Do Pre-pros & AVRs have lower output when too cold?

[AcuDefTechGuy]

I didn't take a picture of the Denon with the fans on top. But here is a picture of the Denon I took a while back. The Denon is sitting on the top shelf. This closet is 8'x12'x10' ceiling. It has it's own HVAC thermostat.

Ignore that flower rug. It's not really there. It's only an illusion.

 

[Irvrobinson]

The fans are literally sitting on top of the Pre-pro and AVR chassis, and they are blowing air upward.

I can't see the configuration of the vents on the top of the AVR, but if they are just little slots stamped into the top, the fans may not be effective at all, and may actually be acting as a heat block of sorts, and may be causing the unit to overheat. I don't know what sort of protection circuits Denon uses, but they may be restrictions on amp output. Fans like that are generally meant for open air use, like to cool a cabinet.

 

[AcuDefTechGuy]

I can't see the configuration of the vents on the top of the AVR, but if they are just little slots stamped into the top, the fans may not be effective at all, and may actually be acting as a heat block of sorts, and may be causing the unit to overheat. I don't know what sort of protection circuits Denon uses, but they may be restrictions on amp output. Fans like that are generally meant for open air use, like to cool a cabinet.

Oh, the fans definitely are effective in pulling the heat up. When I placed my hand over the fans, I could feel the warm air blowing upward. Also, the AVP-A1 and X3100 stayed extremely cool. And I emphasize the world "cool" to the touch. Not warm, but literally cool.

A few weeks ago I did measure the temp of the Denon X3100 with the 4 fans and I think I mentioned it somewhere on the forum. Or maybe I just wrote it on my iPad. But I believe that even after 2 hours of loud 90+ dB music, the X3100's temp was only in mid 80s when I used my laster temp gun to measure the top and side of the chassis.

So I can verify with 100% confidence that these fans are extremely effective in removing the heat.

IOW, if the volume output were either increased or unchanged with these fans, I would have 4 or 5 fans on both the AVP-A1 and X3100. They are that effective in removing heat.

 

[TLS Guy]

It would be good if someone else could repeat this cause-and-effect. The Magnitude of it is Rather astounding! I have no doubt that ADTG is honestly reporting his observations, however there is always a possibility of a unknown variable unique to his situation confounding the relationship of temp vs SPL.

A sample of one should always be looked at with skepticism until confirmed.

Not if the data is correct.

We are talking about gain structure here and not power output per se. In other words increasing volume to maintain power with decreasing gain structure with temperature.

We are not the only ones to have had this discussion.

After doing further reading there is no doubt that the Beta and that means the gain structure of solid state devices is temperature dependent.

So how has this crept up on us.

Well I looked at a paper with a lot of advanced math, and it seems for an output of around a volt going from 50 C to 100 C will change output about 40 mv on average. So if you look at a single device spec. sheet then that is negligible in db. However modern devices have hundreds if not thousands of devices and almost all in series. So the gain losses are all mathematically summed in most cases. This now gets to significant db gain loss. I think we have just not paid attention to the changes in all the series junctions involved.

Now the next point. This is Newton's laws of cooling. Living here I can tell you wind chill matters.

An object cools in direct proportion to the temperature gradient at the interface.

However not with wind or running water. Then we get log decay! So once you add fans you have wind chill and much more effective heat transfer.

So the really only plausible explanation for ADTG's observations, which I believe are correct are: -

Change in gain structure (Beta) of solid state devices with temperature.

The additive series effect of these gain changes.

The much increased efficiency of thermal transfer of forced air over convection.

I have to say I feel rather foolish not having considered these effects more thoroughly before.

I would say to consider the voltage drop of low power draw cooling fans as not in the least bit plausible.

I have to say that in the winter when we return here from Eagan we come into a house at 55 F. That is where we leave the thermostats set. I have long told my wife not to start the rig until the house has warmed up as it does not sound right. I had assumed that this was due to stiffening of loudspeaker suspensions. That may be part of it, but this temperature effect on gain structure more likely has more to do with it. Don't forget that these changes may not only affect linear gain, but would also affect circuits influencing frequency response. So far more likely than not these phenomena can affect frequency response.

 

[TLS Guy]

I have not noticed such difference based on running REW for s few hours, with and without the fans blowing to my prepro and amp. Next time I will pay more attention in case I missed something, though I am quite confident it did not make any significant difference based on the graphs. In my case the difference in temperature would have been between 37 to 47 degrees C., on the top of the enclosure.

The trouble is it is the junction temperatures that count, and that is very hard to measure.

 

[KEW]

Not if the data is correct.

This is just scientific method.
If a cause effect relationship is observed, the next step is to check for repeatability.
You have provided plenty of information/theory to support the instance, but it is still good science to establish repeatability.
I assumed this fell under the domain of your medical science as well as the design of experiments classes I took.
It sounds like you may have linked to other instances where this same cause effect is confirmed, but I wanted to respond beforehand to make it clear that this is not out of doubt of your or ADTG's integrity, just wanting to do it "by the book".

 

[mtrycrafts]

Was this effect with program material or test tone source?
Try the internal level calibration tone with cold amps and set volume control position then at the end of your session without fans, same volume and test signal.

I have not sensed a need to change master volume on BD programs after I installed the large fans. Certainly not a huge 6 dB difference and allowing for the different recording levels of different program materials.

 

[Irvrobinson]

Oh, the fans definitely are effective in pulling the heat up. When I placed my hand over the fans, I could feel the warm air blowing upward. Also, the AVP-A1 and X3100 stayed extremely cool. And I emphasize the world "cool" to the touch. Not warm, but literally cool.

A few weeks ago I did measure the temp of the Denon X3100 with the 4 fans and I think I mentioned it somewhere on the forum. Or maybe I just wrote it on my iPad. But I believe that even after 2 hours of loud 90+ dB music, the X3100's temp was only in mid 80s when I used my laster temp gun to measure the top and side of the chassis.

So I can verify with 100% confidence that these fans are extremely effective in removing the heat.

IOW, if the volume output were either increased or unchanged with these fans, I would have 4 or 5 fans on both the AVP-A1 and X3100. They are that effective in removing heat.

The issue is whether the fans are really pulling enough heat out of the unit, or blocking more than they're pulling. All you know is that the air is warm, but you don't know if the air is as warm as it should be if it was properly cooling the unit. I think it's possible it's the latter, and that some sort of thermal overload explains the lack of output.

 

[M Code]

Where would that measurement best be taken?
Top of case?
Cooling fins?

I would recommend the temperature readings be taken @ (2) points:
a. Above the amplifier's heat sink
b. Above the power transformer

Just my $0.02..

 

[Irvrobinson]

Not if the data is correct.

We are talking about gain structure here and not power output per se. In other words increasing volume to maintain power with decreasing gain structure with temperature.

We are not the only ones to have had this discussion.

After doing further reading there is no doubt that the Beta and that means the gain structure of solid state devices is temperature dependent.

So how has this crept up on us.

Well I looked at a paper with a lot of advanced math, and it seems for an output of around a volt going from 50 C to 100 C will change output about 40 mv on average. So if you look at a single device spec. sheet then that is negligible in db. However modern devices have hundreds if not thousands of devices and almost all in series. So the gain losses are all mathematically summed in most cases. This now gets to significant db gain loss. I think we have just not paid attention to the changes in all the series junctions involved.

Now the next point. This is Newton's laws of cooling. Living here I can tell you wind chill matters.

An object cools in direct proportion to the temperature gradient at the interface.

However not with wind or running water. Then we get log decay! So once you add fans you have wind chill and much more effective heat transfer.

So the really only plausible explanation for ADTG's observations, which I believe are correct are: -

Change in gain structure (Beta) of solid state devices with temperature.

The additive series effect of these gain changes.

The much increased efficiency of thermal transfer of forced air over convection.

I have to say I feel rather foolish not having considered these effects more thoroughly before.

I would say to consider the voltage drop of low power draw cooling fans as not in the least bit plausible.

I have to say that in the winter when we return here from Eagan we come into a house at 55 F. That is where we leave the thermostats set. I have long told my wife not to start the rig until the house has warmed up as it does not sound right. I had assumed that this was due to stiffening of loudspeaker suspensions. That may be part of it, but this temperature effect on gain structure more likely has more to do with it. Don't forget that these changes may not only affect linear gain, but would also affect circuits influencing frequency response. So far more likely than not these phenomena can affect frequency response.

