The sound of silence.
May 18, 2006 9:33 PM   Subscribe

Laptop power supply interfering with PA speakers.

The Problem: My laptop power supply is interfering with my speakers, givng off very unpleasant, persistant interfereance. I know that it's the power supply because when the laptop is running on its battery, the interference is nonexistant. I'd like to be able to plug in my laptop without the interference.

What Works: The PA system has two volume controls- one for the tape in, and a master control as well. When I turn both knobs to around 7, the interference goes down in volume, but is still noticeable.

Equipment chain:

US Grounded Wall Outlet
-----Furman PL-8 Power Conditioner
----------Dell 65w AC Adapter
---------------Dell Inspiron 6000
--------------------Creative Labs 24-Bit Sound card
----------Phonic Powerpod 408 PA Amplifier (120W)
---------------Phonic PA Speakers

The sound card is wired to the PA system using a 1/8" to RCA cable from the soundcard's output to the PA's "Tape in."
posted by The White Hat to Technology (17 answers total) 4 users marked this as a favorite
Have you tried running your lap top and the speakers off separate outlets?
posted by doctor_negative at 9:42 PM on May 18, 2006

Is it a low constant hum? If so, it's a ground loop. You can get a ground loop isolator, also called an isolation transformer, to get rid of it. You can also try switching from using the same outlet to separate outlets, or vice versa, but I've usually found doing that to be fruitless, and have had to rely on an isolator.

There may be another solution. Does your laptop's power supply have a ground plug on it? You can try using one of those 3 prong adapters that you're not supposed to use to disconnect the ground.
posted by zsazsa at 9:58 PM on May 18, 2006

Funny, I have nearly the same experience, with the following:
Dell Precision M90->1/8"to RCA->Vestax PMC 25->XLR->Mackie SRM450 Powered PA speakers.

I have noticed that if I switch to an optical signal path, the interference doesn't go away- because it's being transferred to the optical output device, which is running off another wall-wart.
The only reliable laptop audio ouput I've seen is a PCMCIA->XLR breakout from M-Audio, but I expect that a USB device might eliminate the problem as well. I'll try my old Extigy and see what happens.
posted by wzcx at 9:59 PM on May 18, 2006

Response by poster: Yes. The interference changes pitch but not volume.
posted by The White Hat at 10:00 PM on May 18, 2006

Goddammit, I've asked this question about a million times and people always respond with that ground loop nonsense. It isn't. I even get it on my desktop.

If you turn it up loud enough, can you hear your mouse moving? Can you hear network traffic? Your hard drive accessing stuff?

Someone, please offer a real solution.
posted by redteam at 12:43 AM on May 19, 2006

The most direct solution is volume adjustment. The volume coming out of the computer should be as high as possible, reduce the volume on the speakers until you reach a comfortable listening level - this maximizes the SNR.

If that doesn't work.. Well, Electromagnetic interference is a very hard problem. So, there may not be a real solution.

It might help to add clamp on ferrites to some/all of the cables hanging around your computer. In the right place they probably work wonders, but in my experience it is hard to use them effectively.

You might also try putting the laptop power supply in a metal box. However, we don't know for sure if the interference is coming from the brick itself, or the attached wires, and even small openings in the wrong places will make the shielding box useless.

Sometimes just moving the cables and parts around can help enormously, depending on the source of the problem.
posted by Chuckles at 2:17 AM on May 19, 2006

Also try listening to the audio out directly with headphones. Try to isolate the stage where the noise enters your system.

The sound card is wired to the PA system using a 1/8" to RCA cable from the soundcard's output to the PA's "Tape in."

Often those 1/8" cables, even 1/8" to dual RCA, aren't shielded, so a better cable may help. Don't just go and buy an expensive cable though. It is something to be experimented with, not something to through money at. Whether you hear the noise with headphones will help to determine if the cable is an issue.
posted by Chuckles at 2:26 AM on May 19, 2006

Best answer: Licensed radio engineer here, with a few thoughts.

