Problem with UHF interference [solved!]

[snoopen]

Hi all!

Total newb to this site and still quite novice with electronics.

I've built an LED light timer circuit that is powered and activated by my garage door opener. The problem is that when it's attached the UHF (433.92MHz) based remotes no longer activate the door unless I hold the remote right next to the door opener.

The circuit sits in an unshielded box right next to the door opener. It has one unshielded cable running into the door opener hooked up to power and a relay. The LED lights are about 5m away from the circuit and door opener again with unshielded cable but I've had this disconnected and the remotes still don't activate the door.

The circuit consists first of a buck converter from Pololu (http://www.pololu.com/catalog/product/2102) which powers the light and a 7805. I have a 100nF and a polarised 100uF cap on both the input and output of the 7805. The 5v powers the picaxe which controls the light with PWM via a mosfet.

My suspicion is that either the UHF signal is being picked up and somehow reflected by my circuit (perhaps contaminating ground and causing the signal to cancel out) or that the buck converter is switching about the same frequency and overpowering the remotes.

How do I go about investigating this problem? Are there some simple precautions I should have taken in building the circuit? My first thought is to run shielded cable throughout but there's no easy source of ground to be found so I'm not sure it would be effective.

Thanks,
Scott

[jahonen]

The first thing to do without any specialized equipment, is to eliminate all possible causes of interference, one by one until the culprit is found. Try replacing the buck converter temporarily with something that certainly does not cause any RFI. How are the LEDs driven, is there a switch mode LED driver?

If the converter turns out to be the culprit, you might reduce the noise by putting some cable ferrites (similar than in computer cables) to input and output cables. That is a band-aid solution but often only possible if one does not have specialized equipment.

Regards,
Janne

[sub]

I've built an LED light timer circuit that is powered and activated by my garage door opener.

How heavily are you loading the power supply? Perhaps the simplest possibility is that the voltage is sagging and messing with the receiver gain.  If your circuit isn't connected to anything in the RF path, it seems to me the most likely cause.

[snoopen]

Thanks for the advice Janne and Sub!

The LED light stick has a resistor in it... at least I think I can see what looks like a resistor. I wasn't able to rip it apart entirely. Even with the LED lights disconnected the remotes still failed to work. Given the specs for the light says it runs at 12v 1w I assumed that I could run it off the step down circuit. The door opener claims it can provide 24v and 50mA. So it's very likely that the LEDs were causing the power to sag. But that wouldn't explain why the remotes still failed to work with the LEDs disconnected. Was at least a novel idea while it lasted.

An LED driver sounds like a good idea... however the LEDs only specified volts and wattage. One of these is what I have: http://www.ebay.com.au/itm/3X-Mort-Bay-1W-SMD-10-LED-Tube-12V-30-000-hours-ALU-/160647755592#ht_500wt_1413

I took the circuit down from it's mounting and tried to diagnose what exactly the issue was. Having it positioned differently let the door remotes work from inside, but outside they still failed to work. So I'm still thinking the Pololu step down is noisy around the same frequency as the remotes or perhaps a harmonic some something else crazy that a noob might suspect.

None the less I tore out the step down and ran the entire circuit off a 12v 1A regulated wall wart instead and all works well. I've mounted the circuit back in it's original hiding hole and I still get the remotes working at their original range so I'm happy! Next time hopefully I will have learnt enough to build my own smps!

At a later stage I might end up testing the step down near the door opener if/when I use it in another project. Just glad I used the headers to put it on the board rather than solder it straight in!

I'd probably have installed shielded cable and all sorts of nonsense without your help!

Cheers!
Scott

[IanB]

None the less I tore out the step down and ran the entire circuit off a 12v 1A regulated wall wart instead and all works well.

Regulated wall warts are a perfect solution these days. They are compact, reliable, nicely encapsulated and come with a convenient low voltage output plug.

Next time hopefully I will have learnt enough to build my own smps!

In view of the above, this is probably not time or cost effective. Just start with a suitable wall wart and build your circuit from there.

[Mechatrommer]

maybe its because the wallwart is linear voltage, not smps. wont putting large enough decoupling capacitor to the regulator output solve the issue?

[snoopen]

Next time hopefully I will have learnt enough to build my own smps!

In view of the above, this is probably not time or cost effective. Just start with a suitable wall wart and build your circuit from there.

Would be more of a learning experience challenge for me than anything else Would use a smps controller chip of course!

maybe its because the wallwart is linear voltage, not smps. wont putting large enough decoupling capacitor to the regulator output solve the issue?

You might be right Mechatrommer. If I run out of projects this weekend I might end up playing around with the circuit more to try and pinpoint the exact cause. Tempting since the original reason I wanted to avoid a wall wart is I've now got one less wall socket to plug my tools into