Hi all, I’m working on an electronics project where I’m running multiple LED strips controlled by a microcontroller, and I’m facing possible EMI / noise issues when data lines run in parallel and over longer distances.
I’ve been reading about board level shielding and electronics design techniques and wanted to ask:
- When is board level shielding actually necessary in small embedded systems?
- Is EMI more likely to affect data lines like LED DIN compared to power lines?
- What practical shielding methods work best (copper tape, grounding, enclosure, PCB layout)?
- At what point does shielding become overkill vs proper wiring/layout solving the issue?
Would appreciate insights from anyone who has dealt with similar EMI or signal integrity problems in real-world designs.
This is my opinion only.
IF you are generating EMI to the level it is causing interference, it SHOULD be fixed. Not only can it affect your project, it can also affect other things around it. EMI will be one of those things, that if you want to go to market with a project, it will need to pass EMI testing.
That said, are you sure your issue is EMI ?
If you running long wires (eg. for DIN) then the voltage drop may be causing issues. there are plenty of examples of people runing 3.3V logic out of a microcontroller into a WS2812B LED strip. The WS2812B really should be using 5V on DIN, but over a short distance 3.3V works; but over longer runs expect issues. On hack fix for this is to put one LED on your controller end to “regenerate” the Data stream to 5V. But you would be better ensuring the level are correct with a proper 5V driver.
If its 5V, you can still have voltage drop issues over longer runs and may need to “regenerate/repeat” the signal along the path to restore.
In simple terms, I would verify what sort of issue you actually have, cross talk V signal degradation.
Hi rixcy
Probably never. A wire has to be carrying a fair bit of current to radiate anything much
Would you believe “Steel”. Grounding and PCB layout are important. As for enclosure you will find that to do it properly those aluminium enclosures you see in rack mounted devices that hold many boards actually have a grounded STEEL lining if they are housing sensitive boards. Military and anything that could be subject to covert “sniffing” has to be treated that way. Even connectors going to the outside world are filtered. Yes with tiny LC filters on each pin. Built into the connector (usually sub D types).
I think mostly those metal screening boxes you find on professional boards, particularly test instruments, are meant to keep interference OUT of sensitive stages. Getting back to military stuff we find the opposite where filters are used to keep sensitive information IN to obviously prevent covert snooping. The first line of defence here is to put such devices into STEEL boxes and transport cables in STEEL conduits.
Cabling and wiring techniques can go a long way to preventing unwanted radiation and receiving unwanted interference. Keep “unbalanced” wiring to a minimum and as short as possible. If you need long signal lines try for a “balanced” system. Like I2C (unbalanced) should be very short and is better converted to a balanced line like RS422 or RS485 for longer runs. Long Audio runs should also be balanced lines. Preferably shielded, the shield grounded ONE END ONLY.
There are lots of things you can do and each one will depend on circumstances. It can be a real pain
BUT.
This is a statement worth thinking about. You could worry yourself into some sort of breakdown looking for something that is not there.
Firstly satisfy yourself that you really have an EMI problem. Not quite sure how you are going to do that without some specialised gear and the knowledge to use it properly.
Indeed on this forum there has been a lot of discussion on removing “noise” and other such nasties when the problem the whole time was the incorrect use of an oscilloscope probe like an earth connection too long or connected in ths wrong place. In other words the problem was being put there by the instrument and the technique in using it.
Also heed Michael’s comments re 5V versus 3V3.
Another wise move which says what I said above put in a slightly different way
Cheers Bob
