Home-Made IEMs

Dec 20, 2024 at 11:01 AM Post #16,261 of 16,330
Hey… I have a problem with bass distortion…. i built the setup in the picture....but instead of the 2389 i used twfk30017 (rev pol) with 22uf cross and 38d1xj for 38aj... the curve looking not to bad but there is hearable and measurable distortion of about 3% on 2nd harm.....what can i do????
6driver 38d1.jpg
IMG_1883.jpeg
 
Last edited:
Dec 31, 2024 at 11:23 PM Post #16,262 of 16,330
Hey… I have a problem with bass distortion…. i built the setup in the picture....but instead of the 2389 i used twfk30017 (rev pol) with 22uf cross and 38d1xj for 38aj... the curve looking not to bad but there is hearable and measurable distortion of about 3% on 2nd harm.....what can i do????6driver 38d1.jpgIMG_1883.jpeg
3% is not bad on 2nd harmonic, should just sound a little warm. Try smaller diameter tubing and/or less resistance on 38d1xj for a little more volume, tape the vents if you want less distortion, but you lose about 4db with closed vents, so correct with some added series resistance on the 30017. Watch the phase. What's your circuit? 10k+ looks a bit bright but damping and/or series resistance will tilt the right side of the curve down a bit.
 
Jan 18, 2025 at 4:57 AM Post #16,263 of 16,330
Hello everyone.
After having investigated and collected a lot of information on the Internet and in this forum, I am finally going to start preparing my own ciems.
My list of materials that I have already obtained so far:
Dreve S IO blue transparent.
Dreve Lack 3
Paraffin wax.
For the silicone mold https://www.amazon.es/dp/B0C2ZLPJ27?ref=ppx_yo2ov_dt_b_fed_asin_title&th=1

But I have a couple of questions before I start.
I intend to use my wife's 24w UV LED nail lamp on a rotating platform on a mirror base.
Taking into account the material of the silicone mold, which is not very transparent (rather translucent). In this video he uses a similar product with the same transparency as the one I intend to use and achieves good results

How much exposure time would you recommend as a starting point to start testing the S IO?
You know how expensive these Dreve products are and the idea is to waste as little as possible.
On the other hand, I have read that to finish curing the shells it is advisable to do a glycerin bath for the final curing.
Does anyone have experience with these products that has obtained good results?
On the other hand, doubts about the drivers. I had thought about the Knowles GK 31732 but apparently my ear canal is a little narrow and I'm not sure I can install 2 tubes. Does anyone have experience in this case? Could a single tube be used and if so, with what damper?
Can you recommend other drivers that work well with a single tube?
Thank you very much in advance.
All the best!
 
Last edited:
Jan 18, 2025 at 10:30 AM Post #16,264 of 16,330
I use a rotating mirror with nail light...works fine. Pour a tiny bit on a piece of glass and see how it cures as some of those are slightly off in wavelength. My time with clear and agar molds is about 3.75 minutes...4.5 or even 5 minutes for most tinted. One determining factor is the very centermost part behind the canal...it's usually the last to cure so be sure to rotate your forms halfway through curing or you'll think it looks great but there is an uncured hole in the center. The glycerin bath is a must...it will get rid of the tacky inhibition layer caused by exposure to air. Regarding the GK...it can use one tube stretched over the two...that's normal and is covered here. Check these forums for suggested lengths and dampers...there are tons of designs. I actually like the HE design....I often pair it with a Ci/green damper at BA spout, and a 1 mm to .75 mm ID tube.. tiny tube acts as a low pass without need for crossovers or resistors and the 1 mm is easy to accomodate in the canal....I like the added low freq bump. Good luck.
 
Jan 19, 2025 at 1:59 PM Post #16,265 of 16,330
Thank you very much for the response!
Today I finished sanding the surface of the molds, I dipped them in wax and at this moment the silicone mold is drying so tomorrow I will do the first curing test of the S IO.
With those exposure times that you have mentioned, I get an idea and I will do the test that you mentioned before starting, placing a little and seeing how it cures.
If I have not understood correctly, the glycerin bath consists of submerging the shell in glycerin, in a glass for molds, for example, and curing it for a few minutes, is that correct?
I keep reading on the forum, but often the concepts are a bit confusing.
Thank you very much again!
 
