Electronic Noise and ARDUINO

[Speedy Builder5Participant @speedybuilder5]

Noise upsetting ARDUINO

[Speedy Builder5Participant @speedybuilder5]

What a disappointment, I have been developing a crude engine temperature and RPM application for an ARDUINO UNO. Its been a moment working out 1 Wire temperature measurement, I2C LCD screens and Interrupts, but finally it was time to move the development board into the car and see what we could see.

Mounting the DS1820 temperature probes in various places and finding out the for some reason (unresolved) the UNO would only read 4 out of the 5 probes in spite of all the addresses etc being correct. Oh well, 4 probes working wasn't a bad start.

Then mounted the Hall Effect sensor to measure camshaft speed (2 magnets would give RPM.) OK so far.

Now we get the engine running, and the screen goes crazy – OK, it is an unsurpressed 1932 Austin 7. If I move the UNO about 6 feet from the engine, most things settle down. I want to keep the car pretty original (Un Supressed) . If I shield the actual ARDUINO UNO with a metal housing, would that make much difference, or am I on a hiding to nothing ??

Ps, I do have a fly lead attached to Pin 3 which is an external interrupt pin – could this be an "Ariel" ?

BobH

[Andy CarruthersParticipant @andycarruthers33275]

Are you using screened cables between sensor and Arduino? I suspect pickup on cables

[Former MemberParticipant @formermember32069]

[This posting has been removed]

[SillyOldDufferModerator @sillyoldduffer]

Hi Bob,

Neat job!

Yes that lead on Ext 3 is an aerial, and fairly sensitive too! Microcontroller inputs are high-impedance, sometimes useful as when making a capacitive touch switch, otherwise can be a right pain.

Dealing with electrical noise can be tricky. But the first step as you say is to put the Arduino in a metal box. Quite likely with a noisy unsuppressed car engine it will be necessary to shield the sensor leads as well. If the set-up is powered from an external source like the car battery, decoupling the power line, or filtering same, may be needed too.

A bit like this:

Ordinary microphone cable should be OK for a short run ( a metre or so ). It's high capacity stuff and a long run (several metres) might distort sensor signal signals enough to stop the Arduino detecting them properly. Ethernet cable is good for this job; it contains 4 sets of twisted pair, and the better cables are shielded as well. Probably cheapest way to get a short length is to buy a patch lead from ebay and chop the plugs off.

Dave

[David JuppParticipant @davidjupp51506]

Quite a bit of information around the internet regarding how critical the cable and how it is used can be for 1-wire. If not all sensors are working, that already suggests a problem. Noise on top of that won't help.

[AdrianRParticipant @adrianr18614]

You defiantly need to shield the cables and also decouple the supply with a capacitor as shown by Dave. Another thing you can do is to add clip on ferrite choke to the supply and sensor wires. as close to where they enter the shielded box as possible.

If you are running from a battery and it is inside the box, you would not need to shield it.

The sensor cable heading into the engine bay is most important cable to be shielded and decoupled. Good quality audio coax Sould be fine. Without seeing the circuit I cant say for sure, but you could also put a low value capacitor to ground on the sensor and interrupt.

Of course suppressors on the HT leads would also help no end and also be nice to other road users radios.

Suppressor capacitor on the dynamo would also help.

 

Adrian

Edited By AdrianR on 21/06/2019 15:07:22

[Neil WyattModerator @neilwyatt]

Seems a bit odd to want to install real time engine monitoring but not wanting to fit suppression to the motor!

Suppressed spark plug leads might help and won't look out of place (unless it has screw connections to the plugs)

[Speedy Builder5Participant @speedybuilder5]

Nothing ODD about wanting to know more Neil. 9v DC is from the PP9 in the photo. The DS1820 probes are the waterproof type with no shield, so clip on chokes might be helpful.

Remove the "Aerial" from the interrupt line.

I had a thought, Run it from the laptop and see if it is the Arduino or the display as the code includes lines for serial output tothe laptop screen.

[Maurice TaylorParticipant @mauricetaylor82093]

Look up Speeduino on google ,this is a diy engine ecu based on Arduino mega,lots of info on screened cables etc.I have just built one to fit on Fiesta Mk1.

[Speedy Builder5Participant @speedybuilder5]

Had another go today, at first, it did make a difference with the bonnet closed and doors shut, but not for long ! I then disconnected all the sensors – No difference, so the noise is affecting the UNO or the display.

