Archived Forum

The switch indications looks exactly like what you see when all three switches are pressed at the same time. What happens when you do that and then release them one at a time? If nothing changes, it looks as if you might have faulty switches, or a whisker of solder shorting the contacts - of all 3? - unlikely but possible.

The Up and Down switches are AIO2 and SCK respectively, so those should be 0 V normally and 3.3 V with the button presses. 

0.0C is not the normal value you get from a sensor that is responding, so it looks as if it has either not initialised properly, or it is not responding to the sketch at all. I cannot see anything wrong with your soldering on the front of the PCB, but it is worth checking carefully around the sensor pads.

The GitHub sketch runs correctly on my V1.4, so I think the sketch is OK. 

I erased the first part, because I had the wrong sketch before someone could react ;-)

The soldering around the sensor is correct to my opinion. Ground and positive are measured and showed 3.28 volts. I have also tried to connect one of my other sensors and that didn't help.

It would have been nice if you had found out about the wrong sketch before I wasted my time answering that part.

Did you remove that particular temperature sensor and replace it with another?

Of course I didn't do that on purpose.

I didn't remove it, but assumed that if I connect an encapsulated sensor DB18B20 on the legs of the temperature sensor it should work. If the sensor doesn't work, than the new one should send the temperature digital. Is that a strange way of thinking?

I also have bridged the ground leg of the sensor directly to ground, to see if there was the error.

I have desoldered the temperature sensor, what was a bit tough with the lcd on it. But now I have connected my encapsulated temperature sensor and it gives the same outcome.

"what was a bit tough with the lcd on it"
That is why the build instructions tell you to test it first with the LCD only held in position.

Can you check continuity between the sensor output pin (the middle one) and pin 11 of the processor and the pull-up resistor (the 4.7 kΩ one, at the end nearest the USB connector), just in case there is a PCB fault. On my GLCD, I read approx 3.3 V there, but with a dip every half-second approx. when the sensor is polled and data passes.
 

Data pin of the sensor and processor pin 11 (not digital pin 11) is 0 v.

With library: dallas-temperature-control (initial temperature 0.0 ºC)
​Data pin of the sensor and 4.7 KΩ is 1.5-1.6 V with drops to 1.1-1.2 V

With library: Arduino-Temperature-Control-Library-master (initial temperature -127.0 ºC)
Data pin of the sensor and 4.7 KΩ is 1.2-1.1 V with drops to 0.9 V

This is weird. Those voltages are way too low. If you unplug the processor, what voltage do you get?

Which One-wire library (and version) are you using?

Uplugged the same for pin 11, 0 V. For the  4.7 KΩ and the data pin of the sensor 0.09 V. Have also replaced the ATMega328, but again no difference.

I used the library  from http://www.pjrc.com/teensy/td_libs_OneWire.html. Tested with no difference: V2.3, V2.3.1 and V2.3.2

I have desoldered with succes the LCD and have made a picture of the back of the PCB, which I have attached. 

"Uplugged the same for pin 11, 0 V."

That is where your problem lies. The voltage should be 3.3 V, pulled up by the 4.7 kΩ resistor. With both the processor unplugged and the sensor removed, you should have 3.3 V on both ends of the resistor, on the sensor data pin and on pin 11. Looking again very carefully at the front photo, can I see a solder splash by the resistor? And are those flux or solder splashes on the track from pin 11 to the resistor? Unless the resistor is the wrong value (but the colour code is correct) then that is what you are looking for. If/when you have 3.3 V on both ends of the resistor, short pin 11 to GND with your meter on mA d.c. You should read 0.7 mA.

What you told me, didn't work either unfortunately... 

My patience to get it fixed, is/was completely gone. Probably I did something wrong, because it was also my first soldering project. I ordered direct a new kit, so I will get another chance to get a working EmonGLCD. Now I have even some spare parts. ;-)

Is there information about which values should be found on the pins? 

Thanks for helping Robert!

I don't understand "What you told me, didn't work either unfortunately". Did you check the voltages at both ends of that resistor, and what did you read?

I think there must be a very simple reason why pin 11 is at 0 V, it is silly to throw away a £50 kit when the fault might be rectified by a 10 p component or a piece of wire.

But if you do not want to continue, it is your decision. I have only built two of these, and both worked first time. The only list of voltages that I have are for the GLCD itself, because (as I understand it) an error in the information used to design the emonGLCD showed two electrolytic capacitors the wrong way round, hence the comment in the build notes.

When you build your new unit, please follow the build guide exactly. Before the display is fitted, you can, as I assume you have a serial programmer, see the temperature, light level and switch states on the serial monitor.

It is tricky, but possible, to test the unit without soldering the display itself, but one thing the guide does not mention is, if you lose contact with some of the pins on the display, it will lock up. Cycling the power is the solution. When you have the correct display, with temperature, LDR and switches working, then you can solder the display. It will of course "work" receiving the data from your emonTx and sending the temperature to your base without the LCD to display the data on.

It might be advisable also at that point to load a simple sketch to check that the radio works too. If you need a sketch, I can probably throw one together for you.

Here are the approximate GLCD pin voltages. Particularly important is the voltage multiplier - pins 9 - 12.

1 - 8.53
2 - 7.34
3 - 6.09
4 - 2.48
5 - 1.24
7 - 1.63
8 - 7.07
9 - 4.44
10 - 1.65
11 - 9.68
12 - 10.96
14 - 3.27
15 - 3.27

The circuit diagram is, as always, available on the Wiki.

Both sides gave the same low values what I remember. You are right, that it is a bit silly, but is a small part of my project of 5 EmonTX, 2 EmonTH, raspberry, 30 sensors, etc. Besides that I have to continue, so I took my 2% loss. Of course I kept the LCD, Processor, case, so I threw only 20£ away.

If there is a test sketch for the Radio, it would be nice.

As fast I have received the new GLCD, I let you know if it fits with the approximated GLCD pin voltages.

I have received a new one, soldered it (except LCD) together and the test sketch is working :)

Only I still have to check the radio and the lcd. I have tried to put the LCD into the board without soldering, but no reaction of the LCD (yet).

My voltages are completely different as yours:
1 - 3.26
2 - 3.51
3 - 3.29
4 - 0
5 - 2.58
6 - 0.04
7 - 3.29
8 - 0
9 - 0.72
10 - 0.43
11 - 3.28
12 - 3.15
13 - 3.28
14 - 0.23
 

My voltages are of course with the LCD soldered in and working. And those are the LCD pins, not the processor. If the test sketch works, and it's printing the temperature to the serial port, it's looking good so far. I'll try to throw together a radio test sketch for you before long. (It'll be a standard GLCD sketch but writing the values received to the serial port and not to the LCD.)

Getting the LCD to work not soldered in requires a certain knack and it's easier if you have help. First, the two groups of 3 pins do nothing, you're only interested in the single line of 20. You push those into the holes, then tilt the 2 x 3-pin edge away from the main pcb so that the 20 pins just start to bend. If you're on your own, you need to hold everything in this position and apply power. If you have an assistant to apply power, it's so much easier.

[Edit]

Try this (attached). It reports the data received in bytes, also the power and voltage values are decoded if your emonTx is running the "Discrete Samples" sketch. As usual, set the radio module type and frequency to suit your set-up, and the node ID of your emonTx if you require the data to be decoded. It also transmits the temperature as 2 bytes from node 20, if anyone is listening.

Thanks Robert! It is working, LCD and radio are correct functioning! Now I'm going to focus on the sketch.