First of all, do not plug the laptop into your Furman power conditioner. The Furman provides very little, if any, low frequency filtration anyway, showing only 10db at 10,000 Hz (from their Web site spec sheet), and what filtering it does supply is from AC input to AC output, not between AC outputs on its own back panel. Plugging an electrically noisy laptop power brick switching supply into it isn't going to work well, as you are discovering. The Furman might be OK at protecting your audio equipment from lightning spikes coming from the AC line, but it's not going to be effective at suppressing locally generated noise and interference between individually connected devices on its secondary side ports, and there is no claim or spec for common mode noise suppression on these ports. Basically, the Furman isn't doing what you assume it's doing, because it's not designed to do that. Plug the laptop into the wall, and the Furman into the same outlet, and go from there. If the laptop's power brick is electrically noisy at 60 Hz, or harmonics up to 600 Hz, the Furman won't do much about that, but whatever it will do, you'll be getting at that point, which you can't, the way you are hooking it up now.

Most laptop power supplies are inherently electrically noisy, for a number of reasons, that make them pretty hard to filter. First, the most common type (a "brick" with an AC cord coming out one side, and the DC cable that plugs into the laptop coming out the other) is usually a fairly simple switching type supply, without the big capacitors that filter current in a conventional desktop machine. The supply "brick" generally charges the laptop battery, at a rate that will equal or exceed the operational current draw of the laptop (for "unlimited" usage while plugged in to AC), and the laptop battery acts like a big capacitor for energy storage purposes, unless it's pulled out. So, the brick is switching AC current at line frequency, but in doing so, adds significant harmonics of the AC frequency up to 5 to 10 times the AC line frequency.

On top of that, the cables that connect the brick to the wall, and to the laptop are decent antennas at some frequencies, and the common 2 pole connector on the brick for attaching various AC line cord types for matching local market AC wall connections can become a source of resistance, and additional electrical noise over time. Typically, as there is no ground lug provided by the AC power cord, filtering switching noise to ground effectively isn't possible.

Result: lots of electrical interference due to the combination of poor physical and electrical shielding, very little filtering, and no grounding.

What to do: You can provide a probably less portable, but much better shielded and filtered DC supply [PDF file linked] for your laptop. You may need to fabricate a DC supply cable to your laptop from a spare power brick. The probability of this permanently curing your noise problem is about 99.999% But it is not so cheap, and will take some engineering on your part. There is also some danger to your laptop in operating from such a supply, which is capable of putting out voltages much higher than your laptop is designed to accept, if you don't set the power supply up properly.

Short of this, you could:
1) Be sure your laptop is running a new, high quality internal battery, that is fully charged. Older batteries develop a higher internal resistance, and this effectively forces the laptop to call for more instantaneous power from the brick, and reduces the battery's effectiveness as a filter element.
2) Replace your current power brick with a new one, to eliminate any chance that a loose winding in the power brick's autoformer (or any other bad component in the brick, which takes a beating in being hauled around and used) is causing undue noise.
3) shield the brick in operation as much as possible by operating it in a small Faraday cage. I've made fairly effective ones from a heavy duty metal lunch box. Keep as much of the power cords to the brick wrapped up inside the lunch box as you possibly can.
4) Experiment with adding additional ferrite beads on the power cords of equipment to soak up higher frequency EMI noise components being radiated from the cables. Frankly, this is kind of hit or miss, but the general idea is to put external ferrite beads around cables where they enter or exit equipment, within about 2 inches of the transition point. They are pretty cheap, and just clamp or glue on to the cables, and basically can't do any harm.
5) Do whatever you can to limit or reduce the power demand of the laptop in operation. Cut down the screen light intensity, cut running programs to a minimum, avoid USB devices if you can, etc. The less power you draw, the less noise you'll create.

As for redteam's heated observation, some powered computer speakers I've seen use, (wait for it....) laptop power brick type power supplies for powering their internal amplifiers. Obvious issues there. But it's also possible for a desktop machine's power supply to throw out a whopping lot of EMI, back up the power cord, into the wall socket and house wiring, and to radiate directly from the supply through the fan holes, cable harnesses, etc. Many of the better high current supplies on the market now have Power Factor Correction circuitry added, which is basically bigger filtration, and better switching control of the AC line, and as such, these supplies may be less prone to EMI issues.

