Thoughts re: Auxiliary cooling of AVRs, Amps, etc.

[middlemarch]

I've recently purchased a new receiver, Denon A1H. Without a great deal of thought I installed the AC Infinity fan formerly used with my x6700 on top of it. While the four sides of the receiver are open, there is limited space above it, which the forward exhausting fan now occupies.

A previous receiver of mine, an x5200, which I'd given to a friend, failed after about a year installed in an enclosed cabinet, due to what seemed to be heat related causes. I purchased the Infinity for the x6700 after what seemed to me to be heat related crackling in one of the channels (but of course not proven). However that issue has never returned with the fan installed.

I recently ran across some interior shots of the A1H showing an array of six muffin fans in the bottom of the cabinet, underneath the heat sinks. This started me thinking that, wow, if they ever turn on, I'd bet they flow a bit more air than my Infinity fan sitting in the air flow on top of the unit (rated at 160cfm), and it could possibly be a hindrance in that case instead of a help.

That got me thinking (I know...) that this whole topic of adding additional cooling to our devices seems pretty ad hoc, and I've not seen it discussed much, if at all. In my working career (EE) we paid a great deal of attention to identifying the hottest components on an assembly, then ensuring that available cooling was sufficient to keep the hottest junction temperatures well within limits.

I'm sure that the engineers at Denon, etc. have done the same. But since the cooling ultimately depends on the air flow through the unit, and they have no way to really control that environment, how do we, the end users, determine that we're providing the proper environment? Without some solid data to go on?

Is anyone aware of studies or analysis that's been done in this area? I suppose we could all just go with the conservative approach, plop a fan on top, and hope for the best. It's gotta be cooler than without, right? Can't hurt.

But what if those six muffins should come on in anger, would I be better off without the blockage on the top?

It seems with all the detailed electrical measurements being performed on these units, might it not be useful to somehow get a more data driven picture of cooling requirements, especially as it relates to reliability? I know, big topic, and we don't have the manufacturers thermal analysis, which would be essential.

Anyway, just thinking. Would love to hear other's thoughts on the topic, or else tell me I'm over thinking this, and just shut up already.

 

[BJL]

I think that you should remove your external fan, make sure there is sufficient room for air circulation around the unit, trust that Denon's engineers know what they are doing, and just forget about the cooling issue and enjoy the receiver.

 

[Vincentponcet]

If you want to add a fan , it should be in the cabinet to get outside air in the cabinet

 

[amirm]

I recently ran across some interior shots of the A1H showing an array of six muffin fans in the bottom of the cabinet, underneath the heat sinks.

I have seen these fans in just about every AVR. Alas, they seem to only be there to satisfy UL heat stress test. In practice, the let the unit cook without coming on.

Any kind of forced cooling is far superior to convection. So if you are not bothered by the noise the external fans make, they are likely adding to the life of the product.

 

[Chrispy]

I've had my avrs in small spaces often and do use external fans to help out. The internal fans, meh.

 

[respice finem]

People with biggish home cinema setups tend to have their hardware in actively cooled 19" racks.
These stand in a dedicated small room, so fan noise doesn't matter any more.

Der 19-Zoll-Serverschrank als Technik-Rack für dein Heimkino

Die mit Abstand beste und professionellste Art, deine Heimkino-Technik aufzustellen und zu organisieren ist der 19-Zoll-Serverschrank.

www.heimkino-praxis.de

I'm using a small wall rack for my surveillance system, and the mini-room it's in, for my main PC.
Six fans, five HDD, I hear nothing of it in my "music room", because it's behind a solid wall.

 

[dlaloum]

A couple of observations...

Typically the fans and big heatsinks, are mounted to the power amplifier circuits, which will only really get hot at output levels that most of us will NEVER use in our homes... My AVR is capable of 128W/ch@8ohm but I usually have it idling along at circa 1W up to perhaps 3W when listening loud, and peaks of up to perhaps 16W (based on calculating based on speaker efficiency spec, and MLP measurements of SPL).

So unsurprisingly, the big power amp heatsinks never warm up much!

What does warm up, are the HDMI and DSP chips - which work very hard processing the input signals, decoding streams, mixing and converting them before handing them over in analogue form to the pre and power amp circuits.

Many previous generation AVR's (including my 2008 Onkyo SR876 and 2013 Integra DTR70.4) ended up cooking themselves as a result - although the power amp circuits had large impressive heatsinks, the HDMI and DSP chips had no heatsinks at all, the case above those chips used to get so hot you could not touch it - how hot were the chips underneath given the case above them was so hot!?!? (and note those chips got just as hot, even if you used the AVR as an AVP, bypassing and keeping the power amp circuits in idle!)

