[Sticky] USB-C Microphone (official topic)

I usually drop in once a day or so, time permitting, don't worry about it.

If you haven't butchered the cables (but we all do that...) you can test it by plugging it into your PC audio (headphone) port. You could always bodge together a hookup with some of those god-awful RCA jokes of a plug.

What I can tell you with almost 100% certainty, if your THAT1512 was swinging much above 1V peak, that digitiser will have shuffled off to silicone heaven.

The new designs (I've just finished testing the main sections ) but that still have some experimental stuff to do, have a limiter on the auxillary outputs for digitisers and other equipment such as an Arduino level meter, are hard clipped as described above. I've only got the pro-level audio out (which is an option) to left to test but I'm so behind on other deceased Pathera Leos (dead lions, geddit -- it's a vet joke) that I've been putting it off. 😉

Matt should get his sample boards this week coming (I'll only be making small corrections now) to both parts, but the performance is pretty good and there's less to go wrong since all the tricky stuff has been done at the factory.

In the video he powers the pre amp from the usb c, how would I power it and keep usb A

Welcome @Dog23121.

You have a couple of options here. If you're using USB A (which I assume you mean the older USB 1.x - 2.x standard which uses four wires) you can take power from the red and black wires in the USB cable and patch them to the board. It's messy but not a huge issue.

You can even use a second USB cable, I keep busted charger cables for this very reason. One powers my boards and the other (in this case, the digitizer) plugs into the PC.

Another option is to use an alternative "breakout" - Sparkfun and Pololu have appropriate boards if you want to go that way.

Just drop a search for "USB breakout" and you'll see plenty of alternatives like this:

https://thepihut.com/products/dual-usb-breakout

or

https://www.ebay.co.uk/itm/263886486035

Matt will be receiving his V2 boards in the next couple of weeks, that's been quite a ride but I believe it's a worthy successor to the original design. I've added more (optional) functions and some necessary improvements that have come about as part of the discussions with members. They are available now, but I advise that people hang around for Matt to do the video because he will explain the multiple options visually far better than I ever could. It does use recoverable components from the V1 design (by design) however. Also it's only fair Matt gets to see it first since it was his idea.

One of the options for those worrying it's going to cost a fortune, you can breathe easy. It will cost a more than a stripboard version but (among other things) it can be powered externally from a regulated 5V "wall wart" such as a phone charger. I've reached out to a couple of suppliers but it's still open for someone to order some in from JLC (you get a minimum of five boards, OR three boards, two made up). It's not massively expensive to get five or more made up with all the fiddly components - plenty of them - which can be sold on. Someone will probably run with that ball, I'm not geared up to ship them commercially myself. The actual cost is largely determined by shipping (slower = cheaper) and taxes.

My recommendation for these new designs is to leave the digitiser out of the build (I have another one where you absolutely have to but that's under development at the moment). So for this you use an audio connector (like an RCA female) and a normal patch cable so you don't have to butcher the cables. This other one is an alternative which uses a BM800 donor body and improves it dramatically. Maybe Matt will do a build on that? I don't know, I just do this to keep my hand in.

With some RCA females you can wire the audio from the THAT1512 to the digitizer without cutting the wires off. During testing I discovered that running the original design at high gain without the capsule connected can cause the THAT to overload the digitiser - and destroy it. I would imagine a lot of the early ones blew this way and people just gave up. V2 has protection to prevent most of these.

Hello guys,

My native language is not English, and sometimes I have trouble understanding some of the answers on the forum!

I am self-taught in electronics, and there is a lot I don’t understand, so please don’t judge me harshly.

Please help me figure this out. I assembled the circuit according to my own understanding, as it seemed to me that there was a mistake in the polarity of the capacitors in the schematic from the author, while in the video the capacitors are soldered correctly!

After analyzing (to the best of my ability in both English and electronics) the forum, I see that some people have their capacitors soldered with different polarities than others, and I can’t figure out which option is correct.

Please help me understand! I will attach pictures below, and could you please tell me if the polarities are correct on certain components? And if I soldered everything correctly on my board?

Thank you very much for your answers!

Matt used NON-POLARISED electrolytic capacitors in the original design. While this isn't strictly necessary, it does make for a simpler build.

With "normal" electrolytic you put the positive (+Ve) side to the most positive part of the circuit - so I'm fairly sure you can understand these examples.

All of the voltages I'll discuss are measured relative to the GND (not GNDA) point.

For a more complex example, take a look at the input stage of the V2 design which, due to requests from the beta testers supports switchable "phantom" power. I've left the switching circuit off as it's not needed for the explanation.

First let''s consider the simpler of the two.

15V DC ienters this circuit at the points marked HPWr and CPWr. (Hot and Cold Power).

For clarity I've omitted the rest of this stage, but you can imagine that it's going to the input of the THAT1512.

This means that (reading from left to right) the capacitors C4 and C5 "see" the injected voltage. That means the end of the capacitor is more 15 volts more positive than the right-hand side. So we put the positive end that way around.

In the alternative configuration, per the original design, has +15V on the HPwr rail and -15V on the CPWr rail. Matt's simpler circuit puts 30V across the FET but let's see how that affects the capacitors.

The HPwr rail is at +15V as before, so no changes there. The CPwr rail is at -15V (relative to circuit ground) so in that case we have to flip C5 the other way around -OR- use a non-polarised electrolytic. Another way around this one is to use two polarised capacitors  in series but that means you have to use double the value in uF because capacitors in series "subtract".

So how do you know which way around?

Generally speaking an input, such as the input to an operational amplifier should not have any voltage. In practical terms bipolar input stages do leak a little bit so there is a very small voltage there, but it can be discounted. You can assume for all practical purposes that an INPUT is at or nearly at Gnd (0 volts) which means you put the "negative" side of any capacitor facing the input.

The OTHER side (the output of the THAT in this case) you use the alternative - you put the POSITIVE of the capacitor to the OUTPUT of the device.

There are cases where this isn't as obvious but it's usually in places where we're not using polarised capacitor because the impedances are so large. Recall that f-3dB = 1/ (2 * PI * R * C), we have to use a large capacitor for this input stage because the input impedance is low and we need to pass signals as low as 15Hz to be safe.

Designs like the one here often do away with an output capacitor altogether, but that's because we know that the "differential" output from such a circuit should not have any DC component. In practice it does have a very small offset but the device we're driving (an Audiograbber digitiser in this case) often has its own DC blocking and even if it doesn't, the tiny amount of offset (a few thousandths of a volt) isn't going to affect anything.

