or Connect
New Posts  All Forums:Forum Nav:

Home-Made IEMs - Page 115

post #1711 of 5645

Hi

 

I'm planning on making my first set of DIY IEMs to replace my lost set of westone 4.

 

After trolling through this huge thread I decided on using the easily available Knowles BA drivers, CI, DTEC and WBFK. I have seen people using the combination of TWFK and CI, but since I like rich mids, I rather go the extra mile for the DTECs, plus the westone 4s use the DTEC and TWFK so I figure this will be close enough with the sub-bass boost from the CIs.

 

One thing to note about this combination which I have seen, is that some people comment that the WBFKs are less sensitive than the CI and DTECs, hence the highs sounds recessed or the bass and mids need a resistor in series to tame the response. Could anyone advise me on a suitable crossover network for these 3 drivers (resistor values and cap values would be appreciated, or suitable frequencies as I know how to calculate that myself). Alternatively, should I go for dual WBFKs since they’re really small and easy to fit into the shell?

 

Also, do I need to cascade multiple components to create a bandpass network for the mids? I think I need a minimum of 2 high and 2 low pass to create a suitable network (low pass for subwoofer, bandpass for woofer, high pass for tweeter), but I’ve seen examples with less. Would it be bad to pass the high frequencies to the CI or  low frequencies to the DTEC as that might cause problems with interference?

 

I also considered making an active network in a separate box and hacking a player into it, but that seemed too complicated especially considering the size, but I’ve seen one example in this thread so I might consider modifying it one day to include a suitable network.

 

I also would like to eventually customise this IEM, but for now I’m going to rapid prototype a universal shell since I have access to a really good machine from stratsys with a lovely clear material. Does anyone have any idea where I can find a online model to base my shell on? I’ve done a simple rounded shell in solidworks to see if it fits in my ear, but doing a more complex model without exact measurements of an existing IEM is too hit and miss for me.

 

Thanks for helping!

post #1712 of 5645

why not use twfk instead of wbfk? that gives you a nice 5 way and it will only cost about $20 more total.

post #1713 of 5645
Quote:
Originally Posted by CMOS1138 View Post
 

why not use twfk instead of wbfk? that gives you a nice 5 way and it will only cost about $20 more total.

Considering the local cost of the WBFK being $30, and TWFK being $65, I'd prefer to buy 2x WBFK just for the highs if I went down that road.


The way I thought the crossovers would work would be to cut off tweeters from producing the mid range frequencies, hence there would be no point in paying for the FK mid woofer that we don’t use?

post #1714 of 5645
Quote:
Originally Posted by vector1 View Post
 

Hi

 

I'm planning on making my first set of DIY IEMs to replace my lost set of westone 4.

 

After trolling through this huge thread I decided on using the easily available Knowles BA drivers, CI, DTEC and WBFK. I have seen people using the combination of TWFK and CI, but since I like rich mids, I rather go the extra mile for the DTECs, plus the westone 4s use the DTEC and TWFK so I figure this will be close enough with the sub-bass boost from the CIs.

 

One thing to note about this combination which I have seen, is that some people comment that the WBFKs are less sensitive than the CI and DTECs, hence the highs sounds recessed or the bass and mids need a resistor in series to tame the response. Could anyone advise me on a suitable crossover network for these 3 drivers (resistor values and cap values would be appreciated, or suitable frequencies as I know how to calculate that myself). Alternatively, should I go for dual WBFKs since they’re really small and easy to fit into the shell?

 

Also, do I need to cascade multiple components to create a bandpass network for the mids? I think I need a minimum of 2 high and 2 low pass to create a suitable network (low pass for subwoofer, bandpass for woofer, high pass for tweeter), but I’ve seen examples with less. Would it be bad to pass the high frequencies to the CI or  low frequencies to the DTEC as that might cause problems with interference?

 

I also considered making an active network in a separate box and hacking a player into it, but that seemed too complicated especially considering the size, but I’ve seen one example in this thread so I might consider modifying it one day to include a suitable network.

 

I also would like to eventually customise this IEM, but for now I’m going to rapid prototype a universal shell since I have access to a really good machine from stratsys with a lovely clear material. Does anyone have any idea where I can find a online model to base my shell on? I’ve done a simple rounded shell in solidworks to see if it fits in my ear, but doing a more complex model without exact measurements of an existing IEM is too hit and miss for me.

 

Thanks for helping!

Hey there! Do you plan on measuring the frequency response? What is the material you are printing?

post #1715 of 5645
Quote:
Originally Posted by tomtomtom View Post
 

Hey there! Do you plan on measuring the frequency response? What is the material you are printing?

I would like to, and my friends have plans to measure, but personally I think it's not worth the time as I can't control the variables well enough to actually tune the response. I’ll use a heavy dose of EQ after the fact to smooth out anything the crossover doesn’t do well enough. I think the problems will be heavy spikes at certain frequencies thanks to resonance and the drivers themselves, which a crossover can’t handle unless I squeeze in a notch filter.

