Hearing protection does not count towards the action values
This is the single thing to take away, and nearly every free noise calculator gets it wrong.
Regulation 4(5) permits the effect of hearing protection to be taken into account when working out whether the exposure limit value is exceeded. It expressly does not permit it for the lower and upper action values.
So the 80 and 85 dB(A) figures are measured at the ear without protection. A site running at 88 dB(A) unprotected is at the upper action value with everybody in defenders, and every duty that arrives at 85 arrives anyway: a programme of control measures, hearing protection zones marked and enforced, health surveillance.
A calculator that quietly subtracts an SNR before comparing against 85 tells somebody they are compliant when they are not. That is why this one shows the unprotected figure first and treats protection as a separate question about the limit.
You cannot average decibels
The scale is logarithmic on energy, so the loud bits count for far more than their share of the time.
Two hours at 91 dB(A) and six hours at 85 gives a straight average of 88 and a time-weighted average of 86.5. The actual daily exposure is 87.4 dB(A), which is over the 87 limit that both averages comfortably cleared.
That is not a rounding argument. It is the difference between a day that is legal and a day that is not, arrived at by an arithmetic that looks perfectly reasonable.
Points, because they do add
Exposure points are built so the upper action value lands on exactly 100. Every 3 dB doubles them, so 88 dB for a full day is 200 points and 82 dB is 50.
The useful property is that they add up like ordinary numbers, which decibels do not. Two activities at 60 points each is 120 points, and 120 points converts straight back to a decibel figure. It is the same calculation, without needing logs on a scaffold.
For completeness: 3 dB is not exactly double. Doubling is 10 log₁₀(2), which is 3.0103. This calculator uses the logarithm, so the numbers come out slightly off the round ones you would get from the shorthand.
The peak is a different test
Regulation 4 sets peak sound pressures of 135, 137 and 140 dB(C) alongside the daily averages, and they behave completely differently.
A peak is instantaneous. It is not averaged over the day and it cannot be traded against time, so halving the exposure does nothing for it. One cartridge-operated tool, one hammer on steel, one dropped scaffold board.
Which means a quiet day can breach the peak values. Somebody whose LEP,d comes out at 70 dB(A) and who fired a nail gun once may still have exceeded the upper peak value, and an assessment that only looked at the average will never see it.
More protection is not better protection
Over-protection is a real problem rather than a safe margin, and it is counterintuitive enough that people get it wrong in good faith.
Protection that takes somebody too far down cuts them off from speech, from reversing alarms, from a shout, and from the sound of the job going wrong. Isolation on a working site is its own hazard.
It also comes off. Protection somebody cannot work in gets lifted, pushed back, or worn on one ear, and protection that is not on attenuates nothing. Matching the protection to the exposure is safer than maximising it, and cheaper.
The SNR on the box is a laboratory number
It comes from a properly trained wearer achieving a good seal under test conditions.
Real life is a hurried fit over a hard hat, safety glasses, a hood and possibly a beard, on somebody who was not shown how. The achieved attenuation is meaningfully lower, which is why this tool derates the rating rather than subtracting it whole.
Being wrong in that direction is the right way round: it understates the protection and overstates the exposure, which is the error that does not deafen anybody.
Common questions
Does hearing protection take me below the action values?
No, and most noise calculators get this wrong. Regulation 4(5) of the Control of Noise at Work Regulations 2005 lets you take the effect of hearing protection into account when deciding whether the exposure limit value is exceeded, and expressly not for the lower and upper action values. So if unprotected exposure is 88 dB(A), you are at the upper action value with the defenders on, and every duty that hangs off 85 still applies in full.
What are the six values?
Regulation 4 sets three pairs. Lower action value: 80 dB(A) daily or weekly exposure, 135 dB(C) peak. Upper action value: 85 dB(A), 137 dB(C). Exposure limit value: 87 dB(A), 140 dB(C). Each pair has an average and a peak and they are separate tests.
Why can I not just average the levels?
Because decibels are logarithmic on energy, so averaging understates the loud bits. Two hours at 91 dB and six at 85 averages to 86.5 either way you do it, and the real answer is 87.4, which is over the limit that the average cleared. That is not a rounding difference, it is the difference between compliant and not.
Is 3 dB really double?
Near enough for a site and not exactly. Doubling the sound energy is 10 log₁₀(2), which is 3.0103 dB. So 88 dB for four hours is 99.76 points against the 100 that 85 dB for eight hours gives: the same for any practical purpose. This calculator uses the logarithm rather than the shorthand, which is why the numbers are not quite round.
What about the peak?
It is a separate test and it is instantaneous. Peak sound pressure is not averaged and is not traded against time, so one cartridge tool, one hammer blow on steel or one dropped scaffold board can breach 137 or 140 dB(C) on a day whose average is comfortable. Assessing the average and forgetting the peak is the common gap.
Can hearing protection be too good?
Yes, and over-protection is a recognised problem rather than a safe margin. Protection that cuts somebody off from speech, reversing alarms and the sound of the job going wrong makes them less safe, and protection that people cannot work in comes off, at which point it attenuates nothing. Matching the protection to the exposure beats maximising it.
Is the SNR on the box what I actually get?
No. The rating comes from a laboratory fit on a trained wearer. A hurried fit over a hard hat, safety glasses and a hood achieves considerably less, so this calculator derates the SNR rather than subtracting it whole. Understating the protection is the safe direction to be wrong in.