100Mhz noise

[madshaman]

Hi all.  A while back I noticed some subtle 100Mhz noise on my scope while testing the guts of a meter I'd purchased (HP3478A).

Well, I've seen the same noise on a *different* scope simply plugged into a function gen via a cheap chinese coax cable.  My body seems to couple this noise more strongly by grabbing the cable (doesn't happen with my probes, which are good tek probes).

Last time I saw it I shut down the computer and all mains powered devices, I also put all my iDevices into airplane mode.  This time I did one more thing which was to unplug my UPS (although by definition it keeps running ).

Just thought I'd ask if anyone else has had to track down something spewing or possibly mixing some other signal down to 100Mhz and has an idea of what it could be.

My next step is going to be physically carrying the UPS and any self powered devices to the basement because I'm a bit stumped.  All my main computer and music gear are in the basement and my lab is on the second floor.

Could stuff in the basement or from my neighbours by crawling along the mains or ground to the lab?

I'll probably rule this out next time I hunt for it, but could the circuitry in a iDevice leak strongly enough to cause this (with the radios turned off)? (Placing my iPhone right up to the coax cable doesn't impact the noise)

As an aside, is it feasible to build myself some filters for the mains powering my lab gear to avoid similar issues?

Any guesses as to what's causing this (and yeah I killed the lights of course) to help me avoid going bonkers carrying stuff into remote places of the house?

P.S. scopes involved are a Tektronix 475 and 2465.  Last time I saw it there looked to be a strong 3rd harmonic riding on the 100Mhz, this time it looks like a second harmonic.  I'd take this with a grain of salt because I don't have a spectrum analyser and I'm just eyeballing the waveform which is dubious at best.

[madshaman]

Actually, it's more like 90Mhz, here are some pictures (hope they worked).

[alm]

A frequency of around 100 MHz might suggest FM broadcast radio, although RF typically has a small bandwidth and wouldn't have much harmonics.

[madshaman]

*Palm to face*, don't laugh at me, but being a newb wrt radio electronics, the fact that this frequency didn't ring a bell when I see similar frequencies staring at me from my car radio every day literally made me laugh out loud at myself.

Now, I have no idea if the trace in the photos I posted are indicative of a radio broadcast, or if it's even possible to tell (I don't have a receiver, tuner or any RF gear (yet)).

Sounds like mystery solved if this is what it is.

Out of curiousity:

1) Is it common to see radio broadcasts on scope traces?
2) Any way to get rid of it short of covering my walls with $$$ of copper foil?
3) If I wanted to investigate further, what equipment/methods would I use?  Could I easily find out what direction it was coming from?

[C]

Just to make your day, RFI can be the result of the sum or difference of two frequencies. A poor connection ( a rusty bolt) between something that acts like an antenna for each frequency = mixer. It can be very hard to find. Best hint I can give is, if it shows on your scope, try to make your scope mobile with it still showing. Then you might be able to move scope closer to the source to find it.

C     

[madshaman]

Just to make your day, RFI can be the result of the sum or difference of two frequencies. A poor connection ( a rusty bolt) between something that acts like an antenna for each frequency = mixer. It can be very hard to find. Best hint I can give is, if it shows on your scope, try to make your scope mobile with it still showing. Then you might be able to move scope closer to the source to find it.

C   

Thanks :-), luckily, if there is an implicit mixer somewhere it's not in my gear, at least not anything in my lab.  I see this trace on two separate scopes connected to or probing different equipment.  I guess if I rule out the UPS I could use it to power the scope (so I could carry it further than the powercord will allow or having to worry about tripping on an extension cable); seems a bit dodgy but worth a try.

[C]

I may be seeing things, Second photo looks like you picked two positive cycles in that wave. Do they repeat to two negative cycles and possible two positive again?

[madshaman]

If you mean do I see the same pattern repeated but negative elsewhere?  not that I can see.

Those pictures are probably not the best to post, that's actually the waveform riding on the output of my function gen and triggered by it.  When I get around to it I'll do one better and try to catch it on my digital scope.  It's got a massive sample memory, but only 8-bit ADCs and a 200Mhz BW, I mostly use it for digital circuits.

[C]

Try setting the function gen and scope on a cart. Move them around on cart for max signal. Then try roting the cart. Think dipole antenna, sides of antenna strong signal, ends is weakest signal.  Easer to follow the strong signal but the weak signal from the ends is a better pointer.

C

[madshaman]

Try setting the function gen and scope on a cart. Move them around on cart for max signal. Then try roting the cart. Think dipole antenna, sides of antenna strong signal, ends is weakest signal.  Easer to follow the strong signal but the weak signal from the ends is a better pointer.

C

Thanks, will try that.  Getting it on my MSO-9212 was a total fail (didn't have function gen in the loop, just a loop made from a 100Mhz probe), my faster probes are 10x so that's no good.  It's just not sensitive enough at it's limit (5mV/div), I can *sort* of see the same signal with my eye, triggering on essentially nothing, but the *signal* at this level is quantised to maybe 2-3 bits at best, which makes the FFT total garbage.  Maybe I can find a way to amplify it...

