Question about interference

[kefirtribe]

At page 11 in the document hosted by this page (about how to avoid interference problems) it says earth the screen in both ends. But what am I supposed to earth it in? The chassis?

And to protect the cables to pick up radiated interference, what shall I use? would aluminium foil and plastic pipes work?

[tunna]

Question about interference

Bulldog has used a spray on antiference aerosol available from RS.spray it on the inside of armour,
worked for us -der boyz

[flipperbot]

Question about interference

Best way to stop interference is to keep the radio gear inc aerial away from the main battery and motor wiring/the motors themselfs. Most radio problems come from within your own machine.

If you have to run leads from the radio to the switch gear etc make up your own leads which are 100% screened, you can buy the radio plugs from any good model shop.

[talulah]

Question about interference

To stop interference you really have to design it out in the first place. If you are having problems at the end of the build it is unlikely youll be able to help it much with spray on screens, or by earthing a cable or two.

Have a look at my EMC page at http://homepages.which.net/~paul.hills/Emc/Emc.htmlhttp://homepages.which.net/~paul.hills/Emc/Emc.html, and also some info on the radio page here: http://homepages.which.net/~paul.hills/Radio/Radio.htmlhttp://homepages.which.net/~paul.hills/Radio/Radio.html. Theres also a good document here: http://www.technobots.co.uk/technical/guidance/EMI%20and%20how%20to%20deal%20with%20it.pdfhttp://www.technobots.co.uk/technica...o%20deal%20wit h%20it.pdf

In short, here are the main points, in order of importance (most first):

1. Isolate the radio & low power electronics from the power stuff. That means use separate batteries for each, and make sure the drive signals to the speed controller use opto-isolators. Dont use Battery Eliminator Circuits (BECs).

2. Make sure the antenna is the right length, and is properly connected to the receiver using the correct cable. Dont use any old wire, make sure the antenna is as far as possible away from other metallic parts of the robot. Dont shield the antenna under the metal armour!

3. Keep the antenna and receiver away from cables carrying large currents, and the motors.

4. All the other tips in the documents identified above!

Cheers & good luck!
paul

[kefirtribe]

Question about interference

Well, I havent built it yet, so no problem there. Anyway, thanks for your advices, but no one really answered my questions.

[big_nipper]

Question about interference

To Answer your question you should connect the Earth screen to the chasis at both ends and also connect your Rx battery earth to chasis but dont connect the main battery to chasis. At least thats the way we did it. In effect the Rx & Rx battery is also in a metal box so that the Rx circuit is completely enclosed with a screen from the rest of the wiring, less the wires to the speedos through opto-isolators.

[talulah]

Question about interference

I was always taught in my EMC lectures at university that for protection against low frequency interference (as motor interference is), the screen should be earthed at ONE END ONLY. This is because at low frequencies grounding at both ends achieves nothing, but does introduce a ground loop around which induced currents can flow which can be induced back into the signal wire.

You could try it both ways and see which way works best.

[talulah]

Question about interference

Ive just had a chat with one of the electronics engineers in work here. He has 30 years experience of comms & EMC issues, and he agrees that grounding at one end is probably best. He also says however that it may be worth trying grounding both ends if the single end method doesnt work.

As for what ground is on a moving robot, you have to be careful. If you ground to the armour shield, then the shield should be connected to the negative terminal of the electronic/Rx battery - NOT the power battery. Then all cables that are grounded to the shield should be done at ONE POINT ONLY (star grounded). No way should the power battery be connected ot the electronics battery.

If the armour shield has to be connected to the power battery, then dont ground the cables to it. Ground them back to the electronics battery negative post.

Most of all in EMC issues, make sure you are lucky!

[_jl_]

Question about interference

While we are on this subject, I have a question....

I wired up my motors and relays and stuff on a bench, and tested. It worked fine even though I am using a simple 7805 and all grounds connected together...have not even mounted a single capacitor...

Now...the question...how much worse will it get when I put it into a metal chassis?

[big_nipper]

Question about interference

Ah, now I remember, thats what I was told as well.

Only reason why its earthed at both ends is because the rx Areial is connected through a SMB connector to a coax cable to bring it out of the metal box and hence the box is earthed through this connector. The coax then winds its way past the removable link and all the power wires through the middle of a moving sprocket into the arm. To create a 1/4 wave Dipole Areial the earth must be connected to a ground plate (chasis) at the point where the Areial leaves the coax cable. It would have been better if the ground plane was not already earth but connecting the areial to the rx with 50ohm coax to this areial gives approx (If I remember my areial theory) 50ohm also giving the best matching without additional components. The problem comes from not knowing the impedance of the Rx which is probably not 50ohms so there is a mismatch there with could casue problems.

The Rx after all is designed for a single wire areial. There are many possibilities and basically I used the old EMC trick, guessed, and it worked for us.

[talulah]

Question about interference

Mark - The natural impedance of a quarter wave dipole is 37 Ohms. The receivers are probably matched to this, or maybe 50Ohms. The mismatch between 37 and 50 Ohms isnt that bad and can be ignored for most purposes.

Trial and error is certainly the best way of finishing off the design, but it is important that basic design principles are followed to start with. Trial and error wont fix a basically bad design, but it can improve a good one!

Jonas - following on from that last comment, you may or may not have problems when you come to putting everything in the box. Its impossible to say at the moment. If it all works OK, then congratulations! If not, then dont say we didn;t warn you! Good luck.

[richard_morris]

Question about interference

Back to earthing of the antenna shield - I wouldnt, from experience on robots and other kit. Sure - connect the antenna screen to a metal box around the receiver, but dont link it to the chassis - keep it floating. We do not use the chassis as a return, but you can still get noise on it because there is just so much high current floating around in cables clipped close to the chassis - so my recommendation is to keep the antenna screen off it!

Richard

[jeremy_cuss]

Question about interference

I used a number of tips to increase the effective range of my transmitter from about 2m to 40m. I used capacitors to minimise motor noise, put the receiver in an aluminium box, used a separate receiver battery, used co-ax cable within the robot to carry the aerial signal, earthed the shield to the battery negative (but did not earth the chassis), extended the length of the aerial to ~1.1m, and used sprung steel for the aerial to hold it away from the chassis. Each one of these elements had a positive effect. When at a later time I had to move the on-off switch to make it more accessible I used graphite cores at both end of the wiring. Surprisingly the reception didnt seem to change with or without these cores.

Jeremy

[kefirtribe]

Question about interference

can aluminium foil work as a screen (for cables)?

[richard_morris]

Question about interference

It certainly can! Its just very tricky to solder - so clamp the end for a connection if you need one.