[dhruvmeena96]

Impedance is always frequency related and you should use the impedance at the crossover point to be most accurate. Not the DC resistance.

I zobel everything so everything is DC to me

 

[Farquarl]

I zobel everything so everything is DC to me

Ok... but even with theoretical flat impedance (which is not the case when a zobel is used, it is just more flat) impedance ≠ DC resistance.

 

[dhruvmeena96]

Ok... but even with theoretical flat impedance (which is not the case when a zobel is used, it is just more flat) impedance ≠ DC resistance.

I design zobel in respect to dcr now
Rather than 500Hz

Slight sound change while designing with DCR
But in multi driver, it just works

And works flawlessly

 

[musmecca]

HODVTEC (internal series) L-pad 4.7obm series parallel yellow
ED29689 - green damper - 22ohm series

HODVTEC tube as long as you can fit
ED29689 10mm with green damper at 8mm

The most simple Bass shelf + diffuse feild iem

If you wanna do complex, that can be done too




HODVTEC as above

Sonion 33AJ007(internal series)
10uF 47ohm RC high pass
Red damoer at 10mm
14mm tube
Adjust electrical connection to adjust phase
Technically reverse phase.

ED29689 CT
2.2uF High pass
14mm tube
Green damper at 10mm
Again reverse phase or just check phase in FR graph

WBFK as per taste


Why did I write check electrical phase


Because in certain connection
The 33AJ will dip at 200Hz which is regular in phase with HODVTEC electrically

And uppermid will cancel on 29689 if same phase HODVTEC

As crossover shift phases

So do check, that will give you better idea as how cross behave

So, here is the 1st version i'll call heavy 1. Couple questions:
* You didn't specify diameter of Dtec tube...I went with 1 MM. is that fine?
*You say "yellow" are you referring to a damper on the Dtec? I didn't show one...
*The Dtec is not internally wired is it? Knowles didn't say.... So to make a series, I would take my positive line to the positive terminal, then to negative terminal, to next positive terminal, then next/last Negative terminal where it then connects with the negative side....correct? EDIT...Nevermind, I see they are connected...
*lpad has same values, not two different values...correct? I'm used to seeing two different values.

 

[tedbender84]

I design zobel in respect to dcr now
Rather than 500Hz

Slight sound change while designing with DCR
But in multi driver, it just works

And works flawlessly

Regarding RC filter for still not clear - should I include driver resistance during filter calculation or not?
I spend all day with FR plots from sonion data sheets, trying to find proper hi pass for 38dj and 38aj. From FR plot the best point to hi pass 38aj(I want to use with closed wents to match with two 2389 - around 117db with 0.1v) is 300hz. 38 aj has 118db at 300hz and 38dj has 115db at 300hz with RC hi pass 38aj will have 115db, so out put sum will be 118db. It will give bass boost about 7db from 100hz down to 20hz and flat 118db from 100 to 1000hz. I want to implement two 2389 to increase output by 3db to match 38aj. Which frequency shoud I choose to RC hi pass those two 2389? And I calculate total dc impedance of all drivers and it’s super low - something like 1.5ohm, what is a proper way to increase it - zobel or just one resistor of 15ohm in series with all drivers?
Thanx in advance for any advises.

 

[Bassiklee]

I design zobel in respect to dcr now
Rather than 500Hz

Slight sound change while designing with DCR
But in multi driver, it just works

And works flawlessly

So, correct me if I'm wrong, but my understanding of a zobel is that it's there to correct the impedance peaks/dips. As such don't they have to kind of compliment the actual numbers of what's happening, rather than a dcr number somebody wrote on a page of specs?

