Archived Forum

My friend in Ayrshire has exactly the same problem with Wi-fi.

I think the emonTx runs at maximum power anyway, so all you can do with it is to reduce the bit rate of the transmissions (and of course you need to do the same with all the radios in your set-up), which will reduce the bandwidth needed and so concentrate the power over a narrower band of frequencies.

As reception is marginal, it would be worth trying that first as you might get just enough improvement. You'll need to delve into the depths of JeeLib, or look at MartinR's PLL sketch to see how to set up the radio without JeeLib. The pertinent bit is:
rfm_write(0xC606); // approx 49.2 Kbps, as used by emonTx
you can change it to
rfm_write(0xC657); // approx 3.918 Kbps, better for long range
(according to Martin.)
The RFM12B data sheet has all the details.

In a similar vein, after a long hiatus, I'm starting to look again at my project to monitor my workshop.  When I built it, I inserted huge volumes of Celotex (nasty dusty stuff, especially in a ceiling.....), and my workshop is a wonderfully inadvertent Faraday cage.  So I was wondering whether I should get the emonTx with an external antenna socket, and locate the antenna near a window with line-of-sight to the house.  I'd need about 8m of cable to get from the emonTx to the window.  Is the performance likely to be significantly degraded by the use of such a long cable?

I've not found any data regarding the output characteristics of the RFM12B, so I can only speculate, from the results of tests with external aerials that Glyn & Trystan did a while ago, that the main problem you'll have is not losses in the cable per se, but mismatch between the cable and the RFM12B output. What I did for a friend with 2ft thick stone walls was extend the CT cables and put the emonTx where it had a line of sight. I felt a lot safer running 2 x 12 m 'microphone' cables for the CTs, than running 12 m of co-ax for the RF.

Thanks for replying!

That would be feasible.  Assuming I took this approach,  I have a few follow-up questions.  Could you expand on "microphone cable"?  I presume you mean a screened type - is there anything particular to look for/avoid?  (What did you use?)  Did you cut the jack off the CT and hard wire it to the extension lead,  or would you recommend a terminated lead with a plug/socket?   If I'm making up my own cable, what should I do with the screen??

Use a twin twisted and screened cable, e.g this, and earth the screen at the CT end only, no connection at the emonTx end. I unsoldered the plug and used a 'chocolate block' at the CT end, and re-used the plug at the emonTx end. Connections are to the plug tip and sleeve only. It's best to keep the polarity, but if it's wrong all you need to do is reverse the CT on its cable. The factory wiring of the CT is a bit hairy quality-wise at the plug. The cable screen may or may not have been connected, deliberately or fortuitously, to the plug sleeve; you can safely ignore that whichever, only the red and white cores from the CT are important. If you're worried about pick-up, check that the screen is not connected to the red and white cores, then earth it.

It's best not to have any more plugs and sockets than necessary, and had it been feasible I'd have soldered and sleeved the connections, not used a terminal block.

Thanks again Robert.  Presumably this cable would also be OK?  (Seems cheaper with more CSA, and halves the number of cables I need to support 4CTs.)

Yes, that would be fine. The important factors are the twin twisted cores and the overall screen. A braided screen would be marginally better than lapped, but it's unlikely to make much difference in this application.