Closest thing I have to a cart here is a rollable TV stand; will try that.

[C]

Saying it a different way, You need a bad noise radio. Your Function gen & scope are working for this now. Some part of the combination is acting like an antenna. Worse case, the power cords is the antenna. If it's not the power cords, then by rotating the combo you are rotating the antenna, and most all antennas have some directional properties to them. A setup on an office chair that has wheels and tilt may be better then TV cart. Quick safe move, rotate and optional tilt with out moving the function gen with respect to the scope which would change the radio is my suggestion.
This may or may not help to find the source. I spent a week once trying to find something like this with real fancy equipment. I finally found the problem  miles away and was a poor connection between two pieces of medal.

If you already understood before the above, Great, if not hope above helps.
Good luck hunting,
C

[ftransform]

This kind of shit would drive me insane and I would probably start ripping wiring out of the walls.

Can you maybe rule out the UPS, plug the scope into it and then flip the main distribution breaker to your house? That would save you a lot of effort. I don't see the point in playing wack a mole with a TV cart. After you rule out the house you can put the scope and UPS on your back and climb on the roof and nearby trees!


One day I will have a faraday cage... 

[C]

If I was going to use a UPS to go mobile as stated, I would try first with the USP disconnected from mobile rig and shutdown. The UPS could be the RFI Antenna or the RFI Source.  Off and disconnected would tell a little. Then I would watch what happens each step
1. Connect the UPS to wall
2 Turn on UPS Input
3 Turn on ups output
4. Connect a different load to UPS
If none of the above made a difference then
5. Connect mobile to UPS.
6. Disconnect UPS from wall & remove UPS input cord if possible.
Any of the above could show more or less noise on mobile.
A pain, you bet Tracking down RFI reading the small hints.

Killing the main breaker of house does not rule out all in the house,
A mixer is any non-linear device.
For example:
a wire acting as an antenna for an AM radio station.
a wire acting as antenna for an FM or TV station.
connect the two wires with a diode or any non-linear device
You then have an RFI transmiter.
   
Faraday cage,
Think of what happens with two wifi devices in a Faraday cage. instead of a weak signal an few echo's, you have a strong signal with lots of strong echo's.

C

[ftransform]

If I was going to use a UPS to go mobile as stated, I would try first with the USP disconnected from mobile rig and shutdown. The UPS could be the RFI Antenna or the RFI Source.  Off and disconnected would tell a little. Then I would watch what happens each step
1. Connect the UPS to wall
2 Turn on UPS Input
3 Turn on ups output
4. Connect a different load to UPS
If none of the above made a difference then
5. Connect mobile to UPS.
6. Disconnect UPS from wall & remove UPS input cord if possible.
Any of the above could show more or less noise on mobile.
A pain, you bet Tracking down RFI reading the small hints.

Killing the main breaker of house does not rule out all in the house,
A mixer is any non-linear device.
For example:
a wire acting as an antenna for an AM radio station.
a wire acting as antenna for an FM or TV station.
connect the two wires with a diode or any non-linear device
You then have an RFI transmiter.
   
Faraday cage,
Think of what happens with two wifi devices in a Faraday cage. instead of a weak signal an few echo's, you have a strong signal with lots of strong echo's.

C

Well at least it will rule out active sources.

[SeanB]

Do you have cable internet? Could be poor connections on the cable box, loose earths or corroded wiring makes an excellent mixer, and couples into the house quite well.

[madshaman]

Lot's of great input.  Yes ftransform, it *is* driving me nuts, and you guys are gonna have me knocking on someone's door some day holding a parabolic antenna:

(door opens): "Yes?"
me: "Something in your house is interfering with my equipment..."

Nice call on the cable connection SeanB, that's definitely one of a few things shutting the mains off won't rule out.

What I really need is a highly directional broad-band antenna and a portable high bandwidth spectrum analyser; I don't know how to build such an antenna yet, and building a spectrum analyser isn't going to happen until I finish my frequency counter, signal generator and, god help me, the oscilloscope.

Ever since I've beefed up my lab with an assortment of high precision+accuracy DMMs, and soon a high precision+accuracy time reference, I'm finding myself thinking all sorts of crazy thoughts like filtering and isolating the mains supply to the lab, and yeah, even considering some EM shielding.

I'm a music guy, so I understand a bit about room acoustics, I can't see the inside of a faraday cage really being much different, that leads me to think some of the same techniques can be used for EM suppression inside a shielded lab.  I.e. using dampening and dispersion, the cage itself would do some dampening, but there has to be an EM equivalent to accoustic green goo or acoustic foam which will convert EM energy to heat over a broad spectrum, keeping the inner walls non-parallel is a good start to eliminate any resonance and an uneven fractal-like conductive surface would be even better.

I'm positive this problem must have been solved in labs all over the world.  Maybe I'll try an IEEE search :-).

I'm going to track this noise down though if it's the end of me, already learned a few things trying to find it (like my DSO (obviously not all DSOs) is crap for this task, both the sensitivity and 8-bit resolution contribute to this; made me also decide to test it's *real* front-end bandwidth once I've built my signal generator, curious as to whether it's really 200Mhz as claimed).

