[Sticky] USB-C Microphone (official topic)

Wow, I haven't seen a Weller soldering gun like that one in a long time! Mine must be nearly 40 years old ... no wait, older!

That doesn't look half-bad (I'm not showing you mine or you'd laugh yourself to death and we don't want to lose members like that).

The rubber bits should work to absorb the shock. A little vibration will produce sound at sub-sonic frequencies so that's not an issue. This is really a "suck and see" sort of thing. I can't see how you've attached the carrier to the lamp-stand mount. It's from a BM800 kit isn't it?

The crucial part is that the capsule (and it's little cage) are free to "float" on the rubber bands, springs, what have you if that makes sense.

-update-

Yes, I managed to finish the microphone. If there are friends who can do it, here are a few recommendations:

You can use the upper yellow part of the bm800, it is perfect for this project as it has enough space to accommodate both the image and the PCB inside.

Yes, it is a holder from the BM800 kit because I do not have a brass rod or pipe at the moment. I will do that when I get it. Right now, just holding it is enough. When I supported it with cables from the back, the vibration was exactly what I wanted. It neither shakes much nor remains stable.

That's a good point about the BM800 as a donor "Faraday cage" and dump the main body. The Varee and Sarah capsule adaptors (available soon) were tested like this, but there's nothing saying you can't just solder the FET directly to the capsule itself.

What if we used polarized capacitors instead of the non-polar ones? I can't source the non-polar ones so I tried with the polarized ones and it didn't work. I was thinking of connecting two capacitors in series to diy it but I wanted to know if I did use a polarized capacitor, would it somehow damage any components? Thanks a lot!

The capacitors won't stop it working if you put them the wrong way around. I think Matt used non-polar out of an abundance of caution, but they are just two capacitors in series either way (back to back).

Electrolytics "fudge" large capacitor with a special chemical (an electrolyte would you believe 🤣 ) and that has chemical properties that rely on any constant charge to be a particular way around. You'd have to ask a chemist to explain how it works.

However, it's the input capacitors that matter. Under normal operating conditions, the output capacitor only "sees" a tiny amount of DC and it can even be left out entirely. The positive terminal of the input capacitors face AWAY from the THAT1512.

I'm more concerned that you said "it didn't work" though.

Yep me too haha. Also I was wondering, the out from the PCB is analog, correct? Since 3.5mm audio jacks are also analog would wiring up the "out" from the PCB directly to an 3.5mm microphone jack work? (On a motherboard that supports Line-in mic - 3 ring 3.5mm jack)

There's a "bug" in the original design (easily fixed) that can output sufficiently high voltage (around 26V) that can destroy the input transistors so unless you have a scope handy, I'd strongly advise against that. The new PCB design that should premier early next year. Design is open source and there's a download up the thread somewhere (recently) if you wanted to go that route. It uses the same parts (mostly). You'll need a KiCAD 8 to read the files but the JLCPCB extension will produce the parts BOM and drawings you'll need to order it.

I published a fix for the bug a month or two back in a separate forum post, but in essence you'll need to connect a pair of diodes in anti-phase (one up, one down) from ground to the output. This will (harmlessly) short any voltage going more than about 0.7 to 1V to ground.

This output is LINE compatible, it's far too powerful for the mic input.

I'm a little bit stuffed at the moment as my work PC went to silicone heaven last night and it's blown the power regulator on the main board - which means a new motherboard and since the screen is also cracked, it's not worth repairing. (That sobbing you can hear is me crying into my beer as I give it last rites.) 😉

It was past time for a new one so I guess fate was tapping on my shoulder.

Oh okay, do you think the input transistor on the mic is now dead because of that "bug"? And could you please share the link or name of that thread so I can try and apply the fix?

One more question, should the mic work without this fix too? Asking because it is not working atm as is...

Sorry for the loss, a better PC is gonna be in your life now 🙂

Take your time if you are busy, no issues and thanks a lot!

Thanks. I'm muddling through with a slightly jiggered wireless keyboard and a machine that doesn't have one. But I still have to use the trackpad as a mouse. What joy!

I'm right on a dead lion right now (poor kitty) so I'll see if I can find that link now but if not, I'll do it as soon as I get back.

It's the input transistors fed from the output of the THAT1512 that are the issue. If the mic works first time, that's unlikely to be an issue but if you have a lot of noise, and fairly high gain setting then it's curtains for the input MOSFETs as they can only stand a couple of volts at most.

OK, you should find it easily enough. It's marked USB Microphone addendum.

The mic will likely work perfectly without the fix but if you crank the gain too high, even for a second, it'll destroy the input of an Audiograbber and likely the input of your PC too - and you really don't want to do that. A low-cost oscilloscope is a good idea these days. Even Amazon has usable (if unexciting ones) from China for under £50 and they come with a meter too. Rubbish obviously, but they won't explode if you hit them with 50V P-2-P!