Jan 21, 2025 at 7:40 AM Post #16,266 of 16,330
Hello everyone,
I finally got the desired thickness of the shells. After several tests, 1 minute of exposure was sufficient.
But reading and reading, many of you insist that the ear mold has to be done with the mouth open and I remember that when the audiologist made mine he told me to open and close my mouth several times!! so I'm almost sure that the mold hasn't turned out well. So this morning I had the molds done again.
My insecurity is about now is how to cut the end of the canal.
I've seen like a thousand images and videos and my channel looks very different. Could you please tell me how I can do it?
I feel very unsure how to cut it and I don't want to have to make the molds again.
It seems like I'm back to square one, but I've really learned a lot in the process I've carried out so far. The next steps will be much faster later, but now I'm stuck on this.
I attach some photos.
IMG20250121131611.jpg

IMG20250121131506.jpg
IMG20250121131518.jpg
IMG20250121131533.jpg
IMG20250121131547.jpg
IMG20250121131600.jpg
 

Attachments

  • IMG20250121131611.jpg
    IMG20250121131611.jpg
    3.5 MB · Views: 0
Jan 23, 2025 at 6:42 PM Post #16,267 of 16,330
Hello everyone,
I finally got the desired thickness of the shells. After several tests, 1 minute of exposure was sufficient.
But reading and reading, many of you insist that the ear mold has to be done with the mouth open and I remember that when the audiologist made mine he told me to open and close my mouth several times!! so I'm almost sure that the mold hasn't turned out well. So this morning I had the molds done again.
My insecurity is about now is how to cut the end of the canal.
I've seen like a thousand images and videos and my channel looks very different. Could you please tell me how I can do it?
I feel very unsure how to cut it and I don't want to have to make the molds again.
It seems like I'm back to square one, but I've really learned a lot in the process I've carried out so far. The next steps will be much faster later, but now I'm stuck on this.
I attach some photos.
IMG20250121131611.jpg
IMG20250121131506.jpgIMG20250121131518.jpgIMG20250121131533.jpgIMG20250121131547.jpgIMG20250121131600.jpg
Most of us will make a permanent mold of the impressions. Then, fill that mold with impression silicone. Lastly, use the duplicated impressions for all modification saving the originals.
 
Feb 22, 2025 at 6:05 AM Post #16,269 of 16,330
Эй... У меня проблема с извлечением баса... Я собрал установку, как на картинке... Но вместо 2389 я использовал twfk30017 (rev pol) с 22uf Cross и 38d1xj для 38aj... Кривая выглядит неплохо, но есть слышимые и измеримые добавки около 3% на 2-м вреде... Что я могу сделать????
Hello! I also assembled headphones with the same drivers as in this picture, and was very upset. I don’t understand how you can listen to this at all, it’s not that there’s no bass, there’s no bass at all! Only mid and high frequencies play, I don’t understand how people like it. I watched a video on YouTube, there was a review of several headphones from famous brands. And all headphones had at least one dynamic driver, and I think this is correct. The armature driver cannot produce normal beautiful bass, and now I don’t know which good dynamic driver I should choose.
 
Last edited:
Feb 22, 2025 at 6:07 PM Post #16,270 of 16,330
Hello! I also assembled headphones with the same drivers as in this picture, and was very upset. I don’t understand how you can listen to this at all, it’s not that there’s no bass, there’s no bass at all! Only mid and high frequencies play, I don’t understand how people like it. I watched a video on YouTube, there was a review of several headphones from famous brands. And all headphones had at least one dynamic driver, and I think this is correct. The armature driver cannot produce normal beautiful bass, and now I don’t know which good dynamic driver I should choose.
You can get good bass from this....try removing r1 and r2 ( you can lower the values as well) to increase the bass. Also check that your tube and shell are sealed completely, as any opening will kill the bass from that 38...
 