I have a feeling that, much like a really intractable software bug in a big system, once I discover what it is, I'll slap my head and say "of course!".  Still betting on it being a radio station but trying not to assume that.

I love this hobby.

[kfitch42]

Here's the solution ... looks like it might do well for both RF and Audio 

http://www.eevblog.com/2011/09/22/eevblog-202-emc-rf-anechoic-test-facility-tour/

Installing one in your house might get just a bit pricey

[madshaman]

Here's the solution ... looks like it might do well for both RF and Audio 

http://www.eevblog.com/2011/09/22/eevblog-202-emc-rf-anechoic-test-facility-tour/

Installing one in your house might get just a bit pricey

Holy *!@#&!, I can't believe I overlooked this eevblog video!  Fofl, it looks a lot like an acoustic anechoic chamber.  Gonna watch while I code, maybe it's possible to build something similar for less cost ^^'.

[ftransform]

I feel that a Jim Carrey movie can be made about him as a HAM or EE, going to peoples houses with a parabolic antenna and spectrum analyzer, insisting that they let him analyze their cellphone chargers.
I heard of one situation where someones neighbor had a cheap fluorescent light installed over their garage and someones equipment would go nuts when he turned it on.

He could get progressively more neurotic as the movie goes on.

[SeanB]

The foam they use there is made with a very high carbon particle content, so that it is very lossy to RF. We used this as a test bed for airborne radar units as a test fixture, new foam would smoke and burn for the first hour or so of installation, as the binders burnt off from the 5kW pulses. To do direction finding just use a FM radio, and use the signal strength output ( means either build your own with a chipset that does give a signal strength output or at least a RSSI signal), and use a simple FM radio external antenna ( big but pretty good for foxhunting at this frequency, you just mount it on a broomstick and rotate to get bearings) to do the DF triangulation.

At 90MHz it probably is either a car FM transmitter or one used to broadcast inside the house for music reception.

[C]

SeanB

Here is one for you. Great Lakes is a Navy training base north of Chicago. Back in the 70's, one of the things trained on was very high power ship's radar system. They actually had the complete setup. Big radar antenna mounted on the roof of a building near to the lake. For safety they had limit switches that kept the beam pointed out over the lake.
One free guess what some student disabled.
The antenna did a couple 360's at full power before they realized there was a problem.
Light bulbs exploded or burnt out. It even affected operations a o'hare airport.

C 

[SeanB]

We just used the Radar dummy load as a pie warmer. I used the computer under test quite regularly ( when it was closed up and the 300 capscrews were back in place to hold the lids on) as a cup warmer, it would brew quite a nice cuppa if you turned off the fan in the test area. I wanted a nice power supply, so designed and built one that used the 28V power supply as power, and I rated it to give 0-24V at up to 40A ( the breaker in the wall) but I needed beefier cabling in the wall, I could only get up to 20A as the wiring in the trunking was 1mm only. In the main panel it was 50mm cable and a 100A breaker, and the main TRU was a 28V 16kA power supply. It never moved off the zero on the ammeters. 3 phase transformer, massive fan cooled bridge and almost no capacitance on the output, voltage regulation by a saturable inductor in the power feed. It probably weighed 30 tons in all, along with the 115V 400Hz rotary converter that was part of the room it was in.

[madshaman]

The foam they use there is made with a very high carbon particle content, so that it is very lossy to RF. We used this as a test bed for airborne radar units as a test fixture, new foam would smoke and burn for the first hour or so of installation, as the binders burnt off from the 5kW pulses. To do direction finding just use a FM radio, and use the signal strength output ( means either build your own with a chipset that does give a signal strength output or at least a RSSI signal), and use a simple FM radio external antenna ( big but pretty good for foxhunting at this frequency, you just mount it on a broomstick and rotate to get bearings) to do the DF triangulation.

At 90MHz it probably is either a car FM transmitter or one used to broadcast inside the house for music reception.

Hmm, I'm wondering if it's feasible to roll your own foam of this type; right now it's a bit blue sky, but I can see myself radio-ioslating my lab at some point.  With the long weekend coming up, hopefully I'm going to find the time to track down the offender.  I've got an uncommitted digikey order still, so maybe I'll look into building your suggested rig, any suggestion for a slightly less massive antenna?

[ftransform]

SeanB

Here is one for you. Great Lakes is a Navy training base north of Chicago. Back in the 70's, one of the things trained on was very high power ship's radar system. They actually had the complete setup. Big radar antenna mounted on the roof of a building near to the lake. For safety they had limit switches that kept the beam pointed out over the lake.
One free guess what some student disabled.
The antenna did a couple 360's at full power before they realized there was a problem.
Light bulbs exploded or burnt out. It even affected operations a o'hare airport.

C

ahhahaha, please tell us more.

[kfitch42]

I wonder if something like this might help in your hunt:
http://www.rc-cam.com/forum/index.php?/topic/3891-20-spectrum-analyzer-for-testing-fpv-systems/

For $20 you can't expect spectacular quality (don't see much mentioned about sensitivity/noise...), but might be good enough for DF.