How much DC offset would you consider harmful to HP and after which point would you throw the commercial amp out?

[svyr]

http://www.head-fi.org/forum/thread/500181/musiland-md11-to-be-released/60#post_7424276

I recently remeasured the DC offsets on Line out and HP out of Musilol MD11 and it's not pleasing.

Line out: 70mV, HP out: 150mV (both constant regardless of the volume level).
I think tangent recommended an abs cut-off of 100mv or something (and most people set 15-25 as their personal limit), but seeing this on a commercial amp is just nothing short of horrifying.

How much is actually enough to damage low impedance HP (or does it just displace the diaphragm and distort the sound)? (say for 16 ohm 105 spl ones, or 32 ohm 105 spl ones)

Would anyone else having and MD11 like to measure their unit to deny or corroborate?

 

[High_Q]

Hey, how are you measure it?

  I'm interested in measuring all my sources.  I need a accurate measuring tool, probably need a O-scope.  Don't have one, too expensive.  
 
I would think offset would be harmful since it causes peak of the wave to span out from 0 point since its V=Offset +ASin(wt).  In which increasing the volume would hopefully with a most linear amp, would increase A only not create distortion.
 
Depends on how much peak power the headphone can handle until it blows.  Care to experiement?

 
No experiment necessary for ibuds, it blows from maximum volume of the source(iphone) it comes with. Which makes no sense.

  You should feel ashamed Apple.

 

[svyr]

Hey, how are you measure it?

  I'm interested in measuring all my sources.  I need a accurate measuring tool, probably need a O-scope.  Don't have one, too expensive.  
 
I would think offset would be harmful since it causes peak of the wave to span out from 0 point since its V=Offset +ASin(wt).  In which increasing the volume would hopefully with a most linear amp, would increase A only not create distortion.
 
Depends on how much peak power the headphone can handle until it blows.  Care to experiement?

 
No experiment necessary for ibuds, it blows from maximum volume of the source(iphone) it comes with. Which makes no sense.

errr, with a multimeter in mV-DC mode using a 3.5mm to RCA cable or a torn off earbud cable ..Scope would sure be a nicer bet, and an basic digital one might be relatively inexpensive nowadays but probably still a few hundred. Something tells me for the MD11, it's best not to look at the scope measurements considering the problems I already found, I have a feeling it won't do too well.


> Depends on how much peak power the headphone can handle until it blows.

I thought people were saying something about it overheating the voice coil or something...From DC offset that is, not tearing them up with <-> excessive AC where the DC offset size would probably be insignificant...



> Care to experiement?

don't own any earphones cheap enough to do that

 

[High_Q]

I don't see voice coil heating(probably would  be insignificant) causing damage, if you think about how much current travels through audio system.  

 

[jr41]

I'm interested in an answer to this question also.
 
I too understand that you can measure DC offset using a multimeter, as svyr describes. I use a 1/4" to 1/4" or 3.5mm to 3.5mm cable as necessary (or a 1/4" to 3.5 adapter ( - or visa-versa)). The tip is left, the center ring is right and the "rear-most" ring is ground.
 
Both of those measurements from the MD11 sound very high. I don't know of any definitive guidelines, but for low impedance headphones, I recall reading anything above a few 10s of mV is dangerous. Low impedance headphones seem to be more susceptible to DC offset damage.
 
Apparently DC offset at best degrades sound quality and at worst damages a headphone's delegate wiring. Also, I believe that unless an appropriate capacitor and resistor combination is used in the signal path (to filter the DC offset), any DC offset from the source is multiplied by the amplifier, depending on gain and volume setting.
 
 

 

[FallenAngel]

Purely rule of thumb for me is 10mV after amplification is at which point I start to look into debugging the build.

 

[svyr]

mmm few mv or even few tens of mv doesn't seem that dangerous. (what High_Q said)

It seems like you need the DC offset voltage (or signal + dc offset voltage) to be > max rated input

E.g. for my really cheap woolies in-ears with impedance of 32ohm, and rated max power of 10mw - the dc offset voltage alone would have to be: http://www.ajdesigner.com/phpohmslaw/ohms_law_voltage_power_resistance.php about 0.56v (So even for these flimsy in-ears the 0.15v musiland outputs doesn't seem to be nearly enough, at least alone)

 

[jr41]

 
Quote:

[...]

E.g. for my really cheap woolies in-ears with impedance of 32ohm, and rated max power of 10mw - the dc offset voltage alone would have to be: http://www.ajdesigner.com/phpohmslaw/ohms_law_voltage_power_resistance.php about 0.56v (So even for these flimsy in-ears the 0.15v musiland outputs doesn't seem to be nearly enough, at least alone)

 
That sounds like a reasonable theory, I've never thought to work it out that way.

 

[High_Q]

Check what the max power the phones are rated for
 
P=IV or IVrms
 
Check for Impedance, usually they will till to the max impedance at a frequency, you want to use the lower bound at a given output frequency range.
 
Pmax=Vmax^2/Impedance_low
 
Vmax = sqrt(Pmax*Impdedance_low), just a rough estimate
 
Once you find Vmax
 
Solve for Offset:
 
Vmax = Voffset +ASin(wt)
 
Vmax -ASin(wt) = Voffset
 
This is only for a tone, but ASin(wt) is the peak voltage deviation from the DC offset.  I would use an O-scope to look for peak voltage deviation of the signal.
 
^Actually you can't really solve for that, but O-scope is your friend to find the offset.
 
