[Simon]

Originally posted by rhfactor:
Simon, you're the first person I've run into that has made a scientific statement about this.

Thanks Rich. What bothers me about the whole discussion is that it is highly irrational. We get the guys that come here and say a rocket engine firing up is more quiet than an MDD and then we get the other guys who come in here and say that they can hear a feather fall to the ground next to their MDD.

AFAIC that's all crap.

1. There is no MDD in terms of noise. There are *measurements* that show that different MDDs have different sound spectra and different volume.

2. These comparissons are crap too. If I say it's as loud as my hair dryer that doesn't mean squat because first nobody on these boards knows what my hair dryer sounds like and secondly I didn't say what characteristic of the sound was like what characteristic of my dryer. The frequency? The volume?

3. These boards are full of people of who a) don't care b) care too much c) don't know what they're talking about or d) lie

And I am a scientist (but not an acoustic specialist!) so I hate talking about the grade of something which we can't measure.

In an ideal world we would define a measurement procedure everybody could do for their MDD. We would compare results. Analyze data. Get a bunch of engineers to tell us how to fix stuff and then a bunch of technicians to actually do it. We then would go back to step one and iterate until our MDDs sound like iMacs. (disclaimer for the wize guys: no, I know they will never be that quiet, I was just joking)

I know that for me the high frequency whining pf the PS is way more annoying and grating than a 5-times-louder window airconditioning unit with a white-noise sound.

Yes, 100%. The phon scale here will proove this statement



The page this image is from has some more explanations for those interested.

Aside from google, can you point me to soem good reference material that corroborates your statement about greater sensitivity to high frequency sound?
...
ANyway, I'd like to know more about phon IIRC measurements.

Check your mailbox for more info. I sent you some stuff. BTW, it's just "phon" and you can find exact definitions and formulae in every physics textbook. The IIRC had nothing to do with the unit name (IIRC = if I recall correctly)

[Edit: Finally made my image link work properly]

[Simon]

Something else that bothers me is why Apple can't use some tech to get the problem straight. They have pioneered a lot of technologies so it can't be that they're afraid of new approaches... Maybe it's cost (can't imagine though), I don't know, but anyway, why don't they work with noise cancellation?

It's a pretty simple idea: you have a mic, a DSP chip and a speaker. The mic records the noise, the DSP does a spectrum analysis and calculates for each frequency found in the recording a wave that is phase-shifted by pi. These coefficients are then put together and sent to the speaker as a "counter-wave". This leads to cancellation of a large part of the noise, especially when the noise is a more or less constant waveform. If apple can't fix the PowerMac noise better by using other fans or damping material they (and we) should start looking into this. They could put it in future models and we could maybe one day buy mod kits for existing hardware to cancel at least some of the noise...

The technology is already developed and being used in everyday life. If you sit in a modern car it's already being used. If you sit into an Airbus A320, A330 or A340 series (don't know about Boeing) plane it's already being used.

The way I see it they either haven't looked into it, they don't think it's that a deal or they think it's too expensive.

If a) then their mechanical engineers should get off of the chairs and take a look at it. If b) we should start b!tching more and put on some presssure and if c) their PR department should start to think about the fact that people who pay 2000$ for a G4 are probably also willing to pay 10 bucks more if the machine is very quiet. Hell, Apple should be a Mercedes computer and incorporate all state of the art possibilities necessary to make it "perfect". If Shiller doesn't understand this he should maybe go to a Mac store and listen to all the people complaining about the noise. I recently went to my local dealer and I heard a discussion between a couple potential switchers and an employee. It was pretty embarassing. I think Apple shouldn't have to deal with crap like that. They can do much much better.

[Eug]

Noise cancellation is impractical due to cost concerns, etc., in a cut throat computer industry. Plus it often doesn't work well with higher pitched noises. If you must have it I'm sure it could be done as an aftermarket solution though.

The more direct solution:

1) New CPUs, ones that are not 42% overclocked over max spec. Can anyone buy 1.42 GHz G4 7455s besides Apple? I see them listed nowhere.
2) Quieter fans. Deltas are usually built for performance, not acoustics. Quieter Sunons or Panaflos or whatever would be nice. Furthermore 60 mm fans are not ideal for the noise to air flow ratio. Larger fans move as much air with lower noise, and lower frequency noise. I wonder why they chose dual 60s instead of a single 80.
3) Some simple damping.

Anyways, to the doubters, I have basically finished my homebuilt quiet peeeceee system. Basically all I did was swap out some fans and run them at lower rpms, and got a new hard drive. If I stand 3 feet away, it is almost inaudible. It used to be as loud as the MDD PowerMacs I played with in the store.

So yes, noise reduction is definitely possible. I must admit, I would have had a hard time accomplishing this though if I had to deal with 42% (or even 21%) overclocked CPUs.

[Simon]

Originally posted by Eug:
Noise cancellation is impractical due to cost concerns, etc., in a cut throat computer industry.

I'd doubt that. There are 40$ headphones available with this technology and Apple isn't in the cut-throat part of the industry. They don't make the 400$ machines. Don't forget, their margin on the PowerMacs is about 20%. Making the PowerMac more silent for 10$ is not going to change the amount sold nor their profit IMHO.

Plus it often doesn't work well with higher pitched noises. If you must have it I'm sure it could be done as an aftermarket solution though.

I didn't know about the high-pitched part. I'd like to know more about that. The wooosh of air is around 300Hz. The high-pitched whine of some MDDs is about 1200Hz. Is that such a big difference in terms of noise cancellation devices?

And I like the part about the aftermarket solution. Does anybody here know of a company that would sell us such a box (mic, DSP, speaker) to set up next to the MDD for less than 500$.

1) New CPUs, ones that are not 42% overclocked
...
I would have had a hard time accomplishing this though if I had to deal with 42% (or even 21%) overclocked CPUs.

Eug, I hate to tell you but that 42% overclocking business won't help me squat. You see, my MDD is not overclocked, nevertheless it's too loud. On the other hand a friend of mine has an old dual 1.25 (which according to your theory) should be overclocked. Both are about the same in terms of loudness. Fact is, neither opf our machines are loud because of the temperature-controled CPU fan. The loud guys are the 60mm PS fans and they don't care about CPU - overclocked or not. I can back this up. I unplugged the 120mm CPU fan from my MoBo so it would not come on. I booted and guess what? The machine was excatly as loud as before, i.e. for my MDD the CPU fan doesn't influence any loudness as long as the PS fans are on.

BTW, do we have any proof about this whole overclocking business that Moto is accused of doing? I mean real hard proof - not just some speculation. I know that on Moto's page there isn't any info on 7455s with more than 1GHz but that can't mean all the >1GHz chips are overclocked. It could just as well mean that Moto and Apple have a contract that says only Apple gets >1GHz chips...

[Eug]

Originally posted by Simon:
I'd doubt that. There are 40$ headphones available with this technology and Apple isn't in the cut-throat part of the industry. They don't make the 400$ machines. Don't forget, their margin on the PowerMacs is about 20%. Making the PowerMac more silent for 10$ is not going to change the amount sold nor their profit IMHO.

$10 a computer is a boatload of money. I'm guessing it'd be cheaper to use different fans. Anyways, the $40 headphones with it don't work so well. The ones that do are $300. Mind you, better noise cancellation setups would be needed for headphones than for case noise of course.

I didn't know about the high-pitched part. I'd like to know more about that. The wooosh of air is around 300Hz. The high-pitched whine of some MDDs is about 1200Hz. Is that such a big difference in terms of noise cancellation devices?

Yes, but now that you've reminded me of the frequencies, maybe the 1200 Hz would be partially covered too, or maybe not. Not sure. Those who fly say that the low rumble of a plane is covered well by noise cancellation, but higher pitched noises are not. I'm not sure exactly where the relative cutoff for "higher pitched" comes in. A baby's crying comes in loud and clear. Sudden changing noises aren't cancelled either. ie. People say that normal conversation by others on the plane suddenly become super loud with noise cancellation, because they talk louder to overcome the plane noise. However, since you're not hearing that plane noise as much the voices get very annoying.