Not all buzzing sounds have a common source, I'm afraid. Good for the engineering profession however, as it keeps many of my brethren fully employed, obtaining FCC type certificates for electronic equipment.
posted by paulsc at 3:45 AM on May 19, 2006

A ground loop sounds like this. If your noise doesn't sound like that, it's not a ground loop and you shouldn't do the ground loop voodoo.

Redteam, sorry, some of us are just ground loop nuts, having been bitten by them many times before. Desktops are in fact more prone to them. Anyway. If you hear your mouse moving, etc., that's obviously not a ground loop and is simple electromagnetic interference coming from the computer itself. Are you using an internal sound card or built-in sound? Often the only solution to a problem like yours is to get an external sound card like the Extigy mentioned above.
posted by zsazsa at 5:05 AM on May 19, 2006

Actually, disregard my "if it doesn't sound like that, it's not a ground loop" line. It could be AC noise from the power supply that paulsc describes in such detail. I know of a set of computer speakers with a crap power supply that make just that noise.
posted by zsazsa at 5:08 AM on May 19, 2006

Best answer: paulsc nails it -- it's hash from the DC power supply or the DC/DC converter inside the laptop. How to tell? Easy. Unplug the laptop from the wall. Does the noise go away?

What's happening -- the pulsing DC is appearing on your audio outputs. What you can do.

Capacitors. Caps block DC, let AC through. Problem: your jack on the mixer will have components beyond that have resistance to ground. You're adding an inline capacitor. Resistance + Capacitances = "RC filter". So, you need a large value, *VERY* high quality set of caps, to make sure the filter rolloff is very low in frequency. These aren't expensive, but don't use cheap ones. There will also be a phase shift with the capactiors, which might cause phase effects later in the chain to change.

How to build one. When I built a headphone amp for my PC, I put the caps inside. Mistake, the right way is an insert.

Parts: This assume that your soundcard uses mini-stereo plugs, and that your other patch cords are 1/4".

1) 1/8" / 3.5mm stereo plug.
2) 1/4" / 6.5mm stereo jack.

(These can be swapped to whatever you need, with the caveat that they both need to be stereo or mono.)

3) Two large value, audio grade capictors. Do *not* use ceramics. Do *not* use electrolytics. At worse, use polystyrene caps. I use polypropylene. On my PC amp, I use .33μF, 100V Polyproplyene film caps made by Murata. Yeah, 100V breakdown is overkill, but I needed them for something else as well. Obviously, if you're doing a mono cord, you only need one.

4) Wire and solder.
5) If your clever and the jack/plug work with it, you can make this a cable. Otherwise, a small box to mount them in.

You can get all the parts at Digikey. I can dig out numbers later, if you wish. Don't skimp on the quality of the jacks -- this is part of your audio chain.

Assembly. Easy. You have three connections on the jacks, the tip, which is stereo right, the sleeve, which is ground, and the ring between, stereo left. You want to connect the sleeves directly together with insulated wire. Now, you connect the tips together with the capactor -- in general, you solder one leg of the cap to the correct terminal on the 1/4" jack, solder a wire to the other leg, then solder the other end of the wire to your plug. Insulate the connections. (Heat shrink. Heat shink tubing is proof that Ghod loves us and wants us to hack.) Connect the sleeves with the other capacitor, in the same way.

In the end, you have an insert cable that puts a large value capictor inline with the R&L signal lines, but passes the ground cleanly. The caps will block DC, which should dramatically reduce the DC noise.

Total cost? $10, maybe. Probably $15, since if all you order is one plug, one jack and two caps, you'll get hit by the handling charge.

If you have cheap caps at home of a suitible size, you can try it to see if the noise goes away, but don't be surprised, or put off, by the way the sound changes. Ceramics and Electroylitics aren't good in this application, they'll susbstantially color the sound in ugly ways. Of course, good luck finding a .33μF ceramic.