 

Like I said, I might choose to remove everything redesign around a active filter circuit which I know how to design. I just want to get suggestions on a passive crossover that I can actually squeeze into the earphone. :)


I will likely will print in Veroclear. http://www.stratasys.com/materials/polyjet/transparent . It’s not the most strong, but the tranparency will be nice if I can arrange everything inside presentably

post #1716 of 5645
Quote:
Originally Posted by vector1 View Post
 

I would like to, and my friends have plans to measure, but personally I think it's not worth the time as I can't control the variables well enough to actually tune the response. I’ll use a heavy dose of EQ after the fact to smooth out anything the crossover doesn’t do well enough. I think the problems will be heavy spikes at certain frequencies thanks to resonance and the drivers themselves, which a crossover can’t handle unless I squeeze in a notch filter.

 

Like I said, I might choose to remove everything redesign around a active filter circuit which I know how to design. I just want to get suggestions on a passive crossover that I can actually squeeze into the earphone. :)


I will likely will print in Veroclear. http://www.stratasys.com/materials/polyjet/transparent . It’s not the most strong, but the tranparency will be nice if I can arrange everything inside presentably

I think it'll be hard to design a three let alone a fourway driver system without being able to measure the frequency response, but it'll be interesting to see how they turn out. I`m currently having a similar issue, but I think I'll start of with dynamic drivers and then build up a measuring system for BA drivers.

post #1717 of 5645
Quote:
Originally Posted by tomtomtom View Post
 

I think it'll be hard to design a three let alone a fourway driver system without being able to measure the frequency response, but it'll be interesting to see how they turn out. I`m currently having a similar issue, but I think I'll start of with dynamic drivers and then build up a measuring system for BA drivers.

I understand that, that's why I would like to a) work with known drivers b) get suggestions regarding crossover frequencies. I don't need it to be perfect first time round as eventually I'll remove them from the universal housing and put them into a custom IEM. Hence this crossover is just the beginning :)

post #1718 of 5645
Quote:
Originally Posted by vector1 View Post

 

b) get suggestions regarding crossover frequencies.


Thats most difficult one. You will have to measure it and design your crossover. Crossover components are RC filter. You will need smd caps and resistors. Start with simple 2way design and then  if thats a success you can do more complex designs.

post #1719 of 5645

Some one has uploaded Knowles design guide in other thread. Did you read those docs ?
 

post #1720 of 5645

So I can't just ballpark a figure for a crossover at 150Hz and 2kHz?

 

From what I understand is that the complex (changing) inductance means that the spikes will not be at the expected locations from measurements.

 

However, will this lead to excessive non-linear responses? I would expect a shift in the cut-off point, but not necessarily lead to non-linearity in the frequency response? Furthermore, with passive circuits I don't see a way to actually equalize the response since RC filters will have unity response when not close to the cut-off, and that won't help in smoothing the spikes in the actual response unless notch filters are used.

 

I am thinking that as long the crossover prevents interference of each band from passing to the next, I'll use acoustic filters and EQ to tame the rest of the peaks.

 

Also, I will try to make this modular so I can actually revise the interior circuit design. I would print a special holder for the circuit components that should be serviceable easily, plus help in soldering the SMD components.

 

Regarding the Knowles design guide, if it's the one that has the frequency response of each of the drivers then yes, I have read and extracted that data. I haven't seen any driver specific guide, but I have seen previous examples in this very thread, but mostly using a more simple crossover.

post #1721 of 5645
Quote:
Originally Posted by MuZo2 View Post
 

Some one has uploaded Knowles design guide in other thread. Did you read those docs ?
 

 

Would you kindly post a link?

 

Thanks!

post #1722 of 5645
post #1723 of 5645

Thank you for that thread!

post #1724 of 5645

Sound quality problems.

 

I attempted to finish assembly on my first set of home made IEMs and right away I ran into trouble. 

 

First problem, I can see that the drivers for the left ear have become detached from the acoustic tubing so I did not bother to work any more on that shell. 

 

The right ear looks like everything is hooked up OK but it sounds awful. It is much quieter than I expected it to be and I thought perhaps it just needed to be amped but when I hook it up to my headphone amp it still sounds bad, just louder. I am hoping that the cause of the poor sound quality is that this driver has also pulled away from its acoustic tubing.

 

I plan to start over tomorrow with a new shell (I want to improve the finish anyway) and this time I am going to try attaching the tubing with superglue instead of hot glue. I only worry that I will damage the driver with the superglue. 

 

 

Edit:  looking closer at my right side driver I can see that I was correct and the sound port has become detached from my tubing on that side also. 


Edited by CMOS1138 - 5/30/14 at 10:03pm
post #1725 of 5645
Quote:
Originally Posted by vector1 View Post
 

So I can't just ballpark a figure for a crossover at 150Hz and 2kHz?

Yes you can, but have to have resources to make it work correctly.

New Posts  All Forums:Forum Nav:
  Return Home