Eug, I hate to tell you but that 42% overclocking business won't help me squat. You see, my MDD is not overclocked, nevertheless it's too loud. On the other hand a friend of mine has an old dual 1.25 (which according to your theory) should be overclocked. Both are about the same in terms of loudness. Fact is, neither opf our machines are loud because of the temperature-controled CPU fan. The loud guys are the 60mm PS fans and they don't care about CPU - overclocked or not. I can back this up. I unplugged the 120mm CPU fan from my MoBo so it would not come on. I booted and guess what? The machine was excatly as loud as before, i.e. for my MDD the CPU fan doesn't influence any loudness as long as the PS fans are on.

BTW, do we have any proof about this whole overclocking business that Moto is accused of doing? I mean real hard proof - not just some speculation. I know that on Moto's page there isn't any info on 7455s with more than 1GHz but that can't mean all the >1GHz chips are overclocked. It could just as well mean that Moto and Apple have a contract that says only Apple gets >1GHz chips...

Conjecture on my part about the overclocking. But it does make a lot of sense. The chip is supposed to only reach 1 GHz, which is in line with x86 chips on the 0.18 u process. Even Intel has not attempted to make a 1.4 GHz PIII Coppermine 0.18 u. All their 1.4 GHz chips are 0.13 u. Over 1 GHz the costs go way up, with low yields, etc. Sure > 1 GHz existed, but the real solution was to design the Tualatin 0.13 u and the 0.13 u P4. And even the official > 1 GHz chips out there still require higher volting and put out more heat (and thus require more heavy duty cooling). Now I understand the architectures are different, but still, Motorola itself still says that the 7455 is a 1 GHz part, and I think for good reason.

Now the fan business is a different story. Nowhere did I say it has to be temp/thermistor controlled. They probably aren't but Apple just wants one common part for all the MDDs. So I'm guessing that regardless if you have a 1 GHz or a dual 1.42 GHz, you'd probably have the same PSU. They want to have a PSU that can handle anything in their line, and then some. Also, the CPU fan is not the main issue, it's the heat inside the case that needs to be expelled. Part of that is done by the PSU fans. Having a 232352 cfm CPU fan won't help matters much if all that heat still gets trapped inside the case.

And you confirmed what I said about the larger fans. The 120 mm isn't making the most noise, it's the PSU. I'm not sure going with dual 60 mm PSU fans was the best idea.

But then again, what do I know. I haven't played inside a MDD.

P.S. My PSU on that PC was thermistor controlled. But the control is completely out of whack. At anything beyond about room temperature-ish, the fan speed ramps up. So effectively it always ran at ear-numbing full speed. The Panaflo I replaced it with has no thermistor control, but overall is much quieter. So thermistor control may or may not be helpful, but more useful may be the choice of the original fan setup in the first place. Mind you the Panaflo cost me about US$12 retail, whereas a higher output (and louder and lower quality) cheapie fan costs about $4.

[SMacTech]

Originally posted by Akarso:
Jesus christ! I'm sooo tired of people complaining about the whole MDD noise issue! It's not that bad!! What are you guys doing when you're using your computer anyways? Putting your ear right next to the side of the case?

I am programming on mine and certainly not putting my ear next to the case. Quit being so bloody ignorant. The MDDs are loud, some are louder than others, especially those purchased when they first come out. The majority of the noise is coming from the PSU fans. I have purchased about 6 of them here at work, each one is different, it seems. Mine at home is much louder as I purchased it last August. So loud I was going to return it, but it was so nice and fast, I put it in the server room and I now enjoy peace and quiet.

[Simon]

Eug, thanks for the detailed reply.

Originally posted by Eug:
P.S. My PSU on that PC was thermistor controlled. But the control is completely out of whack. At anything beyond about room temperature-ish, the fan speed ramps up. So effectively it always ran at ear-numbing full speed. The Panaflo I replaced it with has no thermistor control, but overall is much quieter. So thermistor control may or may not be helpful, but more useful may be the choice of the original fan setup in the first place. Mind you the Panaflo cost me about US$12 retail, whereas a higher output (and louder and lower quality) cheapie fan costs about $4.