You CAN use a capacitor at the output but since the input impedance of the audiograbber will be 20K (at least) we can calculate a suitable value from the minium frequency (Fmin) of 15Hz  like this:

This is a HIGH-PASS filter - because we're blocking DC.

C = 1/ ((2 * PI) * R * Fmin)

=1/ (6 * R * Fmin)

=1/(6 * 20,000 * Fmin)

= 1/(6 * 20,000 * 15)

C = 556 nanofarads.

HOLD UP! I can almost hear the screams of 2*Pi isn't 6, Draco!

It's not but by the same token, we can't get a 556n capacitor, nor the 531 nF this works out at either!

Or you can take the easy route these days and drop a link like this:

https://www.digikey.co.uk/en/resources/conversion-calculators/conversion-calculator-low-pass-and-high-pass-filter

Electronics engineering is often compromises. The nearest suitable "preferred" value would be 470 nF but anything around that sort of value is fine 470n would have an f-3dB of 16 Hz. With a modern calculator you can get Pi to 10 digits (or more) but 2Pi = 6 is close enough if you're in a hurry or your calculator lacks the function.

That said this only works when the values (of R in this case) are quite large but experience will tell you when the result looks right.

P.S.

Meant to add, don't worry about judgement - if you've seen some of the mistakes I've made developing the V2, you'd feel a lot better I promise. Some of the more "whoopsie" moments are detailed in this thread from time to time. I blew two digitisers for example - because Matt's design didn't limit the THAT1512's output which is *fine* provided you have everything wired up right.

Simply disconnecting the microphone and using a 50/60Hz AC signal generator (like a finger for example) is quite sufficient, at max gain, to throw the output to within a whisker of 30V peak-to-peak. Into an input that's designed for 1V practically and (probably fries around 3-5V as these are tiny MOSFETs not intended for high-voltage work).

Trust me, if you can break it, I probably have at some point in my career, which is why I tend to stick to writing these days. 🙂

Hi,

I am currently building this microphone because it looked really neat and I am now stuck on the usb interface. I got rid of the 2 RCA cables on mine but kept the right, left and gnd wires of the jack. However, it seems like we need to connect it to only 2 pins : GND and out. I watched carefully the tutorial and didn't clearly see how you managed the wires. I would like to know what I should do with them.

Welcome @karol.

The ground lead is common to everything - it's a reference point for 0V which is where all currents must eventually return to. Ground, therefore, is sometimes called "return".

So you have a few options - ground loops however are a massive pain and there are a lot of workarounds, none of them particularly good. The most common is a ground "Mecca" where all the currents return to by the shortest path and it's what we'll do here.

For the V1 board, I suspect it's better to extend the ground lead by removing the outer insulator and peeling it back. It's only lapped so it's not too hard to work with. So grab yourself a few CMs of the outer copper and wind it nice and tight so it doesn't fray.

Now get some spare 7 of 14 core hookup flex (single core works but it tends to fracture) and solder about a CM to that outer shield. (It's useful to have some in just in case if you're planning this as a hobby.)

Now this is where it really needs @DIYPerks to do a video as it's rather tricky to describe. But what you're going to do is use some heat shrinkable tube (or a good insulation tape) and you going to "bandage" the cut end of the RCA input so that the extension wire is loops backwards away from the inner ("signal"*) core.

You can do the same with the inner cable but that's not usually necessary. Another reason we bandage the outer lapped screen is to stop it fraying and shorting other parts - and that does happen.

You can use the screen from either of the RCA pair and you can use either or both of the inner conductor. The device is stereo but the mic is mono so for many cases it's probably better to solder both inner cables together.

Ground Mecca can be any point  you choose - but I suspect the 0V USB pin on the NMA0515 will be the most convenient. It's been so long since I looked at the original, Matt may have picked a different point and his works so use that instead.

Let me know if this is unclear, sometimes I over-explain, it's hard to guess where everyone is.

* Strictly speaking in a situation like this one both the inner and outer conductors carry the "signal" because it moves in a loop and the ground is part of that loop. I won't worry you with a more in-depth discussion, there are entire chapters of expensive books dedicated to this sort of thing and even professional designers get it wrong from time to time. What most of us are taught in school is pretty vague.

Electricity doesn't really move in the wire - that's just how the fields (the electric and magnetic fields) are affecting the electrons in the conductor. But (and this will bake the old noodle) the energy - that's the stuff people like me belly-ache about - moves in the insulation, it's only guided by the conductors. I don't know of any decent videos on this, most are too simplistic and the rest tend to drop into Maxwell's equations. And they don't cover capacitors either which rely on a quantum effect where the opposite charges attract. This effect is part of how condenser microphones work as condenser is just an old term for capacitor.

If this sounds bonkers, consider that light - visible light - is just energy that we perceive as light - but is really just a wave in the electric and magnetic fields. We can't see radio waves (in fact we can't see very much of the EM spectrum at all) but if we use bits of wire we can trap them and make them work for us. The aerial your household TV (the ones most of us know are Yagus) are just bits of wire that are designed to focus, reflect and trap EM waves and have them disturb the electrons in the wire down to our TV.

If you made it this far without your head exploding, well done. But even if you didn't understand it, you probably understand why getting the wiring right is a major part of a sensitive design like this one.

Thank youvery much for your help, I understand the end, but I am lost at the begining of what you said 😨 . What I meant is that, like DIY perks, I got the usb interface with the stereo jack input and according to the video, he plugs the jac kwires in these connectors (GND and out) but there only are 2 pins, so I am kinda lost on how he did that, which cable he cut/shorted/connected somewhere else

Not to worry, the picture helps enormously. I'm stuck under a cat right now with my ex. wife bending my ear so it might be 24 hours before I can check that wiring. I thought you'd just cut the RCA plugs from the cable. I'll check your picture against one of mine.

You can wire the white and red (left=white and red=right) together and connecting the black to he ground. Using Matt's technique you this gives you a pair of wires. But make sure you follow Matt's polarity. If you're stuck, drop a note here and someone (or I will) will tell you which one is which.

The red/white pair need to go to the single large capacitor fed from pin 6 of the THAT1512. The black can go anywhere marked as GND (although the device is likely to find a ground from the power supply anyway. (Should!)