If I shield the complete project board (except the display), and add a plate bonded to the shield over the UNO, battery and jumper leads etc, would it matter that the edges of the MDF project board were unprotected ?

Also, can the shield be made from perforated metal, or does it need to be solid ?

As there are no physical connections to the car's electrical system, should the ground of the Arduino and the shield be tied to the chassis of the car (Negative earth) ?

Must try my other idea to see if it is the LCD screen acting up or the Arduino.

[Anonymous]

The shield can be perforated. A rule of thumb is that if the apertures are less than a tenth of the wavelength then the shield will act as if it is solid. As the frequency rises the apertures start to act like slot antennas and finally just as a hole.

Andrew

[Speedy Builder5Participant @speedybuilder5]

Thanks Andrew, I will get a rule out and measure the wavelength, seriously, I am well out of my depth here as I have no idea what the interference is, Dynamo, dynamo control box, spark plugs, soil, contacts etc and I might just be chasing my tail. So for a start, I will put it all in a "tin" box, and if that doesn't work, leave it in the box !! But just in case, do I ground the box to Vehicle Earth (negative earth system) or to the GND of the Arduino, or both.

Bob

[AdrianRParticipant @adrianr18614]

Shielding means every part of the electronics has to be inside the metal shield, this includes all the cables to your sensors. Shielded cable wraps the signal core(s) in a metal tube and keeps all the interference out.

So you need to put your electronics inside a metal enclosure, and use shielded cable for all your sensors. The outer sheath of your cables must be connected to the enclosure.

It is normal practice to connect the 0V of the electronics to the shielding. Then the sheath of the cable can be used as the 0v signal wire.

Of course you have to make a hole in your enclosure to see the screen. So you may need to put bulkheads inside your enclosure if you are still getting interference.

The use of ferrite beads or clip on chokes will reduce interference travelling along wires.

[Maurice TaylorParticipant @mauricetaylor82093]

I would remove the fan belt and see if it makes it any better and fit a new condenser across the points and make sure suppressor plug caps are ok, before I made screened boxes etc.

[Anonymous]

I'd disagree, a complete shield, including cable outers should not be connected to 0V. If you do so you simply introduce noise into the 0V line. You might connect the shield to earth, but that's not the same thing as 0V.

A lot of EMC issues are a matter of sucking and seeing. Spent many a "happy" hour in EMC chambers doing experiments. The most common cause of problems for radiated and susceptibility isues are cables.

Andrew

[Maurice TaylorParticipant @mauricetaylor82093]

Could you please explain the difference between the 0 volt line and earth on car. Surely they are the same thing.

[Maurice TaylorParticipant @mauricetaylor82093]

Could you please explain the difference between the 0 volt line and earth on car. Surely they are the same thing.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Maurice Taylor on 22/06/2019 20:36:59:

Could you please explain the difference between the 0 volt line and earth on car. Surely they are the same thing.

.

Most cars are isolated from earth by four rubber tyres !

MichaelG.

[Speedy Builder5Participant @speedybuilder5]

Maurice, the car has an earth / ground – the chassis, and in an Austin 7 like mine it happens to be a Negative earth. The Arduino computer has a 'Ground' or 0 volt. So what Andrew is saying (I think) is that the shield should be connected to the vehicle ground, and not the 0 volt circuit of the Arduino, but by experiment this might change ?

BobH

Edited By Speedy Builder5 on 22/06/2019 21:22:21

[Maurice TaylorParticipant @mauricetaylor82093]

I see now why you have 0 volt line and earth , your 0 volt is on the arduino ,as it’s powered by a separate battery.As I said in a previous post ,I would try to eliminate the interference before doing anything to the arduino.

[Peter BellParticipant @peterbell11509]

Looks an interesting project. I've found it useful to listen to interference like this with a radio off station especially on am. You can also gauge where its from and intensity which may help with siting the application.

Recently found an Arduino nano was being reset when I switched my soldering iron off which contains a conventional transformer for the psu. I could hear the click on my fm radio, addng a supressor to the transformer cured the click and the nano doesnt reset.

Peter

[duncan webster 1Participant @duncanwebster1]

The igniter on a gas hob is another excellent source of interference. Using screened cable and putting a capacitor between the Arduino input and ground solved it

[Author]

Posts