I think you're making more out of this than there is. If there really is a 6db loss in radiated output by the speakers then it's more likely some sort of voltage limiting circuit. Personally, this is one reason I don't like AVRs (I think you don't either) is that thermally they're a bad idea.

 

[PENG]

First of all, for reliable and meaningful readings you should use a test tone, say at 1 kHz, and ear plugs. Then I would measure the output voltage every 10 minutes.

If you are only getting 1 dB difference between time 0 (right after you turn everything on) and 2 hours later, it really is not significant. To have 6 dB difference, something else is wrong, not the beta imo. If the beta can change that much, even with the accumulative effects with that little temperature (say 20 deg F or a little more), how can we trust any SS device, used in other applications such as medical, military, navigation, or anything at all? Surely competent engineers and designers know how to compensate for the temperature effects, even if not perfectly.

I can't argue with TLSGuy at this moment though, only because he has read up on this and I have not, no time today for sure. For now I have to assume he may end up changing his mind, just my gut feeling..

 

[mtrycrafts]

First of all, for reliable and meaningful readings you should use a test tone, say at 1 kHz, and ear plugs. Then I would measure the output voltage every 10 minutes.
...

I would think that is a brutal test on the mid driver only. Perhaps using that internal level matching test tone at the beginning and at the end of a good movie, either spl or voltage.

 

[PENG]

I would think that is a brutal test on the mid driver only. Perhaps using that internal level matching test tone at the beginning and at the end of a good movie, either spl or voltage.

Whatever that can warm things up consistently and gradually.

 

[Swerd]

I'm coming in late on this thread, only reading about it now…

A change of 6 dB volume is large, no question about that. But what caused it? I’m with the others who say the only way to understand this better is to be skeptical. I also find it somewhat surprising that an AVR or prepro cannot adequately dissipate heat while its sides and top are open to ambient air.

To summarize ADTG's primary observations:
Four 120 mm cooling fans are lying on top of a Denon AVP A1HDCI (prepro) and Denon x3100 (AVR). They are arranged to pull air upwards.

Is that 4 fans on each item, or a total of 4 fans on both devices?​

With all fans working, measured SPL is 6 dB lower than when one fan is working on the AVR only. This was repeated 3 times after allowing for a 1 hour warm up period.

A photo shows the Denon (prepro or AVR?) sitting on the top shelf of an open-sided and open-backed shelf system. It was taken before cooling fans were added.

• Were the top and bottom vents in the chassis blocked? Do the cooling fans actually increase passive air flow, or do they block it? Don't assume the fans increase air flow and lower temperature without directly measuring it.
  
  
• Does heat rising from the 3 ATI amplifiers below change the operating temperature of that AVR or prepro?
  
  
• What are the temperatures involved here, both “cold” and “warm”? This leads to the question of just what range of temperatures are expected and normal for the AVR and prepro.
  
  
• Exactly how is temperature measured? What thermometer is used? If temperature readings become primary data in this story, was this thermometer calibrated against a traceable thermometer validated on known sources of heat.
  
  A simple way to calibrate a thermometer is to measure boiling water (100° C, 212° F) and ice water (0° C, 32° F), where some ice is not yet melted in ice cold water. Are the readings correct? And if they are off, by how much? If the thermometer can be adjusted, correct its readings.
  
  One way for dealing with a thermometer that is not calibrated or cannot be adjusted, is to use a second thermometer that works by a different mechanism.
  
  
• Where were the readings taken? Was the air temperature measured, or metal surfaces? With heat producing electronic devices designed for passive cooling, exactly where the readings are taken may be important.
  
  Take temperature readings at several different locations, both on surfaces and air temperature. You may be surprised at how much temperature can vary in refrigerators, freezers, or incubator ovens, depending on location.

KEW talked about the importance of scientific method. I agree.

• Check all measurements and observations to see if they are repeatable.
  
  
• Check the instruments used. Are they reliable? Know what the working range of a thermometer is, such as 85° ± how many degrees? No thermometer is perfectly accurrate and perfectly precise. (Accuracy is defined by comparing the thermometer's reading to a known source, such as boiling water or ice water. Precision is defined as measuring temperature of a known source at least 10 times and establishing the range of variance of the readings.)
  