The value isn't critical, but the lower the capictance, the higher the frequency of the filter's rolloff point. If you used a .01μF cap, that rolloff would be in the 100hz range, depending. .33μF puts mine at about 10Hz. The PC isn't generating anything nearly as low, so for all practical purposes, the filter isn't.

The quality of the cap, however, is critical. Don't settle for anything less than polypropylene -- I'd suggest the BC components MKP series. Murata also makes good caps.

Wait, I've got some digikey number right here for the caps. A .27μF 63V MKP Polypro is BC2064-ND, it costs $.81 per, or $.69 per for 10 (or $.30 per, if you buy 10,000. ;) )

A good quality 3.5mm jack is CP-43502PM-ND, it costs $3.69 per, or $3.19 per if you order 10. A decent 1/4" hack is SC1123, it's $2.48 per, or $1.86 per for 10. Both of these are panel mount jacks, not inline jacks.

I don't have any plug numbers handy, but if you're an audio guy, you've got sources for plugs, if not a small stock in your parts bin.
posted by eriko at 5:12 AM on May 19, 2006

Response by poster: Many thanks for the answers, guys- especially eriko and paulsc. Time to fire up the old soldering iron.
posted by The White Hat at 6:35 AM on May 19, 2006

HI, Had exactly the same problem with a Dell going into pa desk myself. Bought one of these and alakazam, the bad noise went away.

Sorry to all the "it's not a ground loop" guys, but this totally worked for me and involved next to no effort!


posted by merocet at 8:38 AM on May 19, 2006

Considered using a pro-quality external (USB or Firewire) audio output? I would not count Creative Labs product as pro-quality even if it is 24-bit, and in any case it sounds like it's inside the computer, which would let it pick up much more interference.

You might also consider just running the laptop from an add-on battery, one of the big lithium ones that sits under it and will run the computer for hours and hours.

Both of these are expensive so you'll probably want to try massaging the power supply first, but if that doesn't work you could consider them as alternatives.
posted by kindall at 9:20 AM on May 19, 2006

You can provide a probably less portable, but much better shielded and filtered DC supply ... There is also some danger to your laptop in operating from such a supply, which is capable of putting out voltages much higher than your laptop is designed to accept,

A very good idea. I don't think you should let the following stop you, but..

There is some danger from the current produced too. Electronic devices sometimes rely on the limitations of power bricks to protect internal components from overheating, but an improved power supply won't have those limitations. See Am I using the wrong power adapter for horror stories. My answer there isn't as well written as it could have been, specifically:
low voltage: Think of transistors as voltage controlled resistors. The higher the voltage, the lower the resistance of the resistor. Power dissipated = I^2xR. So, if for some reason the current wanted to stay the same but the voltage was lower, Now, consider some transistor in the middle of a complex circuit. If the voltage controlling it droops, it's resistance goes up, but the current flowing through it doesn't necessarily drop accordingly. You would see more power dissipated, which could overheat the part.
Regardless, check out the list of other threads discussing power brick and battery charging issues.

I don't think there is much to be done about it.. Theoretically you could stress the power brick to the point of thermal shutdown and then try to mimic that output characteristic with your improved supply - doesn't seem practical.
posted by Chuckles at 9:43 AM on May 19, 2006

Sorry to all the "it's not a ground loop" guys, but this totally worked for me and involved next to no effort!

It would be awesome if you cracked that open and told us what is inside. It's not just a simple ground lift, you'd either get nothing, or get very weird effects from using the other conductor as a return for a signal line.

I'd bet a pint of your fave that there's a couple of inline caps in there.

Of course, the "don't fuck with a working part" rule is in order -- and I wouldn't be surprised if it was potted in epoxy.

If you do crack it open, sketch out the schematic and parts, so we can build it ourselves.
posted by eriko at 11:29 AM on May 19, 2006

eriko, while I haven't cracked mine open, all the DIY ground loop isolators I've seen have been simple 1:1 transformers. I'll try to pop mine open tonight.
posted by zsazsa at 3:12 PM on May 19, 2006

« Older Maybe moving there would be easier?   |   How can I make dosh in WoW? Newer »
This thread is closed to new comments.