[mckpaul]

A long time ago I developed my own UV PCBs using inkjet traces on transparency film. Not sure if that would be practical for something of this size though. If you have an FDM printer, have you thought about getting a laser for it and cutting the copper that way? The lasers are very cheap and should have good accuracy but I've not tried it

I have a NEJE laser engraver, much larger than needed for this kind of work, but I figured I may use it for other projects. I've used it on various shell colors created from Fotoplast. For the record, if you ever want to use a laser engraver on translucent UV created shells, the power and speed changes from color to color, and even isn't consistent on two shells the same color if one is slighter thicker and darker than the other. And the blue laser has a hard time on blue shells. But the results were pretty consistent on black shells. I found a simple solution that generates the same results one any color shell for small paint fillable text, that being spray paint the shell being engraved black (or at least the engraving area, but being so small you end up spraying the whole shell). The spray paint is easy enough to remove with fingernail polish remover.

But that wasn't the point. I thought about using the laser to etch away the copper from the board, since the board is so thin I could probably use the laser to cut the tiny boards out as well as etch the copper. I guess I have a new experiment to try, but if anyone has etched copper from a pcb with an LED laser please share.

 

[swtnate]

I have a really simple question but one that has a lot of different answers, what effect does different sized of acoustic tubing have on sound ? I have 1.5mm ID and 2mm ID but the 2mm tubing is quite a squeeze in the canal when using two tubes with a gk.
Any smart people answers on the effect of the tubing size ?

This is such a wide conversation. The short of it, wider tubing is great for filtering out low end. Smaller ID filters out high frequency. Tube length can amplify and filter certain frequency ranges. You will have to account for resonances and reflections. Its a simple question with a novel’s worth of potential responses.

 

[tonybuild]

I use pro3dure pouring resins. If you are using the Ali nicefit resins, you will probably do best with either the Egger or Dreve. Pro3dure cures at a bit different UV wavelength. For lacquers I use the Pro3dure Lacquer and the Detax Shellac. I have purchased all my liquids through warner.

I use detax shellac too . Like it a lot .
Dreve gets yellow fast .


also can i have you advise? I want to try this driver . But I don’t know if I should pair it with sonion 33 or 38 ?

 

[piotrus-g]

I use detax shellac too . Like it a lot .
Dreve gets yellow fast .


also can i have you advise? I want to try this driver . But I don’t know if I should pair it with sonion 33 or 38 ?

This is tweeter so you need something in between like ED/2300. While E50DT can produce 2-3kHz its not a good range to use it due to very very high thd in that area. If you really have to use only one out of 33/3800 and no ED between then 33AJ007/9i would be my choice to produce a rather flat (actual flat not harman or DF flat) looking response on graph, but be wary of what's happening around 6kHz with 3300, as it is prone to sibilance

 

[PeaksZA]

Another possible noob question, but I'm not getting any high end with my knowels GK. Im using two 1.5mm tubes with a red damper on the CI and grey on TWFK , but there is no presence and sounds quite lofi subsequently. But good bass and mid response. Ive done a few experiments but still seems like there a great lack of high frequencies. My next thought is that it could be my canal design , I have my molds go to the first bend but seems like that could cause a problem. I made duplicate shells with RAB and 2mm tube and still the same problem of great lack of high end above 6-7k .
Any suggestions?View attachment IMG_20231021_190031.jpg

 

[Touchowl]

Anyone know what would cause bass leakage after front vent has been covered, and coupler is sealed?(first pic compares against Truthear Zero Red measured with same tips to show that there is a seal, second pic shows just the bass DD, comparison with front vent covered/uncovered.)

Alrighty, I've gotten the response out of "something here is clearly broken" territory and into "gold star for effort":

I thought the lack of bass was from the bass driver not having a properly vented back cavity, but I saw someone on here say that wider tubing can filter out low end, and all of my tubes meet in a 4mm wide cylinder cavity before reaching the nozzle, so maybe that's the problem?
Either way, I'm happy with my first try. I even got a successful crossover! Mid BA vs Bass DD here:

The 5mm DD tweeter is laughable, though. Barely audible, if it affects the FR at all it's not in a way my 711 coupler can detect:

 

[IEMbiker]

Alrighty, I've gotten the response out of "something here is clearly broken" territory and into "gold star for effort":
I thought the lack of bass was from the bass driver not having a properly vented back cavity, but I saw someone on here say that wider tubing can filter out low end, and all of my tubes meet in a 4mm wide cylinder cavity before reaching the nozzle, so maybe that's the problem?

The bass problem is normally due to the seal issue. A good seal is essential, as the vents are for tuning and pressure relief. many factors can still cause this: crossover, phase cancellation, tubing length, and size. It helps to go through all the pages for starts. This thread is my DIY iem bibble.