I was told by an Apple dealer friend of mine (well actually he's a tech guy in the back not a salesman) that the new MDDs actually have temp dependent PS fans. I'm anxious to see how they sound and how they cycle. Anyway, I was told by him and I also get the impression on these boards that the new MDDs are really better. Surely not perfect and not close to the tweaked self-built silent cooling some of you guys have done on your PCs, but certainly better than the old MDDs in a sense that you can really hear a difference.

Well, there have been one two people who claim that the new MDDs are actually as loud as the old ones. But I doubt those statements. There are not many 14 year olds who can afford two MDDs.

[Simon]

Originally posted by rhfactor:
ANyway, I'd like to know more about phon IIRC measurements.

Rich, I saw you posted the phon business on your own forums. Thanks for the compliments btw.

I wanted to try to contribute over there, but I didn't get the registration confirmation mail (check your log, a guy wanted to register as "Simon" from xx.218.57.178 ). Hope you can fix this. I think we need to get up and moving on measurement procedures. I think your forums would be an excellent place to start doing so.

[danengel]

About overclocking: Does anybody know why in a production line there are some chips that run at higher speeds than others? Production faults are clear, some chips work, others are spoilt because of dirt on the masks, but how can one chip be faster than another?

[Chun Hsu]

Simon,

I did noise analysis a few months ago on my MDD and posted some instructions on the web:

http://home.centurytel.net/chunhsu/mac/noise/index.html

The gist of it is that there is a nice piece of software called Amadeus which you can use to analyze the noise. I have found it to be very useful for isolating the problem and measuring the effects of a fan mod. Also, by using the denoising option, you can get a preview of what your MDD would sound like without power supply whine.

In any case, I haven't kept my web pages up-to-date, but maybe it will be of some help to you.

[rhfactor]

Originally posted by Simon:
Rich, I saw you posted the phon business on your own forums. Thanks for the compliments btw.

I wanted to try to contribute over there, but I didn't get the registration confirmation mail (check your log, a guy wanted to register as "Simon" from xx.218.57.178 ). Hope you can fix this. I think we need to get up and moving on measurement procedures. I think your forums would be an excellent place to start doing so.

Ooops, I just now saw this. OK, let me go check it out. It should be automated but sometimes it requires a manual push!

Your information is so RIGHT-ON! You've totally cracked open a whole new sphere that I knew, intuitively, must exist, but my not being an engineer, I didn't know the scientific measurements. This is great. I would love to have you guide a series of tests and let's get some scientific data on this pheonomenon.

Thank you, Richard
p.s. here's the thread -- if anyone here is interested...

http://www.g4noise.com/forum/index.p...=ST&f=18&t=289

[russellb]

Hi, just bought a new G4 Tower MDD, 1Ghz ... sitting under my desk and I can hear it at all ... would not even know it was running.

Not sure how different the Dual CPU models are ?

cheers

[flojoe]

I modded my first gen dual gig mdd g4 over the weekend. I bought 2 stealth 60mm fans and replaced the loud delta ones in the ps. I then used heavy alluminum foil to make an airscoop to force the air that is wasted because the 120mm fan isn't aligned with the heatsink. And finally, I took out the front speeker, and again out of alluminum foil made a funnel that takes in fresh air from the front of the case and is pushed through the ps. Is silent and works good. I will buy an electronic thermometer to measure the output temperature just in case.

[Simon]

Originally posted by Chun Hsu:
I did noise analysis a few months ago on my MDD and posted some instructions on the web

Chun, I've seen your web page and the work you've done. This stuff is outstanding! Compliments. You have been able to identify 1.2kHz PS fan "whine" noise and we have now already seen that there are new MDDs that don't have the 1.2kHz spike in the analyzed spectrum. I'd say that's already quite a good job.