If you get these the wrong way around you'll be shorting the THAT to ground but it's got a protected output and the capacitor stops DC so it shouldn't hurt anything. It won't work but it won't break! 🙂

The V2 - due sometime after Christmas (don't hold your breath, I've been developing it over the last two years) re-uses almost all of the parts from this build but it's far easier to make as almost everything is pre-soldered at the factory. You can get "bare" board but take it from someone who's done it - don't! 😉

Marc

@marcdraco Thank you so much for the help !

ive just finished the microphone but when i plug it in nothing happens, i have checked that everything that is supposed to be connected is, and everything that isnt, isnt. the nma512 is giving correct volts, i might have connected things wrong but idk

Hi Dawg, welcome to DIY Perks. Let's see what we can do to get you moving.

So you have a multimeter - that's a great start. I'll assume you don't have access to any other test gear such as an oscilloscope, sig-gen etc. It's astonishing how cheap this stuff is these days with Chinese imports and modern technology. (I have portable scope, multimeter combo that cost less than couple of evenings out. Not amazing by any stretch but it's handy for a quick measurement or low-frequency measurement.)

OK, let's start from the top down and see if anything has gone pop or it's just being awkward.

There are quite a few things that can potentially go wrong and catch out the best of us.

In order we're all on the same page, I want you to follow my instructions as best as you are able. If you're unsure, please shout up - other people might face this issue in future and we can all learn. I learn things too - including what's tripping people up so I can design around them where possible.

First off, I want you to disconnect the "Audiograbber" (ADC) from the circuit board. You can leave it connected to the USB-C adaptor if you've done that already. If not, leave the end with a USB plug to make it easy to hook up to the PC.

And that's our first test.

V1 had a feature that allowed full-swing voltage to appear at the output and we need to make sure that any testing we do doesn't blow the ADC sky high. I blew two myself because I didn't even think to check! (D'uh!)

Given that, we'll test the ADC first.

I'll assume for this discussion that you have some recording software like Audacity (it's free https://www.audacityteam.org/) and excellent for our purposes. It's better to record what we're doing at the moment than listen to "live" output. Audacity will show you the signal as it's been received but you won't risk getting an earful of static or worse.

When you're set up with something and you've checked to make sure it's working we'll move on. I say this because it's very easy for Audacity, etc. to pick up the wrong recording device and you might find yourself recording from the internal mic by mistake. 😬 😳 

When you've done checked that you're connected to the correct input device, start recording and make sure the recording level is at max. Audacity has level meters too which can help. I'll assume you have Audacity for this discussion, at the price it's unbeatable.

If you've just left the input leads on your desk (recall you must disconnect them for these tests), watch the level meters and waveform display.  When you start recording the blue LED on the Audiograbber will start flashing to indicate that it's transmitting data to your PC. (If not, you'll need to deal with that issue first.)

Now grab hold of any one of the input leads by the copper part. (It won't give you a shock unless your PC and house wiring are already wired so badly they would have killed you by now.)

OK, so you should see the level meters bounce or jump and sit in the red area. It doesn't matter - we're not looking for music, we're looking for a signal and using our bodies as a "meaty signal generator" is often all we need. It's quick, very dirty and only works on high impedance inputs, but that's often where these issues lie. If you have a sig-gen you can do the same thing with that, only I would assume if you did you would have already thought of that one.

Assuming that everything is OK with the ADC, let's check some voltages.

To do this I need you to find a ground (reference point) that we're all agreed on. The ground point on the USB-C (Pololu) breakout is useful, but anywhere you can find you can get your black lead to and hold it while you do voltage tests.

The output from the NMA0515 is about 22 volts when the THAT1512 isn't present so we can test for that first. Carefully remove the IC from its socket and read the DC voltage at each pin. I'd expect to see the following with the board powered. There are a few ways of powering the board without re-connecting the ADC. It's the audio leads that need to be left disconnected though.

1. 0 volts.
2. 0 volts.
3. 0 volts.
4. -18 to -23 volts (approximately).
5. 0 volts.
6. 0 volts.
7. +18 to +23 volts.
8. 0 volts.

If anything is off here, the board won't work. But I'll assume it's OK for now, so moving on, plug the THAT1512 back in the socket and continue to read - from the IC pins!:

1. 0 volts.
2. 0 volts.
3. 0 volts.
4. -15 volts (approximately).
5. 0 volts.
6. a few millivolts DC.
7. +15volts.
8. 0 volts.

Similarly check the voltage at the on the Drain and Source pins a the capsule. That's a little tricky so be very careful when you do this. The exact voltages don't matter here but the Drain should be more positive than the Source pin. (FETs are a bit weird hence the vaguery here.)

The alternative for this particular test, which you might find is easier is to check there is a voltage drop across the channel - black lead to Source and red lead to Drain. That should be a positive voltage of several volts (if it's negative, the source and drain are wired back to front but FETs aren't usually fussy* .

That should give you enough to chew on for now - and the Director (Walter) is getting antsy so I'll check in with you later, he's patting me and his claws are sharp.

* JFETs are voltage controlled devices. No current (technically a little current) flows through from the Drain to the Gate. This makes them very simple to use and quite robust but the difficulties in biasing them have relegated them to specialist usage such as capacitor microphones and some medical devices like ECGs.

Hello everyone, I decided to make this microphone but I have a problem. Unfortunately, the audio capture card is not sold in my country and I cannot bring it abroad. What kind of solution can you offer?

Only

https://www.n11.com/urun/usb-video-dvr-capture-kart-video-yakalama-karti-easycap-15968973?magaza=aghbilgisayar&utm_source=comp_akakce&utm_medium=cpc&utm_campaign=akakce_genel

There are similar ones. Can I use it if I buy it in my country or how can I connect it? I would be glad if you could help me

Hi Joseph and welcome to DIY Perks.

I've tried that digitizer with this project and the performance is fairly drab. It works, but not well as it's mainly targetted at video and seems to lack sensitivity. The "Audiograbber" is specifically designed for audio and it's the best I've used at that price point by far.

The "next" best is going to be something like the Focusrite series (Focusrite Solo) which is a super-high quality, custom chip featured in Focusrite's professional desks (according to the company "blurb") but it's not "cheap" even if you can find one of the older ones on eBay; and you can't fit them into a smart case which is the centrepiece of this project.

