Hello I am trying to choose a psu to power my class d module, and i am interested in finding out if there is any meaningfull diference between the 3 stated psu. More specificaly the psu in question are: a generic laptop power brick, a mean well smps (MEAN WELL RS-100-24 Switching Power Supply SMPS 100W 24V 4.5A - Audiophonics) and a connex smps made for use with class D amps: SMPS240QR ( SMPS240QR | Connex Electronic ). Especially given the diference in price 20 euro for the laptop psu, 30 for the mean well and about 60 for the last one. i knwo there is a diference in voltage but that not my main concern. They are to be used for a powered- active one way speaker with an indigraded class d amp propably one from sure electronics/wondom/dayton audio. The amp is going to be conected with a pc and i fear of forming a ground loop especialy sience the pc is grounded as well as the psus u mentioned, expecpt from the case where i use an extremely sceatchy ungrounded cheap laptop psu (or i dont connect the groudn wire inthe other two). Would the psu play a role in noise and ground looping given that it's a speaker ment for critical listening but not very high end.
Companies like Dell or HP arent going to risk a call from the FCC or UL, so their laptop bricks probably are decently engineered and pass fairly rigorous testing. Meanwell ia a good company and knows their way around power supply design and builds solid products, IMHO. I dont know anything about the Connex one; I assume the 60 isnt just for the name. Personally I use analog +20V supplies on my two class D amps. I was lucky to get them for a song on eBay.
Regarding ground loops, there's pretty much just one thing you can do, as all switch mode power supplies I'm aware of need to be grounded at the AC input for conducted emissions purposes. Your PC is the same way.
1. You can make all the cabling as short as possible, plug both PC and SMPS into the same outlet / strip.
2. You can use a laptop with a 10 hour battery life and run it on battery when you listen.
3. You can get a BT connected amp and "suffer" the available bitrates of that wireless connection.
4. You can get a USB isolator and put that between the grounded PC and amp.
5. You can buy / build an analog DC power supply, let the secondary side ground to the PC through the USB cable. Note: this is potentially unsafe from an electrical fault perspective, as I'm sure the USB cable ground wont carry the current from an inadvertent hot line connection - til the fuse breaker blows. My analog supply has the chassis connected to ground. I believe like on the Meanwell SMPS, you can connect either + or - to chassis on the secondary side. I let that float. I use BT to connect to the amp, so there's no ground loop issue to begin with.
Hope this helps,
Regarding ground loops, there's pretty much just one thing you can do, as all switch mode power supplies I'm aware of need to be grounded at the AC input for conducted emissions purposes. Your PC is the same way.
1. You can make all the cabling as short as possible, plug both PC and SMPS into the same outlet / strip.
2. You can use a laptop with a 10 hour battery life and run it on battery when you listen.
3. You can get a BT connected amp and "suffer" the available bitrates of that wireless connection.
4. You can get a USB isolator and put that between the grounded PC and amp.
5. You can buy / build an analog DC power supply, let the secondary side ground to the PC through the USB cable. Note: this is potentially unsafe from an electrical fault perspective, as I'm sure the USB cable ground wont carry the current from an inadvertent hot line connection - til the fuse breaker blows. My analog supply has the chassis connected to ground. I believe like on the Meanwell SMPS, you can connect either + or - to chassis on the secondary side. I let that float. I use BT to connect to the amp, so there's no ground loop issue to begin with.
Hope this helps,
To be honest I am not that confident with electric stuff, especially since we are talking about 220v ac here. What is an analog DC psu? You mean instead of using transistors like an smps to rectify the ac to DC it's using a different method? How do I search of an analog psu in the net for me to buy? Does the power supply really makes a difference to noise and hiss? I know that ground looping is a different problem I was just curious if it would help. I already have a class d connected to a round of the mil Chinese loptop power brick, I did got ground loop sience I am using a pc as a source but I fixed that first with Bluetooth and then with a ground loop isolator (just a signal transformer in between the line out and line in) but I still have a lot of noise (low Snr) and I thought the Chinese power brick would be the problem and maybe with a better smps in my next project I wouldn't have this problem. (as a side question do you know why the connex smps has a +20,0,-20 volt out put? I don't get where one would need the - 20)
I did mean "linear" - sorry for the confusion. Type "24V linear power supply" into ebay's search. You'll see a flood of Chinese products, as they realize the efficacy of such designs for audio use. Then choose the "Business and Industrial" catagory. There you may find a unit like;
International Power IHN24-3.6 24V Power Supply | eBay
This is similar to what I use. I'd assume you could search the 'net likewise.
Yes, a linear supply uses a different "method" for providing 24VDC from 220 AC. The biggest effect of the differing methods is efficiency, which is basically why switch-mode power supplies came about. You pay for that efficiency via switch-mode with the problems that design presents, which is basically the attendant noise such circuitry generates during the conversion process. This noise shows up on both conducted (wired) and radiated (wireless) emissions.
