22mm capsule

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Yeah…that thread and the banzai build thread are entertaining.
SDC capsules is a rich subject, and so unsexy compared to a U47 we are kind of stuck with mediocre options.
I dream of someone designing a honeycomb backplane for DIYers. The shell, spacers, and diaphragm of these ali express capsules seem a fine starting point.
Warts and all, mic-parts is still the flattest one I know about.
I am looking forward to trying out some of the gzt capsules in the next month. I hope they’re cool. Bbecause GZT sells both depths, the Banzai 84 users can get some variety too.
There was somebody on eBay selling adapters that made the MK-012 capsules work with Rode NT5 bodies. If more of those get made it could be an option. Found it: Adapter Rode NT5 -> Octava MK-012 - Use All Octava Capsules with Rode Body | eBay

Some of the stuff I see from Mic-Parts is just nuts. Now that the Chinese "CK12s" are available for $30-50 each in single order batches this statement was posted on the site:
Screenshot_20230116_215332_Chrome.jpg
Funny thing is that boost sounds almost like a description of the RK-47.
 
Last edited:
Some of the stuff I see from Mic-Parts is just nuts. Now that the Chinese "CK12s" are available for $30-50 each in single order batches this statement was posted on the site:
He is loosing ground, and tries to squeeze the last drop out of his stale business model.

He has done some pretty remarkable work when it comes to cataloging the mic/capsule info. But i can't see how he couldn't predict this happening.

It is also beyond me why people, while they are still on the run, don't develop some unique stuff. He had the knowledge and contacts, and the best he could do is exploit cheap capsules and mics with built in eq?

What he's saying is an utter lie. That capsule design originates from China, and had been used by Rode long before he started selling his fairy dust sprinkled version. RK-12 existed in his Mic Parts shop long before he started the Roswell mic company. He used the original name which comes from RayKing. Hence RK-12.

In the end none of these have anything to do with CK12, so it's all just BS.
 
Last edited:
He is loosing ground, and tries to squeeze the last drop out of his stale business model.

He has done some pretty remarkable work when it comes to cataloging the mic/capsule info. But i can't see how he couldn't predict this happening.

It is also beyond me why people, while they are still on the run, don't develop some unique stuff. He had the knowledge and contacts, and the best he could do is exploit cheap capsules and mics with built in eq?

What he's saying is an utter lie. That capsule design originates from China, and had been used by Rode long before he started selling his fairy dust sprinkled version. RK-12 existed in his Mic Parts shop long before he started the Roswell mic company. He used the original name which comes from RayKing. Hence RK-12.

In the end none of these have anything to do with CK12, so it's all just BS.
What's funny is the Roswell Pro Audio site even says that that company came after Microphone-Parts. So by his logic on the RK-12 page he's claiming to have named his capsule after a company that he made years later.
 
Ok, kingkorghave a look at these measurements and let me know if I should continue with the rest of the capsules.
I added the 2 Takstar measurements with the same conditions, and overwrote the file on google drive. Here is the new link.
I appreciate your insight


-edited out of order by a dunce
 
Last edited:
@Chrisfromthepast please use @ in front my name so i get notification. Just like I did here, i saw the post by pure chance. Can you please do the measurements of the two stock Takstar cm60 so i'm sure you are on the right track. Purely because i have those and i know what they should look like. These look better for sure. There is still some timing difference between the measurements. In Make measurement page set timing to - no timing reference, set t0 at ir peak. But do do the Takstar measurements the way you did these, don't change anything yet. It seems to be ok.

Your measurements confirm my suspicions, that MP capsule is produced by 3u audio, as the seem to be almost identical. Also that there indeed is a dip before boost in MP capsule, so it isn't flat. I've never encountered a flat capsule with that backplate, it is pretty much impossible.
 
Last edited:
He is loosing ground, and tries to squeeze the last drop out of his stale business model.

He has done some pretty remarkable work when it comes to cataloging the mic/capsule info. But i can't see how he couldn't predict this happening.

It is also beyond me why people, while they are still on the run, don't develop some unique stuff. He had the knowledge and contacts, and the best he could do is exploit cheap capsules and mics with built in eq?

What he's saying is an utter lie. That capsule design originates from China, and had been used by Rode long before he started selling his fairy dust sprinkled version. RK-12 existed in his Mic Parts shop long before he started the Roswell mic company. He used the original name which comes from RayKing. Hence RK-12.