Have you been able to identify the frequency of the CPU fan? Multiples of 60Hz I'd guess. 120Hz maybe? And what about the fan for the optical drives? Do you have relative dB for these different noise sources?

And now, most important, have you tried transforming your noise in dB units to noise in phons for the different noise sources? It would be very intersting for us to compare. I get the impression that most people suffer from the whine whereas I'm also sure tht the dB of the whine aren't so much higher than those from wooshing air. But the phons - they'd go to show that the PS fans Apple used are really a threat to our sanity.

Congrats on your work!

[markizack]

Originally posted by Simon:
Thanks Rich. What bothers me about the whole discussion is that it is highly irrational. We get the guys that come here and say a rocket engine firing up is more quiet than an MDD and then we get the other guys who come in here and say that they can hear a feather fall to the ground next to their MDD.

AFAIC that's all crap.

1. There is no MDD in terms of noise. There are *measurements* that show that different MDDs have different sound spectra and different volume.

2. These comparissons are crap too. If I say it's as loud as my hair dryer that doesn't mean squat because first nobody on these boards knows what my hair dryer sounds like and secondly I didn't say what characteristic of the sound was like what characteristic of my dryer. The frequency? The volume?

3. These boards are full of people of who a) don't care b) care too much c) don't know what they're talking about or d) lie

And I am a scientist (but not an acoustic specialist!) so I hate talking about the grade of something which we can't measure.

In an ideal world we would define a measurement procedure everybody could do for their MDD. We would compare results. Analyze data. Get a bunch of engineers to tell us how to fix stuff and then a bunch of technicians to actually do it. We then would go back to step one and iterate until our MDDs sound like iMacs. (disclaimer for the wize guys: no, I know they will never be that quiet, I was just joking)



Yes, 100%. The phon scale here will proove this statement

[image]http://www.sfu.ca/sonic-studio/handbook/Graphics/Phon.gif[/image]

The page this image is from has some more explanations for those interested.



Check your mailbox for more info. I sent you some stuff. BTW, it's just "phon" and you can find exact definitions and formulae in every physics textbook. The IIRC had nothing to do with the unit name (IIRC = if I recall correctly)

before you go saying there is no problem, go to


www.g4noise.com

[Simon]

Originally posted by markizack:
before you go saying there is no problem, go to www.g4noise.com

Buddy, you're new here, so learn the first important lesson right now: Don't read the last two posts in a thread and then decide to post nonsense. If you go back and read the whole thread you will see that I know Richard's www.g4noise.com site and that I have even subscribed to his forums. I'm not negating anything here, I'm highly interested so watch it.

[markizack]

Originally posted by Simon:
Thanks Rich. What bothers me about the whole discussion is that it is highly irrational. We get the guys that come here and say a rocket engine firing up is more quiet than an MDD and then we get the other guys who come in here and say that they can hear a feather fall to the ground next to their MDD.

AFAIC that's all crap.

1. There is no MDD in terms of noise. There are *measurements* that show that different MDDs have different sound spectra and different volume.

2. These comparissons are crap too. If I say it's as loud as my hair dryer that doesn't mean squat because first nobody on these boards knows what my hair dryer sounds like and secondly I didn't say what characteristic of the sound was like what characteristic of my dryer. The frequency? The volume?

3. These boards are full of people of who a) don't care b) care too much c) don't know what they're talking about or d) lie

And I am a scientist (but not an acoustic specialist!) so I hate talking about the grade of something which we can't measure.

In an ideal world we would define a measurement procedure everybody could do for their MDD. We would compare results. Analyze data. Get a bunch of engineers to tell us how to fix stuff and then a bunch of technicians to actually do it. We then would go back to step one and iterate until our MDDs sound like iMacs. (disclaimer for the wize guys: no, I know they will never be that quiet, I was just joking)



Yes, 100%. The phon scale here will proove this statement

[image]http://www.sfu.ca/sonic-studio/handbook/Graphics/Phon.gif[/image]

The page this image is from has some more explanations for those interested.