Matt and I are beavering away at fixing these issues. My end is done but it will take some time for Matt to write and edit the video. I couldn't even guess how long but it will hopefully dodge some of these issues. One of the capsule PCB designs I produced during the development was tested on a Focusrite Solo and it works a treat. Trouble is it's also quite expensive due to the extra complexity... I haven't even tested the "beefed up" version myself yet (so I've sent it to Matt instead) [sorry Matt]!

As to this issue, I'm afraid that places you in a bit of a predicament (and I don't speak Turkish) so unless you can shoot me some other examples I can't be a whole lot of help.

The alternative (thank goodness for choice!) would be to use a standard microphone digitisier. And no, I haven't just  "taken leave of my senses" (that happened decades ago) this can actually be an option if you really want to output to a USB-C port OR you can just drop that part of the build entirely and wire the output to a standard 3.5mm jack and pop it in the "line in" port on your PC if it has one.

The mic input on a modern PC is usually quite decent (the technology is well established and, therefore, cheap). But this way you don't get the ease of switchable gain which is one of the major selling points on Matt's design. You can still drive a mic input with this board but it means changing all the gain resistors which might be a step too far. (The output is 100x greater or more than a mic input expects.)

Let me know your thoughts and we'll see what we can cook up.

@marcdrac hello again!

the reason I asked this was because I bought the jlı 2555 microphone today and I really don't have any idea how to make the circuit most of the parts on the pcb are either not available in my country or are incredibly expensive so I'm looking for the cheapest way to do this
even if I find the components the usb audio interface conversion part is very expensive. while browsing the internet I came across a microphone called lennox bm800 it uses an electroid microphone so I thought if I remove the pcb inside can I connect jlı 2555? (probably not)
features:
BM800 Professional Studio Youtuber Recording Microphone Concert Frequency Response: 30 HZ
18KHZ Sensitivity:
36dB+
2dB Output impedance: 252 Load impedance: > 10002 level level: 7dB
i've seen people doing this on the internet so I wanted to ask.
thanks for your answer...

I know the BM800 intimately, too well you might say. It’s a donor body I’ve used with some of the upcoming designs. 

It’s passable in performance (the analogue one is; the USB sounds like you’re talking in a blizzard. 

Sounds like electronic parts are expensive there which rather puts a dampener on everything.

The JLI capsule will fit but you’ll need a replacement saddle (or you’ll have to use glue and risk jiggering it). 

The guts are fairly simple a stripped down version of a classic design by Schoepes Mikrophone back from the 1960s during the change from valves to transistors. It’s a wonderfully elegant design and the front end design is tough to beat (and trust me, I’ve tried). 

The Chinese don’t spend any effort biasing the JFET - something that is required by even the Schoepes design so performance is passable but not awesome. I adapted a circuit developed by others for one of the new ones as a result but I kept the basic output design the same. 

JFETs are weird fish but so far as noise is concerned they are pretty quiet so the (penny-pinching) circuit inside a BM800 should work fine. The FET doesn’t care what sort of capsule is there so long as (in this case) it’s a pre-biased electret like the JLI2555

The capsules in those mics are … kinda nasty … so the JLI would give better performance although I suspect it’s a little quieter. 

That could be an option but it’s usually cheaper to get them from Ali Express if they will ship to where you are. They look great but the build is iffy in places, most notably in the XLR socket in the base. 

you can pick them up quiet cheaply with a pop screen and even a desk mount that I wouldn’t try to hang a moth from, let alone a recording mic but they have to make saving somewhere. 

I can certainly help you if you did decide to go that way. 

Hello

I will probably go with bm800 because it costs around 30 dollars but if I build the circuit it will cost me 70-80 dollars (especially that 1512, NMA0515SC and usb audio editor only cost 60 dollars). If I buy b800 the microphone itself costs 17 dollars and the power provider costs 23 dollars, totaling 30 dollars, so it is not very expensive for me, besides bm800 is available in my country, I will only buy the phantom 48v power provider from abroad. But there is one thing I am curious about, how is the sound quality level and if I remove the bm800 microphone and solder it to jli2555 will I have any problems, I am sorry to bother you a bit, here is what I will buy:

https://www.trendyol.com/lennox/bm800-profesyonel-studyo-youtuber-kayit-mikrofonu-p-69349877?boutiqueId=61

https://tr.aliexpress.com/item/1005005486695822.html?src=google

After I get these, I will start working on the case (of course) and I will probably try to stick to the original design. Thank you for helping.

Joseph, you're very welcome and no bother at all. It's a pleasure discussing this with you. These questions help us to help others in similar positions although this thread has gotten rather unwieldy now and I expect it's going to keep on getting longer as this is such a popular project. I think the only one close is the artificial daylight.

The only folk who cause a little annoyance (save for the SPAMbots) are those who ask a question, or demand Matt does something for them (!) and then run away never to be seen or heard from again. The minority of course, but we're not here to help people create things that could hurt them or others (high voltage, high vapour pressure etc.).

The BM800 is "hated" by the pro-audio types because it's cheap Chinese tat. The metalwork is poor (good enough for the body but not amazing) but the electronics inside isn't entirely awful because it's essentially based on a design that stood the test of time.

We have a saying here, "If it's not broken, don't fix it."

This is the output stage of a typical Schoepes microphone - this is my "version" of it but they are all basically very similar. The signal comes from the left as as balanced pair (required for noise rejection). This is the same trick Matt used in his design just a more "powerful" version of it designed to run along very long cable runs with "phantom" power.

The vertical line just to the left of those two PNP transistors is the "power tap" for the 48V phantom power. We don't need 48V any more (for a number of reasons) but it's a long-established standard so it's what we work with. I've done designs that work all the way down to 5V, although not using a phantom system, the lowest we can realistically get is about 10V.

I think this is why Neewer also offer(ed?) a USB version of this mic which does work on 5V and doesn't require phantom power. But it's also utterly awful as I've said. So I've designed a better one but given the difficultly for you getting parts in your part of the world, it won't be a whole lot of use to you.

One issue I have with the Neewer 48V converter is that it's noisy. Not massively noisy but audibly so. There are multiple ways of generating the 48V supply and some are better (quieter) than others. Neewer is designed to a price so they've used a simple CMOS oscillator based on a 74HC series hex Schmitt inverter. (It's simpler that it sounds from that description I assure you).

But that's off topic for where we are night now.

The BM800 body will get you going and we can sort you with something to get rid of that awful power supply later - as an upgrade that won't break the bank.

