Quieting down SMPS fan noise

[Bruce Morgen]

I power my TPA3255 chip amps with a typical Mean Well clone SMPS that has an always-on two-wire cooling fan. In an effort to limit its cooling fan noise, I bought a supposedly quiet Vantec PC case fan that was too thick to mount inside the case, so I mounted it outside, and the resulting setup is just as loud as the original inside-mounted fan. Since current demand from the amps is relatively sporadic with occasional peaks and the SMPS runs pretty much ice-cold, I'm wondering if it's a sensible risk to run it without a fan, perhaps slow the fan down with a simple series resistor, or even concoct some sort of physical muffler arrangement. Any and all idea appreciated in advance!

 

[RayDunzl]

I have a couple of these to blow air at my amps...

https://www.amazon.com/Antec-TriCool-92mm-Cooling-3-Speed/dp/B00066ISVQ

 

[Count Arthur]

What size are the original fans?

PC fans come in a range of standard sizes, so you should be able to find a better quality, quieter fan that is a drop in replacement for the original.

Noctua are nice quality and include a "Low-Noise Adapter" cable: https://www.amazon.com/Noctua-NF-A9...eywords=15mm+noctua+fan&qid=1599899925&sr=8-5

 

[Webninja]

Just installed a Noctua case fan on my pc, and with the rubber mounts running at about half speed it’s dead quiet. The stock fan had a high pitch whirr that was audible from under the desk.

 

[Bruce Morgen]

Solution found. I had a DOA 36V 10A Mean Well clone which the AliExpress seller replaced hanging around, so I harvested its (essentially new, since the unit was dead) fan from it -- and lo and behold, it was pretty quiet despite being a cheap sleeve bearing model, perhaps because it's rated for a lower current draw (0.10A vs. the noisier fan's 0.15A) and therefore spins considerably slower. So, maybe it's just a matter of swapping in a 0.10A fan rather than any particular brand and part number -- but anyway, I'm a happy camper with a much quieter and still cool-running SMPS powering my amps!

 

[restorer-john]

A small linear adjustable voltage regulator will allow you to slow the fan and greatly silence it while still maintaining the airflow you need.

 

[Bruce Morgen]

A small linear adjustable voltage regulator will allow you to slow the fan and greatly silence it while still maintaining the airflow you need.

Good idea -- but easier to simply swap in a slower spinning fan with the same form factor.

 

[AnalogSteph]

BTW, please note that for a sleeve bearing fan, not all orientations may be OK long-term. If possible, make sure it's oriented vertically (blowing sideways).

https://www.digikey.com/eewiki/pages/viewpage.action?pageId=85295261

As can be seen from the following photos, operation with the hub pointing downward is likely to result in the lubricant promptly leaking into the hub for many sleeve bearing construction styles, resulting in prompt failure.

Sure enough, failed sleeve bearing fans usually seem to occur where they are mounted blowing upwards, e.g. on graphics cards in ATX tower cases or in ATX power supplies (for 120-140 mm sizes).

Some time last year I needed a 2-pin fan for some extra air circulation inside a 19" 1HE managed switch whose fans had all failed, so I grabbed an old 80x80x25 sleeve bearing fan out of some old power supply that wasn't exactly a spring chicken or dead quiet. Peeling off the label revealed a rubber plug, so I flooded the bearing with 3-in-1 / sowing machine oil (5W synthetic, I think?) and crossed my fingers. It's now blowing down vertically, slightly tilted, and has been doing so for months now, perhaps close to a year, 24/7. It wasn't a super high RPM model to begin with, so that probably helps, but I'll keep my eyes on it. The main fans were replaced by some Noctuas, which should make this one of the poshest DLink DGS-1248Ts around. (I have no idea why this model was even bought to begin with back then, it's way oversized and an unmanaged 16-port would do the job just the same... It was getting a bit toasty with all fans out though, thank goodness this already was the newer 3-fan model with power saving features...)

