. The result is a 2V voltage drop across the diode, and hence, Vout. Since I didn’t have a zener diode, I thought I would use a regular LED in its place, but placed in forward bias (so that it lights up).

When I tested the circuit however, I realised the voltage drop across the LED was less than 2V. I missed an important point: the current-voltage curve.

Observe the LED current-voltage curve on the left and the zener diode current-voltage curve on the right. You will notice that the zener diode has a steeper curve compared to the LED. Apparently to avoid blinding myself, I had limited the LED current to about 10mA. I missed the fact that at 10mA the voltage drop across the LED is now only ~1.8V. To get 2V across the LED, you need at least 20mA, which would be too bright as an always-on indicator.

This also answers a question I previously had of whether I can use an LED in place of a zener diode for crude voltage regulation. One of the techniques used by VUSB devices to limit the D+ and D- voltage is to use a pair of 3.6V zener diodes (solution B). The SparkFun AVR Stick uses blue LEDs so that they also double as activity indicators, but from the user comments, it is clear that this causes problems.

If you want to use an LED for voltage regulation, remember to check the current-voltage curve.

I had a new ceiling fan installed yesterday, and the installation guy came over with a power drill fitted with the Phillips screwdriver bit to do his job, and he used it on everything. The fan came with the blades and motor hub separately, so some assembly was required. Just have a look at the result.

Here’s a view from the top. If not for the fact that this screw has a groove into which you can insert a flathead screwdriver, you probably won’t be able to remove this at all.

I actually needed to unscrew the bottom cover (which they term the canopy) to check if there’s a serial number printed on the motor hub. I’ve tried 3 different Phillips screwdriver and none of them can grip the screw properly.

The thing about using a power drill is it’s powerful – it really jams the screw in there tight. But if you abuse or misuse it like this, you will probably have a hard time trying to remove the screws, and when you do, you have to find a replacement screw.

So, why couldn’t the guy just have used a regular Phillips screwdriver? I think he was just fucking lazy. How would you feel if your factory-assembled products came with stripped screws like that?

Charlie Miller reverse engineers the Mac battery firmware updater, sniffs battery communications on the SMBus, writes an IDA processor plugin (in IDAPython) for the CoolRISC 816 processor in the bq20z80, and mucks around with the its firmware.

All the source code and presentation materials are provided.

[via Dangerous Prototypes]

After I had my very first set of PCBs manufactured, I was thinking of making my next set of PCBs in a different colour. However I found the price difference to be quite big and might just stick with plain old boring green instead.

Now that they’ve announced their new offering, I’d definitely go for a different colour for my next batch. They’ll be offering 8 boards instead of 10 (which is not a problem at all for me), and fixed thickness of 1.6mm and HASL finish – all for an extra US$5 more than their base PCB service. Their available colours are blue, white, yellow, red and black.

[via Dangerous Prototypes]

With so many tabs open, it’s definitely a good thing to have desktop notifications (something like Growl for Mac), especially since I use Google Chat on the Gmail page sometimes. I was hoping that Gmail detects a capable browser based on the availability of Javascript objects, instead of some stupid sites that detect the User Agent string, so that the functionality could at least be injected into the page using a Greasemonkey script or a Firefox plugin.

It turns out that someone has already written a Firefox plugin to implement desktop notifications, and that Gmail does in fact detect the presence of Javascript objects instead of blindly relying on the UA string. It’s called ff-html5notifications. With this plugin installed, Gmail will now ask if you want to enable desktop notifications, just as if you were using Chrome.

Chrome implements the Web Notifications API (draft specifications at this point) via the window.webkitNotifications object. It is unclear whether Firefox would support this API in the future, but the ff-html5notifications plugin provides a good solution for now.

Edit the /System/Library/CoreServices/Springboard.app/K48AP.plist file and add a boolean key multitasking-gestures in the capabilities dict, and set its value to true. You can add both display mirroring and multitasking gestures to the iPad 1 using this method.

That’s it!

Alternatively you can use Cydia or redsn0w to do this for you.

“Wait a minute… then what have you been using?” I hear you ask. Previously I had access to an Arduino Duemilanove, and used it to burn the Optiboot bootloader onto an ATmega168 that I had. The Duemilanove board used an FTDI chip which had additional pins brought out to an unsoldered header marked as X3. Following this guide by Kimio Kosaka, I downloaded the precompiled avrdude for Windows and used it to program the ATmega168 via the X3 header.

The end result was a bare-bones Arduino that ran on a 12MHz crystal. The reason for 12MHz was because that was the maximum “safe” clock speed at 3.3V, which I used for a university project and have been using ever since. You can see it below, driving a HD47780 parallel LCD.

However, it’s a real chore to hook this up, especially since everything on the breadboard like the crystal, FTDI header and reset button and pullup resistor are all inserted like any other component. It is more convenient to have a ready-built board for prototyping, where you don’t have to worry about the Arduino components.

If you haven’t gotten an Arduino board yet, I highly suggest you get one. If you would like to hook it up on the breadboard similar to what I did, at least get the Really Bare Bones Board (RBBB) Kit.

Anyhow, I bit the bullet and read through the ctypes documentation and wrote AMDevice.py which exposes some simple classes to handle connecting to an iPhone. I only implemented the minimal set of functions required to download and upload files to the iPhone, as I wrote this primarily for my iPhone SMS import script.

As I use the iPhone SMS import script on a regular basis (the Samsung Jet frequently runs out of SMS memory), I got tired of opening iPhone Explorer, copying the file out, running the script, and then copying the file back. If only the bookmark feature actually worked, I wouldn’t have to navigate to the directory every single time.

With this bit of code (and a jailbroken iPhone) the script will now download the sms.db file, insert new SMSes from Samsung New PC Studio, and upload the file back into the iPhone. Very convenient.

On top of that, I can also use the Python wrapper for other purposes. If you need it, you can find AMDevice.py in the Bitbucket repository for my import script.

The process managed to shave off about 315 μA, which boosts battery life quite a bit:

The overall effect this has on the product is that the 24 hour time between battery changes can be upped to 38 hours. That is a pretty good power savings for the day.

It’d be great if this was incorporated into the products.

I do disagree on one thing though, which is cutting away the data points used for the PWM. Unless the output was measured with an oscilloscope and produces the same breathing pattern, I would very much keep the data since the additional space gained is not used for anything else.

If you don’t know what iCufflinks are, watch the video below.