A word of caution though. The Chinese are masters at making stuff that looks amazing in a photograph and often looks pretty good in person.

Then you scratch the surface and the paint comes right off! 🙂

What Neewer did with this is use a unbalanced (single core cable) rather than the balanced one that you would normally associate with those XLR connectors. This isn't "illegal" but it's deceptive and relies on our assumptions that if it's got XLR then it's professional grade.

Even their XLRs aren't wonderful.

It's cheeky but it's also clever because we "feel" as if we're using professional gear and only notice when we compare like with like and say... WAIT A MINUTE.

That XLR patch lead - it's a single core with a 7-strand centre core and lapped screen: as cheap as it comes and about the same quality you'd get in a very cheap RCA patch lead that you might find on an old HiFi. The real stuff has two beefy conductors running through a braided screen and some even include an extra screen for the case only (it's not grounded). That's only for cases such as studios and live concerts were there is an abundance of electrical noise though.

They "fake" the balanced output with a impedance trick (if they even bother at all).

I have a solution that will appeal to your particular use case in Beta at the moment actually. It only needs a JFET (one intended for microphone use) but there are a few to chose from - and a little soldering to hook up to a standard PC microphone connector. It should also drop right into your BM800 body and will run the JLI2555 like a boss.

It's actually the one I'm recommending people start out with and work up to the more advanced designs as budget and needs dictate. I should have a couple spare as part of the Beta program if you're interested. They should be with Matt and I in a couple of weeks as I didn't have them sent over by courier.

You don't get the gain control with this but, with Matt's case, you can hide a simple USB mic adaptor and you're laughing. The wonderful thing about this project is that it's got loads of upgrade paths for people who are learning and don't have the bottomless pockets some mic building websites expect. Sometimes I feel like screaming, "If I had a coupe of grand to drop on a mic, do you really think I'd build it myself?" That sort of money could buy some seriously nice kit like the Russian made https://soyuzmicrophones.com/023-series are just gorgeous. (Worth mentioning these guys are not part of the current conflict.) Or a German made beauty although ultimately, music is art and the "natural" law applies as the price increases, the quality does at lesser and lesser rates.

The difference between a $1 mic and a $50 mic is noticeable. From $50 to maybe $500 it's audible if you have good hearing (and definitely measurable) but beyond that you're getting into the realms of material and design aesthetic, not actually improving the sound. Matt managed to effect a beautiful aesthetic with a pretty decent sound which is a neat trick that's taken a lot of work to beat (electrically anyway).

We're all on a budget and most of us are hobbyists with precious little experience so there's no point making stuff that's out of everyone's reach. It's supposed to be fun and educational. I'm grateful to Matt for letting be a part of this and get the old grey matter working again.

If your JLI2555 didn't come with a saddle (that's the little plastic mount "C" shaped mount) you will need to get one to use it in a BM800. It's not an ideal fitting but it can be done (if I can do it, anyone can).

Will it sound better?

Yes and no.

The JLI2555 is loved by the mic builders community with good reason. It's an excellent mid-range electret capsule on the 25mm diameter LDC (large diaphragm condensor) format. JLI also do a larger 34mm version which probably has slightly better low-end performance but 25mm is ideal for vocal work.

The difference depends a lot on which capsule the underpaid, overworked lady in the factory pulled out of the parts bin. I'm not joking. These things are made by the million and they cost next to nothing as they are made by machines and are actually very simple indeed. Your mobile phone will have at least a couple, many have four and perhaps more. But they are tiny MEMS devices another remarkable piece of tech.

The truly barmy (waves hand) even try making them at home. That's a story for another day.

But Neewer will get whatever it can - cheaper the better and although these things are simple, the quality of the materials does make a world of difference. I've compared the "cheap" (>$10) 25mm capsules to the JLI2555 and they are like chalk and a nice mature cheddar.

The basic idea is the same but you can feel the difference the materials. You get what you pay for.

You might get a "decent" capsule, you might not. I've gutted a few of these during the research for this project and it amazed me to see the variations. Even in the "same" product from the same supplier. It's a bit of a crap shoot.

The JLI2555 and similar mics are more consistent in this regard. That can be important if you're working in stereo or Ambisonic*  where it's absolutely vital.

The other issue with the Neewer and the bad one is what I mentioned earlier. That power adaptor is horrible. The Focusrite Solo (although bit expensive for most) is, for all purposes, silent whereas the Neewer one can be heard as a nasty whine running through the quiet bits of your recording. This is especially problematic if you're working at higher gain or trying to capture say, bird song.

The JLI2555 with a properly balanced cable design can do all of that and more beside. And do it well.

It's an upgrade path and each stage will make it that little bit better.

Marc

* Ambisonic is a British invention for 360 degree sound. It produces four channels in a special format which requires post-processing to produce a useful audio signal. It's something I hope we can do in future time allowing. Four capsules have to be mounted the right way and then there's the mixing of the signals to recover the sound format you want from mono (why?) to superb stereo and even all the way to live surround sound that will blow your mind. Soyuz has one on their website but it's probably not something you want to get lost in right now. 😉

There's another system which has as similar sounding name that works differently but I can't recall it now. Here's a starter if you're interested. Maybe someone else can remember - they have similar sounding names but seems like Ambisonic is the more popular one with the mic companies.

https://en.wikipedia.org/wiki/Ambisonics

Hello,

First of all, I am surprised that you have put in so much effort, I was really surprised when I saw the article, thank you very much. Actually, my purpose in buying the bm800 is to use only the circuit inside and not to use the rest, I will do the rest by staying true to the original and when I searched for the phantom power supply in my country, I saw some places that produce it, generally people commented positively, I can return it if it arrives broken and I have to wait for 1 month. What I am wondering is what should I pay attention to when choosing this phantom 48v converter? I also paid around 25 dollars for the jlı 2555 (normally 5 dollars) of course, since there is tax and shipping fee, do you think it would be better to cancel it or change the one inside the bm800? I want to get a nice sound. Other than that, thanks for your help again.
Best regards

Hello again

While browsing the internet I found a Se2200a microphone. The capsule inside was faulty, so the seller put it at a price of $43. If I put a JLI 2555 in it, would it work?

Or  must I buy this capsule?

https://tr.aliexpress.com/item/1005007767573357.html?spm=a2g0o.productlist.main.67.4818wxqiwxqiiN&algo_pvid=84afdb0a-314b-4bd3-b094-70b4bb010d77&algo_exp_id=84afdb0a-314b-4bd3-b094-70b4bb010d77-33&pdp_npi=4%40dis%21TRY%21843.37%21430.06%21%21%2123.16%2111.81%21%40210386d117299570196343040ee489%2112000042136337919%21sea%21TR%216059718901%21X&curPageLogUid=zr7nxRt3aVPp&utparam-url=scene%3Asearch%7Cquery_from%3A

Hi Joseph, nice microphone but no the JLI2555 isn't suitable for that and the capsule from Ali isn't either (although they do carry ones that will fit).

Short answer:

You'll need something with TWO diaphragms for the Se2200a and they will be unbiased too. This is where more recent condensers come into their own vs. the original idea. I've touched on capacitor theory elsewhere in this thread but that might take some time to find. Dual diaphragm capsules are more complex (not much more) but that does push the cost up. You COULD get away with the one you've picked but you'd lose all the pattern selection.

WAY, WAY longer answer:

Condenser is an old name for capacitor, so a condenser mic. capsule is really nothing more than a very large capacitor. Physically very large. Practically speaking their actual value is in the picofarads (sometimes called "puffs") and the capacitance of a condenser mic capsule is not much greater than a couple of bits of wire twisted together! Most capacitors use two large foil plates separated by a thin layer of an insulating material. In a capacitor microphone the insulator is just air (and that's sealed at the factory).

Typically they have a capacity of around 40-50pf.

Capacitors store electrical charge using a quantum force (weak electromagnetic IIRC, but don't quote me). When we apply power to a capacitor electrons are pulled from one side (plate) and moved to the other so one plate becomes slightly negatively charged and the other slightly positive.

If you've done this with magnets you'll know that opposite poles attract and that's what's happening here. The opposing charges are attracted to each other - the effect is incredibly weak so the distances involved are very small. That's what makes it hard to make a practical condenser capsule at home. It can be done but you'll need a lot of specialised tooling to make a decent one.

The distance between the plates governs how much charge can be held - the closer the plates are to one another, the more charge can be stored. The practical limit is about an atom or two but in the real world no one does anything like that close. I don't know the exact distances used but it's in the order of a fraction of a millimetre. If the two plates (which are metal or more usually gold-spattered plastic touch even for an instant that charge is lost.

If we alter the distance between the plates we vary the amount of charge that can be stored. Hence the movement of air molecules caused by sound moves the plates closer and alters the charge. The JFET's job is to convert that incredibly small variation into an electrical signal we can amplify - and enjoy as music, speech, the roaches breaking wind etc.

So far so good.

The first condenser mics (and as of this writing, the best ones) require a biasing voltage to charge the capacitor when it's being used. The voltage is important in these mics because what may not be immediately obvious, the more voltage we apply the stronger the force "trying" to pull the two plates together becomes stronger until the insulation resistance breaks down and the capacitor flashes over (discharging and often breaking something else along the way). It's important to get the right rating for this reason. Air isn't an amazing insulator compared to long-chain polymer plastics but it's slightly compressible.

At low voltages say, 10-20V the amount of charge is so feeble you need a lot of amplification but as the voltage increases the action of the two plates pulling at each other tends to tension the front plate (the diaphragm), the back plate is fixed. So over the years the designers have settled on bias voltages of 50-100V (I expect early ones used much higher voltages). Any given capsule will have its own spec but most are quite happy around the 48V mark which (surprise!) is why phantom power is set there. Mic designers can tweak the supply voltage to affect the high-end ringing and other characteristics but that's beyond the scope of this question.

The faulty Se2200 is going to need a DUAL diaphragm capsule because it's a multi-pattern mic - you can see this from pictures of them opened up or just checking the specifications. I expect that wouldn't be a cost-effective purchase as the capsules of that type, even on Ali, will run you as much as the "used" body did. It's also the single most expensive part of the whole device.

Some of these capsules are no longer manufactured and there are actually places dotted about that will refurbish an old capsule that has been damaged from misuse or just deteriorated with age. All at significant cost as it's done by hand.

Enter the Electret

Now some good news and an observations of why people in the pro-circuit get "sniffy" about us upstarts with our JLI2555s.

Almost every Chinese microphone and a good number of fairly decent "everyday" condenser microphones use a material called an electret (hence the term "electret condenser") which is a material that's pre-charged. In other words, it works without needing a biasing voltage. The charge, it's claimed, should last >100 years so they won't ever wear out in our lifetime.

So no need for a 50-100V supply!

The downside - and why the professional lads look down on them is they are "common". Once you remove the need for a high-voltage bias circuit, you simplify the overall design dramatically. It also means, no high voltage, so they can be made smaller. Much, much smaller. Modern electrets are so small you'd need very powerful microscope to see the "moving" parts. These MEMS mics are in everything from mobile phones to Bluetooth earbuds and usually reside behind a hole the size of a pinprick. And done right, they can sound really rather good.

Now you might reasonably ask, if this is so amazing, why anyone would bother faffing around with the older system. Well (to abuse the phrase) "all the sputters is not gold".

Electret material lacks some of the qualities found in some modern plastics and there's the real issue (for the people who pretend to care or care because they don't know any better). And it's a fact that while electrets, yes, even the ones you might find in the bottom of a box of cornflakes as you holiday in North Korea (sorry guys) are very good indeed but they lack a certain "something". Material science is a bit outside of my field but some plastics are more suitable for diaphragms than others and electret isn't ideal. Good yes, ideal no.

So the big boys use very specialised, ultra-fine plastic films that are treated with pure gold to make them into a conductive plate that is flexible but resistant to mechanical damage from being dropped. By the same token, brushing this stuff with a fingernail will leave a mark and damage irrevocably. Gold is used because it doesn't tarnish over time like copper or other metals can. I suspect the JLI and other quality electrets use something similar but the the older system is still preferred for the high-end gear.

This also makes these capsules far less suitable for home-builders to work with as a simple slip of a screwdriver or a cat poking its claws around on the desk will destroy them.

JLI, Panasonic and others have this off to a fine art now and in all honesty most of us would be hard pressed to tell the difference. The size of the diaphragm matters far more than the material it's made from. Larger diaphragms are more sensitive to lower frequencies but less sensitive overall because the mass of the material affects how far it can "wobble". The microscopic electret mics in your typical earbud is amazingly sensitive as as result but also a little "peaky" and "tinny" (studio guys describe this as "bright").

