A cheap passive am4 cooler?

[LabRat 891]

If you can hear an active SSD, it has problems. You should never hear an SSD. If you do, it is typically caused by the piezoelectric effect due to faulty assembly (bad solder joints) or when faulty components expand and contract due to heat. Either way, it is not normal.

The comment was rhetorical (though factual), to make a point:
Arrangements of metal and tiny passive phase-change loops (AKA: heatpipes), don't make (audible) noise, all on their own (aside from thermal expansion/contraction; typ. inaudible or a creak/pop)

Which, is what myself and at least 2 others in the thread had recommended:
-Get big black heatsink, remove the fan, and use the existing (active) case ventilation to keep the heatsink from totally heat soaking.



Regardless, I am far from alone in noticing the effect in SSDs.
A quick search in any search engine, using the terms "SSD" and "noise" or "sound", reveals this isn't 'odd' or 'rare'; it's merely unnoticed by most people.

Piezoelectric effects are the cause, yes. However, according to Dell. it is not indicative of a fault.
(specifically, when caused by lower C-states)

The Solid State Drive (SSD) on my laptop makes a strange squeal/buzzing noise | Dell US

There's also paper(s) on RTN/Burst Noise in 3D NAND,

with Stanford Research paper(s) describing the audible qualities of RTN.

Wayback Machine

Note: I don't much care for Quora but, this answer talks about what is causing those piezoelectric effects in a more broad and friendly manner.

Dell and Mark Sin's overviews on the respective causes, along with the research from Stanford and DEIB del Politecnico di Milano lends me to believe
-ancillary components on a SSD can interact with and (effectively) amplify (in-die) electronic noise into something (barely) audible.
At the very least, audible piezoelectric noise from rapidly switching components, under both load and low-power states, can and does cause audible noise.

It can be an indication of fault or malfunction. However, there are far more than mere anecdotal accounts that it is not implicitly indicative of a fault.

Personally, It has occurred across every SSD I've ever had in an arrangement allowing of bringing my ear to it. Some, much quieter than others.
USB Adapted, Riser-extended, Bench Tests. SATA and NVME alike, even Optane.
Models I've noticed it occur in, off the top of my head:

Intel Optane M10 16GB (dozen+), Intel Optane P1600X 58GB(x2), Optane P1600X 118GB(x4), Samsung PM963 960GB (x2), Solidigm P41 Plus 2TB (x2), Sabrent Rocket 4.0 1TB, Sabrent Rocket Q4 1TB, Samsung 840 Evo 250GB, Crucial M4 256GB, OCZ Vertex 3 240GB.

Don't get me wrong, it is (typically) inaudible unless brought RIGHT UP to one's ear.

 

[Bill_Bright]

Not sure why your big defensive post. I did not say you were wrong, or that it didn't happen. I said if your SSD is making noise, it has problems because it is not normal.

There are no moving parts in a SSD. That is a simple fact.

Sound is created by something moving, sending sound waves through the air. That is also a simple fact.

Therefore, since there are no moving parts in a SSD, if you hear sound coming from the SSD, something is moving that is not supposed to be moving.

Ergo, that SSD has problems.

 

[LabRat 891]

Not sure why your big defensive post.

I drive the same way...

I did not say you were wrong, or that it didn't happen. I said if your SSD is making noise, it has problems because it is not normal.

You claim it a 'fault'. Both Industry and Academia disagree.
My anecdotal experiences also makes it seem unlikely (10+ years of SSDs, across SLC, MLC, TLC, QLC, and 3DXpoint) that it's 'abnormal' or 'faulty' behavior.

There are no moving parts in a SSD. That is a simple fact.

Only in the conventional sense. Piezoelectric (vibration) effects and RTN are implicitly a part of switching microelectronics. The effects must be engineered around or, in-hand with.
(Typically, the go-to solution, is to 'tune' the noise outside of human-hearing. See: SMPS' common PWM frequencies)

Sound is created by something moving, sending sound waves through the air. That is also a simple fact.

Correct.

Therefore, since there are no moving parts in a SSD, if you hear sound coming from the SSD, something is moving that is not supposed to be moving.

Ergo, that SSD has problems.

That is untrue, and it is not my opinion.
It's the (hopefully, factually-based) opinion of computer manufacturers like Dell.

 

[Bill_Bright]

You claim it a 'fault'. Both Industry and Academia disagree.

Oh? Not according the Laws of Physics. I would be interested in seeing where it is "normal" for inanimate objects to make noise.

My anecdotal experiences also makes it seem unlikely (12+ years of SSDs, across SLC, MLC, TLC, QLC, and 3DXpoint) that it's 'abnormal' or 'faulty' behavior.

Well, I won't hold your obvious youth and lack of experience against you.

Again - I am NOT saying it doesn't happen. Even power lines can "sing", and transformers and coils can hum. I am just saying it is not "normal" for SSDs. Nothing should be "vibrating", thus make noise.

I will concede that my choice of the words "problems" or even "fault" may not have been the best because I do not mean to suggest anything that might indicate the SSD has a performance issue, or that its life expectancy is affected. So my apologies for that.

