This post details the wireless connection to the RPi B+, along with some other voltage checks.
In order for the eLDERmon to be most useful, it needs to be untethered from the ethernet connection.
We've removed the wires to the sensors, so its only logical to remove all the wires at the other end ... well the power might be an issue.
I decided to use a WiPi to connect the Raspberry Pi to the interweb thingy. .. as a certain British motoring show host calls it.
I plugged mine in and unlike the older models, the B+ didn't reboot ..
The first step is to backup your interfaces file ...just in case
In the terminal window, type sudo cp /etc/network/interfaces /etc/network/interfaces_original and press ENTER.
This copies (cp) the file 'interfaces' that lives in the '/etc/network' directory.
I recommend using a fixed IP address, rather than letting the router assign some number that you then have to hunt for.
You could force the router to reserve a number, but this is just as easy.
You'll need to know the IP number you can use, the network/lan address and the gateway used to get out from your house to the world.
Type ifconfig -a and press ENTER and you should get something similar to this
This tells me that the network is 192.168.0 and the gateway is 192.168.0.20
To add the wireless settings type sudo nano /etc/networking/interfaces and press ENTER.
(sudo gives you root privileges and nano is an editor that I prefer over some others)
My interfaces file for my test network looks like this, BUT you need to add a space and the name of your network after the 'wireless-essid'
ie wireless-essid minion
Once you're happy, press Ctrl and X, then follow the prompts to save the file.
You need to force the interfaces to pick up the new setting, and you can either reboot, or reset the interfaces by typing
sudo /etc/init.d/networking restart and press ENTER.
You should be able to confirm the wireless (wlan0) is working by typing
ifconfig -a and noting the new settings
As the above shows both the Ethernet and Wireless are connected and working
There has been issues with getting WiFi to work, and for this test network I don't run security.
You'll find plenty of setups on the internet if you do use security, or have issues.
You can test it's working by pinging your gateway address (192.168.0.20 in my case)
ping -I wlan0 192.168.0.20
The inclusion of the -I wlan0 forces it to use the wireless
having confirmed everything is working, you can pull the ethernet lead out, and reconnect using the wireless.
Just be aware the latency is slightly higher on wireless, so some things may be a bit slower.
So now that works correctly, another item has been ticked off my list.
I was keen to ensure that I had an adequate supply for my RPi, Wireless, Arduino and my mystery item, and luckily I found this supply.
I'm not sure of its history, and one of the problems with older supplies is that the capacitors in them expire by drying up.
This means they either don't work, have poor regulation or even worse overheat and melt or catch fire.
We tend to ignore power supplies, and expect that they just keep on going, no matter how poor the ventilation, or how old they are.
This is rated at 4 Amps and I need about 2.5 Amps, so it should work .... but will it?
I could pull it apart, and check the capacitors by measuring their value and the ESR to see if they have worn. (ie dried out/changed)
However the best way to find out is to test it using a resistive load that simulates the expected load you will apply when its in service.
Many years ago I rescued some very large resistors to make an audio load.
These are 2 ohms and 100W rated, and are perfect for this test. (or are they? ... see further down)
I = E/R so 5/2 = 2.5 Amps.
The supply has a 2.1mm DC plug, so I attached some wires attached to a socket and wired the resistor across the socket.
You could measure the voltage with a voltmeter, but this only gives you part of the story, so an oscilloscope is your best tool.
Connect the Tektronix Scope
Note the earth is connected to the negative side.
First test is to check the voltage
For this test I set Channel 1 to 1.0v/ division with DC coupling, and then plugged the socket onto the DC lead (hence the little stumble)
As you can see it's steady at 5v with a 2.5 A load.
It's also gratifying to note that the voltage went straight to 5v with no overshoot due to suddenly dumping the load onto it.
Next check is to measure the ripple.
Ripple is the variation from the nominal voltage, due to losses in wiring, worn parts, switching spikes, or even poor design.
In an ideal world that would be zero, but in reality needs to be appropriate to the task at hand.
This requires a higher resolution (or is that lower) setting of 20mV/div, and AC coupling.
Again I started the trace and then plugged the socket onto the DC lead.
The spike is the capacitor inside the scope stabilising after the voltage change.
One thing I haven't got used to with this scope is that it stores the captured data, and you can then add things later, or even change some of the range settings.
Rather than guess the ripple, I plugged the USB stick in and selected the cursors to measure it.
This measurement is of the ripple across the load with no volts supplied.
It includes some noise spikes, but is 8.8mV.
This measurement is with the voltage applied, and again includes some switching noise.
The cursor difference shows it to be 15.2mV.
This means that 15.2 - 8.8 = 6.4mV of ripple at 5v is only 0.128%
Obviously one of the other things to watch for is heat.
We have had supplies at work melt their cases, while the screen is still working.
Its not ideal and often results in the cases becoming brittle.
If they are moved they can open and expose potentially lethal voltages to unsuspecting persons.
A couple of large cable ties around the case is added insurance if you are in any doubt, since it may not be you that touches it.
For this test I left it running for an hour or so and while the resistor was extremely uncomfortable to touch (only 12.5 w) , the supply was barely above ambient temperature.
With theses tests completed, I have no concerns leaving it at 'Grannies' house.
So another part ticked off the list.