  
• As far as cause and effect, don’t jump to conclusions. Don’t fall in love with your first idea.
  
  
• Assemble a list of all possible causes, include ones you like and dislike. Work on eliminating them, one by one, until fewer or only one cause remains. This is where the hard work begins.

 

[AcuDefTechGuy]

The issue is whether the fans are really pulling enough heat out of the unit, or blocking more than they're pulling.

I measured the temp with and without the fans yesterday.

With the fans, the top of the chassis at the rear of the Denon AVP was 35C degrees.

Without the fans, the same spot measured 50C degrees.

Physically, when the 4 fans are blowing, the chassis feels very cool - like room temperature cool when you touch a piece of metal.

Without the fans, the chassis feels significantly warmer.

 

[AcuDefTechGuy]

I would recommend the temperature readings be taken @ (2) points:
a. Above the amplifier's heat sink
b. Above the power transformer

Just my $0.02..

I took some measurements yesterday at the top of the chassis. The rear of the Denon measured the hottest. The front of the Denon measured the coolest. After 4 hours without the fans, the front top measured about 37C, while the rear top measured 50C.

 

[AcuDefTechGuy]

Anyone else measured the SPL at both "Cold" and "Hot" temperatures of the Pre-pro/Amps/AVRs?

My new HT equipment closet has 2 dedicated 20-Amp circuits vs a single 15-Amp circuit at my old house. Both have the same components (ATI amps, Denon AVP-A1HDCI, RBH SX-T2/R).

My new HT closet has it's own HVAC zone so it stays a lot cooler. My old HT room was always too hot in the Summer time because it was made to be an office, not a HT room.

The new HT room is a lot bigger and I sit a lot farther away from the speakers, about 17-18' listening distance vs 10' listening distance at the old house.

It's possible something is just plain weird with my new house and it requires a 60 minute warm up. But this makes sense with people saying they like to warm up their electronics before listening.

Yesterday, I measured the SPL again without any fans.

The Denon and ATI were powered off since 10PM the previous night. Hot SPL yesterday was 75dB for all speakers.

The Denon and ATI were turned back on at 4PM, which is 16 hours later.

The Cold SPL (Denon chassis measured 78F/26C degrees) was 72dB for the Front L/R, and 69dB for the Center.

After 60 minutes of playing movies, the Denon AVP-A1 measured 95F/35C degrees. The Hot SPL was back up to normal levels of 75dB for all speakers. The temp continued to rise to 50C degrees and stayed there until I turned off the Denon at 10PM.

My conclusion is that the 4 fans kept the Denon AVP's temperature below optimal operating temperature and all the speakers' SPL stayed below the normal numbers of 75dB.

Without the fans, it took 1 hour for the Denon to warm up and produced good SPL.

 

[Irvrobinson]

My conclusion is that the 4 fans kept the Denon AVP's temperature below optimal operating temperature and all the speakers' SPL stayed below the normal numbers of 75dB.

Without the fans, it took 1 hour for the Denon to warm up and produced good SPL.

I've never seen or heard of anything like this, nor can I think of a single electronic reason why it would be the case. In our previous home, we often saw interior temps of 60F, and I never heard of or experienced any negative effects from various kinds of equipment, including a Sony AVR. You have a real mystery on your hands.

ICs like the cold. Not just coolness, the cold. There's a moderately sized market for liquid CPU coolers for gaming machines.

 

[lovinthehd]

Interesting but would like more information on just what you're measuring and how...the avr's internal test tones? External? Is there a difference? Maybe its just the test tones...

 

[AcuDefTechGuy]

Interesting but would like more information on just what you're measuring and how...the avr's internal test tones? External? Is there a difference? Maybe its just the test tones...

In the past I think I've always powered up the Pre-Pro/Amps to warm up about 30 min before watching anything and I've never placed the fans atop the Denon AVP.

The reason I checked the test tones was because when I watched some movies, I had to turn up the master volume way up to around -5.0. The master volume is usually about -20.0. So it didn't make any sense why I had to turn the volume up that high.

And I noticed this happened after I placed the fans atop the Denon AVP.

So now without the fans, the volume is back to -20.0.

It does seem very strange.

I still wouldn't leave my Denon/ATI powered on 24/7.

But I sure will power up my system and let it warm up at least 30 min before watching anything from now on.