help with GSM TDMA noise on board

the noise is coming from a cell phone which we use to transmit. No noise when its not transmitting. I located the actual chip on the board, an op amp, that injects the noise into the signal. Since it is from the cell phone GSM TDMA accessing (4.6 ms packets, looked it up on scope, 217 hz and its harmonics), I guess it is electromagnetic radiation. When I change to WLAN the noise completely goes away.

I am working with audible frequencies, the noise does not affect the other signals as we have a lot capacitance between power and ground at all parts of the board. I've already completely shielded the power source ( a battery several feet away ) using shielded power cables, and even wrapping the whole board and power cabling in aluminum foil which is grounded. all no luck.

There is no 217 hz and its harmonics on the power and ground from the opamp that creates the noise, so i am completely dumbfounded. I even added 220 uF capacitance (more than enough for 217 hz and above) and it doesnt affect it. I've even added capacitance to all the capacitors in the circuit surrounding the op amp, and none help (some just cut the sound completely off). The trace is a couple inches in length which COULD be a problem and I could try unsoldering the chip that it goes to but it would be a ton of work, or I could cut the trace with an exacto knife. However, even shielding it with aluminum foil and grounding it completely does not fix it!

Since everything else I've measured doesn't pick up the 217 hz it is not the oscilliscope cabling.

I tried disconnecting the microphone such that basically ground was going into the op amp. The noise very well could be coming from the line attached to the opamp output but I don't know how I can deal with it. There was still noise with none on ground/power.

I will try out some better than aluminum foil insulation and shielding methods between them. We place these devices within several inches from the cell phone in order to have it small in the woods.

This happens across over 50 boards with the same layout so it can't be the opamp. it only happens when the gsm phone is transmitting.I am going to try increasing the resistance in the feedback loop but I really see no 217 hz in the input of the opamp at all. so is it possible that the output trace which is several inches in length and whatnot could be picking up signal? What are some better ways to shield it?

formatting link

Does anyone have any ideas?

Reply to
Johnny Chang
Loading thread data ...

I'll assume that the op amp is the microphone amp or part of the tx audio chain. Somewhere in that area, you're rectifying RF. The trick is to find out WHERE it's entering.

Find a piece of wire that's 1/4 wavelength long at either

800/900/1800/1900 or wherever your transmitter is operating. Hold the wire perpendicular to the circuit board using a well insulated pair of plastic tweezers, soda straw, wooden stock, or anything non-metallic. Touch each of the pins and major entry points around the op amp (including the power supply). When you find a point that creates a large *INCREASE* in GSM noise pickup, you've found the entry point.

I can't offer a general solution because it depends on the entry point. If it's the power supply, install a bypass capacitor directly across the op amp with extremely short lead and trace length. If it's coming in on the inverting input, decrease the value the 15K resistor to lower the input impedance. Also add an RF bypass to ground. If it's entering via the microphone, UNGROUND the electret microphone ground at the other end of the coax cable to disarm the ground loop. Incidentally, I suspect the last is the most likely cause.

Meanwhile remove all the added shielding. Shielding is what you use to obtain the last few dB of isolation between sections and stages. It's totally unsuitable for gross pickup and leakage problems as you're currently experiencing. Note that many handsets have no shielding of any kind. What little shielding is present, is for meeting FCC Part 15 incidental radiation requirements.

Also, another trick is to put your handset into the non-GSM mode. Then take a 2nd GSM handset, connect a coax cable and short antenna to the antenna connector, and use it as a probe to see what sections of the handset are most sensitive to local interference. Actually, it would probably be best to use a signal generator with 217Hz AM modulation for this test. You may have additional areas that are picking up RF. Watch out for excessively high input impedances to just about anything.

Good luck and welcome to the magic world of RF.

Jeff Liebermann
150 Felker St #D
 Click to see the full signature
Reply to
Jeff Liebermann

What kind of opamp is it?


Reply to
John Larkin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.