Help with new capsule project

[And]

Hello everyone,
I’ve designed a large diaphragm condenser microphone capsule, but I’m encountering a few issues with the first assembled prototypes. Does anyone have any advice or suggestions?

The capsule is edge-terminated with a large diaphragm. It works, but it lacks high frequencies, making the sound quite muffled. I’ve tested it with both transformer-based electronics (similar to the KM84 style) and transformerless electronics.
Based on calculations (considering the area and distance between the diaphragm and backplate), I was expecting a capacitance of 100pF. However, when I measure the assembled pieces (only 3 so far), I’m getting 300pF, 350pF, and 390pF.
I suspect the lack of high frequencies is due to the high capacitance (the cutoff frequency is too low). I’m trying to understand why the capacitance is so high. It could either be due to parasitic capacitance or because the distances vary from the design during assembly.

Here’s a bit more detail to help anyone with experience on this chime in with their opinions...
My capsule, though different, is built similarly to a CK12. The brass backplate is housed inside a plastic support, and the outer ring that holds the diaphragm is screwed onto the plastic support. I’ll attach a couple of sectional images showing the construction (indicative drawing, not to scale).
The diaphragm is 5µm mylar with an aluminum coating, sourced from a supplier I’ve seen mentioned here on the forum.
I used 50µm kapton as the diaphragm spacer. The plastic material for the prototype is 3D printed PA12 nylon (MJF), and the brass backplate adheres to the design tolerances.
For the electrical connection to the backplate, there’s a small extension that protrudes on one side (as illustrated in the second image).

Could it be that the materials used (the PA12 plastic support or the kapton) don’t provide adequate electrical insulation, leading to parasitic capacitance? Or do you think parasitic capacitance could be forming elsewhere?
Another possibility I’m considering is that during assembly, the kapton spacer might be getting compressed, so instead of 50µm, the actual distance is less.
Please note, as shown in the first image, the plastic support is slightly thinner than the backplate (to avoid plastic tolerance issues), but there’s only a gap of about 0.2mm. I did this so the distances would rely only on the brass parts, where I have better control and precision.

Thanks in advance to anyone who can offer some insights or advice!

 

[kingkorg]

This part is causing excessive capacitance. This is not the source of the HF roll off however. You haven't posted the details about your capsule which determines the amount of HF content. We also don't know how far are the screws from the metalized part of the diaphragm, and what is the size of the diaphragm/backplate area if you don't post hole and backplate dimensions.

 

[And]

This part is causing excessive capacitance. This is not the source of the HF roll off however. You haven't posted the details about your capsule which determines the amount of HF content. We also don't know how far are the screws from the metalized part of the diaphragm, and what is the size of the diaphragm/backplate area if you don't post hole and backplate dimensions.

Thank you very much, Kingkorg!
I apologize for the missing details; I'm trying to provide you with more dimensional guidance in the attached image. I’m sorry if the images aren’t precise; I hope they are still clear enough.
If it’s not clear, the backplate has a feature: there’s a protrusion to facilitate the electrical connection. Usually, you enter with the screw; I did it this way so that it can be connected without side holes. From a production perspective, it would simplify things, but if this is causing my problems, I will definitely revise the design (I would like to understand if this is the issue).

Regarding the HF roll-off, I haven't taken measurements or made graphs, but it’s already clear "by ear."

At this stage, I would like to adjust the capacitance value, and if the HF roll-off is still present, I will revisit other aspects, such as air paths and the perforations in the backplate.

Thank you very much!

 

[kingkorg]

Thank you very much, Kingkorg!
I apologize for the missing details; I'm trying to provide you with more dimensional guidance in the attached image. I’m sorry if the images aren’t precise; I hope they are still clear enough.
If it’s not clear, the backplate has a feature: there’s a protrusion to facilitate the electrical connection. Usually, you enter with the screw; I did it this way so that it can be connected without side holes. From a production perspective, it would simplify things, but if this is causing my problems, I will definitely revise the design (I would like to understand if this is the issue).

Regarding the HF roll-off, I haven't taken measurements or made graphs, but it’s already clear "by ear."

At this stage, I would like to adjust the capacitance value, and if the HF roll-off is still present, I will revisit other aspects, such as air paths and the perforations in the backplate.

Thank you very much!

Ok, i get a bit more of the picture now. That diaphragm connection is not causing that much capacitance increase. I thought it goes all the way around.

In general capsule follows the same rules that apply to capacitor design. The difference is that you want close proximity of the electrodes just where the diaphragm can move. That is why k47 and k67 have just the dot metalized. You could go with fully metalized diaphragm, but you get too much parasitic capacitance under the tension rings.

High end is controlled by the air behind the diaphragm, pattern control delay network, diaphragm tehickness, and overall diameter of the capsule.

 

[pasarski]

My non-expert tl;dr: I think you need to move to the hard part, holes/chambers/other means to achieve damping and acoustic resistance.

EDIT: Super cool you're doing this and kudos for getting this far! Excited to witness how this ends up.

 

[And]

Ok, i get a bit more of the picture now. That diaphragm connection is not causing that much capacitance increase. I thought it goes all the way around.