 

[Touchowl]

The bass problem is normally due to the seal issue. A good seal is essential, as the vents are for tuning and pressure relief. many factors can still cause this: crossover, phase cancellation, tubing length, and size. It helps to go through all the pages for starts. This thread is my DIY iem bibble.

Yeah I don't think it's the seal. I got a consistent result from 2 different tips as well as by sealing the entire coupler with putty. Those graphs from my first post are what happens when there's a leak, while proper sealing allows it to get *some* sound below 500Hz.

I am trying to read through this thread, but man can it be tough. A lot of discussion predicated on knowledge I don't have and don't really know where to find, and there is a lot of discussion.

 

[swtnate]

Yeah I don't think it's the seal. I got a consistent result from 2 different tips as well as by sealing the entire coupler with putty. Those graphs from my first post are what happens when there's a leak, while proper sealing allows it to get *some* sound below 500Hz.

I am trying to read through this thread, but man can it be tough. A lot of discussion predicated on knowledge I don't have and don't really know where to find, and there is a lot of discussion.

have you looked at any of the THD, electrical phase testing, acoustic phase testing, and/ or the proper solder temp specs were met AND the proper solder type was applied. These micro devices are not as hearty as the big boys.

 

[Touchowl]

have you looked at any of the THD, electrical phase testing, acoustic phase testing, and/ or the proper solder temp specs were met AND the proper solder type was applied. These micro devices are not as hearty as the big boys.

Ooo I don't know how to check any of those things, but you might be right on that front. The larger DD and BA seem fine, but the smaller DD only producing sound(and very quiet sound) below 6kHz might suggest the driver itself is damaged or improperly soldered. Will have to do tests with a fresh driver.

 

[Touchowl]

Diving back into shell design, trying to ape InEar's Prophile shell to some degree

 

[tonybuild]

Hi guys .
Im trying experiment with twfk ,which I never worked . My customer asked me for 4 driver setup .this is what im working right now ,and what to know your opinion about this scheme ?
Thank you

 

[swtnate]

@tonybuild all three cross components look like great starting points. Couple of things I always watch for when working with these two very different type of drivers. the 3300:i9 has TONS OF juice on the deep sub freqs. Tons of punch but can get a bit reckless, meaning it can do one of two things, muddy up the overall soundstage OR (which is more common) leans toward a much darker sounding IEM. ways I try to combat the bully driver; 1) I always run my positive lead from the connector to the negative pad on the 33i9. 2) connect the negative pad on the 33i9 to the positive pin on the connector. Doing it this way in my testing always yield a cleaner more controllable response that made the lower mids much more responsive to the Passive components. 3) I have stopped using thick film SMD resistors in every build design I make. In their place I use SMD MELF resistors. The tolerances are much more accurate (within 1%) and there is considerable less THD. 4) Over time thick film resistors seemed to always get a little crispy or chippy. Meaning, they would always begin to breakdown and cause some low level unwanted frequencies to trickle in. 5) I prefer L-Pads on the i9 as it causes the first resonate peak to lay down a bit in the same peak position. Whereas a series resistor will always push the resonant peak up the EQ leading to NOTICABLE distortion. 6) Im loyal to the 1mm ID for the I9 when paired with the TWFK. When the i9 is paired with the 2389 or ED using 1mm ID, the result almost always sounds like an old tube radio. BUT, with TWFK, I dont hear this and the charts bare it out.
TWFK is extremely easy to work with. 2mm ID tubing at arlund 15mm with a grey damper 3-4 mm from the sound bore. In this build, dont strangle the TWFK output pontential. Without any RC components, TWFK only reaches 110dB. Real world though, its more like 107-108. You need most of that passive draw. If not, you’ll find yourself chasing your tail jn final tuning.

TL;DR version:
3300 9\i - reverse polarity, calculate your L-Pad for -8 to -12 dB (if Im remembering right thats like a 5.2 Ω MELF (or thick film) in seriez with a 6 Ω in parallel. Also, wire the 2 subs of the 33 parallel. 1mm ID x 2.5 OD at 15mm tube length with a yellow Damper positioned betwen 8-12mm
The TWFK, 2mmIDx3mmOD @15-16mm in length and a white damper 3-4mm from the sound bore.