Feb 23, 2025 at 11:51 AM Post #16,272 of 16,330
Hello! I also assembled headphones with the same drivers as in this picture, and was very upset. I don’t understand how you can listen to this at all, it’s not that there’s no bass, there’s no bass at all! Only mid and high frequencies play, I don’t understand how people like it. I watched a video on YouTube, there was a review of several headphones from famous brands. And all headphones had at least one dynamic driver, and I think this is correct. The armature driver cannot produce normal beautiful bass, and now I don’t know which good dynamic driver I should choose.
If you assembled it right it should sound very good. Balanced armatures can produce very good bass. If you have absolutely no bass at all, then something went wrong. You should 1) check if you wired everything correctly, and the components you used are of the correct values; and 2) check if your tubes are sealed correctly (probably a more common mistake) - when you put on the iems do you feel there is a proper seal in your ears or do you feel a leak and can hear the background noise?
 
Feb 27, 2025 at 3:45 PM Post #16,273 of 16,330
Вы можете получить хороший бас из этого... Форма уберите r1 и r2 (вы также можете уменьшить значения), чтобы увеличить бас. Также, что ваша трубка и корпус полностью запечатаны, так как любое отверстие убьет бас, проверьте это 38...
Корпус моих наушников не был запечатан, крышку корпуса я не приклеил. Не думал, что корпус может быть главным в этом случае, так как арматурный динамик и так является корпусом, похоже, я ошибся, попробую приклеить крышку, а потом послушаю, что получится. Сразу ли мне снять R1 и R2? Или поэкспериментировать позже? Может, мне поставить 2 динамических динамика huayunxin CN10? Подскажите пожалуйста как мне выйти из ситуации, у меня слуховой проход маленький и три трубки диаметром 2 мм не помещаются в корпус наушников. Пришлось часть корпуса стачивать, чтобы в отверстие поместились три трубки, правильно ли я сделал? Что делать в такой ситуации?
 

Attachments

  • 1.4.jpg
    1.4.jpg
    113.6 KB · Views: 0
  • 1.5.jpg
    1.5.jpg
    86.4 KB · Views: 0
  • 1.2.jpg
    1.2.jpg
    87.7 KB · Views: 0
  • 1.3.jpg
    1.3.jpg
    88 KB · Views: 0
  • 1.1.jpg
    1.1.jpg
    79.7 KB · Views: 0
Last edited:
Feb 27, 2025 at 9:04 PM Post #16,274 of 16,330
Hello! I also assembled headphones with the same drivers as in this picture, and was very upset. I don’t understand how you can listen to this at all, it’s not that there’s no bass, there’s no bass at all! Only mid and high frequencies play, I don’t understand how people like it. I watched a video on YouTube, there was a review of several headphones from famous brands. And all headphones had at least one dynamic driver, and I think this is correct. The armature driver cannot produce normal beautiful bass, and now I don’t know which good dynamic driver I should choose.
If you are not getting a bass signal from the sub armature its more than likely an issue with your fit. If your canal is not completely occluded, there will be zero bass. Either that, or the sub armature was over heated during soldering.

In any case, with the proper low driver, you can FOR SURE get just as much low frequency as a dynamic.
 
Feb 28, 2025 at 2:26 PM Post #16,275 of 16,330
Hi everyone, reading this thread made me think that some of you might not know how to properly make a crossover so here it is.

In making a RC (resistor-cap) crossover circuit you should know first how every components work. Every positive and negative connection or complete loop to a driver is called a circuit. Connecting a resistor parallel with the driver, decreases the resistance of the circuit. While connecting it in series increases the resistance. Meanwhile, the capacitor determines whether the crossover will be a low-pass (only allows frequency below the cut-off frequency) or a high-pass filter (only allows frequency above the cut-off frequency). Frequency cut-off (fc) is where the filter starts to drastically curve down the frequency sent to the driver. You should connect the capacitor in series with the driver if you want it to be a high-pass filter, or connect it in parallel if you want it to be a low-pass filter.

Now in making a crossover circuit, you should determine which frequency should you cut each of the driver in your iem. This should be the same or close to the cut-off frequency (fc) of the next driver. For example, a low-pass @220Hz for the subwoofer and a high-pass @200Hz for the midrange driver. It's better for those two fc to overlap than to have a gap. That's because of the phase shift that's happening the more you get far from the fc. I won't dwelve to much with that but that's just the way it is.

Now, this frequency cut-off for an RC filter has a formula, Fc=1/(2πRC). R being the overall resistance in a loop/circuit including the impedance of the driver at the cut-off frequency you wanted (see the graph in datasheet, you should also consider if there is too much fluctuation at your desired frequency cause this will affect the frequency roll-off of the circuit), and C as the capacitance of the capacitor you need to use. You need to keep your resistor lower if you want your iem to be sensitive rather than to make it a power hungry beast. That means you should mostly adjust the capacitance than to resort on using higher resistor.

