[castleofargh]

I do it with REW. I guess the signal can't be of fixed voltage because the most sensitive IEMs wouldn't appreciate and too small a voltage would be a problem for less sensitive stuff. so I guess setting the output to something like 90db at 1khz for the IEM and just adjusting the gain/sensi of the ADC like I do, might be what most people do?
 
should the IEM simulate normal use or be lying down with a straight cable? like should the IEM feel some occlusion effect as if in the ear? that can definitely have some impact on the measurement of some IEMs. it got me very confused when the math wouldn't add up between impedance changes measured without occlusion(adding resistors in the path for the lolz), and FR measured with occlusion from getting a proper seal for my coupler. for the longest time I just assumed my measures were utter crap because sometimes they would fit electrical rules and sometimes not.

should the cable be pulled straight? on rare occasions the way I place the cable changes very significantly the impedance, I think the worst was with some fitear IEMs. just making a little loop with the cable  while putting it down on the table, or keeping both drivers close, would make it look like I was measuring a different IEM.

 
 
anyway, even if there isn't a rule to doing those stuff, maybe some who do IEM measurements have a routine they'll be willing to share? I get by just fine for my own uses, but I'd rather have a kind of standard for when I wish to show the impedance of an IEM to others and it happens to be a tricky earphone.
 
 
to the 3 guys who will read this. thank you ^_^

 

[100VoltTube]

Tyll does all of his measurements with the IEMs in the measurement head (I think) I have no idea about anyone alse

 

[CraftyClown]

Just so you know, I'm always one of the three guys who reads the posts, however I'm rarely one of the guys who fully understands the question, let alone has an answer

 

[pinnahertz]

  I do it with REW. I guess the signal can't be of fixed voltage because the most sensitive IEMs wouldn't appreciate and too small a voltage would be a problem for less sensitive stuff. so I guess setting the output to something like 90db at 1khz for the IEM and just adjusting the gain/sensi of the ADC like I do, might be what most people do?
 
should the IEM simulate normal use or be lying down with a straight cable? like should the IEM feel some occlusion effect as if in the ear? that can definitely have some impact on the measurement of some IEMs.

Simulate normal use, always, for measurements.

it got me very confused when the math wouldn't add up between impedance changes measured without occlusion(adding resistors in the path for the lolz), and FR measured with occlusion from getting a proper seal for my coupler. for the longest time I just assumed my measures were utter crap because sometimes they would fit electrical rules and sometimes not.  

Impedance measurements must be taken with a known source impedance, and it will be higher than the driving amp of necessity. That means the sense resistor must be there, and the total Z of the generating system, including the R, must be known for the measurement to be accurate (you do that during calibration).  From the REW manual, "Good results can be obtained using a headphone output to drive the load, with a 100 ohm sense resistor. If a line output is used the sense resistor typically needs to be much larger as line outputs have high output impedance and limited drive capability, try 1 kOhm but note that the results will have much higher noise levels."
 
There's no particular reason the impedance curve would match FR, and quite a few reasons it won't.

should the cable be pulled straight? on rare occasions the way I place the cable changes very significantly the impedance, I think the worst was with some fitear IEMs. just making a little loop with the cable  while putting it down on the table, or keeping both drivers close, would make it look like I was measuring a different IEM.

This sounds like a test setup problem, connection, impedance, etc. With typical IEM impedance the cable should have no effect. I'm not saying your observations are wrong, but to do that it sounds like a connection problem. Or the worlds worst cable (unlikely, though). IF your source resistor is quite high and signals are low those IEMs become microphones, so you have to balance source and load Z to keep things sane.

to the 3 guys who will read this. thank you ^_^

4.

 

[castleofargh]

oh yeah I've struggled for like 2 years with R sense because the resistors I had to play with would significantly change properties depending on the signal passing through(I dissipate how I want!)

. not too long ago I realized that there were some resistors just for such exercises so I asked and got told they cost real money. which of course is a no go, I measure like 10 impedances per year out of curiosity. but since, I've found a crappy looking 100ohm resistor that is way more stable than the ones I had from some old electronic kit. not perfect but I'm happy with that for now.
I'm using my O2 as source that's the lowest impedance I have. I measured 0.58ohm.
 
anyway I honestly never asked those questions because I didn't expect much accuracy from cutting corners and prices so often. so no reason to sweat the small stuff. but last week I got bored of having to make a blob of fiio LO cables and croco plugs each time I wished to measure the impedance of something, so I went and demonstrated my pitiful inability to solder anything and made impedance measurements great again. with better(more stable) results, some small stuff now feel like they might matter a little. but again my concern is more about sharing measurements done with some expected method if there is one. for myself, I measure stuff in the conditions I need and don't care about standards.
 
 
 the cable anecdote is just that, I often get small variations but there were only like 2 times with really craycray behaviors, even my hand near the cable was significant. I wish I still had them to show you measurements and pics of the cable side by side for fun. but now I almost only own single driver IEMs and tend to run away from anything with real low impedance, so cables being important in any way are less likely for the time being. I sometimes get some impact from the magnets if I leave the 2 drivers on each other, but that's obviously not even worth a question. nobody will put both drivers in one ear ^_^.
 
and when I mention the math not working I meant like adding 5ohm to the IEM and getting the proper FR change predictable from the impedance curve(and known source). after trying to measure impedance with the IEM still sealed in the coupler, the results were much more like expected. which is kind of obvious, same conditions, same data. but that too is IEM dependent.