 

[dhruvmeena96]

So, correct me if I'm wrong, but my understanding of a zobel is that it's there to correct the impedance peaks/dips. As such don't they have to kind of compliment the actual numbers of what's happening, rather than a dcr number somebody wrote on a page of specs?

i mean, for super accurate build, you have to measure every driver
but it works with spec sheet value too

just ask knowles for inductance, wait for reply and they will give you some drivers sample inductances
average them all up and then work on your driver

it works, just not as neat, there would be 0.5 to 1ohm dips here and there(often at where the impedance peak was)

overall, its flat

if you measure your driver with your own rig, then it reaches damm close to flat line

but i remember once what @piotrus-g told
it the ratio of impedance points more than impedance itself

if you can get near 1:1 ratio on highest and lowest point. its good

 

[Farquarl]

i mean, for super accurate build, you have to measure every driver
but it works with spec sheet value too

just ask knowles for inductance, wait for reply and they will give you some drivers sample inductances
average them all up and then work on your driver

it works, just not as neat, there would be 0.5 to 1ohm dips here and there(often at where the impedance peak was)

overall, its flat

if you measure your driver with your own rig, then it reaches damm close to flat line

but i remember once what @piotrus-g told
it the ratio of impedance points more than impedance itself

if you can get near 1:1 ratio on highest and lowest point. its good

I get that you simplify but why take the DCR value and not the average impedance you get of the datasheet?
Take this knowles CI-22955 curve for example. DCR would be roughly where i painted the red line (20Ohms). If you cross before 300-400 hz sure go ahead and use the DCR (if so you don´t need a zobel here anyway) but it is hard for me to imagine that a simple RC zobel can flaten this impedance peak out sufficiently.

 

[dhruvmeena96]

I get that you simplify but why take the DCR value and not the average impedance you get of the datasheet?
Take this knowles CI-22955 curve for example. DCR would be roughly where i painted the red line (20Ohms). If you cross before 300-400 hz sure go ahead and use the DCR (if so you don´t need a zobel here anyway) but it is hard for me to imagine that a simple RC zobel can flaten this impedance peak out sufficiently.

Just to remove hassle
I mean while crossing we can control many sound parameters

So a DCR zobel is kinda fool proof

For single driver iem

I use classic 500Hz data though

Plus the inductance of CI 22955 is 11.3 to 13mH(varying per sample)
As per knowles

 

[Farquarl]

Just to remove hassle
I mean while crossing we can control many sound parameters

So a DCR zobel is kinda fool proof

For single driver iem

I use classic 500Hz data though

Plus the inductance of CI 22955 is 11.3 to 13mH(varying per sample)
As per knowles

Allright, i guess where are on the same page than.

 

[Aldo40]

Hello if it helps to be more accurate for your designs, I sent a shdh to his website some used driver impedance that I measured with a DATS V3.

 

[mattmatt]

Hey guys just want to share something i whipped up some time ago.

This is as flat as I can get. It's a 2-way, 3BA driver set up. fairly simple.

 

[Farquarl]

Hey guys just want to share something i whipped up some time ago.

This is as flat as I can get. It's a 2-way, 3BA driver set up. fairly simple.

That's crazy flat!

 

[JEHL]

Speaking of target curves, by the time I forced my BL-03 to sound flat to my ears, attempting to add the EQ to existing measurements seems to turn its curve into a non bass rolled HD600 lookalike.

To my knowledge the Sennheiser HD600 isn't targeting ANY curve at all. And yet it apparently would sound very flat to my ears.

And yes I'm comparing that EQ'd BL-03 to a headphone because I never found an IEM equivalent of said headphone.

 

[Farquarl]

Speaking of target curves, by the time I forced my BL-03 to sound flat to my ears, attempting to add the EQ to existing measurements seems to turn its curve into a non bass rolled HD600 lookalike.

To my knowledge the Sennheiser HD600 isn't targeting ANY curve at all. And yet it apparently would sound very flat to my ears.

And yes I'm comparing that EQ'd BL-03 to a headphone because I never found an IEM equivalent of said headphone.

Doesn´t suprise me at all, our ears are quite different. The best approach to see what would sound flat to you in my opinion the Griesinger Method with pink noise.

" Using a known flat/accurate loudspeaker positioned in front and up close (to minimize room effects), alter the loudness of pink noise at different frequencies (every one-third of an octave) until they sound similarly loud to a pink noise at a reference frequency (500Hz). Repeat for the whole audioband. Then repeat for headphones. "