Somewhat Off Topic but....Laptop Computer Power Supply Filtering

[dano12]

I've received several emails from people who were interested in my simple Huminator. Seems that lots of laptops have either crappy or non-existent power-supply filtering.

The problem goes like this:

Laptop on batteries hooked up to stereo: sounds great
Laptop on AC adaptor hooked up to stereo: lots of noise.

A quick google shows loads of people with this problem. One of the proposed solutions is lifting the ground on the AC adaptor, but I'd sure like to avoid that route.

So I'm wondering if my simple power supply filter will work to help solve this issue.

Problem is, I'm not sure what rating to use for the parts. Here's the current schematic:



So an average laptop power supply yields 9-20 volts at about 100 watts.

What ratings would I need on the resistors and caps, or is this even feasible?

Thanks!

[jjs]

More filtering would probably reduce the noise, but only if it is actually generated in the external power supply. Maybe some other switching circuitry is responsible (e.g. for charging the battery).

As for the part values the biggest challenge is probably the R1 resistor. As for now it is way to large with 100 Ohm. Remember it will drop 1 V for every ampere current that goes through it. For a typical notebook that would be about 1 - 2 A minimum at 20 V. Even a 1 Ohm resistor would drop 2 V.

Only chance is to increase the value of C1. I could imagine that too high a value could potentially damage the power supply as they surely design it for the specific load of the notebook and a large capacitor would suck a lot of current when initially charging.

[oskar]

Yeah! The laptop and it's power supply are complex things so I have no idea if this'll work or if there is other switching adding noise in the laptop. The filters I've worked on was in cameras drawing some 200mA and they had a filter going something like:

input --- L --- R ---- output
                 |      |
                 C     C

The current drawn from them was something like 200mA and I don't remember the values of the components but it could have been something like R~0.1R, C~100uF
I don't know what damage the power supply could take from this...

[R.G.]

A couple of things come to mind.

- There are two kinds of noise from that power supply: differential mode noise between the + and - sides and common mode noise where both + and - wiggle up and down together.
- Your filter works exclusively on differential mode noise, and has zero rejection of common mode noise.
- I would use an inductor instead of R1 for low DC power loss.
- I would use about three caps for C2: a 01.uF, a 0.01uF, and a 0.001uF, all multilayer ceramic. All capacitors have a lowest-impedance point where the capacitance starts getting overwhelmed by ESR and ESL. ESL is worst. The only way to keep the impedance going down is to use a yet-lower ESR/ESL cap in parallel, and that almost always means a lower value cap. Ceramic has a better ESR than many film caps.
- For common mode noise, put in a common mode choke.
- It is not simple or easy to come up with a filter that will preempt all the noise modes that a switching power supply can produce, and even harder to make one that will work interchangeably with different units. What works for one may be useless for the others.
- You may still have noise, as the AC power connection on the battery charger for the laptop and the inevitably AC power connection on the audio equipment still are vulnerable to low frequency noise from the  AC power line and neutral offset voltages.

[Auke Haarsma]

input --- L --- R ---- output
                 |      |
                 C     C

If we assume 16V DC input and a current of max 1A (more likely max 500mA), what value for L would I need?

- For common mode noise, put in a common mode choke.

Can you elaborate on this a bit? What value of a choke would be good? Is there a rule of thumb?
I can get small (resistor-like) chokes with a value of 1uH rated at 700mA. Values go up to 100uH, but the mA rating drops in that case to like 60mA (seems too low to me, but I'm lacking knowledge here...)