About admin

Posts by :

Variable Speed Cooling Fan for Raspberry Pi using PWM (video#138)

Variable Speed Cooling Fan for Raspberry Pi using PWM (video#138)

, 27

In video #131 we pimped a Raspberry Pi. One of the new features was a cooling fan which switched automatically on if CPU temperature was too high. Viewers suggested to make the speed of the fan variable. And this is, what we will do today. We also will learn a few things about Pulse Width Modulation, PID controllers, MOSFET transistors, and even a little bit of magic.

So, let’s start: in video #131 I used a simple concept: A MOSFET transistor to switch the fan on and off and a Python program which measures the CPU temperature and switched the MOSFET on, when the temperature was too high. Very simple. I used a beefy IRLZ44N transistor, because it is logic level. In video #137 I showed, that these FETs switch on at a voltage of below 2.5 volt. This is, why they are ideal for microprocessors with 3.3 volt logic.

Usually, you should add two resistors to drive a FET: R1 to discharge the gate capacitance of the MOSFET, and R2 to limit the current which is created if you change the signal at the gate. R1 is usually a few k Ohms, and R2 a few 100 ohms.

A diode should be connected in parallel to the fan, because it has an inductance and inductors create high “reverse” voltages when switched on and off fast. In our case, R1 and R2 are built into the GPIO of the Raspberry and can be omitted. The diode is not mandatory for a small fan, but it is better to include one.

The python script has two tasks. It has to be started at boot. I described how this can be done in a earlier blog and in video #131.

The python script is at the same place. You can get it with the following command:

wget http://www.sensorsiot.org/transferFiles/Pi_Zero_MQTT_Broker/fanShutdownPWM.py

Please rename it to fan_shutdown.py:

 mv fanShutdownPWM.py fan_shutdown.py

Or adjust your startup file accordingly.

As a result, you should get a quite constant temperature of 45 degrees centigrade:

You can change this temperature on line 15 in the python file:

desiredTemp = 45 # The maximum temperature in Celsius after which we trigger the fan. If you use a different fan, it might be necessary to change the parameters of the PI controller:

pTemp=15
iTemp=0.4

 

 

 

 

Short tutorial for translators of YouTube videos

Short tutorial for translators of YouTube videos

2

After a few days of translating my videos, we found out, that there are basically two different ways:

  1. Translate may script using google translate, correct the translation, and then add the timing
  2. Using the transcript with the timing and translate the text manually
With Google Translate
  1. You download my script from: https://github.com/SensorsIot/Scripts-for-my-videos and translate it with google translate
  2. You correct the text in your text editor till you are satisfied
  3. You add the timing and convert your file into a SBV file using this service: https://kurienzach.github.io/subtitle_converter.html
  4. Then, you create your language
  5. And upload this SBV file to YouTube
  6. At the end, you have to adjust the timing to the spoken words
  7. Please translate also the title and the video description, because this is, what your people will see, if they browse YouTube

Advantage: Google translate results are better than with method 2
Disadvantage: The timing is only rudimentary and needs quite some adjustments.

 

Manual Translation
  1. You download the English transcript (not English (Automatic)). DFrop me a note, if this language is not there for your video and I will create it.
  2. You get an SBV file which can be edited with every text editor. The timing is already included and should be ok
  3. You translate the text manually
  4. Then, you create your language
  5. And upload this SBV file to YouTube. Please check the timing. It should be ok.
  6. Please translate also the title and the video description, because this is, what your people will see, if they browse YouTube

Advantage: No timing needed

Disadvantage: Everything has to be translated manually, because Google translate does not work properly on these short sentences

Release

As soon as you release your version, I get a request to publish it. I will check if everything is ok, and publish it.

Steps:

 

 

Where do you find the translation menu?

Check, if your language is already translated

Choose “English”

If not, create your new language

Download SBV file for your translation

 

Upload file

Check and adjust timing

Safe Shutdown for Raspberry Pi with Super Capacitors (#133 )

Safe Shutdown for Raspberry Pi with Super Capacitors (#133 )

19

Idea of a Shutdown-Protector

Create a signal for a GPIO of a running Raspi as soon as the supply voltage is interrupted to trigger a software which immediately starts a controlled shutdown. And to build some sort of “energy source” for a few seconds, until this shutdown is finished.