This noise shows up on the AC input cord. You may have noticed a large ferrite bead permanently attached to AC input or DC output cords of some power bricks, this is an attempt to stop these noise emissions by reducing them to a level where the product can be sold. All switch-mode power supplies of decent design have input filter circuitry with inductors and capacitors, designed to suppress conducted emissions from getting out the AC cord. These filters require a ground connection to be most effective.
The type of power supply may make a difference in the noise and hiss you're hearing. However, that may be due to the amp design itself; you could spend $100 on a linear and find no change. I've read that analog input class D amps (TPA types) have great immunity to power supply noise, while the digital input (TAS) amps not so much. I use the TAS type, so I'm "extra careful" on the power supply for it.
Now I know why the connex PSU costs double - you pay for that extra -20V output that you wouldnt be using...
If you have access to a 4-5 18650 lithium batteries, you could do an experiment to suss out the noise culprit. Connect them in series to make a 16 - 20V string and power your amp with the batteries. If the noise is something you hear with no music signal applied, can you still hear it when powered by the batteries? If so, going to a linear supply wont help your situation. (Careful when using 2, 12V lead acid batteries, as that series string may exceed the amp's max input voltage - when fully charged) It may be an experiment worth doing and should be pretty safe to do - compared to messin' with AC line powered stuff while not being particularly confident with it.
For transformer audio isolation, I suppose the final sound depends on the transformer. Audio is like the second law of thermodynamics - everytime the audio signal changes "form" - you lose something. Going through the transformer it changes from electical to magnetic and back to electrical. Audio is actually even worse; a signal going through anything (cable...) not even changing form - you lose something. Or so they say. Thinking in this context, something like a miniDSP is a terrible thing to do - analog input form to digital, then digital back to analog - I wonder what that does to the sound?
I realize some computers (Apple) have optical digital outputs; my lowly HP PC business desktop didnt come with that feature. Of course, so they say, converting to optical and back again isnt ideal, "something is lost" though it's hard to believe this happens to "bits" in a stream of digital words that comprise the music signal at that point.
Designing an audio playback unit is an adventure in systems architecture . Choosing a power supply, just where a digitally encoded piece of music changes to a form you can listen to, what that form goes through before it hits the speaker or headphone driver - it's all a juggle and knowing some of the myriad tricks and trade-offs. On top of that, it's personal - I'm 63 and have committed to low, CD res because... I admit I probably couldnt hear the difference anyway. The only "analog" part of my two systems is the power supply, the speakers themselves and the speaker-wires. No line level wires, no ground loops, no crossovers, no transformers or capacitors in my signal path. That's just one way to architect it's a nice fit for me. Looking around the forums, everyone's certainly got their own ideas they're pursuing - it's amazing to me the number of ways people come up with to do it - like how many ways you can build a house or automobile - particularly the range of effort between the best and the worst in how these things are put together.
Last edited:
smps have transformers but they are much smaller and cheaper than linear supply.
i recently used an 20 watt smps on a jlh96 project.
it was £9.
Anything that does not require burst power works very well with a quiet SMPS. I've had very good luck with Meanwell, but have used cheaper supplies on Class A amplifiers with no issues at all. I've not had much luck with Class D though, all of them seem to have some residual noise that I can't fix even with a high quality linear PSU.
Same here - i am testing a meanwell LRS350-48 for a class A amp - so far it does seem to be quite outstanding. There is a tiny bit of ripple at the outputs around the switching frequency tho at about the same levels as the omnipresent mains harmonics from the environment.
Do you use any post filtering techniques or add a lot of capacitance to the meanwells you use ? Their internal current limiting should take care of the cap charge surge ...
Do you use any post filtering techniques or add a lot of capacitance to the meanwells you use ? Their internal current limiting should take care of the cap charge surge ...
I do use a small LC filter to clean up HF ripple and where performance is absolutely critical I would use a cap multiplier instead of that. Wurth has some nice common-mode HF inductors with up to 5A capability.
Large cap banks seem to hurt switchers more than help, I see a rise in output noise when connecting large caps. Meanwell supplies seem to have very little output capacitance but lower residuals.
I've never had much luck with Push-pull Class A, or Class AB amps with switchers - at least not the ones we get here. I'm sure there may be better options, just that I can't seem to access those.
Large cap banks seem to hurt switchers more than help, I see a rise in output noise when connecting large caps. Meanwell supplies seem to have very little output capacitance but lower residuals.
I've never had much luck with Push-pull Class A, or Class AB amps with switchers - at least not the ones we get here. I'm sure there may be better options, just that I can't seem to access those.
There is a slight lack of clarity in the midrange when using switchers. I think it's the inability to keep up with drastic changes in energy requirements in the midrange. I thought bass would be an issue but it's not, and neither is the top end. I hadn't measured any of this, but I generally stay away from this combination unless cost becomes an issue (professional amps are one area where I might use it, with a cap multiplier).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.