 

[Bruce Morgen]

BTW, please note that for a sleeve bearing fan, not all orientations may be OK long-term. If possible, make sure it's oriented vertically (blowing sideways).

https://www.digikey.com/eewiki/pages/viewpage.action?pageId=85295261


Sure enough, failed sleeve bearing fans usually seem to occur where they are mounted blowing upwards, e.g. on graphics cards in ATX tower cases or in ATX power supplies (for 120-140 mm sizes).

Some time last year I needed a 2-pin fan for some extra air circulation inside a 19" 1HE managed switch whose fans had all failed, so I grabbed an old 80x80x25 sleeve bearing fan out of some old power supply that wasn't exactly a spring chicken or dead quiet. Peeling off the label revealed a rubber plug, so I flooded the bearing with 3-in-1 / sowing machine oil (5W synthetic, I think?) and crossed my fingers. It's now blowing down vertically, slightly tilted, and has been doing so for months now, perhaps close to a year, 24/7. It wasn't a super high RPM model to begin with, so that probably helps, but I'll keep my eyes on it. The main fans were replaced by some Noctuas, which should make this one of the poshest DLink DGS-1248Ts around. (I have no idea why this model was even bought to begin with back then, it's way oversized and an unmanaged 16-port would do the job just the same... It was getting a bit toasty with all fans out though, thank goodness this already was the newer 3-fan model with power saving features...)

Interesting that sleeve bearing fans blowing upward are pretty much standard for most of the low-cost Mean Well lookalikes on offer via AliExpress and eBay. I know the real Mean Well units, e.g. the LRS-350-48 that's often recommended for TPA3255 applications, have a fan thermostat, but I have no idea what sort of bearings their fans have. Anyway, 60mm fans are cheap and I'm going be experimenting with a couple of ball bearing models over the next couple of weeks. Meanwhile, I've installed a mini-toggle switch to control the fan manually and added another plain aluminum heat sink outside the back of the case where the unit's main hot spot is. I hope I can come up with a fan that I can recommend confidently, because once you need more than the 7.3A the LRS-350-48 provides, real Mean Well units start to get pretty pricey.

 

[Solveit]

Interesting that sleeve bearing fans blowing upward are pretty much standard for most of the low-cost Mean Well lookalikes on offer via AliExpress and eBay. I know the real Mean Well units, e.g. the LRS-350-48 that's often recommended for TPA3255 applications, have a fan thermostat, but I have no idea what sort of bearings their fans have. Anyway, 60mm fans are cheap and I'm going be experimenting with a couple of ball bearing models over the next couple of weeks. Meanwhile, I've installed a mini-toggle switch to control the fan manually and added another plain aluminum heat sink outside the back of the case where the unit's main hot spot is. I hope I can come up with a fan that I can recommend confidently, because once you need more than the 7.3A the LRS-350-48 provides, real Mean Well units start to get pretty pricey.

Did you find any good replacement fan for Mean Well LRS-350-48?
I run two TPA3255 and the fan starts sometimes even at very low volume.
The LRS-350-36 didn't do that.

 

[Bruce Morgen]

Did you find any good replacement fan for Mean Well LRS-350-48?
I run two TPA3255 and the fan starts sometimes even at very low volume.
The LRS-350-36 didn't do that.

I have no direct experience with actual Mean Well PSUs and have no idea what sort of fan -- 2-pin? 3-pin?, 4-pin?, exact size? -- would be needed to quiet one of them down.

 

[Deleted member 9685]

Just buy LRS-200-36. LRS-200 series doesn't have a fan. You won't need to worry about fan noise.
LRS-350-36 is also good because it turns fan off when it's not hot.
Be aware that since LRS series doesn't have PFC components, power factor is about 0.65.
A non-PFC power supply must be able to supply at least 1.5 times the actual power needed by an amp.
With LRS-200, each speaker can output 65W at maximum because (200W * 0.65)/2 = 65W.

If your speakers are close to you, you won't need more than 8W for each speaker.
LRS-200-36 could do the job for you.