 

 

[qusp]

 
need a scope for measuring dc offset? er.......no, a dmm is usually fine depending on the signal you are using. if you wanna get crazy you can use a voltmeter and play a known voltage sine wave, whatever the heating value over what is expected after gain and converted to dc is your offset. of course as scope can be used to good effect, but isnt needed imo. 
 
as far as i know, you dont need to go over max rated input to cause issues, that is not what its about except in the extreme case where you will overheat the coils, its more about excursion of the driver and therefor dynamic range being effected. your 110mv amount is pretty high in this regard

 

[High_Q]

ignoring qusp... As, I should be on his ignore list

 
I mentioned headphone power rating because we cannot know hen the headphone will blow, all we have is the manufacturer's power rating to go by or our worse case scenario because we don't really know what the maximum worse case is unless we find out for ourselves

.  Also, I don't know how accurate a DC meter would be vs a O-scope, and O-scope is easier to spot and more accurate.  I would rely on O-scope.  Anyway, offset cannot amplify beyond the saturation point or Opamp supply voltage, and the amplified signal will get clipped if saturation is reached.  When a signal goes into a amplifier there is a decoupling to separate DC(or offset) and AC. The output offset voltage cannot be larger than Opamp supply.
 

 
I found an interesting video from MIT that goes over the math of the amps(it great if you're an extreme amp geek, here is some school for you qusp, you may need it)
http://videolectures.net/mit6002s07_agarwal_lec10/
 
For example, my CMOY is supplied by 9v(+_ 4.5) and offset cannot be greater than 4.5.  If I increase the gain of the small signal, it will start clipping at a certain point.  So, if the supplie is much greater, and we know it will headphones will blow at a certain point(voltage), it would be an issue if gain is increased to max out the swing of the signal to reach the high supply voltage.  So, knowing the max power, thus max voltage is helpful.
Quote:

I'm interested in an answer to this question also.
 
I too understand that you can measure DC offset using a multimeter, as svyr describes. I use a 1/4" to 1/4" or 3.5mm to 3.5mm cable as necessary (or a 1/4" to 3.5 adapter ( - or visa-versa)). The tip is left, the center ring is right and the "rear-most" ring is ground.
 
Both of those measurements from the MD11 sound very high. I don't know of any definitive guidelines, but for low impedance headphones, I recall reading anything above a few 10s of mV is dangerous. Low impedance headphones seem to be more susceptible to DC offset damage.
 
Apparently DC offset at best degrades sound quality and at worst damages a headphone's delegate wiring. Also, I believe that unless an appropriate capacitor and resistor combination is used in the signal path (to filter the DC offset), any DC offset from the source is multiplied by the amplifier, depending on gain and volume setting.
 
 

wire can be damaged if there is enough current to create enough heat to melt the insulation etc... Typically for high current application, larger cross section is used.
 
R=pL/A  
 
Larger the area, less R, which relates to heat given off by power equation:
 
P=IV=I^2R=V^2/R
 
 
but again, its audio signal(not enough to create heat issues, look at unnecessary thick ALO cables made out of silver or gold LOL)
 
 

 

[qusp]

 
 
 
yeah well i'm ignoring you already. i forgot you know more than the rest of the forum combined
 
making people think they need an oscope for measuring offset on their amps is irresponsible and needlessly overcomplicated for the basic measurement needed to know if dc is higher than it should be. dc not only damages headphones before you go past the power rating, but causes lower audio quality WELL before that. an oscope is the perfect tool for it, but it is not needed in the slightest, stop flexing your brain already, its just not that complicated

 

[nattonrice]

Quote:

its just not that complicated

This.
You don't need elementary algebra to figure it out.
Take a digital multimeter and measure DCmV from the output to ground.
If it is higher than 10mV and it is a diy job, fix it.
 
If it is a commercial product... wait this is the diy section! Go build something!
 

 

[svyr]

> You don't need elementary algebra to figure it out.

relax. The ohms law re-arrangement was to see how much DC is deadly to dodgy in-ears.

As for an oscilloscope, sure you don't need one, it's also probably easier to measure the DC offset using a DMM (however looking at the O-scope plots is potentially a lot more informative than noticing a DC offset shift). Still, if there are cheapie $200-ish oscilloscopes around (the PC only types, with a low-ish hz-khz range) it's probably not a bad buy compared to an average DMM.
Speaking of that, any recommendations and do those exist, or did I just imagine them ?


> If it is a commercial product... wait this is the diy section! Go build something!

It's a DIY related question . Plus I'm a bit over DIY audio after fixing my SRM-1/mk2 with the help of the big two ... (yes, you can build what you want, how you want it, with what you want...No it takes a lot of time to find/order parts, not burn down the house if you're a nub like me, and then solder it up and debug. Case work connectors, panels and knobs are also potentially a pain. )


>Also, I don't know how accurate a DC meter would be vs a O-scope, and O-scope is easier to spot and more accurate. I would rely on O-scope.

The DC accuracy is quoted in the manual for any DMM you can buy. Generally it's higher than AC and sufficient to know whether you're below 10mv ...


>If I increase the gain of the small signal, it will start clipping at a certain point. So, if the supplie is much greater, and we know it will headphones will blow at a certain point(voltage), it would be an issue if gain is increased to max out the swing of the signal to reach the high supply voltage. So, knowing the max power, thus max voltage is helpful.

I see, thanks. That looks to be the upper bound for DCoff + Ac signal, since above that it'd be clipped?

 

[High_Q]

True.  I said DC, but actually it is DC + Audio signal Max Peak
 
 
If DC + Audio Signal Max Peak > Vcc(Amp supply voltage), there will be clipping.
Quote:

I see, thanks. That looks to be the upper bound for DCoff + Ac signal, since above that it'd be clipped?