I built an amplifer using the following:
1) 36vdc switchmode power supply (see photos)
2) TDA7498E Class d amp
3) ne5532 + lm1036 pre-amp
4) FM/Mp3 decoder board from AliXp
5) 12vdc (power for pre-amp)
6) 5vdc (power for mp3 decoder board)
NOTE: Mp3 decoder board is powered through a seperate 5vdc adapter to remove ground loop noise
Everything works fine except the FM radio. I spent days, trying to figure out the component which was creating the noise. Finally I got myself a portable handheld radio(with the little antenna) and narrowd down on the 36vdc power supply.
I took the power supply out and ran it with a resistive load(nichrome wire) and sure enough, it was the guilty part. The noise would increase as I increased the load.
The power supply has all the filter components in the primary side so why is it still noisy?
photos attached
schematic of power supply
WX-DC2416 PSU -
EasyEDA
1) 36vdc switchmode power supply (see photos)
2) TDA7498E Class d amp
3) ne5532 + lm1036 pre-amp
4) FM/Mp3 decoder board from AliXp
5) 12vdc (power for pre-amp)
6) 5vdc (power for mp3 decoder board)
NOTE: Mp3 decoder board is powered through a seperate 5vdc adapter to remove ground loop noise
Everything works fine except the FM radio. I spent days, trying to figure out the component which was creating the noise. Finally I got myself a portable handheld radio(with the little antenna) and narrowd down on the 36vdc power supply.
I took the power supply out and ran it with a resistive load(nichrome wire) and sure enough, it was the guilty part. The noise would increase as I increased the load.
The power supply has all the filter components in the primary side so why is it still noisy?
photos attached
schematic of power supply
WX-DC2416 PSU -
EasyEDA
Attachments
Last edited:
The SMPS radiates RF (through the air like a transmitter). Like many of them do... but many people like them as they are cheap. Since they are standard many don't even know classic PSU's. In some cultures cheap is seen as good and it adds up that SMPS are newer and meanwhile standard technology so "SMPS are good". A transformer-rectifier-filter cap combo may prove to be superior but it is swimming against the stream.
Last edited:
Not yet
Just a few dozen nasty shocks over the years hehe
I've read a lot about the rfi/emi noise in switch mode supplies. I couldn't figure out why?
The ncp1200p60 chip switches at around 240khz fixed. So why am I getting noise in the mhz range(fm band). It's like the filter components do nothing. I get noise 5-8 feet in radius.
Linear supplies are good but too big and heavy for my liking. Also I live in india where the weather is unfriendly and elements deteriorate much faster. The plastic bobbin inside the tf usually goes brittle and cracks.
I'd hate to junk this amp after spending so much time assbling it. It's much cheap for me to get a linear premade amp with fm and stack it over this amp. I thought someome here could guide me and help fix the noisy ps.
Last edited:
I have seen lots of SMPS in pro amplifiers, It doesn't mean that they are bad but need some extra care where it comes to filtering. Nobody mentioned that there should be a copper screen between the windings for safety and emi blocking. The input filter should be screened also. Sometimes a single stage filter is not enough and has to be a double stage one. I also see screening of all the smps to block any interference. I usually find a non regulated smps which happen to be more quiet than a regulated one due to the square wave is not chopped. These are all my findings trough out the years. If all is made as it should no ill effects shall ensue
High power, small size, light weight and high efficiency are the main keys manufacturers are choosing to use SMPS
High power, small size, light weight and high efficiency are the main keys manufacturers are choosing to use SMPS
Last edited:
A switcher may produce harmonics ranging up to 100MHz and even more. This specially applies to hard switching smps like the ones inside desktop PCs.
CE marking never has been a guarantee for noiseless smps. Even measured within the allowed limits UKW-FM may be disturbed upto a distance of several meters.
Your Class-D amp as a switching device may disturb UKW-FM as well.
All your LED illumination will transmit rf-noise.
One of my tiny MP-3 Players incorporated an FM-tuner as well with the ear-plug cable serving as the antenna. Pretty useless by the way, the harmonics of internal microprocessor produced lots of "noise stations" masking any reception. You won't believe it, it was an expensive TEAC.
CE marking never has been a guarantee for noiseless smps. Even measured within the allowed limits UKW-FM may be disturbed upto a distance of several meters.
Your Class-D amp as a switching device may disturb UKW-FM as well.
All your LED illumination will transmit rf-noise.
One of my tiny MP-3 Players incorporated an FM-tuner as well with the ear-plug cable serving as the antenna. Pretty useless by the way, the harmonics of internal microprocessor produced lots of "noise stations" masking any reception. You won't believe it, it was an expensive TEAC.
Last edited:
Yes so true with all cheap stuff that you buy from China. You get what you pay for. I bought that kind of smps to make a small battery charger, it only costed around 8 euro or so including shipping.
I think I have such a supply somewhere (Daniel sent me one when experimenting various simple amps configurations).
I may have a look, and see if there are some "low hanging fruits" regarding EMC. Don't hold your breath though: this type of consideration normally has to be included from the very start in the design, and it is rarely possible to add one or two miracle components to fix the problems.
