Archived Forum

Pics..

Please click for bigger versions.. 

Should add when you plug the clamp into the board it does see it as it sends the data to the GLCD but it looks like noise as it shows 8kW+ then drops down to 1kW then 500W (example). 

I've had a little problem with corrosion on the barrel plugs - unscrew them to see if they look okay.

Thanks for the idea Stuart..

but there is no corrosion inside the barrel plugs they look good as new.. Also this does not explain why both of them work in CT3 plug.

 "the board it does see it as it sends the data to the GLCD but it looks like noise as it shows 8kW+ then drops down to 1kW then 500W (example). "

That is most likely the digital filter on the input settling. If that happens when you power up the emonTx, it doesn't help. If it happens when you plug in the c.t., it might indicate the analogue input is at least reading something. 

When you say CT 3 "works", I assume you mean it reads a sensible value, thus ruling out a fault in both c.t's and their cables and plugs?

Have you reloaded the sketch? (Just in case it's 'lost its marbles'?) I presume you actually loaded another sketch to test C.T.3?

What voltage do you read on the burden resistor? - it should be 1.65 V or thereabouts.

Write a 'one-line' sketch to read the analogue inputs and print to serial. If you read 512 or thereabouts, the analogue input is probably OK. For a more complete test,  wire a potentiometer (any value 1 k < R < 100 k) across the battery connection points (3.3 V) and connect the wiper to the tip connection of the c.t's socket.  If varying the pot. from one end to the other gives you readings from 0 to 1023 (or thereabouts) the analogue input is OK. Else you have a dry joint or your chip is dead.

A dry joint is a possibility (especially after a year or two - you don't say how many months). You need to touch all the soldered joints with a hot iron and a tiny amount of flux (or new solder if you are using the flux-cored variety - who doesn't?).

What is the temperature in the fuse box?

 Robert Thanks for the reply..

The unit has been in operation since end of March 2012

Im using the standard   emonTx_CT123_Voltage 

I Admit its a bit older than the current one on github but iv never had to update it in the months iv had it installed..

This has the 3rd ct enabled (I forgot to disable it..) and when pluging in the import / export CT or Solar CT i get the correct readings from this plug reported to Vis.openenergymonitor.org and to the serial port..

Iv current not tried to reload the sketch.. WIll give that a go tonight and see how that works..

Ok now im going to show how stupid i am.. 
Where is the burden resistor?  Iv read lots about this but to be honest no idea where it is... Im using the standard kit off of the website.. Noting fancy.. The only hacks are to the GLCD.

Will also try the wiper with a 1k pot etc.. 

Will try the dry joint option but it seems weird that both CT1 and CT2 dropped at the same time on the same day..looking at the graphs all was fine and then suddenly 0 all the way.

The temp of the fuse box is a few degress lower than our house.. Its in the hallway but its got a wooden box round it.. Looking at the alertme graphs never above 25c and never below 15c 

"Ok now im going to show how stupid i am..
 Where is the burden resistor? Iv read lots about this but to be honest no idea where it is... "

You need to look at the circuit diagram and layout of the pcb. The burden resistor is the 15 Ohm or 18 Ohm right across the c.t. - it's R6, R3 & R16, there are 3 groups of 3 resistors, 2 x 470 k and the burden, close by the c.t. sockets.

"Will try the dry joint option but it seems weird that both CT1 and CT2 dropped at the same time on the same day..looking at the graphs all was fine and then suddenly 0 all the way."

I agree, most unlikely that 2 dry joints have happened together, so probably not worth resoldering at this stage. How do you get your mains electricity, is it on poles and overhead lines or underground? Were there any thunderstorms about at the time? If the answer is yes to either or both, and the pot. test is negative, I would tend to favour the theory that the c.t's have seen a spike of current that has blown the ADC inputs of the processor. Time for a new processor. 

And if it is the processor that's blown, I don't really know the answer to the next question, "What can I do to stop it happening again?"  You could consider something like this: http://uk.farnell.com/fairchild-semiconductor/p6ke11ca/tvs-diode-600w-11...  in parallel with the burden resistor, though it may well be cheaper to just replace the processor.

Under ground cables. 1930's house average surban area.  (big green shed, (transformer) 35+ houses down, another 15 up the street also before main road.

We have had thunder in the area but not at 1603 (according to vis.openenergymonitor.com) My computer was on also (working from home currently have an ear infection) and did not notice anything happning. 

The high readings only happen when the emonTX is on and you unplug and plug a CT back in.. 

Will try looking at it all again in a bit.. sorry side tracked by work.. Gerrr 

 Well after reloading the sketch 

all the problem was, was the 3.5mm sockets.. bendin the contact in them back in slightly works..  Weird how both went at the same time.. Maybe there is a better connector out there?

 Hi Robert,

Sorry to hear of the problem. We are aware that the 3.5mm sockets currently used on the emontx are not the best. Iv been looking at alternatives. Iv recently just finished a prototype design for an emontx arduino shield board. On this board iv gone for using the same 3.5mm jack as on the raspberry pi. Its a bit bigger but has a nice plugging action and doesn't cost too much more. I'll leave look into fitting this to future emontx versions but it would neccesitate making the PCB bigger. The emontx smt will also use a better jack plug

Sorry about this, it sounds like it was a tricky issue to track down. More often than not the sockets work fine.