Hi there,
I'm trying to design my first complete headphone DAC/amplifier and I would like to know where I can find, in the general case, the maximum power for an headphone, according to his nominal impedance.
I've seen some amplifiers providing more than 500mW @ 600ohm, while some headphones cannot support more than 100 or 200mW at the same impedance... I really don't know what information I can trust and I'll be pleased if it exists a document which talk about maximum power of some headphones.
Thanks,
BT
I'm trying to design my first complete headphone DAC/amplifier and I would like to know where I can find, in the general case, the maximum power for an headphone, according to his nominal impedance.
I've seen some amplifiers providing more than 500mW @ 600ohm, while some headphones cannot support more than 100 or 200mW at the same impedance... I really don't know what information I can trust and I'll be pleased if it exists a document which talk about maximum power of some headphones.
Thanks,
BT
I think the question is really what voltage do the headphones need, rather than "power".
I did some tests when designing a headphone amp and found that it was under 1 volt peak/peak for low impedance phones. 500mw into 600 ohm is 17 volts RMS so that means an amp running on over 50 volts dc in total... which has to be way over top.
I did some tests when designing a headphone amp and found that it was under 1 volt peak/peak for low impedance phones. 500mw into 600 ohm is 17 volts RMS so that means an amp running on over 50 volts dc in total... which has to be way over top.
1st headphone amp design Q: for which headphone?
sensitivity, Z and target SPL are good things to know how to relate
I don't think there is any standard for max power ratings - most headphone coils will mechanically bottom at low frequency at still "safe" electrical drive levels - long before heating softens glues, damages insulation, melts wires
given the wide range of sensitivity, Z among headphones a particular amp's I,V capability will likely be inappropriate in the sense of being potentially dangerous to some while not even being able to drive to reasonable SPL some other headphones
sensitivity, Z and target SPL are good things to know how to relate
I don't think there is any standard for max power ratings - most headphone coils will mechanically bottom at low frequency at still "safe" electrical drive levels - long before heating softens glues, damages insulation, melts wires
given the wide range of sensitivity, Z among headphones a particular amp's I,V capability will likely be inappropriate in the sense of being potentially dangerous to some while not even being able to drive to reasonable SPL some other headphones
Last edited:
Thanks a lot for your answers.
Yes, indeed. The main aim of my question is to know if there is a lot of differences in volt to have the same audio level for two headphones with the same impedance.
If not, it's easy to calculate the maximum voltage of my signal and the output impedance of my amp.
For eg, if I need 3V to reach 120dB with a 32ohm headphone and 11V for a 600ohm headphone, i can calculate that I need a 13V signal with a 106ohm output impedance. So the audio level will be the same for this two headphones, and i'll not have to change volume if i change headphone
I've just done some mesurements with my 32ohm HD448. 500mVpp if more than enougth (i usually listen music around 100mVpp with this one).
But I have no idea of the voltage value nedded for a high impedance headphone.
Yes, indeed. The main aim of my question is to know if there is a lot of differences in volt to have the same audio level for two headphones with the same impedance.
If not, it's easy to calculate the maximum voltage of my signal and the output impedance of my amp.
For eg, if I need 3V to reach 120dB with a 32ohm headphone and 11V for a 600ohm headphone, i can calculate that I need a 13V signal with a 106ohm output impedance. So the audio level will be the same for this two headphones, and i'll not have to change volume if i change headphone
I've just done some mesurements with my 32ohm HD448. 500mVpp if more than enougth (i usually listen music around 100mVpp with this one).
But I have no idea of the voltage value nedded for a high impedance headphone.
Beyerdynamic T1
These are 102dB/1mW at 600ohms beyerdynamic T 1
sqrt(1mW*600) =0.775V; So we need 0.775V to get us to 102dB -very, very loud directly into our ears; On these headphones, this gives us an idea we should be listening at <<0.775V which is not even half of what typical digital sources' line level outputs provide
Extended levels of 87dB is deemed to cause hearing loss by OSHA. 120dB is at the threshold of pain for humans.
These are 102dB/1mW at 600ohms beyerdynamic T 1
sqrt(1mW*600) =0.775V; So we need 0.775V to get us to 102dB -very, very loud directly into our ears; On these headphones, this gives us an idea we should be listening at <<0.775V which is not even half of what typical digital sources' line level outputs provide
Extended levels of 87dB is deemed to cause hearing loss by OSHA. 120dB is at the threshold of pain for humans.
I've built a couple of headphone amps,with around 1W of output per channel. I can tell you-that's a stupid amount of power for headphones. WAY more than is even safe/sane. (ESL 'phones,etc. might be a different matter.)
As a ballpark guesstimate,a couple of volts (assuming low-ish Z 'phones) is probably plenty. It will depend on your particular set of cans.
Whatever you do,be extra careful with headphones/headphone amps. Test the amp with a small speaker or something first. NOT with headphones on your head! If something were to go wrong,a loose connection,oscillation,etc. it could easily damage your hearing.
As a ballpark guesstimate,a couple of volts (assuming low-ish Z 'phones) is probably plenty. It will depend on your particular set of cans.
Whatever you do,be extra careful with headphones/headphone amps. Test the amp with a small speaker or something first. NOT with headphones on your head! If something were to go wrong,a loose connection,oscillation,etc. it could easily damage your hearing. Having a high supply voltage of more than the usual +/- 12-15VDC has a few advantages, especially when working with discrete amplifier circuits.
Higher headroom - amp can be designed to never clip.
Lower distortion - due to the higher headroom the output signal will never be getting even near the supply rails which should help keep the distortion lower.
Better linearity - a discrete headphone amplifier, especially ClassA, usually have some kind of largish BJT(s) or FET(s) in the output stage and most of those like a certain amount of C-E or D-S voltage across them to perform at their best.
Now, what I usually do with my +/-12-15VDC discrete ClassA designs is that I adjust the gain to suit my own personal goal of about 115dB maximum output power into my headphone of choice(Beyerdynamic DT 880 PRO, 250 Ohm, 96dBmW).
I'm using a self designed ES9023 DAC with a line level output of 1.9VRMS, my discrete ClassA headphome amplifier is configured to a gain of x 2.5 which in combination gives me an output level of 4.75VRMS.
4.75VRMS into 250 Ohm equals 90.25mW of output power which in turn equals a maximum headphone SPL of 115.55dB.
The result is no possibility of clipping, ever, and very low distortion all the way up to 115dB SPL.
To make it short, having very high supply rails is not much of a problem if you know exactly what headphone impedance and sensitivity you are designing for as well as how much power you need to get the SPL you want. You can always change the gain of your headphone amplifier to suit your needs.
Higher headroom - amp can be designed to never clip.
Lower distortion - due to the higher headroom the output signal will never be getting even near the supply rails which should help keep the distortion lower.
Better linearity - a discrete headphone amplifier, especially ClassA, usually have some kind of largish BJT(s) or FET(s) in the output stage and most of those like a certain amount of C-E or D-S voltage across them to perform at their best.
Now, what I usually do with my +/-12-15VDC discrete ClassA designs is that I adjust the gain to suit my own personal goal of about 115dB maximum output power into my headphone of choice(Beyerdynamic DT 880 PRO, 250 Ohm, 96dBmW).
I'm using a self designed ES9023 DAC with a line level output of 1.9VRMS, my discrete ClassA headphome amplifier is configured to a gain of x 2.5 which in combination gives me an output level of 4.75VRMS.
4.75VRMS into 250 Ohm equals 90.25mW of output power which in turn equals a maximum headphone SPL of 115.55dB.
The result is no possibility of clipping, ever, and very low distortion all the way up to 115dB SPL.
To make it short, having very high supply rails is not much of a problem if you know exactly what headphone impedance and sensitivity you are designing for as well as how much power you need to get the SPL you want. You can always change the gain of your headphone amplifier to suit your needs.
Last edited:
So 2.2Vpp
Thanks for this info.
Don't worry, i'll remember your advice !
Thanks for your explaination.
On this website, they give us results of headphone tests : Headphone Data Sheet Downloads | InnerFidelity
So now I can calculate the voltage of some headphones for a specific audio pressure.
(With measurements I've done with my headphone, I can determinate that I usually listen music between 86 and 90dB. According with Blues, it's a little too much...)
Thank you for having answer my question
BT
Attachments
So, if i've understood, you adviced me to design my amp to drive headphones at 120dB ?
Even if i'll never reach this level (for continuous audio level), there can be some audio pic which need more power to be well listened ?
For exemple, with my HD448 (35ohm), 0.1mW for 90dB, it will be 100mW for 120dB, so 1.9vrms. My amp should deliver 5.3Vpp
I've no problem with electronic. I'll only use op amp for reducing distortion. Output stage will be designed with JFET (self-design AB class), so i'll not have output curent problems
http://headwize.com/articles/hearing_art.htm => 404
http://gilmore2.chem.northwestern.edu => server cannot be found
Even if i'll never reach this level (for continuous audio level), there can be some audio pic which need more power to be well listened ?
For exemple, with my HD448 (35ohm), 0.1mW for 90dB, it will be 100mW for 120dB, so 1.9vrms. My amp should deliver 5.3Vpp
I've no problem with electronic. I'll only use op amp for reducing distortion. Output stage will be designed with JFET (self-design AB class), so i'll not have output curent problems
I'm very interesting in reading your articles, but all the links are dead :/jcx said:
http://headwize.com/articles/hearing_art.htm => 404
http://gilmore2.chem.northwestern.edu => server cannot be found
the solution to offline/outdated links is often to try the WayBack Machine archive.org
HeadWize - Article: Preventing Hearing Damage When Listening With Headphones (A HeadWize Headphone Guide)
HeadWize - Article: Preventing Hearing Damage When Listening With Headphones (A HeadWize Headphone Guide)
I've plugged my headphones into speaker amps that can output 400W. It's fine. You probably are only using 0.01W with most headphones.
I like a gain of 4x with low impedance headphones and 20x with high impedance headphones. It seems high but there's always those recordings recorded at low level and it's nice to have extra gain on hand.
I like a gain of 4x with low impedance headphones and 20x with high impedance headphones. It seems high but there's always those recordings recorded at low level and it's nice to have extra gain on hand.
Yes, that is more like it for me too...78-84dB average and peaks to about 90-96dB. Remember these are right-on-the-ear levels...great quality sound need not be ear-drum splitting be it rock or symphonic.
QED!
These paragraphs for me stand out most from the HeadWize link...
"In loudspeaker reproduction, sounds must travel several feet before reaching the listener's ears. By the time they arrive, a portion of the high frequencies have been absorbed by the air. Low frequencies are not absorbed as much, but they are more felt through bone conduction than actually heard. With headphones, the ears hear all frequencies without any attenuation, because the transducers are literally pressed against them. Thus, when listening to headphones at the same effective volume level as loudspeakers, headphones may still transmit louder high frequencies that are more likely to cause hearing damage.
Another hearing phenomenon that seems to be more noticeable with headphones is a decreasing sensitivity to sound levels over time, as the ears adapt to loud sounds. The listener perceives a gradual drop in loudness even though the volume control setting hasn't changed. The acoustic isolation of headphones tends to highlight this dulling effect. It is all too easy for headphone listeners to turn up the volume to the point where hearing is at risk. Interestingly, most people find it difficult to distinguish between 85dB and 100dB SPLs, despite that the latter is more injurious to hearing. Therefore, it is important to avoid listening fatigue by resting the ears in silence after long sessions with headphones and to fight the temptation to turn up the volume."
The 2nd paragraph might explain for some heavy headphones users' preference 100dB or more levels.
Aside from low sensitivity planar mags and high voltage requirements of electrostatics, I don't see why we would need to introduce a gain stage into the equation.
ps: GO 'Hawks!
Congratulations to the Falcons...
Last edited:
Yes, though in your calculations, you have to take into account that average levels are at -12dB. So instead of working with a 2V source, you're looking at -12dB of 2V. And then there's those recordings that simply did not have enough microphone gain. So, it is better to have some gain on the headphone amp.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.