Note // By the way, active PFC increases power factor at the cost of leaking electrical noise from appliances into house wires. Low power factor is a buffer that prevents electrical noise from leaking into house wires. Passive PFC decreases electrical noise on the other hand because it is essentially just a low pass filter. If you want to reduce electrical noise going into your audio appliances, avoid active PFC power supplies in your house. An active PFC computer power supply can increase electrical noise significantly in all house wires.

 

[Solveit]

I lowered my LRS-350-48 to 46v and now it just have started once with heavy volume.
Im happy with that.

 

[Deleted member 9685]

I lowered my LRS-350-48 to 46v and now it just have started once with heavy volume.
Im happy with that.

Do you actually listen at heavy volumes?

 

[Solveit]

Do you actually listen at heavy volumes?

Just for short whiles but very annoying to have the fan running after that in low volume.

 

[Deleted member 9685]

Just for short whiles but very annoying to have the fan running after that in low volume.

You may want to try LRS-200-36 with a lower amp volume so that you don't push it too far.

 

[Solveit]

You may want to try LRS-200-36 with a lower amp volume so that you don't push it too far.

I have LRS-350-36 and the fan never started but the LRS-350-48 sounds less distorted so I prefer that.
No way I goe with an lower 200w psu.

 

[wwenze]

I generally advice against changing the fan of the power supply (Have heard people talking about since the Pentium4 days). Firstly you will pretty much void the PSU's warranty by opening it up. Secondly you can't guarantee the new fan will be able to safely cool the PSU under stress or fault condition.

With these kind of power supplies, you can cool it well with an external fan. You can see the indentations and screws where the transistors bolt onto the chassis so the entire bottom plate of the chassis is a heatsink. Cooling this will greatly reduce the need for the internal fan to turn on. In this arrangement, removing the top cover and replacing it with the fanless version (honeycomb) further helps. Or you can run it without the top cover but you lose one layer of safety against electrocution so either practice good spacing practices or hide your equipment where the insurance guy can't see.

Oh by cooling externally I mean at least a 80x80x25mm fan so it is much quieter. Would be counter-productive if you just took the same small fan and relocated it outside.

 

[Bruce Morgen]

I generally advice against changing the fan of the power supply (Have heard people talking about since the Pentium4 days). Firstly you will pretty much void the PSU's warranty by opening it up. Secondly you can't guarantee the new fan will be able to safely cool the PSU under stress or fault condition.

With these kind of power supplies, you can cool it well with an external fan. You can see the indentations and screws where the transistors bolt onto the chassis so the entire bottom plate of the chassis is a heatsink. Cooling this will greatly reduce the need for the internal fan to turn on. In this arrangement, removing the top cover and replacing it with the fanless version (honeycomb) further helps. Or you can run it without the top cover but you lose one layer of safety against electrocution so either practice good spacing practices or hide your equipment where the insurance guy can't see.

Oh by cooling externally I mean at least a 80x80x25mm fan so it is much quieter. Would be counter-productive if you just took the same small fan and relocated it outside.

A small Class D audio power amp presents a very different load to a PSU than a PC does. Peak current demands tend be extremely brief in such an application, making the likelihood of a conservatively rated (e.g. 7+ amps for an A07) SMPS overheating almost non-existent when reproducing music or even a bombastic movie soundtrack. My fancy variable-voltage 10A SMPS is set for 46V, and its thermostatically-controlled fan has never even turned a single revolution running an A07 turned up about as loud as my ears can endure into nominal 4 ohm speaker loads. My other 10A SMPS had an always-on fan, which I rewired with a switch so I could turn it on if/as required -- and I've never once had to touch that switch under those same conditions. As for warranty issues -- well, frankly, who cares when we're talking about a $25-30USD (shipped!) generic SMPS? The return shipping to China alone makes sending such an item back for repair or replacement a non-starter anyway, so I say experiment away as long as you can do it safely!