Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

atx power supply problem

Status
Not open for further replies.

gun-powder

New Member
i use an atx power supply to power up some signal interface crcts for an automotive research before analyzing them on pc through data acquisition card.

i logical thing is that of course the atx power supply have much more better rectification than that made in china 12v dc adapter.

unfortunately that atx power supply sometimes adds a very high noise in the signals i read but when i turned it off and used that cheap adapter it totally disappears.

the most two strange things i see now are that:
-some times this atx power supply dosent show this problem & sometimes it does.
-noise is generated only when the power supply is turned on even if dc power is not plugged in the crcts like if it generates a magnetic field or something.

someone will tell me why didn't you use those cheap adapters if they work well?

i think can't use 8 or more adapters to power my crcts that will make my wiring sucks.

does any one find a reasonable explanation for this problem?
 
Well, I'm not an expert but here's my opinion:

There are a lot of 12V adapters. You should specify what are you using. When you say "cheap adapter" I guess you are talking about a wall wart.

I hate cheap wall warts: The delivered power is not the specified on the label, the construction techniques are poor, efficiency is awful. Maybe the model you're using is a better built wall wart, or a (non switching) power brick.

The ATX is a switching power supply. Switching PSU's are good for delivering very nicely regulated voltage and high power, on small volume /weight. Their CON is a bit of interference on the output, that depends on the PSU design, so using a decent brand like Termaltek should improve the noise condition.

Anyway, your question should be: "how to filter the ATX PSU output?" instead :)

I hope it helps
 
Last edited:
i think you didn't understand me sorry
i was saying that the atx power supply generates noise even if its not connected to the crcts & the noise get even worse when its connected.
while wall adapters show a very clean signal with out any noise despite it should be less regulated than atx power supply that is used in PCs.
 
Yes, the noise is a consequence of the switching operation. Also note, some ATX PSU's doesn't work well or don't even turn on unless there is a significant load on their terminals
The wall warts show a cleaner signal because they are not switching mode

Some of the forum bosses may tell you how to filter the ripple on the ATX PSU, but I don't see any of them online right now. Be patient
 
You may not have enough load on the ATX power supply. Which rails are you using? Try adding a load on the unused rails.
 
Hi there,


To add to the previous posts a little here...

Switching power supplies use rectangular waves to convert energy from one voltage level to another. Simple cheap wall warts do not do this.
Rectangular waves generate harmonics way up there in frequency which can propagate just like a radio wave. Sensitive circuits pick this RF up as noise and this could happen even with the circuit not connected. An AM radio for example may pick up this noise on one or more channels and that could block the channel due to the noise level being higher than the signal to be detected.

In addition to that problem, they also produce a small amount of ripple at the switching frequency. It's not too hard to get rid of this however, as a balanced inductor capacitor filter can get rid of that noise to a great degree. One idea then might be to use two inductors and a capacitor or two to filter the DC so it provides a cleaner voltage to the circuit. Even small hand wound coils can help knock the noise down quite a bit.

For the problem of RF noise, you can try moving the power supply farther away from the circuit itself, but you'll also need the balanced filter on the output. If that is not practical, you could experiment with shielding the circuit inside a metal can or box. Another idea would be to design the circuit so that it is less sensitive to noise, which can come as easy as reducing the impedance of some of the more sensitive circuits sometimes.
 
Last edited:
Hi there,


To add to the previous posts a little here...

Switching power supplies use rectangular waves to convert energy from one voltage level to another. Simple cheap wall warts do not do this.
Rectangular waves generate harmonics way up there in frequency which can propagate just like a radio wave. Sensitive circuits pick this RF up as noise and this could happen even with the circuit not connected. An AM radio for example may pick up this noise on one or more channels and that could block the channel due to the noise level being higher than the signal to be detected.

In addition to that problem, they also produce a small amount of ripple at the switching frequency. It's not too hard to get rid of this however, as a balanced inductor capacitor filter can get rid of that noise to a great degree. One idea then might be to use two inductors and a capacitor or two to filter the DC so it provides a cleaner voltage to the circuit. Even small hand wound coils can help knock the noise down quite a bit.

For the problem of RF noise, you can try moving the power supply farther away from the circuit itself, but you'll also need the balanced filter on the output. If that is not practical, you could experiment with shielding the circuit inside a metal can or box. Another idea would be to design the circuit so that it is less sensitive to noise, which can come as easy as reducing the impedance of some of the more sensitive circuits sometimes.

thanx for this answer made things clear a little bit
i still can't explain why this problem some times disappears by its self.
So i tried out something recently i brought several atx power supplies and the old one was connected to the crcts and turned off i brought a new one and plugged it into the ac power source and placed it over the old one i was just about to transfer the wiring from the old one to the new one then if found that when the two bodies of the atxs touch each other the noise raises up again i know that the body is the ground so i finally realized that if the ground of the atx supply touches the power wiring ground noise come all over the crcts readings.
what does this mean?
 
Hi again,

Try measuring the AC voltage potential between the atx ground and the circuit ground. See if you can measure some voltage between the two.
It could be that the atx conducts some current into the second circuit ground through the input filter capacitors, but im not completely sure.

ADDED LATER:
I just tried this with two different ATX power supplies i had laying around and it did not show any voltage difference, or very little like 27mVAC.
 
Last edited:
i use an atx power supply to power up some signal interface crcts for an automotive research before analyzing them on pc through data acquisition card.

i logical thing is that of course the atx power supply have much more better rectification than that made in china 12v dc adapter.

unfortunately that atx power supply sometimes adds a very high noise in the signals i read but when i turned it off and used that cheap adapter it totally disappears.

the most two strange things i see now are that:
-some times this atx power supply dosent show this problem & sometimes it does.
-noise is generated only when the power supply is turned on even if dc power is not plugged in the crcts like if it generates a magnetic field or something.

someone will tell me why didn't you use those cheap adapters if they work well?

i think can't use 8 or more adapters to power my crcts that will make my wiring sucks.

does any one find a reasonable explanation for this problem?

Really depends on the power supply in question, some of them are notoriously inefficient (mainly the cheap ones) and have dodgy rails so in the end they may be no better than an el-cheapo Power Adaptor
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top