I may have a look, and see if there are some "low hanging fruits" regarding EMC. Don't hold your breath though: this type of consideration normally has to be included from the very start in the design, and it is rarely possible to add one or two miracle components to fix the problems.
I was referring to the fact that the AC feed is being taken directly "off-line" -- OK for an OEM, not for an amateur or DIY'r.
If you did the PCB on your own, likely that you have some spare nano-Henries lying around -- when the current changes over time you get a voltage proportionate to the product of inductance times di/dt
If you did the PCB on your own, likely that you have some spare nano-Henries lying around -- when the current changes over time you get a voltage proportionate to the product of inductance times di/dt
> why is it still noisy?
Interference is coming out ALL OVER that board. And radio is all about really weak signals. Your construction is like an apartment building with no walls. There's noise everywhere, no quiet spot. Walls help. But must completely surround the source, the destination, or both.
And in audio, distance is by far the cheapest shielding. Put the power supply and power amp in one box, radio and preamp in another box some distance away.
Interference is coming out ALL OVER that board. And radio is all about really weak signals. Your construction is like an apartment building with no walls. There's noise everywhere, no quiet spot. Walls help. But must completely surround the source, the destination, or both.
And in audio, distance is by far the cheapest shielding. Put the power supply and power amp in one box, radio and preamp in another box some distance away.
Attachments
Even if you put the SMPS in a different box, both the mains lead and the output lead from a SMPS can act like transmitting antennae at radio frequencies. If you mount the SMPS in a metal box, it would be good practice to make sure you have additional filtering where the cables go through the chassis.
Sometimes even a clip-on ferrite choke can make enough difference as ling as it it close to the output of the SMPS.
Another trick is to use a shielded cable from the power supply to the device being powered so that the cable doesn't act like an antenna broadcasting hash everywhere. To avoid introducing ground loops and hum noise, connect the shield to ground at the supply end only.
Sometimes even a clip-on ferrite choke can make enough difference as ling as it it close to the output of the SMPS.
Another trick is to use a shielded cable from the power supply to the device being powered so that the cable doesn't act like an antenna broadcasting hash everywhere. To avoid introducing ground loops and hum noise, connect the shield to ground at the supply end only.
Here's another failed experiment I did yesterday.
I got a linear powered amp from a neighbor. Installed it near the window where the radio signal is crystal clear. Zero static. Got a 15ft shielded audio wire. Connected one end to the output of the window amp and the other to the aux input of the indoor smps amp. Now I have two amps with broken fm.
The signals arent even faint. It is total static noise. The audio cable radiates rfi and noise is introduced to any room where it runs through.
I read somewhere that these chinese power supplies were actually aimed at driving led strips. I think I'll just get another amp(linear powered) for fm for now. Maybe I'll replace the smps with a toroid in the future.
Thank you all for sharing your thoughts.
I got a linear powered amp from a neighbor. Installed it near the window where the radio signal is crystal clear. Zero static. Got a 15ft shielded audio wire. Connected one end to the output of the window amp and the other to the aux input of the indoor smps amp. Now I have two amps with broken fm.
The signals arent even faint. It is total static noise. The audio cable radiates rfi and noise is introduced to any room where it runs through.I read somewhere that these chinese power supplies were actually aimed at driving led strips. I think I'll just get another amp(linear powered) for fm for now. Maybe I'll replace the smps with a toroid in the future.
Thank you all for sharing your thoughts.
My amp chassis is metal body itself. It should act as a faraday cage. I connected the chassis to -36v gnd but no reduction in noise. Using portable habdheld radio to check rfi. I then removed chassis gnd and did chassis earth. Still no difference.
I ran the 36v output through(3 loops) a ferrite ring(see photo in first post) and still no difference.
I ran the 36v output through(3 loops) a ferrite ring(see photo in first post) and still no difference.
But if the noise source and the radio are in the same "room", there's no shielding one to the other. Like when I shut my bedroom door, but the snoring dog is under my bed.
MY HP desktop all-in-one SMPS supplied blew up and a genuine secondhand replacement was .$100. I made the mistake of buying a cheaper non-genuine one, but I couldn't use it because it radiated so much crap into everything else in the house.
The fake SMPS had all of the European, US, Aus and NZ EMI/RFI approvals stamped on it (U/L, PCT GS, SA, CE, C-tick, etc, etc) but obviously because industry is self-regulating these days no one is checking and people can stamp any approval mark they like on any product, tested or not.
The fake SMPS had all of the European, US, Aus and NZ EMI/RFI approvals stamped on it (U/L, PCT GS, SA, CE, C-tick, etc, etc) but obviously because industry is self-regulating these days no one is checking and people can stamp any approval mark they like on any product, tested or not.
All in all FM-listeners are a shrinking lobby nowadays. And besides EM-radiation there are a multitude of regulations to comply with covering safety, efficiency, material...
I assume about every second imported smps violates regulations. Makes me think how bad this stuff would be produced in the absence of any regulations.
I assume about every second imported smps violates regulations. Makes me think how bad this stuff would be produced in the absence of any regulations.
Last edited: