Data Centers

Data centers may be hazardous to your hearing

The noise levels data center workers are exposed to could lead to hearing impairment. Read about one way these workers can protect their hearing.

Tinnitus or "ringing in the ears" is my constant companion due to spending long hours in shipboard engine rooms during my six years of military service. A friend also has tinnitus, but his diagnosis is very different from mine. An ear, nose, and throat doctor told my friend (who has been a contract network engineer for 25 years) that spending long hours in data centers could very well be the reason for his tinnitus.

I remember my first time in a data center; I could not imagine spending eight hours a day in that environment. Still, until my friend mentioned his tinnitus, I assumed noise levels in data centers were acceptable.

Current OSHA standards

When it comes to noise-exposure guidelines in the workplace, two government agencies get involved: the Occupational Safety and Health Administration (OSHA) and the National Institute for Occupational Safety and Health (NIOSH). NIOSH provides recommendations about occupational safety and health, whereas OSHA creates and enforces the regulations.

The slide in Figure A, from this paper (PDF) by Patricia A. Niquette, AuD provides recommended and permissible exposure times to different noise levels.

Figure A



The 85 dBA range (circled in red) is important to those who work in data centers for a number of reasons. The fact that OSHA requires monitoring of workspaces where noise levels approach 85 dBA implies that it is a significant threshold. Yet, that significance seems moot compared to the disparity between OSHA's Permissible Exposure Limit and NIOSH's Recommended Exposure Limit at 85 dBA -- 16 hours vs. 8 hours, respectively.

The bottom of this OSHA report lists how other organizations regulate noise exposure, including the World Health Organization. It appears most countries have noise exposure limits similar to what OSHA requires.

Tad Davies, executive vice president of The Bick Group Inc., a company that designs and builds data centers, recalls in this Computerworld article, only one time when an IT manager asked for noise-level measurements in a Bick Group constructed data center. The readings ranged from a low of 70 dBA to a high of 79 dBA.

Data indicates hearing loss is cumulative

In Niquette's paper, she cites a 2009 study by Sharon G. Kujawa and M. Charles Liberman that questions previously held beliefs about our ability to recover from noise trauma. The scientists found that certain parts of the inner ear (i.e., the basal region of the cochlea) do not recover and exhibit dramatic degeneration when exposed to high noise levels (Figure B).

Figure B


 Image: Weizmann Institute of Science

Kujawa and Liberman also point out, "This kind of damage is undetectable using current test methods, and damage is not seen until weeks or months after the exposure." Niquette put it more bluntly, "The implication of this research is that noise can produce subclinical damage that goes undetected, progresses unnoticed, and finally manifests itself long after the fact."

No cure for noise-related hearing loss, but it is preventable

Hearing disorders caused by medical reasons (even tinnitus) can be cured by removing the underlying cause, but that is not the case if prolonged exposure to excessive noise caused the hearing impairment. My friend doesn't want to make his tinnitus worse so now he wears noise-cancelling headphones when working in data centers.

The technology used in noise-cancelling headphones has been around since the 1950s, with notable advances being made recently. One example is Brigham Young University's Acoustics Research Group under the auspices of Dr. Scott Sommerfeldt and their development of Active Noise Control (ANC). From the group's website:

"Active Noise Control (ANC) is based upon the principle that two sounds of equal amplitude and opposite phase cancel each other out, leaving silence."


 To get an idea of how ANC works, consider the noise pattern at the right (image courtesy of Using a microphone, ANC captures the noise, creates a duplicate noise pattern of equal strength but 180 degrees out of phase, and then sends the cancelling noise to a speaker --all in real time.

Dr. Sommerfeldt has used ANC to quiet everything from operator cabins on heavy earth-moving equipment to cooling fans used in computers.

To help those who work in data centers, Dr. Sommerfeldt and the Acoustics Research Group have been working with C7 Data Centers, an outsourcing service with several large data centers, affording Sommerfeldt and his team ample opportunities to test ANC (PDF).


 Image: C7 Data Centers


Bottom line

Tinnitus sucks, so please err on the side of caution when it comes to hearing loss. If OSHA decides to rethink and decrease exposure limits tomorrow, it won't help those of you dealing with noise in data centers today. 

Share your experiences

Does your company monitor the noise levels in its data centers? If you work in a data center, do you suffer from hearing loss? How do you try to prevent hearing loss and/or slow hearing loss progression? Post your thoughts in the discussion.



Information is my field...Writing is my passion...Coupling the two is my mission.


One clarification ... Tinnitus is not necessarily associated with hearing loss.  I'm in my 60s and have fairly bad tinnitus (as does my wife) and I have no hearing loss.  In fact, I have as good a hearing as anyone.  The ringing noise does not seem to interfere with the ability to hear at all.

My wife and I have no idea why either of us has tinnitus, not from working in a data center.  While exposure to high decibel levels does aggravate it, other things besides noise also aggravate it, and I have no reason to believe that I acquired it from exposure to noise in the first place.  Not much is known or documented about tinnitus, but it shouldn't be assumed to be related to data center noise.


The US standard allowing the exposure to 85dBA for up to 16 hours without hearing protection is ridiculous.

The EU standard of 80dBA for 8 hours is much MUCH more reasonable. 

Part of our really reckless noise level standards is that we focus on the narrow bandwidth of speech intelligibility rather than noise levels which are destructive of normal hearing.

Normal, healthy hearing has a bandwidth of 20Hz to 20KHz, though few adults can hear the upper reaches of that these days. An Audiologist is unlikely to test anything above 8KHz, and may not test above 5KHz, 'cause they are focused on speech intelligibility, the enjoyability of music and environmental sound is not within their interest.

As a programmer, I haven't spent a lot of time in server rooms, but every time I have been in one, my ears feel abused. I suspect this is because the noise in server rooms approximates white noise, which has a LOT of energy at high frequencies. 


Anyone in IT would know that data centers can result in hearing loss.  As for your "Tinnitus or "ringing in the ears" is my constant companion due to spending long hours in shipboard engine rooms during my six years of military service", the only people to blame is you.

In possible hearing loss area, the USAF always issued ear protection, and it would stand to reason the Navy or Coast Guard would.  The one time that the USAF didn't issue ear protection was because things were on back order (for about 1.5 weeks), I bought ear protection for myself.

Since I've been out, if the company didn't issue ear protection, I bought mine.  It should be noted that time I've had to spend in Data Centers was greatly reduced the last six years in the USAF, and limited since I got out.


I have been in lots of Data Centers and the noise is not that bad; however, noise is not monitored to my knowledge. One thing most people do not realize is that the human hearing frequency range is from 20 Hz to 20 KHz; however, they are glued to the cellular telephone 24x7 and cellular phones frequency band is in the 800 MHz range - that alone is far more damaging than the noise level of any Data Center. Another point to consider is that servers are monitored remotely from the system administrator's office and a visit to the Data Center is done when working on the hardware which is not a daily basis routine.

One of the things I hate about doctors is that as a rule they are very ignorant and they steer the patient in the wrong direction. The following are the causes of tinnitus:

  1. Correlation with any degree and type of hearing loss, common with sensory deficits
  2. Noise exposure – noise exposures at work, recreational activities, hobbies (loud music/bands, target shooting, hunting, racing, carpentry/power tools, etc.) 
  3. Physiological/pathological, ranging from simple wax impaction or ear infection to more serious types of growths, tumors, or neurological involvement.
  4. Side effect of certain medications
  5. Vascular/circulatory
  6. Idiopathic (cause unknown)

So, do not blame it on the Data Center as most people are exposed to noise far more damaging than the noise in any Data Center such as the high frequencies in cellular phones and other sources of high pitch frequencies.

Jibran Mirza
Jibran Mirza

I use noise cancellation ear phones and do some pushups and drink hot coffee before stepping inside that cold and loud place.


Noise cancelling technology is great. However, it's also expensive.

Back in the 1970s, we worked in noisy environments (many minicomputers, with large fans, in one room), and we were issued ear protectors - basically headphones with padded, over-the-ear covers (and no sound reproduction hardware). While they didn't provide the opportunity to talk to other people, they were a LOT cheaper than modern solutions

William Blake
William Blake

A little too late for some of us. I started wearing hearing protection when our air conditioner broke and Management decided to use large fans to cool out data center instead of fixing the cooling unit (note the management here is not very bright). Now we've moved and my office is actually in the data center so I wear hearing protection all the time.


This pattern of NIOSH calling for lower limits than OSHA enforces seems common enough.  The cause, of course, is lobbying by industrial interests....

HAL 9000
HAL 9000 moderator

It's not only the level of noise but the type of noise that you need to guard against.

But 85 dBA seems a Universal point where damage begins. I remember years ago when Racing Motorcycles where measured with a upper limit of 110 dBA being established and your exposure to that level was measured in minutes.

But there is different noise. A 4 Stroke producing 110 dBA will adversely impact on your hearing but a 2 Stroke making considerably less noise can cause Brain Damage with very short exposures. Like the time spent on a Starting Grid before the Team Staff have to vacate the Race Track. So the time spent behind a modern racing 2 Stroke warming it up and removing the tire covers before vacating the track is sufficient to leave you brain damaged. Of course the guys on the front row of the grid are not as adversely impacted but those further back are and it covers everyone involved from the riders down tot he Pit Crew. Also put simply just sitting on the bike and participating in the race has the ability to cause Brain Damage to all participants with the damage being more likely the more of those 2 strokes you have in front of you. Be that the guy running last or the leader coming up to Lap Slower Riders.

So where I have a problem with Blanket Limits they just cover the actual noise not the type of noise which could cause considerably more damage and still be within the stated limits.


Michael Kassner
Michael Kassner


You may have missed this paragraph: 

"Hearing disorders caused by medical reasons (even tinnitus) can be cured by removing the underlying cause, but that is not the case if prolonged exposure to excessive noise caused the hearing impairment. My friend doesn't want to make his tinnitus worse so now he wears noise-cancelling headphones when working in data centers."

I might suggest seeing a doctor if you have not seen one specifically about tinnitus, as it may be a symptom of another issue.

Michael Kassner
Michael Kassner

How much time do you spend in the data center? I would be curious to learn if there is any record of the noise levels in the data center have been checked. 

Michael Kassner
Michael Kassner


I have been looking into white noise and if our hearing is damaged by noise at frequencies that we are unable to hear. Thank you for your comment.

Michael Kassner
Michael Kassner


Well, I had ear protection and I wore it. As for getting our own better protection that was not allowed. 

So, you can blame me all you want. I am proud of my service and would not change a thing.

Michael Kassner
Michael Kassner

I would curious to learn how well they work for you? Are they the ear bud style of the type that completely cover your ears? 

I remember the cold as well. Brrr.

Michael Kassner
Michael Kassner


That is true, Mike. I wore those same kind of protectors when I was in the service. The problem is that they do not cancel out all the noise, as Hal 9000 alluded to.

Michael Kassner
Michael Kassner

@HAL 9000  

Thanks, Col. What I am surprised at is the lack of any information regarding data centers. I suspect that is because data centers are supposedly below 85 dBA where required monitoring starts. 


@Michael Kassner - You're implying here that tinnitus is a hearing disorder, which isn't the case in everyone, but I guess this was outside your article's purpose.  I have been checked medically, and there seems to be no underlying cause, but like I said very little is known about tinnitus and it's causes.

Editor's Picks