This is the 5 driver one I've been tuning. The >10kHz part changes dramatically depending on the position of the WBFK driver which has made it a nightmare for me to tune. Also I can't seem to get rid of the ditch at 3kHz...but I think that part isn't too bad since the ear canal resonance frequency is ~3-4kHz?
Home-Made IEMs
- Thread starter Bilavideo
- Start date
This is the 5 driver one I've been tuning. The >10kHz part changes dramatically depending on the position of the WBFK driver which has made it a nightmare for me to tune. Also I can't seem to get rid of the ditch at 3kHz...but I think that part isn't too bad since the ear canal resonance frequency is ~3-4kHz?
Almost good, you need to find a driver and reverse its polarization to reverse what you are seeing between 1kHz and 4kHz - you need peak at around 2,2-3kHz not 1,5kHz. Deep at 12kHz could be better but I don't know how "true" your measurements are in terms of IEC711 so I'm not sure if the deep would actually occur.
I'm guessing WBFK isn't causing the 3kHz dip right? I have also CI22955 and ED23147 as full range, with two ED29689 in first order (reverse wired). I couldn't find which driver was causing that dip, since reversing any driver will cause other problems...
If I push the earphone in a bit, the graph will look quite different in the 10kHz+ region. There will be a much smoother decay and better extention. I'm guessing the region isn't very accurate.
If I push the earphone in a bit, the graph will look quite different in the 10kHz+ region. There will be a much smoother decay and better extention. I'm guessing the region isn't very accurate.
vector1
Head-Fier
- Joined
- May 28, 2014
- Posts
- 85
- Likes
- 11
On one hand it sounds like an awesome idea if you really want to customize a frequency response to someone's ear, but there might be other factors like hearing loss/difference in sensitivity between ears that make it more complicated. I think the additional cost might be better spent in terms of actually auditioning and calibrating from a baseline though like the Ultimate ears personal reference monitor...
Perhaps this can be a unique selling point for very high end earphones, where each ear's frequency response is taken into account. I can imagine that it might significantly improve the sound if the person has a unique/very different canal shape between ears.
To make it faster though, I think some kind of 3D printing technology would work great for this since you can scan and rapid-prototype out the canal with the right plugs/distances and a consistent material. Creating a mold seems a little iffy since you still have to build up the part connecting to the eardrum? The tube might end up affecting the sound even more than the material used, since I think the material will play some part in the resonance effects.
vector1
Head-Fier
- Joined
- May 28, 2014
- Posts
- 85
- Likes
- 11
You make me very eager to test out my own earphones to see what's wrong/right with them now ^^.
HODVTEC/CI/GQ are on the way with an extra set of EDs to put into my current testing set! Waiting for the shipment now...
I also want to ask everyone, what do you think of a form factor like the beats earphone for my next testing set?
The left side is my current design updated to fit the GQ in, and angled the nozzle even further to fit my ear properly.
The right is a design form factor I'm just looking at to see if I can make this into a conventional looking straight cable earphone, but with awesome sound. This will fit girls/guys with smaller ears as I've seen that my current design doesn't fit well into those. Moving one driver outwards will reduce the size directly in the ear, and give more length for the GQ design to develop a long channel before reaching the eardrum to maximize the response around 2kHz.
I'm worried that the WBFK might mess up the phase somewhere, since it seems to be quite sensitive to positioning...
What are your tube lengths at each ED driver? I'm guessing your tubs are too long and are causing frequency shift. You should be operating at around 20-25mm for CI 16-13mm for ED, 16-13mm for WBFK.
ED23147: ~18mm, ED29689: ~11mm and 15mm, WBFK30019: ~8mm, CI22955: ~14mm
I tried shifting the lengths by a few mm but they didn't really do much :\ The WBFK changed a lot with shifted position though, but that could have been the inaccuracy of the 711 above 10k
Oh I just changed the crossover of the ED29689. Both are in series connected by center tap to a 2.2uf cap. Do I need to reverse wire this?