Instead of batteries, we use small super capacitors of 10-15 Farad.

Diagram for the “Protector”

The two super caps have parallel resistors to “bleed” them in case of power loss. If the power is away for a few minutes, the Raspi  boots if power is applied again.

Some viewers also  suggested some sort of over current protection from the USB plug to the super caps. This can be done with a small resistor. However, thin USB cables have a similar effect.

Link to another proposal: http://www.hackerspace-ffm.de/wiki/index.php?title=Raspi_EDLC_UPS

Creating subtitles for my YouTube channel

Creating subtitles for my YouTube channel

6

If you are a translator, please look at the “tutorial” blog here: http://www.sensorsiot.org/short-tutorial-for-translators-of-youtube-videos/

In video #132 I asked for help to create subtitles in different languages to make it easier for non native English speakers to follow the videos. I asked for two things:

  1. For translators
  2. for somebody who would write a small tool to support the translators

The response was overwhelming for me. In just a few hours I got offers for the following languages (in alphabetic order): Arabic, Chinese, German, French, Dutch, Croatian, Polish, Portuguese, Russian, Spanish, Hungarian

I got also a few offers to write a tool for the translators. But I got more: The (usual) discussion about the need of subtitles in general and, also many tips for an improved workflow.

I took these tips as a start for a little investigation in the matter and present now my findings. This, of course, is only a starting point, and I would be glad to open a discussion. I also have to admit, that this article is written without a lot of experience.

The problems
  1. Getting the transcription of the video in English (with a Swiss Accent)
    1. Use my original script
    2. use YouTube transcription of my spoken language
    3. Listen to the video
  2. Get the timing of the sub-titles
    1. Download the timing from the YouTube transcript
    2. Use a simple algorithm to approximate the timing
    3. Do the timing manually
  3. Translate the text into the different language
    1. Translate it manually
    2. Use Google Translate (or another translator software)
  4. Get the timing and the translated text together into a file which can be uploaded to YouTube
First impressions

I used one of my last episodes to test the different possibilities (#132 Pimp my Raspberry). First, I downloaded the transcript from youtube and compared it with my script. Here is the result:

Comparison between my script and automatic Google translation

The result is quite good, but there are still many differences. Not all of them are important, but they can influence the automatic translation later on. Because I have a script, I would opt to use it, because it is not a big effort for me to distribute it.

Of course, if you want to do all manually, you can also listen to the original video and enter the translated text directly into YouTube. Then, you would avoid all intermediate steps. I am pretty sure, this would create the best results, however, it might also be the most time consuming variant.

The next is the timing of the sub-titles. Here, the transcript is spot-on, because it really listens to the language and the timing of the video. Also the length of the texts fits well to the page width (in average 37 letters). Unfortunately,  if we use this text for our translation in google, the result is worse than if we use the original script. Sometimes, it is even hard to understand it. You can see it in the example below (forgive me, that I used German for that test). But you can do this test also in your language. Actually, this is also good news: Google translate really translates whole sentences, and not just words.

But for our purpose, for the sake of translation quality, it is better to use the plain script, and not the transcription with the timing.

If you do not want to use Google translate to start, you can use English instead of “Automacit English”. This language can be created by using my script. YouTube then automatically synchronizes my script with the really spoken text. To create this language, if it is not yet there, you do the following:

Create “English” from “Automatic English”

You create a new language “English” and press the button “auto transcribe and sync”. Then, you enter the text of my script (copy-paste), and press “set timing”. After a few minutes, your transcript is ready: The timing from the spoken words and the text from the script. Now, you can download and translate it manually.

I took the other route and therefore have to compare the timing of YouTube with the proposed simple formula. Fortunately, Kurien Zacharia already created a service which I was able to use (https://kurienzach.github.io/subtitle_converter.html). If I used it with an average of 36 letters per subtitle (Google used 36.9, but without punctuation marks), I got nearly the same number of sub-titles, and the differences in timing usually was only a few seconds. So, I would vote here for the automatic translation of the script and the formula rather than have the exact timing and a less good translation. It is probably less of an effort. But I am not sure, and it might differ from language to language. So, it is up to the translators to try both.