At this point, this topic has nothing to do with SSDs - so best we just move on.

 

[the54thvoid]

@Bill_Bright - the member you're arguing with posted a link from DELL with regards to SSD noise. Namely this:

The Solid State Drive (SSD) on my laptop makes a strange squeal/buzzing noise | Dell US

Which, from DELL, says this:

Some portable systems equipped with solid state drives (SSD) emit an audible squeal when the system is idle or under light use. This is normal behavior.

That you are arguing the point is beyond ludicrous. It's a factual piece of data. If you want to continue like this, please do it on DELL's forum, because it's their say so that says SSD noise is normal.

Now, with that said, please talk about passive AM4 coolers.

 

[Bill_Bright]

Fair enough but please note in that same article, just a couple lines below what you quoted, it says,

The noise comes from the processor’s power circuit

 

[Chrispy_]

Tbh I already have a noucta d15 somewhere at home but I should find it
Although it does less noise, it's still noisy XD

Noise is down to the fan RPM and nothing else - you just need to create a fan profile that's quiet.

A is the highest speed you can't hear. It doesn't matter what speed this is, just find the RPM that is so quiet you can't hear it or don't care about it.
B is going to be about 80C, because you'll occasionally get short burst loads that heat up the CPU package, but it takes a decent full-TDP load a few seconds to hit 80C
C is the highest speed you can tolerate. It doesn't matter if this is 35% fan or 70% fan, it's entirely subjective. The higher you go the more performance you'll get.
C and D are at the "panic temperature", which is 100% fan at 92-94C to prevent aggressive CPU throttling at 95C.

Basically, if you hit 100% fan for any reason, you know your CPU is overwhelming the cooler and you have two choices.

1. Manually set PBO to reduce the peak power draw of your CPU. You'll lose a bit of performance but it'll be quiet.
2. Flatten the line at C so that the fan can't possibly spin any louder and simply accept that your CPU will throttle under sustained full loads.

Personally, I'd pick option 1 and spend a little time with prime95, OCCT, or Cinebench to work out how much PPT I can get away with for the amount of fan noise I'm willing to put up with. Doing that will get you the quietest PC you can possibly have with the hardware you have available, and shouldn't compromise speeds at full sustained loads by very much. Cinebench's 10-minute multithreaded tests is a great way to find out how much performance you've lost and for my HTPC I set a 115W PPT (so about a 20% power reduction) and I lose 7% of the performance that I get at the stock 142W PPT. That 7% loss also applies only to fully-threaded multicore workloads. Single-threaded and low-threaded stuff runs just as fast as "full" power

 

[lyx]

I was looking at the alpine passive am4 cooler

Alpine AM4 Passive | Silent CPU Cooler for AMD AM4 | ARCTIC

The ARCTIC Alpine AM4 Passive is a passive and thus completely silent cooling solution for AMD AM4 CPU.

www.arctic.de

But it's out of production and was stocked at -20€, fairly cheap, in the used market no luck at finding it

Anything else? I'm fine with used, but not fine at spending ≥50€ for a cooler

There are also the Dynatron A37 and A43. But you have to check the capacitors placement on your motherboard and make sure the bottom heatsink plate is not touching or crushing them.

Aliexpress also have another model similar to the A43, but with fewer heatpipes.

 

[sLowEnd]

There are also the Dynatron A37 and A43. But you have to check the capacitors placement on your motherboard and make sure the bottom heatsink plate is not touching or crushing them.

Aliexpress also have another model similar to the A43, but with fewer heatpipes.

The A37 is a tiny 1U server heatsink. It's not meant to be actually used passively, but instead it relies on case airflow, usually provided by extremely fast and loud fans.
e.g. Like this kind of layout

File:1U rack mount server cooling system fan heat-sink.jpg - Wikimedia Commons

commons.wikimedia.org

The A43 is a bit bigger as a 2U heatsink, but it still has a lot less surface area than a typical tower heatsink. I wouldn't count on it for passive operation of anything but extremely low power CPUs.

 

[Bill_Bright]

The A37 is a tiny 1U server heatsink. It's not meant to be actually used passively, but instead it relies on case airflow

I agree the A37 is designed for rackmount servers. In fact the product's webpage clearly says that here.

The following is likely obvious to some (apologies if so), but for others, and just to clarify, all heatsinks depend on case airflow. Even if the device's heatsink has a fan, that fan is useless if there is no (or an insufficient supply of) case airflow to move the heated air away from the device and out the case.

It is the case's responsibility to create a flow of cool air through the case. The device's heatsink need only toss the device's heat into that flow. And it is the user's responsibility to ensure the case can and is fulfilling that duty.

Now whether that flow is created by case fans, or simply facilitated by case design to support the flow created by the "convection" process is another story. But if 100% passive (no fans at all), there MUST be room around the heatsink (ideally above - since heat rises) to allow the heat that's been "conducted" through the heatsink materials to escape out of the case. The heated air that escapes creates a space that MUST be refilled. So (in a properly designed case) cool air is drawn in to fill it. Hot air out, cool air in creates that flow so the cooling process continues.