Know Electronics

Part 1: 1. Run command: sudo apt install python3 2. Run command: sudo apt-get install python3-gst-1.0 3. Run command: pip3 4. If not installed go to step 5, else go to step 6 5. Run command: sudo apt-get install python3-pip 6. Run command: sudo apt install libcairo2-dev pkg-config python3-dev 7. Run command: sudo apt install libgirepository1.0-dev 8. Run command: sudo pip3 install pygobject

Part 2: Install gstreamer on Ubuntu Run command: apt-get install libgstreamer1.0-0 gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-plugins-ugly gstreamer1.0-libav gstreamer1.0-doc gstreamer1.0-tools gstreamer1.0-x gstreamer1.0-alsa gstreamer1.0-gl gstreamer1.0-gtk3 gstreamer1.0-qt5 gstreamer1.0-pulseaudio

Part 3: Install Tornado Run command: pip3 install tornado

Part 4: Install git Follow instrunctions from this website: https://www.liquidweb.com/kb/install-git-ubuntu-16-04-lts/

Part 5: sync a local folder with git 1. create the folder where you want the files to reside. 2. run command: git init 3. run command: git remote add origin <location of the git> 4. run command: git pull origin master 5. run ls command to verify the files are in the desired folder.

Part 6: gstreamer installation testing: 1. Test the installation of gstreamer using this command: gst-launch-1.0 playbin uri=rtsp://184.72.239.149/vod/mp4:BigBuckBunny_175k.mov

BeagleBone Black is a God Sent for Embedded Developers. It provides unique features which were unthinkable a few years back.

Imagine 1 GHz processing, and one feels that embedded developers who were like very happy at even 72 MHz went in a tizzy!

You can find a lot more about BBB (as I prefer to call the BeagleBone Black). The authentic source will always be Beagleboard.org, but as with everything Open, the information abounds and your ability to Google determines if you can find it.

And this ability, or rather lack of MY ability is the subject of this post.

Well we had a BBB setup which essentially is controlling a Slave controller via Serial Port.

Now, know this, I know as much about Linux, as Newton knew about Theory of Relativity. Which means zero. Oh OK!, not zero but very close to it!

So, as mentioned above we were controlling a slave controller (MSP430) using BBB. Initially we were able to receive data from the controller on UART0 (or known as TTYO0 in the linux file system)

Many attempts of Transmission of data from BBB to MSP430 were leading us to dead ends. I looked and re-looked at the serial code but to no avail.

Looking at the help on various forums, I decided to power it with external power supply. Then without any modifications to the code the serial communication was established and the entire master slave interface ran beautifully.

So far, So good!

This system was for an Australian customer, so we packed up and sent the system to them.

This is how it looked. As you can see, we had hooked three wires from J1 (J1.1,J1.4,J1.5) to a connector on our PCB. Before dispatch the circuit was running continuously for extended period of time, without any issues.

So a couple of days later, my customer received the PCB, switched it on in this precise configuration. Opened the software in Cloud9 IDE, as we were at our end, and tries to run it in manual mode. Only 1 transaction went through, then nothing! And by nothing I mean, after 1 transaction there was no successful transactions, inspite of power cycles, hard resets and what not!

We tried different exe’s with different timeout delays, but no success.

So we decided to debug this by connecting the BBB to the PC via TTL-232 converter. Which looked like this:

For serial debugging, we used the tool XCTU from Digi, our favorite tool for serial debugging. (Try it out!)

This problem persisted for multiple days and nothing happened. What we were seeing was that garbled data appeared after transmission of the master frame. Sometimes, the BBB also used to reset.

We were doing communication at 9600 baud and thus we decided to try different baud rates to view this data.

Then suddenly at baud of 115200 the data made sense. It was being sent by BBB and had details about user name and password. There was a lot of stuff there but unfortunately, I have not saved the image of the same.

Further scavenging of the Internet, I came to know that ttyO0 is the default debug port! If you are interested you can find further details here.