I tried shifting the lengths by a few mm but they didn't really do much :\ The WBFK changed a lot with shifted position though, but that could have been the inaccuracy of the 711 above 10k
Oh I just changed the crossover of the ED29689. Both are in series connected by center tap to a 2.2uf cap. Do I need to reverse wire this?
- Joined
- Jul 15, 2008
- Posts
- 1,150
- Likes
- 95
FitEar also uses a similar method. Though I wouldn't recommend it for the cost involving isn't suitable for a DIY project, unless you wanna go all out.
This is the 5 driver one I've been tuning. The >10kHz part changes dramatically depending on the position of the WBFK driver which has made it a nightmare for me to tune. Also I can't seem to get rid of the ditch at 3kHz...but I think that part isn't too bad since the ear canal resonance frequency is ~3-4kHz?
It looks pretty good. You just need to fix the upper mid region as well as sub 100hz frequencies.
Ear canal resonance is within 7khz~10khz. What you were referring to is the effect of head & torso acting as filters.
Sub 100Hz looks like that because the ear plug that fits on the shell keeps popping out of the 711 
It always moves slightly outwards during the recording and can break the seal. But when I press it down it looks better. Upper mids...I'll try reversing the CI and later the ED to see if it helps... 
It always moves slightly outwards during the recording and can break the seal. But when I press it down it looks better. Upper mids...I'll try reversing the CI and later the ED to see if it helps... Alright, decided to just push it in and give it the proper seal. This should be from a deeper insertion than normal, but the graph changed quite a bit. The tubes were adjusted slightly, but then the 3k issue is still there...there seems to be a 4k issue too now o.O
Alright, decided to just push it in and give it the proper seal. This should be from a deeper insertion than normal, but the graph changed quite a bit. The tubes were adjusted slightly, but then the 3k issue is still there...there seems to be a 4k issue too now o.O
This is perfect man! It should be spot on. I'd another WBFK for the sake of high frequencies level. Other than that it's really really good.
- Joined
- Jul 15, 2008
- Posts
- 1,150
- Likes
- 95
If I were you I would disconnect the WBFK, reverse the polarity of the 2 ED's. There's only 4 permuations so it wouldn't take more than 30 min. When everything is good try connecting the WBFK again. If all fails play around with tube length.
Btw, always start with a fixed length for all tubes. That would make it easier to troubleshoot later.
Btw, always start with a fixed length for all tubes. That would make it easier to troubleshoot later.
Really? This is what a good frequency response should look like?! It looks pretty nasty...
I usually seal the mouth after putting the tubes in, but leave a bit of around 5mm sticking out for adjustment in case anything screws up.
Edit: That's a really bad idea to do. After getting the previous response, I cut the tubes to fit it back into the shell and now there is a phase cancelling issue at 5k, really bad issue too
@tranhieu there's actually less than 4 permutations since the 2 ED29689s are connected in series through the center tap
I usually seal the mouth after putting the tubes in, but leave a bit of around 5mm sticking out for adjustment in case anything screws up.
Edit: That's a really bad idea to do. After getting the previous response, I cut the tubes to fit it back into the shell and now there is a phase cancelling issue at 5k, really bad issue too
@tranhieu there's actually less than 4 permutations since the 2 ED29689s are connected in series through the center tap
- Joined
- Jul 15, 2008
- Posts
- 1,150
- Likes
- 95
Really? This is what a good frequency response should look like?! It looks pretty nasty...
I usually seal the mouth after putting the tubes in, but leave a bit of around 5mm sticking out for adjustment in case anything screws up.
Edit: That's a really bad idea to do. After getting the previous response, I cut the tubes to fit it back into the shell and now there is a phase cancelling issue at 5k, really bad issue too
@tranhieu there's actually less than 4 permutations since the 2 ED29689s are connected in series through the center tap
No what I'm saying is, you measure all 4 scenarios and look for the one with least cancellation.
If you are half wiring, try reversing the polarity and see if that helps?
If you have some spare caps with values being lose to the one you are using lying around, try swapping them as well
Users who are viewing this thread