The manual way to copy-paste each part of the translation and place it below the video is clearly less comfortable.

The next is the translation: Here, there is not a big difference between the two timing variants. In both variants, you correct the proposed text inside the timing file. I do not think, it is appropriate to adapt the timing here. This has to be done anyway later and it is easier to do that in YouTube itself or in one of the other proposed programs.

If you go the complete manual way, you have to enter the translated text directly into YouTube along with the translation and the timing.

If you did not chose the manual way, you should have now a .SBV file ready for upload to YouTube. And after the upload, you can adjust the timing here and there.

Other tools

I got proposals for the following services:

http://subtitle-horse.com

http://amara.org/en/

http://www.aegisub.org/

http://www.nikse.dk/subtitleedit

I looked at them, and they are all offering the synchronization of prepared subtitle files with the video, or the direct entry of translations of the video. The functionality is similar to the one which is offered by YouTube. Maybe a little more comfortable, or on a bigger screen. But none offered a solution to automatically take the translated script and place it somehow on the time line. At least, I did not find this functionality. Please correct me, if I am wrong and I am more than happy to correct this statement.

With my today’s knowledge, I prefer to use YouTube for this function, because you do not need another tool. But if somebody likes to work on a big screen, these tools are all free for usage. And the end-results are also uploadable subtitle files for YouTube.

The script files are ready for download for the Translators on Github (https://github.com/SensorsIot/Scripts-for-my-videos). Not all are there, but the newer ones. Just contact me, if you want another one.

Please also make sure, you do not translate a video which is already translated in your language! Just chose another one.

So, I have to thank you all for your support. As soon as I have the first requests, I will start the page for all translators here on this address. Just drop me a note with a small description, a photo, and of course, which episodes you translated.

Title and comments

I recognized, that YouTube presents the titles of the videos in the login language, if the title and the comment is translated. I think, this is a nice gesture to non-native English speakers. So, please, also translate the title and the comments. And if you do not have a lot of time, just that 😉

Current Status of translations (sorted alphabetically by Language)

Viewers, which want to help create subtitles  (sorted by Language alphabetically)

(names taken from YouTube comment)

LoRa Gateway as Wemos Shield

LoRa Gateway as Wemos Shield

, 19

Recently I read, that the Dragino Single channel gateways do not work anymore (because of a change of the backbone). But I read also, that the small ESP8266 gateways still work. The problem seems to be solved. So, I built such a gateway. It can be mounted on a Wemos and you can also mount a small OLED display on top. Cool.

Now, the display shows, when traffic arrives and also the time it arrived, the RSSI,  the SNR and the length. Of course, you cannot display the content or the sender. This information is encrypted…

In addition, two Neopixels show the state of the gateway.

Look at the size of the Gateway (right) compared with the Arduino/Dragino node. it is really cute!

I am now preparing the video to show you, how to build it.

Diet Pi Supporting Material (videos #126 and #128)

Diet Pi Supporting Material (videos #126 and #128)

, , , , , 30

In videos #126 and #128 I introduced a SD card image file based on dietPi and a script from Peter Scargill. Here, you find some material and links for this topic:

How to get Peter’s script:
wget https://bitbucket.org/api/2.0/snippets/scargill/94yxL/99a359e843f0997c6b9247f36ba4848d5b322f5e/files/snippet.txt
SD card image:
http://www.sensorsiot.org/transferFiles/Pi_Zero_MQTT_Broker/DietPi_MQTT_Broker_8G_New.zip
If you use the SD card image, the following users and passwords are used:
For the operating system:

root / dietpi
pi / password

For Node Red:

for the Flows page: admin / admin
for the dashboard: user / user

You can change the usernames/password with this script:
wget http://www.sensorsiot.org/transferFiles/Pi_Zero_MQTT_Broker/nodeRedUserReset.sh
Change the passwords in the script
nano nodeRedUserReset
username="user"
userpass="user"
adminname="admin"
adminpass="admin"
and run the script:
bash nodeRedUserReset.sh
Flow file for video #128

http://www.sensorsiot.org/transferFiles/Pi_Zero_MQTT_Broker/Flow%20enhanced.txt

Pimp my Raspberry Pi 3 (video #131)

Pimp my Raspberry Pi 3 (video #131)

12

In my last mailbag, I got an ordinary Raspberry Pi Model 3 with an acrylic housing, a fan, and a battery power supply. All nice things, but not good enough for me. I will pimp my Raspi!

