Hi Guys,
We have been using Moteino R4/R6 in our prototype devices. Where Moteino is connected to a 5V power supply and have RX/TX going to ESP8266 module similar to one on the link below.
https://www.amazon.com/HiLetgo-Internet-Development-Wireless-Micropython/dp/B010O1G1ES
Basically now the product is moving into production phase, hence we need to plan for EMC testing. Moteino has RFM69HW which I think are already CE/EMC approved same goes for ESP modules they have CE/FCC logos on them.
Moteino is using 868MHZ.
So my question is how do I prepare our device for EMC testing. Product will be sold in the UK mainly, maybe EU. Are there basic tests we could do in-house before booking the authorised LAB. We have simple tools like Oscilloscopes(not every high end). Also we have a competent team so we could test the device. I have spoken to a lab, they willing to rent us tools for pre-complaince testing. So Just want pointers basically where to start and what to test. I have read a lots of blogs/articles online, they all seems to give very abstract information.
You will likely need to cert your product as a whole in any case.
The Moteino is not CE/EMC certified. If it were, perhaps it might help speed your final product certification.
If the wifi module is already certified, it will be interesting to know if the frugal Moteino will add any challenge to certification, since its so low power and max TX currents are much lower than wifi.
Your main source of EMI will be from cables and interconnect, and primarily from common mode radiation. There may be issues with conducted EMI on the power cables too, so be prepared to add some common mode filtering.
I would suggest all I/O cables from the Moteino comes from one side of the module only, otherwise you can get a dipole antenna effect with the Moteino acting as a noise source between the two cables. The same argument goes for any other device.
The basic rule of thumb is that when a cable exits it should have a chassis (or digital if it is the same) ground as close as possible to the exit point, with low inductance lead surface mounted capacitors from the signal to that ground to 'short' any RF energy back into the chassis, and use ferrite beads just prior to that to form an LC filter. Alternatively use a ferrite bead followed by an ESD protection diode right at the exit point (the ESD diode has intrinsic capacitance, so that kills two birds with one stone). The goal is to keep the reference ground on the module as clean as possible, and to remove RF on the signals exiting by making those as low impedance to that ground at RF as possible. Bear in mind even a tiny wire or thin lead a few mm long at 1GHz will have a significant impedance.
It will need to be tested in a representative setup, while running representative code.
BTW it is actually possible to use a very low cost spectrum analyzer with a current transformer to detect conducted EMI on cables. I have used the wide band RFExplorer with an additional ultra-low noise preamp to remove conducted noise from power supply noise by placing capacitors between the power signals and the cable shielding. It's good for showing relative improvements in filtering and will get you a good idea as to what is likely to work at the test house.
http://rfexplorer.com/
Mark.
Thanks for your reply Perky.
I was looking at low cost LISN devices like the one below.
https://www.tekbox.com/product/tekbox-tblc08-50uh-lisn-cispr-16/
And a lowish cost spectrum analyser like Rigols's 815
https://www.rigol-uk.co.uk/Rigol-DSA815-TG-Spectrum-Analyser-p/dsa815-tg.htm
But this particular analyser has range upto 1.5Ghz which is Ok for RFM69 which has frequency of 868 Mhz, but it won't cut the mustard for Wifi which requires range of 2.4Ghz ?
Finally you mentioned 'current transformer' and RF explorer, with current transformer I guess you mean LISN device ?
Quote from: malik on November 12, 2018, 07:25:30 AM
I was looking at low cost LISN devices like the one below.
https://www.tekbox.com/product/tekbox-tblc08-50uh-lisn-cispr-16/
Yes, I should have been clearer. I used a home made LISN, they are very simple and circuits are out there. You'll need one to give a defined source impedance on the power supply when measuring conducted noise.
Quote from: malik on November 12, 2018, 07:25:30 AM
And a lowish cost spectrum analyser like Rigols's 815
https://www.rigol-uk.co.uk/Rigol-DSA815-TG-Spectrum-Analyser-p/dsa815-tg.htm
But this particular analyser has range upto 1.5Ghz which is Ok for RFM69 which has frequency of 868 Mhz, but it won't cut the mustard for Wifi which requires range of 2.4Ghz ?
I took the view that to get absolute accurate measurements the setup environment needed to be very well controlled and accurately calibrated. I decided that it wasn't really necessary during the experimentation phase to reduce the EMI, it was possible to do that with a much lower cost analyzer and less rigourous setup, as long as the setup was consistent between measurements and the analyzer had a resonable degree of accuracy (I actually calibrated it with a known high end spectrum analyzer first, and the RFExplorer was remarkably accurate for it's cost, at least for my needs). The RFExplorer/3 ranges from 15MHz to 2.7GHz.
Quote from: malik on November 12, 2018, 07:25:30 AM
Finally you mentioned 'current transformer' and RF explorer, with current transformer I guess you mean LISN device ?
Sorry, I actually used the home made LISN for measuring conducted noise, and a simple toroidal current transformer over the cable for measuring radiated. This latter one wasn't quite as accurate as the bought ones but definitaly shows what works and what doesn't. The dynamic range was increased by adding a ultra-low noise amplifier (HEMT based), that gained an extra 18dB at the low end.
Do some searching, there are articles showing how to use the RFExplorer as a DIY compliance tester, that's where I got the plans for the current transformer and the LISN. Below is a picture of the basic kit.
Mark.