There's also a choice of which plate is pre-charged - front or back with each having benefits and drawbacks. The front plate is the diaphragm the back plate is fixed - the JFET registers the changing distance between the two. Without looking I'd guess the JLI2555 is a back electret design which likely means a better quality diaphragm and that accounts for the excellent reviews. (As I've mentioned it's also beautifully machined compared to the cheaper copies).

This is why Matt tracked down that particular capsule - but since JLI don't sell to the general public we have to get them via resellers which accounts for the inflated price.

If you followed me this far you'll probably be getting a sense of why some of these beasties cost an awful lot of money. Neuman (another massive name in this space) have produced some of the best sounding unbiased capsules but in all honesty, they're not suitable for this project and not really for the home builder unless you're a little unhinged like me and (unlike me) have nerves of one-inch thick titanium plate.

A Word from Our Sponsor About Patterns

You can make a multi-pattern mic with two of these mounted back to back (electrets that is) but that's not useful in a microphone that's being used for vocal work except in very specific circumstances and it's never going to be as good as the real thing due to the physics of getting the actual diaphragms as close as possible. (Sound waves are tricky little blighters - something which Matt alludes to in the original build video when he didn't vent the capsule cover in his first tests.)

These capsules are fairly directional which means they "hear" sound mostly coming from directly in front of them and don't react as well to sound from the back. In order to do this they need to have an open access port at the rear to the pressure to equalise correctly. I won't cover directionality here - but the standard vocal mic pattern (called cardioid) allows a singer to be heard while "rejecting" sound from the sides and rear (the audience). There are multiple variations on this feature alone if you care to read about it from "omidirectional" all the way down to the super and hyper cardioid which reject more and more sound from the back. Hypercardioids are found in those big "balloons" you see sound guys carrying around. The balloon is there to keep the wind out.

If I get some time to sit down with Matt I'll look at making a practical hypercardioid somewhere down the line - but for now, this is more than enough to keep everyone going. I've got some crazy stuff in development which have absolutely nothing to do with sound. I just picked this one because it inspired me - and Matt's skills in everything from mechanical design to video production make me look like a unranked amateur with no fingers and one arm tied behind my back. Trust me, no one wants to put me in front of a camera unless it's for a Halloween prank.

The Scary Bit

This talk of Halloween reminded me you did ask about 48V phantom adaptors. ("Phantom" power - get it?)

Truth is I don't know of any "good" ones that won't cost a bomb (relatively speaking). The Chinese ones are going to be as cheap as possible and that means a square-wave oscillator which, even if it's running well beyond the audio range produces so much "hash" on the power rails that the noise "folds back" into the recording. A pure sine wave won't do this but it's a lot harder to generate - by harder I mean more costly in parts. The professionals use a whole variety of techniques - the most popular one now would be DDS (Direct Digital Synthesis) but that requires a microcontroller (to make a clean sine wave) and then something to boost the voltage from the 3-5V up to 50ish volts. This stuff kept me up at night - or maybe that was my fondness for cheese toasties? I can't be sure?

There multiple ways to boost the voltage from a classical transformer (inefficient, bulky) to some extravagant MOSFET circuits that are really only practically realisable on-die (that is, on a custom-designed IC).  This is, I expect, what the Focusrite designers have done and they do it with incredible elegance. But their own chips are used in huge numbers of desks, not just the little personal ones like the Solo. I did look at this for a discrete design but the number of transistors required per stage renders it pointless.

So the rest of us are stuck with more modest affairs but fear not, even Rode use Schmitt oscillators in their battery operated 48V pencil mics so it can be done. Except that there's really not that much point because, as I've mentioned, the law of diminishing returns rears up and this is more an exercise in futility for a one-off design intended for the maker space.

Be Kind, Rewind

So let's rewind a bit and ask WHY 48v is needed. For an electret mic, it's simply a waste of time. I've made a 48V capable PCB that's just 34mm diameter and fits on the back of a JLI2555 or the larger 34mm electret LDCs (referred to earlier as Varee) but it can work down to the 15V Matt has available from those little NMA0515 inverters. That started out as a much simpler affair, meant to make it easier to mount the JFET vs. having to solder it direct to the capsule which is potentially tricky (and worse, disaster if you snap the wrong leg off!).

Over time it got bigger and split into the two final ones we have now - the 15-50V Varee and the 3-15V Sarah (mentioned earlier). The Sarah is simple to work with and offers multiple options so it's the go-to for converting donor bodies like the BM800 into fully functional handsets for Matt's original all the way through to a PC mic input. It also has space for the *required* 1G resistor that we often overlook. Leakage around a hand-wired circuit is usually enough to provide a 1-10G ohm path for the JFET but (as luck wouldn't have it) as I improved the design of that board, it became increasingly difficult to get the JFET to self-bias - so I added space for the resistor (although a small amount of grease or pencil "lead" (graphite) works a treat. (Just don't tell the professional lads or they will be chasing me all over Yorkshire with U87s attached to the end of pitchforks.)

The reason for these long answers isn't to show how smart I am (I'm really not that clever) but as a way for others to grasp what's really going on here and you've asked some great questions.

Who Turned the Heating Up?

I'll probably take heat from some quarters for the sin of self-biasing because it's a bit naughty but we have to balance the simplicity of a self-biased JFET against the extra noise produced by the extra resistors required to bias it "properly". Rod Elliot will hate me for that, sorry Rod. (You can find Rod over at Elliot Sound Products. He forgets more over his morning cornflakes than I'll ever know and his articles are absolutely first rate. He even sells kits too and is very approachable.)

That noise is a bit tricky to explain (the exact amount is largely dependent on the value of the resistor. The larger the resistor, the greater the noise but the calculation is bit... mathematical and I have trouble counting the change I have for the bus ride to the pub, let alone the journey back.

And despite what I've heard people say about microphone design, this IS as much of an art as it is a science. That's not to say we can just guess everything (that's a sure way to faceplant into the pot of molten solder) BUT there's an art in determining the frequency response and other factors because we're recording music/voice etc. and "flat" mic (not that such a thing exists in the real world) sounds, well, boring.

That's the arty bit. Manufactuers like Rode, Schoepes, Neuman and others spend inordinate sums perfecting something that sounds just right. That's a large part of what you're paying for as the physical parts are (for the most part) not all that expensive.