So with this information, I started on the trail to find how to disable this. Finally I found my answer on one of the Google Groups. Link to that group is here. In there look for the answer by guy/girl with user name as jgold.

The answer is that you have to run the command systemctl mask [email protected] to disable the default debug port. The system ran perfectly at my end as well at my customer end.

Now you may ask me how did it run at my end and did not run at my customer end, INITIALLY. Well, honestly, I am at loss here! If you know the answer please educate me!

Remember this!!!! Always!!!

We are coming back to PHP Extensions.

Ok you find one example on the internet which is quite abundant and that is  this:

ZEND_FUNCTION(DoubleUp)

{  

  long theValue = 0;    

  if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC,"l", &theValue) == FAILURE)

     {    

       RETURN_STRING("Bad parameters!", true);    

  }    

  theValue *= 2;    

  RETURN_LONG(theValue);    

}

This example is quite abundant and you could find it in many locations.

This link we have already mentioned:

http://blog.slickedit.com/2007/09/creating-a-php-5-extension-with-visual-c-2005/

And this link will make this blogpost useless:

http://devzone.zend.com/317/extension-writing-part-ii-parameters-arrays-and-zvals/

Or will it?

Ok. So by the above mentioned example you will know how to write an extension with one parameter. But what will you do if you have to pass two or more parameters:

Well here is the solution:

ZEND_FUNCTION(AddUp2Params)

{  

  long firstParam = 0;

     long secondParam = 0;    

     long theValue = 0;  

     if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC,"ll", &firstParam, &secondParam) == FAILURE)

     {    

       RETURN_STRING("Bad parameters!", true);    

  }  

theValue = firstParam + secondParam;

     RETURN_LONG(theValue);    

}

So what have we done to pass two parameters?

We have declared two variables firstParam and secondParam. So that did it? No. We continued to get errors telling that our function supported only one parameter.

So the real solution is

if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC,"ll", &firstParam, &secondParam) == FAILURE)

See that enlarged “ll” that’s the solution!

You don’t believe it do you? We also were shocked!

But it is quite straightforward and simple!

The datasheet of TL431 will fool you into believing that it is a simple IC. Well it is if you understand it correctly.

TL431 Voltage output equation explained

 (Well I am not able to  paste an image here. So just refer to the Figure 2 in TI’s TL431 datasheet!)

  The output Volatge is Vka

The reference Voltage is Vref.

 Lets Call the Current through R1 as Idiv.

 Then Vka – Vref = Idiv * R1   ----- (eqn. I)

 Idiv Splits into two Currents.

One as shown in the Diagram is Iref. Now Iref has been specified in the datasheet and has a current in the range of few micro-amperes. Not that it matters to our task in hand!

The Second Current is Ir2. Ir2 is the current flowing through R2.

 So Obviously Idiv = Ir2 + Iref

 But, Ir2 = Vref/R2

 Thus Idiv = Vref/R2 + Iref

 Substituting, equn derived for Idiv in eqn. 1

 Vka = (Vref / R2 + Iref) * R1 + Vref

 Vka = Vref * (R1/R2) + Iref * R1 + Vref

 Vka = Vref( 1 + R1/R2) + Iref * R1

 This equation exactly matches the one in the diagram.

 Since Iref is very small, Iref * R1 is very small and thus its contribution is negligible.

The equation can thus safely be simplified to:

Vka = Vref(1 + R1/R2)

Important Points about the TL431 equation:

  The     programmed output voltage is not dependent on the input voltage.  Of Course it is expected that the output voltage should not be dependent on the input voltage, since we are building a regulator.

It’s our take and we may be massively wrong. But we believe that the electronics and Software industry has entered the serene patch of the development river. The white water patch may be behind us.

The industry it seems will now take it’s time to focus on the development of the existing technologies rather than look for new ones. So we will see more products being developed on IoT, New ARM Processors, more integrated Analog Front Ends, and stuff.

When the Big Boss of ARM says IoT is in it’s infancy, it indicates that he is not getting the successes he desired.