The formula in finding the right value of your capacitor is C=1/(2π x fc xR). The numbers x10-⁶ or 10^(-6) in the result means you need a capacitor in microfarad (uf) and x10-⁷ or 10^(7) if you need it in hundreds of nanofarad(nf). But most of the time it would be in microfarads. Just use the nearest available capacitor in your kit or just add the capacitance of different capacitors by connecting them parallel to each other (not in series). Just make sure that the result would be close enough.

Now, the resistors in a crossover circuit are primarily used to formulate your desired cut-off frequency. If you want to decrease the volume of the driver, either you use a higher resistor (connected to the driver in series) or you use an l-pad.

L-pad is used to decrease the volume of the driver while maintaining the resistance R across the circuit. It uses both a parallel and series resistor. Here are the formulas and an example from chatgpt:
1000005278.jpg

1000005279.jpg

Remember, the resistance R in the formula fc=1/(2πRC) is the overall impedance. With a series resistor, the R is the sum of impedance plus the resistor.

For using a parallel resistor, the formula is R=(Id x Rr)/(Id + Rr). Id being the impedance of the driver @ the desired cut-off frequency while Rr is the resistance of the resistor. With a resistor in parallel, you cannot just add the resistance together like in a series connection. When you add a resistor in parallel, you add another path for the current to go. It's like opening another gate. Thus, it decreases the overall resistance and also lowers the power received by the driver making it sound a little quieter.

Though resistor in parallel is commonly used in high-pass filter and has the benefit to stabilize the impedance in the circuit, it's not always necessary. You can just use it to lower the overall resistance R if the impedance of the driver is high enough for you to achieve your desired cut-off frequency and use a resistor in series if the resistance is too low. Just add it after the capacitor.

But still resistors are not specifically used to decrease the volume of the driver. It will only work like that if you don't have a capacitor connected, thus it is not a crossover circuit but merely an attenuator. But of course, the general rule of thumb is that using higher resistors will still lower the volume of any driver may it be in a crossover circuit or not. That's why you can see some manufactured iems use a large tantalum capacitors in their iems. They're keeping the resistance lower and just using higher capacitors to compensate.

"But how about if the impedance of the driver itself is already good enough for the equation?"

Good question, if the impedance (Ω) and volume of the driver itself is good enough, then resistors are not necessary anymore. Meaning, you can have both high-pass and low-pass filters without any resistors. Resistors are just used to alter the computation. So, you would only need a capacitor to be connected either in parallel or in series with the driver. Except of course if you want to lower the volume of the driver. Then you would use an l-pad to keep everything in place.

Here are some examples of my own crossover circuits:
1000005170.jpg

3.3 Ω resistor in parallel and a 10uf in series for a high-pass filter to be used in a tweeter. The impedance of the driver at that frequency is 7Ω. The cut-off frequency is close enough to 7kHz. I just did trial and error from my available resistors (2.2Ω and 1.1Ω in series to produce 3.3Ω).


1000005169.jpg

10Ω resistor in series with the driver with a impedance of 27 at 200Hz. I used a 20uf capacitor to be used as a low-pass filter for the subwoofer.

1000005168.jpg

4.7Ω resistor in parallel with a midrange driver with a 15Ω impedance at 200Hz. I used a 220uf tantalum capacitor to be used as high-pass filter. I used a large capacitor to retain a low resistance across the circuit. But here, as an example, we can just use a 50uf capacitor without any resistor connected to still produce a cut-off frequency of around 200Hz.
1000005286.jpg


Hope this helps, if there were anything wrong on what I said just feel free to correct me. For those who haven't still learn how to use crossover circuits, use this technique together with other techniques like tube length and diameter to see further improvements in your frequency graphs. Lastly, if you want to make a dip in your frequency, just search the formula in getting the tube length and diameter of the tubes to be used. I know there is, I have read that. Just use some AI like chatgpt to search for it. I know crossover circuits would take up space, but still if you can fit it, it would be better. Hopes for your improvement. That's all thank you!
 
Last edited:

Users who are viewing this thread

Back
Top