The other gotcha with those BM800 donor bodies is the saddle - which is solid plastic. That means that (even if you use a suspension) they will pick up extraneous vibration. A "rubberised" mount is the order of the day here and failing that, a suspension arm as you see in the original project is required. This is something else that the Varee and Sarah capsule adaptors have - they're designed for suspension on springs or rubberised bands to make it easier to adapt for people like me who have all the metalworking skill of a worker ant. That's what will help make V2 of Matt's original even better than the first and make it more accessible for the rest of us. (I hope.)

Exciting stuff, so watch this space for more.

Hello, first of all, I want to thank everyone who gives their energy to the development of this project!!! THANK YOU!

I think a lot of people are just waiting for the v2 video where Matt will take us by the hand and push us to throw it all in to ultimate success
Please open the gate, is there already a v2 video in the works?
and secondly, have you tried to use AI to improve the PCB idea?

All I can tell you Matt has the V2 preview boards. There is another “part” on way from JLC which is for one of the lower cost options but the project is at the stage where we can let Matt work his magic. I’ve sent him some simple instructions but the new one has a lot of new features that weren’t possible or practical on Veroboard.

I’m just a contributor so I don’t have any insight (not that I’d confirm or deny if I did have insider info).

JLC have done a stunning job. They look gorgeous yet still offer enough work for you to be able to show you made it. It’s just the PCB is far easier to work on.

Options range from simpler than the original (to construct) through to some very advanced techniques for bragging rights. Performance is equal to or better than the original, esp  in terms of noise rejection.

The “advanced” capsule adaptor can be used, stand alone, with a professional 48v phantom powered system but it’ll have Matt’s special design. (No pressure mate!)

Assuming there are no hiccups they designs are licensed under a 2-Clause BSD license so anyone will be able to order boards and re-sellers make up kits, but we’ve worked hard to keep the costs reasonable.

It’s going to cost more than the original for a factory built PCB but they are almost ready to go and only need the larger, through-hole parts fitted.

By forum user request I’ve created a simple drop in replacement for the THAT1512, with acceptable performance for areas where the 1512 and similar chips are unobtainable. Not ideal but better than nothing. It’s plug in so can be replaced easily. 

None of this would exist without Matt so I’ll refrain from dropping spoilers. 

My role is to help out if things don’t work out and the PCB has gone a long way to address any hidden “gotchas” that we discovered along the way  

I’ve had to give the magic smoke genie the rest of the year off. Poor chap was run off his … oh wait  he doesn’t have feet. His chemtrails? 

Available now. Probably the best entry-level capsule adaptor for 25mm and 34mm LDC microphone capsules.

Compatible with both V1 and V2 of DIY Perk's ultimate USB-C mic.

And usable with those nasty PC mic inputs (at a pinch).

If anyone is interested in selling kits for any of these parts for the project, drop me a PM or reply to this post. We're not geared up for that, but this is a game-changer for anyone looking to work with LDCs for these projects or their own.

Can also be ordered from board houses like JLC in several configurations complete with SMD JFET and other parts to make a variety of easily usable mics: common drain, common source, FET only and more. Can be mounted to most common LDC capsules directly or screwed inside some donor bodies like the BM800.

Hello again!
I know it's a bit late to reply, I waited to make the module and bm800 to write this, first of all, thank you again for your help, I installed the module bm800 and it's working fine now, only the outer part is left to do.

Thank you for helping.

Excellent Joseph. Well done. Don't forget if you're able, post some pictures and get some kudos!

@marcdraco Don't doubt, I will share it as I did. 🙂

@marcdraco I have been away from this project for a long time due to various reasons but would love to give the v2 board a try but haven’t managed to actually find the file am I crazy? Have I been looking in the wrong place?

Hi Reaper, I haven't published the Open Source designs yet simply because this is really Matt's project and we don't have a supplier to ship out single boards and JLC, etc. will only do a minimum of two (part) assembled boards and that could run you a fair amount with shipping, imports and what-not. It's also got a fair amount of "optional" bits although you can start from essentially Matt's original (minus the large PSU caps) and work from there. Here's a ZIP with the V2.x design in KiCAD8.x format. It won't open in earlier versions. This is currently under BSD two clause licence which means you can do as you like provided the original authors are credited.

This was my from latest pre-production run. Matt has a board at his end but his production schedule is almost as glacial as my own due to time constraints.  The "right" side of the PCB is basically the ginned up version of Matt's original, re-imagined for SMD with a minium of through-hole parts for space and noise reduction. Electrolytics are required in a couple of places as per THAT Corps. suggestion. I could have used tantalum but somehow he bigger ones look "better" ... Solid SMD caps can be substituted if you know just the basics of KiCAD.

This board also has protection from the "Murder Bug" that caught me out on one occasion and I expect has others too. Built in "one" go and wired as Matt suggested, everything will work but remove that Mic capsule, and hold an input wire... and poof. The THAT can rocket all the way to its 26ish volt peak swing and there goes the neighbourhood (digitiser).

This side works with both Matt's original capsule design and the Sarah one described above. Sarah is a greatly improved carrier board which is primarily for this project although it can be used on any number of others by swapping "links" (0 ohm resistors).

Varee which split from the original is standalone for P48 systems and (without the 431 regulator fitted) works well with this too, but it REQUIRES (as of this writing) the second, galvanically isolated NMA0515.  It's not beyond the wit of man to adapt the Michelle board to work with a single NMA0515 but that extra supply finds use for a separate balanced output to drive a professional-grade mixing desk.

Note: due to advanced PCB design, Varee needs a 1-2G resistor and many Sarah configurations will too. There isn't enough leakage left for the FET to self bias without it. As noted somewhere way up there you can use a little bit of graphite from a pencil across the pads to do that.

If you can read a circuit diagram this should be fairly easy to follow, if not, probably best wait for Matt to do a video (or two!) to explain what developed over the two years or so this has been crawling along like the proverbial tortoise in the tortoise and the hare story.

hello again!

The upper part of the microphone is considered finished, only the microphone will be placed and its cable will be made. I need a suggestion. The microphone shakes significantly when I touch the table in any way. I think the reason for this is the rubber tires I used (I used all the ones I could find since I wasn't at home). Do you think I should connect the microphone from the back or should I buy a new hard rubber and try it? I know the soldered parts look bad but that's what I could do.