A small fan will keep the Raspi 3 cool

A simple hardware switches a fan on and off

A Python program reads the CPU temperature and switches the fan
So, we have a cool Raspi without too much noise
I will auto-launch Python applications during boot (including error logging)

Build-in a nice “battery fuel gauge”

 

Hack the gauge to provide a “battery low” signal for automatic shutdown

Useful commands:
Create Scripts directory:

cd
mkdir Scripts

cd Scripts

Python program:

wget http://www.sensorsiot.org/transferFiles/Pi_Zero_MQTT_Broker/fan_shutdown.py

Automatic Startup

cd

mkdir logs

cd Scripts

wget http://sensorsiot.org/transferFiles/Pi_Zero_MQTT_Broker/launcher.sh

chmod 755 launcher.sh

sudo crontab -e    (select “nano editor” if asked)

Fill in:
@reboot sh /home/pi/Scripts/launcher.sh >/home/pi/logs/cronlog 2>&1

Stress test:

wget http://www.sensorsiot.org/transferFiles/Pi_Zero_MQTT_Broker/powerTest2.sh

Additional links

http://www.instructables.com/id/Raspberry-Pi-Launch-Python-script-on-startup/?ALLSTEPS

https://www.element14.com/community/docs/DOC-78055/l/adding-a-shutdown-button-to-the-raspberry-pi-b

 

Extending a dietPi file system (video #126)

Extending a dietPi file system (video #126)

, , , 6

Download image of the SD card

Here, you find the image of an SD card for the Raspberry Pi Zero. It only works for the Pi Zero. If you want to create a MQTT broker for another platform, you have to download the normal OS and then apply Peters script

Simple way

Look at https://learn.adafruit.com/resizing-raspberry-pi-boot-partition/edit-partitions

  1. If you have a other Linux system with a graphical interface, you can install gparted and mount your newly created SD card with the (too small) image into a USB card reader and the card reader into your running Raspberry.
  2. You have to enter your password twice
  3. You start gparted and search for your SD card. Then, with the right mouse, you resize/move your ext4 file system. Enter zero in the field “free space following) and hit “resize/move”
  4. Then, you apply all operations.
  5. Done!

Complex way (all commands in bold)

  1. Mount your newly created SD card with the (too small) image into a USB card reader and the card reader into your running Raspberry.
  2. Sudo fdisk -l shows you the name of this device (usually /dev/sda). It has two partitions, SDA1 and SDA2. And SDA2 is very small
  3. Sudo umount /dev/sda (usually, it is not mounted)
  4. Sudo parted /dev/sda
    unit chs
    print
  5. The numbers in red show the max number of cylinders, heads, and sectors of your SD card (actually, all numbers are too big). The yellow shows the actual beginning of your second partition
  6. Now, we have to remove the boundaries of the second partition: sudo rm 2
  7. And create new boundaries: mkpart primary 8,138,2 1963,254,62. We use the two numbers from before and subtract at least 1 from each (red number). You might get the error message “…is outside of the device /dev/sda”. Then, just experiment till your values are accepted (I found no better way)
  8. Now we quit the utility
  9. Next, we have to clean the newly created partition: sudo e2fsck -f /dev/sda2
  10. And resize the partition: sudo resize2fs /dev/sda2
  11. If we look at our partitions (sudo fdisk -l), we see, that the partition is bigger than before

 

This is the way it worked for me. No guarantee, that it works for you! I am a Linux beginner.

 

Source: http://elinux.org/RPi_Resize_Flash_Partitions