In general capsule follows the same rules that apply to capacitor design. The difference is that you want close proximity of the electrodes just where the diaphragm can move. That is why k47 and k67 have just the dot metalized. You could go with fully metalized diaphragm, but you get too much parasitic capacitance under the tension rings.

High end is controlled by the air behind the diaphragm, pattern control delay network, diaphragm tehickness, and overall diameter of the capsule.

Could I try to increase the distance of the diaphragm? For example, by doubling the Kapton spacer or using one that is 50% thicker. But I'm afraid that if there are parasitic capacitances, they would then be even greater than the capacitance of the capsule...

I still don't understand whether I'm dealing with parasitic capacitances or not... Is there a way to find out? That is, when I measure the capacitance of the capsule (I'm using an LCR bridge, not a tester), if there were no parasitic capacitances, should I expect to read a value that keeps changing (due to surrounding noise)?

I should also add that aside from the loss of high frequencies, the capsule sounds good; the bass is deep, and the volume is consistent with other commercial capsules. If I had issues with parasitic capacitance, do you think I would have problems in general, even with the bass or the volume?

Thank you again.

 

[And]

My non-expert tl;dr: I think you need to move to the hard part, holes/chambers/other means to achieve damping and acoustic resistance.

EDIT: Super cool you're doing this and kudos for getting this far! Excited to witness how this ends up.

Thank you for your interest! I'm doing a lot of work on the backplate for this very reason... it's exciting but also very challenging! In fact, now that I'm starting to do the first tests, I'm also facing some initial frustrations. It's inevitable.

I was asking about parasitic capacitances (which are a bit difficult for me to predict) to understand whether it's an aspect I can exclude and focus on the backplate, or if I need to first address the capacitive value issue...

 

[kingkorg]

Do you have any holes in the backplate, and what is the overall diameter of the capsule?

If you increase the distance, you will lose signal level.

 

[And]

Do you have any holes in the backplate, and what is the overall diameter of the capsule?

If you increase the distance, you will lose signal level.

The overall size of the capsule is indicated in image 3; it is 36mm (active area 28mm).

Regarding the holes, keep in mind that the backplate is composed of multiple carefully matched parts. The first layer (under the diaphragm) has 124 holes, and the subsequent layers have different hole patterns to create damping and acoustic resistance. I will definitely need to adjust this last part (and it could be the real cause of the HF loss). But as I mentioned earlier, I also wanted to understand why I'm measuring such a high capacitive value...

Thank you, you are very kind.

 

[soliloqueen]

This is definitely very very wrong for the construction but the construction itself seems solid from your drawings. Either something unexpected is going wrong or your method for measuring the capacitance is wrong. Are you measuring charged or uncharged?

 

[And]

This is definitely very very wrong for the construction but the construction itself seems solid from your drawings. Either something unexpected is going wrong or your method for measuring the capacitance is wrong. Are you measuring charged or uncharged?

Thank you for stepping in! But sorry, I didn’t understand your response... What exactly is wrong? I’d be happy if you could explain it better.

As for the measurement, I’m doing it with an uncharged capacitor (not connected to anything) using an LCR bridge (not a multime

 

[soliloqueen]

Thank you for stepping in! But sorry, I didn’t understand your response... What exactly is wrong? I’d be happy if you could explain it better.

As for the measurement, I’m doing it with an uncharged capacitor (not connected to anything) using an LCR bridge (not a multime

The measured capacitance is wrong for the design. How flat is the plate?

 

[And]

The measured capacitance is wrong for the design. How flat is the plate?

The backplate is accurate, measured with a micrometer... But I’m not sure I understand what you mean...

 

[soliloqueen]

The backplate is accurate, measured with a micrometer... But I’m not sure I understand what you mean...

how flat is it. There's a critical flatness to both the plate and the diaphragm's suspension off the plate that need to be met

 

[Tim Campbell]

Have you tried reading The AIP Handbook Of Condenser Microphones? It would answer a lot of your basic problems that it seems you must solve first or at least do a search of this forum by your basic questions. All of these questions have been answered here before. It would help greatly if you actually show us your design.

 

[pmfalcman]

And - How precise is the plastic support/spacer? 3D printing is not so perfect as to deliver the same height in each and every point. I'd go for precise turning on a lathe. Teflon, for example.
Second possible issue - as far as I know the nylon (at least the nylon used for 3D printing) is hygroscopic and the filament must be kept in a dry container (with dessicants).
Absorbed humidity may cause problems like charge leakage.

 

[molke]

If you start over, you could try reducing the (inner) backplate diameter. Your drawings seem to indicate that the metal backplate and the metal clamping ring overlap (with just the spacer distance in between—even if it’s just 0.2mm radially). Under idealized conditions (in particular, no holes), you could go as low as 81.6% of the membrane diameter (about 23 mm in your case) since the diaphragm movement decreases closer to the clamping ring. The Debenham capsule (strongly recommended design for starting out) has about 94%.

A plastic support straight off the printer most certainly won’t work, but I don’t see why it wouldn’t with some grinding. For isolating material, PA12 seems to better than PA6 or PA66 regarding water absorption. Teflon is probably almost impossible to glue to anything.