Check your mailbox for more info. I sent you some stuff. BTW, it's just "phon" and you can find exact definitions and formulae in every physics textbook. The IIRC had nothing to do with the unit name (IIRC = if I recall correctly)

before you go saying there is no problem, go to


www.g4noise.com

[Chun Hsu]

Originally posted by Simon:
Have you been able to identify the frequency of the CPU fan? Multiples of 60Hz I'd guess. 120Hz maybe? And what about the fan for the optical drives? Do you have relative dB for these different noise sources?

If you look at the analysis page on my web site, you can see that my measurements say it's in the 600-800Hz range. The exact frequency is likely to vary from machine to machine.

And now, most important, have you tried transforming your noise in dB units to noise in phons for the different noise sources? It would be very intersting for us to compare. I get the impression that most people suffer from the whine whereas I'm also sure tht the dB of the whine aren't so much higher than those from wooshing air. But the phons - they'd go to show that the PS fans Apple used are really a threat to our sanity.
Congrats on your work!

No, I haven't. Let me explain. My goal from the beginning was to try to identify what differentiated noisy MDD's from non-noisy MDD's. A simple spectrum analysis without any weighting should accomplish that. My suspicion has always been that quiet MDD's would lack the upper frequency noise. I have one FW800 measurement that seems to confirm that, but I would love to get more numbers. Few seem willing to do these measurements for some reasons.

Perhaps I am not understanding, but it seems like a conversion to phons will help to explain why the power supply whine is so annoying. It doesn't really help to say whether or not your MDD is louder than normal. I always felt that the key would be to come up with a procedure to prove that one MDD was overly loud compared to the rest. I think the spectrum analysis already does this. The only problem is that I don't have enough data to work with.

In any case, I think it might be appropriate to move this to the G4noise forum. I am sure the group would like to hear your thoughts on this.

[markizack]

well its plain to see that apple is not doing the right thing, these machines cost alot of $$$, and some people feel like $3000 beta testers for apple.

[Simon]

Originally posted by Chun Hsu:
If you look at the analysis page on my web site, you can see that my measurements say it's in the 600-800Hz range. The exact frequency is likely to vary from machine to machine.

Ah perfect. Found em. Thanks.

Perhaps I am not understanding, but it seems like a conversion to phons will help to explain why the power supply whine is so annoying. It doesn't really help to say whether or not your MDD is louder than normal. I always felt that the key would be to come up with a procedure to prove that one MDD was overly loud compared to the rest. I think the spectrum analysis already does this. The only problem is that I don't have enough data to work with.

You're absolutely right. We are looking into two different aspects of the problem. You're work has prooved that some machines are really loud and that we can distinguish between different kinds of MDDs from the noise point of view. You're work has also introduced a measure how to qualify different mods.

The stuff I'm looking into takes off from there. I believe that maybe some people who get a spectrum that doesn't show a loud machine in dB maybe still have a very bad machine, since they could be suffering from noise in the higher frequencies. But this noise could be - in dB albeit - lower than many other machines. Looking at dB is good for comparing hardware, but also looking at phons could help us understand why some people can't work with a machine that seems not even that bad if you just look at dBs.

Also I believe we need the phon measures to get Apple to realize that a loud woosh of air (like in some cheap PCs) is still much better than a PS whine - even if it has less dB. Apple needs to lower the dB but also the Hz.

Therefore, I think our work on the subject goes very well side by side.

In any case, I think it might be appropriate to move this to the G4noise forum. I am sure the group would like to hear your thoughts on this.

Absolutely. I have registered at G4noise.com and I plan on summing up the thoughts I've stated here before in one post. From that on we could maybe do some measurement. I just haven't had enough time to compile all this info yet. But, don't worry, bear with me, it's getting done.

[booboo]

Originally posted by iChristopher:
I'm not being sarcastic here, but how quiet do you expect a dual processor workstation to be? Is it possible your expectations exceed what's reasonable?

I don't see why it shouldn.t be only twice as noisy as say a 1GHz iMac or even an upgraded dual processor cube, I really don't.

But then I'm not an apologist.

And I just sent back a DP867 for a full refund.