In the end none of these have anything to do with CK12, so it's all just BS.
he started roswell in 2013 and the rk12 didn't pop up in his web store until a bit later. not saying they're not just stock AYM capsules (they are, and they cost about $20 at volume), but he's probably not lying about how he came up with the R designation.
 
Last night was my window of time to measure, so I went ahead and took a bunch of traces.
@kingkorg, here is the link

@Chrisfromthepast has done great job taking these measurements. The Audix cal mic works great as a reference mic, so you can view these as accurate absolute frequency response curves of these mics. I have some of these capsules and they match my measurements. Also dips and bumps are consistent with typical deviations of these when construction is sub optimal.

My conclusion. 3U audio and MP capsules are of the same origin. They are not flat compared to something like CM4. But they are flatter than most capsules of this type. Banzai capsules provided with those km84 kit are pure rubbish, inconsistent, and hove nothing in common with km84. Waste of time and money.

 
Last edited:
he started roswell in 2013 and the rk12 didn't pop up in his web store until a bit later. not saying they're not just stock AYM capsules (they are, and they cost about $20 at volume), but he's probably not lying about how he came up with the R designation.
So is Rayking just rebranding the AYM capsules?
 
@Chrisfromthepast has done great job taking these measurements. The Audix cal mic works great as a reference mic, so you can view these as accurate absolute frequency response curves of these mics. I have some of these capsules and they match my measurements. Also dips and bumps are consistent with typical deviations of these when construction is sub optimal.

My conclusion. 3U audio and MP capsules are of the same origin. They are not flat compared to something like CM4. But they are flatter than most capsules of this type. Banzai capsules provided with those km84 kit are pure rubbish, inconsistent, and hove nothing in common with km84. Waste of time and money.


Thanks for making these!
 
Ok, back to business.
to recap:
the Banzai capsules have 1 good match, one dark capsule, and one bright one and they all share a resemblance in terms of the shape.banzai 84.png

The micparts capsules all purchased separately are very close to one another, and vary from perfection by a small amount in roughly the same general shape as the others., but they are much better than a stock capsule from mxl. the pink reference capsule is from a monoprice 60700, which has the same MXL body and pcb as a 603s.micpartsCom.png
My first impressions of the 3U capsules are very good. they feel good, threads are quality, just like the micparts and Banzai capsules. The matching is pretty good. the shape is that familiar dip before rise found in so many 22mm chinese SDC mics.
The cm60's are Takstar CM60 mics which cost less, but perform pretty well. Notice this curve is closer to the nady CM90 we will see later.

3u and cm60.png
The matching, and the shape of the matched pair of 3U omni capsules is quite good. I haven't used them in any serious way yet, but Im hopeful about these. The purple trace is barely even visible under the orange one. omni capsules.png

This brings us to the reason for this quest in the first place - the dreaded 603s capsule.
Here I have a couple stock ones, and a couple with face flattened on a grinder, and one with a fancy restretched diaphragm. Notice that nady has a different shape from the rest of the crop. curious.603s.png
Speaking of curious, I bought some 22mm capsules from ali express, and they are out of phase with the rest of what we have tested. the red one is the same capsule body, with the 603s guts. interesting change in top end shape due to the tapered opening I suppose. that tapered opening is not enough to pull high end out of a bad backplate/chamber arrangement.ali.pngali phase.png
 
On the left is the 603s backplate and air chamber combination. this is separated from the diaphragm, which is grounded by the shell of the capsule, by a thin washer of mylar.
On the bottom left is the backplate for the AliExpress capsules, which sits in a keyed position on the other side of that plastic structure, so that the air pocket between the diaphragm is a different air chamber for the delay network. positive pressure should still create positive voltage, right? Does that mean that the polarity is the same, but the 2 chamber arrangement is 180 degrees late? Does having a completely separate delay network give the DIYer more tweaking options?
I posted an exploded view of the Ali Express capsule in an earlier post on this thread
 

Attachments

  • backplates.jpg
    backplates.jpg
    3.3 MB
On the left is the 603s backplate and air chamber combination. this is separated from the diaphragm, which is grounded by the shell of the capsule, by a thin washer of mylar.
On the bottom left is the backplate for the AliExpress capsules, which sits in a keyed position on the other side of that plastic structure, so that the air pocket between the diaphragm is a different air chamber for the delay network. positive pressure should still create positive voltage, right? Does that mean that the polarity is the same, but the 2 chamber arrangement is 180 degrees late? Does having a completely separate delay network give the DIYer more tweaking options?
I posted an exploded view of the Ali Express capsule in an earlier post on this thread
Btw those Aliexpress capsules are actually electrets! Applying additional polarization won't hurt them, but in some cases depending on how they are prepolarized the additional voltage might fight that, and play tricks with capsule output.

Not sure how they are prepolarized but it is possible someon flipped something in assembly, so you get the phase shift.

PRO TIP SECRET! THIS IS IMPORTANT FOR ANYONE WILLING TO READ AND MAKE SOME EFFORT!

In case you didn't know, you can use those absolute measurements i posted here, export them in TXT format, open in text editor, get rid of the phase column and unnecessary info, rename file extension to CSV and load in Voxengo CurveEQ. Invert the response and you get correction curve to get those mics flat! Before exporting make sure you adjust the level of the measurements to 0db at 1K, otherwise you'll just confuse the CurveEQ. There is trial version of the plug in. 😉😉😉
 
Ok, electret explains quite a bit.
I am going to be bothered by some my misnomers and will have to clean up some language throughout this thread.
I think I can try it in the proper circuit in a few days just to see whats up, and if we can make some useful mics out of them.

The curve EQ correction eq with a csv file sounds very handy, thanks for the tip!

I stopped at 30 traces, but there are more capsules, more experiments, 90 and 180 etc. What process did you use subtract the Audix reference mic from each mic trace to get the absolute measurement?
 
You simply open a measurement you are interested in, and audix measurement. In trace arithmetic (find that function) you divide the fr of the first mic by audix.

There are sevaral more steps i used to bring the cm60 into the equation since i have it to be sure the measurements are correct, and then use my own cm60 measurement to get those smooth lines. But that's tricky, and some of it i keep as proprietary part of the process, sorry for that.

But that is generally how you do it, you divide the DUT mic by reference mic, and that gets rid of the impact of the speaker/room in between. The reference mic has to be flat, or have calibration file which again corrects the deviations in that mic. Your audix seems to be flat without cal file.

You can smooth out the curves with smoothing fuction to make them pretty, but that can be decieving, as it can hide issues.
 
In case you didn't know, you can use those absolute measurements i posted here, export them in TXT format, open in text editor, get rid of the phase column and unnecessary info, rename file extension to CSV and load in Voxengo CurveEQ. Invert the response and you get correction curve to get those mics flat! Before exporting make sure you adjust the level of the measurements to 0db at 1K, otherwise you'll just confuse the CurveEQ. There is trial version of the plug in. 😉😉😉
Wouldn't editing the CSV in Excel / Libreoffice Calc make it easier to remove the phase column (skipping the text file stage altogether? As opposed to deleting the end of each row (out of tens if not hundreds) 😁
 
Wouldn't editing the CSV in Excel / Libreoffice Calc make it easier to remove the phase column (skipping the text file stage altogether? As opposed to deleting the end of each row (out of tens if not hundreds) 😁
Sure, it's just that REW exports them in TXT, but use whatever you like. And use 1/12 smoothing at export from REW, there's limited number of points CurveEQ can load.
 
I purchased what appears to be a k40 from a dubious ebay sale.
It had a little bit of foam that needed to be cleaned out, and a little evidence of impact to the vents, but it seems to operate just fine.
I think this test might reenforce KingKorg's assessment that the conical chamber behind the capsule is a big part of the problem with this form factor. I wasn't expecting to see the same familiar curve with this capsule.

IMG_4531 2.jpgIMG_4531.jpgk40.png
I put this capsule onto 2 flavors of Banzai 84 with 2 different transformers, and the Takstar for reference.
The GZT transformer performs well.
I want to point out that NTE 10:3 SEEMS to tame the bright capsules, but at a considerable cost.
the THD measurement is very high.
This is my first time dividing one trace from another, so here is the raw data, in case anyone wants to tell me how to improve these traces. here is the drive:
 
What's behind any capsule at close proximity is key part of the response. No way around it. It is as important as the front of the capsule. No conspiracy theory there, just common sense.

To account for this Neumann does this to newer capsules. There's fixed plate separating the capsule from the rest of the body, and vents are in between the backplate and the plate.

A tiny bit of conspiracy theory: I'll repeat my findings. No km84 i've ever tested had flat response. At this point i suspect they "faked" the claim of flat response. Or, at the time, this could be viewed as flat response since there weren't many truly flat cardioids.
 

Attachments

  • Screenshot_20230130_192446.jpg
    Screenshot_20230130_192446.jpg
    283.8 KB
Back
Top