Archived Forum

If you have another board (arduino) you could try swopping the chip and programming it in the other board. That may rule out a faulty chip, and if it program's OK return it to the emontx.

Paul

Good idea Paul, I will give that a try later to eliminate one more variable.

Thanks

So I took the processor out of my emonTX and dropped in the one from the temperature module, uploaded a sketch with no problem so I guess its back to the drawing board and time for the magnifying glass and multi-meter. I dropped the now programmed processor back into the temperature node on the off chance that it might just be a problem with the link from the header to the chip but I still get nothing out of the remote temperature board.

If anyone has suggestions where best to start I'd appreciate it but if not I guess it will be a case of strip down and start again.

UPDATE.......

OK, so once I started looking more closely I noticed I had put the processor back in upside down, I flipped it round and put the batteries back in and suddenly I have two new feeds on my emoncms. So the board itself is working fine, but I can't upload programs or monitor the serial output. This narrows down my search a bit. Will post more info later if I discover what is wrong.

Keith

SOLVED - found a faulty track on the PCB.

I had no connection from the 10k resistor to pin1 on the ATMEGA. I just soldered on a jumper wire and now it is working.

Don't know if its a general problem with this batch of boards or if I was just unlucky. For reference the board is an emonTX v2.2 marked made in UK, Designed in North Wales

Hello Keith, thanks for the heads up on the faulty PCB, Just looked over the current batch here and they all appear ok with continuity so it must have been just an unlucky PCB, happy to send a replacement PCB if it would be of use.

Edit: Was there a surface scratch on the track, as in the picture?

Hi Trystan,

the board is working fine now I have a jumper applied. No obvious signs of damage that I can see. No need for a replacement this time. Many Thanks.

Keith

I've just spent a frustrating but ultimately successful couple of hours solving the same problem (the out of sync error in IDE). The weird thing was, when I first completed the board I uploaded a blink sketch to test it and everything worked perfectly. Then it stopped, and occasionally came back for a few minutes.

During my search, I found this handy test to prove everything from the IDE up to and including the FTDI is working:

http://forum.arduino.cc/index.php?PHPSESSID=hus37jspuja3s479s388vj4010&t...

Mine passed the test, so it had to be something on the board (emonTX). Tried another 328 incase I'd mess that one up (which was risked ruining two chips, I realised in retrospect!), but that made no difference.

I found a cople of slightly ropey solders, which also made no difference, but I realised that I'd pushed the ceramic caps in too far and the insualtion/casing material that extends down the legs had gone partially into the PCB hole, and was preventing the solder flowing through and making a good join between the two pads on opposing sides of the board. I pulled them out and corrected this and it works perfectly. See attached drawing if that doesn't make sense.

I wondered if it might be worth mentioning this in the build guide? It had not occured to me that it was important to connect pads on opposite sides of each hole, so other novices might value this nugget of info.

Hope that helps someone sometime!

I've just checked on several bare V2.2 emonTx PCBs, and the solder pads on opposing sides of the board all seem to be firmly connected together.  The internal face of each hole appears to have been reliably coated with solder.  Some connections between the two sides do not involve a component being soldered there, so the plated-through technology must allow for this.

By allowing your capacitor(s) to extend too far into their holes, it is possible that one of their legs was failing to make contact with the surrounding solder? When slightly withdrawn and re-soldered, connection would then be made and normal service resumed.

A favourite bad-joint of mine is caused by pushing fine insulated (Kynar) wire too far through stripboard.  The protruding insulation can cause the solder to 'pop' away from the end of the wire are the iron is withdrawn.  With my eyesight now not what it used to be, I find it helpful to inspect each joint with a x10 "loupe" before applying power.

FWIW, the tracks between R1 and the Atmega's pin 1 are fine on these boards too.

As calypos_rae says, the holes should have been plated-through, and it is not necessary to solder both sides. If the holes were not plated right through from one side to the other, all round, then that's a manufacturing fault. It might help to look at unused holes to see if any others show the same fault - if you shine a light from the back of the PCB and angle it so that the light reflects on the plating, it's usually it's very easy if it is present.

Problem too with a new Low Power Temp Module... Temp: 32.4, Volt: -12.70 fixed, not moving...

Uploaded last firmware emontx_lowpower_temperature.ino ....

Your opinion ? Missoldering something or a firmware compilation issue ?

Thanks

Did you set the correct processor in the Arduino IDE?

Have you done the things in the comment at the top of the sketch - link PWR and Dig7 on JeePort 4, and no regulator?

What voltages do you measure from your battery?

How are you reading Temperature and Voltage?

I setted "Uno" in Arduino IDE.

I soldered PWR to Dig7, yes
3.12v from batt.

Those are temp and volt read from emoncms's feeds.

Node seems to transmit every 20sec temp & volt.
Tells me perhaps not a firmware upload issue.

Dead DS18B20 ?

I soldered PWR + Dig7, do i need to mdify the .ino part:

pinMode(7,OUTPUT);                                                    // DS18B20 power control pin - see jumper setup instructions above
  digitalWrite(7,HIGH);

??

You should see the battery voltage in mV in emoncms, ± 9% due to uncertainty in the actual voltage of the internal band gap reference. I think you are reading the voltage as Temp = 32.4 (it should be 3.24) and temperature as Volt = -12.70.

-127 is the "failure" temperature reading, so that is saying the DS18B20 is not responding. Do you get an error printed to the Serial monitor when you first start up "Unable to find DS18B20 Temperature Sensor"?

What happens is, in order to save power, the temperature sensor is fed from PWR (normally 5 V, as the normal default) but by omitting the regulator and linking PWR to DO7, the processor can switch PWR and therefore the sensor on and off inside the sketch with digitalWrite(7,HIGH) etc.

I can see what you mean about the through-plating, so maybe there was another generally dodgy solder that I corrected in the process.

Thanks for your help anyway!

Exact ! enable to find DS18B20, checking my solderings...

A way to check the DS18B20 itself ?

teo770:

You might like to look at this:  http://openenergymonitor.org/emon/node/2516#comment-14563

"Exact ! enable to find DS18B20, checking my solderings..."

enable or unable ?  enable = to make able; unable = not able, impossible

"A way to check the DS18B20 itself ?"
No, it sends a digital value as a stream of 0 & 1 on one wire.

If you have an oscilloscope, you should see a short pulse ~3 V for a few milliseconds on V_DD as the reading is taken.

UNnable... sorry.

Up to your link, i should try this 2.2K resistor test ? Wich one is to replace 4.7 -> 2.2K ? The R1, right ?

Voltage seems right in emoncms, Temp is returning -127 always.

No, the resistor to change is R9. It is the one that lies parallel to the Atmel processor next to the DS18B20. The easy way for a quick test is to wire a 3.9 kΩ in parallel with the 4.7 kΩ:

(You have the DS18B20 where I have a jack socket for a remote sensor)