[DJ The Rocket]

I'm using a procedure that I made up. I'm certain I'm not the first person to think of this, but I'm pretty sure I am the first to post about it online. I'd like to share my technique here in case anyone finds it helpful, or someone blasts holes in my methodology

The results all match my observations though, so I think I have it right. My method works better for planar magnetic drivers, since their impedance curve more closely resembles resistance. For dynamic drivers it's still an accurate enough estimate to be useful, and will typically lean more toward the worst case.

It's pretty simple really. Grab a pair of cans, and on your digital multimeter's lowest setting, touch one lead to the ground on the plug and the other to the tip. Note the reading, then move the second lead to the plug's middle ring.

Take the difference between the two drivers, then divide it by the driver's impedance. For instance, my AKG Q701 measures 62.2 ohms on one side and 61.4 on the other, so the difference is .8. Divide .8 / 62 to get a percentage, 1.29% or .0129. Now multiply that by the intended listening level. I use 85 db, but it doesn't make much difference if you use 80 or 100. So .0129 x 85 db = 1.09 db. Therefore my Q701 has a rather large 1 db imbalance, which is indeed audible.

(I read that 1 db is intended to be the smallest perceptible change in volume, which seems about right to me. I received an ATH-EC700 today, and my measured .8 db imbalance has been driving me nuts!)

A couple of things I feel I should note:

*I've tested every pair I own, and I've noticed nearly every "high end" pair is matched to 1/3 of a db or less. The worst is an old Pioneer SE L40, and after removing the parallel resistor in the earcups I got 16 ohms on one side and 14 on the other! Yikes!

*Most multimeters will only display three (and a "half") significant digits, so I'm not sure what you can do for anything over 200 ohms. It's possible that tenths of an ohm might be negligible for numbers that high. I don't own any to test.

*Multimeters can't measure impedance, only DC resistance. Yes, DC is bad for headphones, but on the lowest setting a multimeter isn't going to provide anything near what it would take to do any damage, even on sensitive earphones. I don't own any BA drivers though and I haven't tested one, but I don't think there's any plausible chance of damaging them either. If it can play music at an audible level, it can handle the one second it takes to measure.

 

[castleofargh]

can't say that I'm sold on this at all.
for starters, you assume some more or less even imbalance over the frequency response, which is a special. then you assume that the resistance(and not impedance!) is going to show imbalance, but that's not going to show any imbalance created by acoustic causes. and on most drivers it's not even going to be a complete electrical estimate. so here are even more special cases.
but most of all, your calculations say that the louder we listen, the louder the imbalance??? did I misunderstand something here?

 

[MindsMirror]

Take the difference between the two drivers, then divide it by the driver's impedance. For instance, my AKG Q701 measures 62.2 ohms on one side and 61.4 on the other, so the difference is .8. Divide .8 / 62 to get a percentage, 1.29% or .0129. Now multiply that by the intended listening level. I use 85 db, but it doesn't make much difference if you use 80 or 100. So .0129 x 85 db = 1.09 db. Therefore my Q701 has a rather large 1 db imbalance, which is indeed audible.

Your calculations are wrong, Decibels don't work like that. The difference in your measurements is a factor of 1.013, which translates to only 0.11dB

 

[DJ The Rocket]

@castleofargh I indeed made a bunch of assumptions that I knew to be incorrect. But if @MindsMirror is right, it suggests these estimates are still useful, or am I still missing something?

I hadn't considered that there could be more factors than electrical ones that have real effects on balance/imbalance. That would suggest there may be ways to fix a volume imbalance besides swapping drivers. Trying to figure out how to fix my Q701s is what got me thinking about this in the first place.

Edit: I should note that I replaced one of my Q701 drivers with one I bought in the for sale forum here (I broke the original). So I knew that a new driver caused the volume imbalance. I made the only measurement I have the equipment for, and the result was exactly what I predicted. Obviously I'm not equipped to be exact, but is there a reason more exactness is important, for this kind of problem?

For example, if I were to buy another Q701 driver that matches the DC resistance of either of mine to a tenth of an ohm, shouldn't I expect that to fix it?

 

[castleofargh]

the only good point I see here is that it doesn't require any special plug or adapter. and that's about it. I would sooner trust a microphone, even a cheap one(cellphone?), so long as we can manage to ensure a repeatable placement for both drivers to be measured under the same conditions(or as close as we can achieve). then play some noise and look at a RTA or spectrum app. that would probably provide better information. and at least the imbalance measured is the total imbalance, including even the amp. my O2 with the loudness set randomly will easily have 0.5dB of imbalance on its own. my other amps have way better matching, but it's still something to consider for the average consumer.

about modding, it would probably be limited to specific issues. acoustic changes are unlikely to give an even attenuation over the entire audible range. but depending on the sort of imbalance, it might be a solution. I have the double combo of being lazy and having butter fingers, so I will tend to solve everything through EQ first, and only go for something else if I have no other choice ^_^. but some people seem to do real well with small modding (I would suggest a microphone for that too, having a way to control what's going on is really great help).

 

[MindsMirror]

You're still missing something. Higher resistance does not automatically mean lower output. For instance, higher resistance could mean that the driver has more turns on the coil, which could potentially mean greater efficiency. Sensitivity is not solely an electrical property, it is heavily dependent on the physical properties also. Variations in the manufacturing process could lead to drivers which have the same electrical properties but different sensitivity due to physical variations.

You should be able to match them closely enough by ear using software balancing. Assuming you can hear as low as 1dB difference, in theory I think you can match the channels to within 0.5dB by comparing them in normal vs reversed orientation. After you know the adjustment needed, it is possible to add resistance to electrically change the sensitivity, however the resistance the two driver after the adjustment would not necessarily be the same.

 

[71 dB]

Take the difference between the two drivers, then divide it by the driver's impedance. For instance, my AKG Q701 measures 62.2 ohms on one side and 61.4 on the other, so the difference is .8. Divide .8 / 62 to get a percentage, 1.29% or .0129. Now multiply that by the intended listening level. I use 85 db, but it doesn't make much difference if you use 80 or 100. So .0129 x 85 db = 1.09 db. Therefore my Q701 has a rather large 1 db imbalance, which is indeed audible.

(I read that 1 db is intended to be the smallest perceptible change in volume, which seems about right to me. I received an ATH-EC700 today, and my measured .8 db imbalance has been driving me nuts!)

That's so wrong it's hard to read! Let's assume the output impedance of your amp is 2 Ω, left channel is 62.2 Ω and right channel is 61.4 Ω.

Attenuation on left channel due to voltage division: 20*log10 (62.2 Ω / (62.2 Ω + 2 Ω)) = 0.275 dB
Attenuation on right channel due to voltage division: 20*log10 (61.4 Ω / (61.4 Ω + 2 Ω)) = 0.278 dB

In other words the level difference due to impedance difference is less than 0.01 dB. The smallest perceptible change in sound pressure level is about 0.5 dB.

Now, the two drivers are probably different and can have much bigger difference, but you can't calculate it out of the measured impedances.