PSA - On lower driver count builds, Its good practice in my build experience to never share pads with other drivers. All wires connect to driver pads the. led to the the connector and not another driver. Makes any potential future repair easier to diagnose.


sorry for the novel. You caught me just about to redesign my in-home entertainment room and rebuild my wifi network for faster cloud performance. Let me know if you have any questions. I fell in love with doing this by doing this exact 4 driver build at least 40 times. If you can get this dialed in and consistently reproducing the same soundstage from the breadboard. The higher driver count builds make more sense and the biggest challenges become engineering enough space to them as opposed to chasing your sound carrot on a 48 hour bender.


gotta go. THE ADHD is kicking in and if I dont follow the dopamine rn and get back to designing, I will most like start a new favorite hobby by noon.

im here to help anytime!

 

[tonybuild]

@tonybuild all three cross components look like great starting points. Couple of things I always watch for when working with these two very different type of drivers. the 3300:i9 has TONS OF juice on the deep sub freqs. Tons of punch but can get a bit reckless, meaning it can do one of two things, muddy up the overall soundstage OR (which is more common) leans toward a much darker sounding IEM. ways I try to combat the bully driver; 1) I always run my positive lead from the connector to the negative pad on the 33i9. 2) connect the negative pad on the 33i9 to the positive pin on the connector. Doing it this way in my testing always yield a cleaner more controllable response that made the lower mids much more responsive to the Passive components. 3) I have stopped using thick film SMD resistors in every build design I make. In their place I use SMD MELF resistors. The tolerances are much more accurate (within 1%) and there is considerable less THD. 4) Over time thick film resistors seemed to always get a little crispy or chippy. Meaning, they would always begin to breakdown and cause some low level unwanted frequencies to trickle in. 5) I prefer L-Pads on the i9 as it causes the first resonate peak to lay down a bit in the same peak position. Whereas a series resistor will always push the resonant peak up the EQ leading to NOTICABLE distortion. 6) Im loyal to the 1mm ID for the I9 when paired with the TWFK. When the i9 is paired with the 2389 or ED using 1mm ID, the result almost always sounds like an old tube radio. BUT, with TWFK, I dont hear this and the charts bare it out.
TWFK is extremely easy to work with. 2mm ID tubing at arlund 15mm with a grey damper 3-4 mm from the sound bore. In this build, dont strangle the TWFK output pontential. Without any RC components, TWFK only reaches 110dB. Real world though, its more like 107-108. You need most of that passive draw. If not, you’ll find yourself chasing your tail jn final tuning.

TL;DR version:
3300 9\i - reverse polarity, calculate your L-Pad for -8 to -12 dB (if Im remembering right thats like a 5.2 Ω MELF (or thick film) in seriez with a 6 Ω in parallel. Also, wire the 2 subs of the 33 parallel. 1mm ID x 2.5 OD at 15mm tube length with a yellow Damper positioned betwen 8-12mm
The TWFK, 2mmIDx3mmOD @15-16mm in length and a white damper 3-4mm from the sound bore.

PSA - On lower driver count builds, Its good practice in my build experience to never share pads with other drivers. All wires connect to driver pads the. led to the the connector and not another driver. Makes any potential future repair easier to diagnose.


sorry for the novel. You caught me just about to redesign my in-home entertainment room and rebuild my wifi network for faster cloud performance. Let me know if you have any questions. I fell in love with doing this by doing this exact 4 driver build at least 40 times. If you can get this dialed in and consistently reproducing the same soundstage from the breadboard. The higher driver count builds make more sense and the biggest challenges become engineering enough space to them as opposed to chasing your sound carrot on a 48 hour bender.


gotta go. THE ADHD is kicking in and if I dont follow the dopamine rn and get back to designing, I will most like start a new favorite hobby by noon.

im here to help anytime!

Thank you. I also sent you my corrected pic of the scheme

 

[stegeoc]

Another possible noob question, but I'm not getting any high end with my knowels GK. Im using two 1.5mm tubes with a red damper on the CI and grey on TWFK , but there is no presence and sounds quite lofi subsequently. But good bass and mid response. Ive done a few experiments but still seems like there a great lack of high frequencies. My next thought is that it could be my canal design , I have my molds go to the first bend but seems like that could cause a problem. I made duplicate shells with RAB and 2mm tube and still the same problem of great lack of high end above 6-7k .
Any suggestions?

Personally I had a lot of problems with GKs too. The built in crossover just didn't work for me whatever I tried. The natural mid-range dip always seemed to be around 1kHz or higher, which just sounds wrong to me. After a lot of messing about I separated the drivers (they come apart easily) and redesigned the crossover with a proper low pass on the CI and high pass on TWFK. Tubes were <1mm on the CI and > 2mm on the TWFK (no dampers).

Personally I ignore everything above 6-7kHz on my measurements as my setup doesn't seem to measure that very well. It could be that you do have it, but can't see it. Perhaps your tubes are pointing in the wrong direction out the end? If not going around the bend you probably need to be more careful about what direction the sound is coming out in