Ideally, all those chips should have had their own heatsinks, and perhaps even their own fans.

If you look at some of the current designs, that is exactly what has happened - we now see small heatsinks on the HDMI/DSP boards, and frequently we also see a fan dedicated to the processing circuit boards.... and as a result, if you feel the case above the processing boards on current generation AVR's it is often no more than slightly warm - AS IT SHOULD BE! (in an open air setup...)

With a forced air setup such as the AC infinity, the goal should be to keep the electronics at that "warm" state and no hotter...in an enclosed cabinet, that warmth can build up even on the best designed current AVR's with the best heatsinks and fans - so you need to ensure proper air inlet and egress... and the AC infinity can help achieve this.

 

[Chr1]

This sounds exactly right to me. I find it astounding that it's taken this long for AVR manufacturers to address this properly.
Most AVR failures I have read about are due to cooked HDMI boards and DSP chips etc.

 

[dlaloum]

This sounds exactly right to me. I find it astounding that it's taken this long for AVR manufacturers to address this properly.
Most AVR failures I have read about are due to cooked HDMI boards and DSP chips etc.

I think the engineers were familiar with the heat generated by power amps, and therefore engineered ample margins for that, but especially during the early generations of complex DSP (early 21st century) they were not familiar with keeping processors (CPU's / GPU's) cool... they therefore did what they usually do with generic silicon... engineered based on the manufacturers heat tolerance specifications. - What they did not do is provide additional margins based on their own engineering experience. (probably got rolled by the beancounters)

So all the HDMI and DSP chips ended up running close to their manufacturer tolerance levels... theoretically not an issue - but it leaves no margin for other factors! - when Onkyo had a bad batch of capacitors, which did not live up to their manufacturer temperature spec, the HOT HDMI/DSP chips cooked the nearby capacitors causing a huge rash of AVR failures ... they ended up replacing a heap of HDMI boards.

Later TI produced a flaky DSP chip, which had a nasty tendency to warp/twist if it reached a high temperature - a temperature that SHOULD have been within spec... a small heatsink, even without fan, would have fixed the problem, but the engineers depended on the manufacturer specs, and without the depth of experience to lean on with regards to margins of error... got rolled by the beancounters - and for the lack of a 50c heatsink, an entire generation of AVP/AVR's suffered from DSP driven failures.

With a decade of these issues providing the actual actuarial support for additional safety margins... the engineers seem to have beaten the beancounters in the current generation... AVR/AVP's of the 2008 to 2015 period (roughly) all tended to run seriously hot regardless of manufacturer ... current models tend to run merely warm (which is the way we want them to!) - collectors of AVR's from that period, will probably find it hard to find intact examples... whereas our current ones, the lower temperatures involved bode well for a long service life...

 

[wwenze]

Maybe those fans already turn on without you knowing? The fins are widely spaced which is designed for either passive convection or very low airflow.

Generally with temperature-controlled fans I am not worried about longevity up till warranty period. Any extra external cooling is a plus, especially with fans that shut off completely.

But what if those six muffins should come on in anger, would I be better off without the blockage on the top?

In majority of cases, if the vents themselves are not very dense to begin with, no.

Cascaded fans are seen in many places requiring higher pressure - which can translate into higher CFM: think of pressure as voltage and CFM as current

A stopped AC Infinity there will have minimal effect on blocking the airflow, while a moving AC Infinity will boost the airflow even further since it still creates a pressure difference on top of what is being created by the internal fans.

Add: Here's the fan, 12V 0.12A, 2000rpm 28CFM.

 

[middlemarch]

Lots of very good information here, thank you! I found the part number and spec for that fan but couldn't find the CFM rating. Now we have that, thank you.

You're correct, I personally don't know if the built in fans have ever come on, at those ratings they must be pretty quiet. I did key in on Amir's comments that in his experience they only come on in extreme conditions, I suspect he stresses these things a bit more in his testing than most of us would ever do.

I found dlaloums comments of failure rates over time really interesting. The only unit in my experience that failed thermally, that x5200, was built in 2013. That was confirmed since it died slowly and could initially be revived if allowed to cool, until it finally gave up.

I definitely intend to keep the Infinity in place, cooler is better. I feel for the engineers designing these things, trying to anticipate all the operating conditions customers subject them to. I suspect cost is one inhibiting factor, but weight would be another. Nothing like a really big thermal mass to solve a cooling problem.