We are running a series on PHP extensions and are interrupting the series midway for this post, to signal the sense of urgency as well as one of despondency.

IoT can be successful only if there is a strong leadership. IoT is about Things. ARM, Intel, Microsoft may think that they know about Things but frankly they don’t.

The persons who know about Things are the factory managers, floor supervisors and the project managers of companies which make wall papers, bathtubs, ceramics, bulbs and what not.

Where are they? Who is asking them what they want? Today in almost all companies, it is always felt that IT department works in isolation. Even today it is used for Sourcing, Billing, Scheduling, Planning. It is not used for monitoring what program is running on the CNC machine no. 23. It is not used to find out to find out how much Kilo-Watts of energy was utilized for producing component A.

IoT has to be about machines and components talking to each other as well as to the Internet. Each company is bringing out its own flavor of things. There is no standard and this is Parochial thinking!

These are the three main sources of information we used when we started on the PHP extension for C++.

This is the main source. This is the site which gave us the success, so it finds the top mention:

http://blog.slickedit.com/2007/09/creating-a-php-5-extension-with-visual-c-2005/

This is the second most useful site according to us. Also the guy has a very friendly face! ( A self proclaimed: Sofware Developer, Dad, Hiker, Foodie, Advocate and Tech Dabbler)

https://jojokahanding.wordpress.com/2014/05/26/writing-php-extensions-for-windowsintroduction/

This should also have worked for us, but we left using it midway:

http://www.dreamincode.net/forums/topic/244215-introduction-to-creating-a-php-extension-for-windows/

At this point let me tell you frankly that we are not 100% successful. Before you stop reading and move on to some other website let us tell you what we have achieved:

1. We are able to call a C function from the PHP code and the result of the same is displayed in the browser.

2. We are able to call a C++ class inside the C function and it runs without any issues.

What we have not achieved is:

1. Calling a C++ class function directly in PHP.

2. We (truly and honestly) are yet to understand what this Zend thing is all about.

3. We have not tested the functions with different types of parameters passed as arguments (like strings especially)

Final note before we sign off. If you fail initially to do the extensions, do not blame yourself too much. The documentation is sparse, many times illogical, and there is much of Windows vs Linux (close vs Open) tussle when you ask questions. So if it does not work, well be persistent.

Okay... Now we had mentioned earlier that our posts mostly would focus on IoT and then we disappeared! Well there is a reason to it.

During the study of REST protocol and stuff we realized that we could achieve a lot on the embedded platform but we were all near (or rather total) zeroes when it came to Internet technology. This meant that for IoT we knew about ‘T’ (or Things or Embedded Systems), but we were weak in the ‘‘I”. So we started reading up about JavaScript, CSS3, HTML5, and PHP.

On the other front we are working on a Test Equipment for a Gas Valve Manufacturer. We have given him PC controlled test equipments earlier too. Some of them were implemented in Labview and others were implemented in Visual C++.

Now, Labview is a pain in the wrist. Once you have worked on C++, this visual programming thing seems a little too basic. Also overall feeling was that the code was not that readable. Now these are our feelings and thus everyone has the right to their own views. All negative points mentioned, Labview has a UI to die for. It is absolutely ideal for industrial environment. Similar results no doubt can be achieved by Visual C++, but they are too cumbersome. There is also a provision for Integrating Labview Frontend components with Visual C++ code. But again, that too is an expensive proposition.

While reading up about HTML5 we became aware of the power that this language has. Especially the Canvas element. It is so well thought, so powerful and yet so basic that it really is an eye opener. Along with this there is a wealth of Information on the Internet. If you have a problem, or you think that there could be a problem, then 99.9999% of the time somebody already has asked it and the solution is quite nicely presented.

At this point the idea of using HTML5 as front end and Visual C++ as back end began to take hold.

The user interface (or the front end) could be easily tackled with HTML, JavaScript and CSS3. But to link the real time code from VC++ to PHP turned out to be quite a challenge.