Okay so I definitely ruined the input on my PC then (and likely my laptop too lmao). The Gain was set to maximum as I was just testing it the first time and I saw some high voltages on the multimeter too but I didn't know then. Since you have pointed out the diode thing, I have ordered a kit and I will test it again with those installed.

The preamp is likely fine/working since I could at least hear taps when it was connected via a USB-C Dac (the one made for phones) but I got nothing more than that, No voice when I was speaking into the mic. I think adding 250ohms on the gain alongside the diodes should do something at least. :p

On that note, do you think something like this would work?

"USB Audio Video Converter, VHS to Digital Converter, Video Capture Card Digitize from Analog Video VCR VHS DVD, for Windows 7 8 10"

Ouch on the laptop front, that's a bind. My work laptop, the one that created almost all of the designs discussed here and home of my writing finally went to silicone heaven last week. The screen was cracked and the spacebar had almost gone but it was still trucking... and then the power supply decided it would play silly and ... blew the regulator. Thank goodness for the IT resellers, I grabbed myself an i7 based Lenovo which is now my pride and joy. Lenovo are known for their excellent keyboards so that part mattered more than performance.

Cheap laptops and desktops are a great way to test out stuff and you won't be crying in your beer if you break it! 🙂

I've actually used that exact digitiser myself and it's nowhere near as good as the Audiograbber Matt used in the original design. While it's a little more expensive, it's quieter and more sensitive and the only one I'd recommend.

If you wanted to go really daft (!) you could always go the Focusrite Solo route (if you can find an earlier model on eBay etc. they are reasonably priced now). I used a Solo to develop the Varee P48 capsule adaptor, which is based on Matt's original idea, but is capable of driving balanced lines many tens of metres in length. I'm hoping Matt can do a version of the mic with this one as it's my own favourite design - but with that level of quality and sophistication the cost rises quite steeply.

Focusrite Scarlet Solo 3rd Gen

So I got the diodes but I am not sure how exactly to wire them up. I did find and read the post you pointed to, it's just that I am not too great at reading the electrical circuit diagrams unfortunately :p. If we go off the schematic for the preamp, would you be able to mark where exactly the diode goes and how it connects?

I would really appreciate it, thanks!

ps: I have noted the focusrite as a later addon when my relative comes from the UK later this year. At least If I get this working :p

Edit: Do you mean to put the diodes in parallel in between GND of output and the resistor connected to the voltage amplifier? I am asking because I am just gonna solder the diodes on the back of the board so I needed to know where they go exactly

Yeah, not so easy on Vero but let's see.

I think I'd invest in a small selection of "heat shrink" tubing for this. This is a soft thermoplastic material that can be slipped over leads and even whole components and then tightened down by applying a little heat. (I use a soldering iron for close work, some people use matches but naked flames are not advisable.).

Since this is an "after market" bodge ( 🤣 ) it's not going to be pretty but we can make a reasonable job with a little care.

So first of (regardless of if you do insulate this) you're going to solder the diodes in anti-parallel (I think I've said antiphase before which is wrong). You can see from the diagram that it simply means the pair are wired in opposite directions.

You can check you've done this correctly with a multi-meter on diode test and you should see that you get a "diode" whichever way around the probes are.

When you've done that, I'd suggest you solder a about 5-10 cms (a couple of inches) of hookup wire - anything thin will do - to the the diodes and then insulate the whole thing idelly with heat-shrink but if you don't have any, electrical tape will do in a pinch. Sellotape and similar aren't suitable since they are too rigid and won't bind close to the assembly.

Either end of the adapter goes to any convenient ground. There's a ground connection on pin 5 of the THAT1512 so you can pick it up there or anywhere along that track.

The other end goes to pin 6 of the 1512 and you're done.

Here's a very rough indication of what I'm talking about. Now you don't *have* to use pin 5 - you can use a meter on "continuity buzzer" to locate another one.

Here's an incredibly terrible drawing but it's the best I'm capable of at the moment. :/

Dude! It finally worked! I ended up sticking one of the legs of the diode directly alongside the IC in the socket and soldered the other to a ground point on one of the resistors. It worked! Thanks a bunch.

It would be nice to have this point added in the google drive document so people don't end up breaking their devices :p

Apart from that, I have still been directly using the Samsung Type-C to 3.5mm adapter. It wasn't that loud when I tested with the Gain at 100ohms. (Pointing this out since you told me it might be unreasonably loud without the interface but it might be better because of the Samsung mini-DAC)

On that note, do you think the quality is sub-par because of that and if yes should I consider the dongle I sent you in the previous message? Do you think the circuit of that thing can be improved somehow?

I can't find the exact model of the one Tom used in his Video since I am in India so I was thinking of using a USB Microphone instead :p Do you think that would work reasonably well or do you think the preamp circuit would be integrated alongside the audio interface in those mics? Here are a few examples:

One is a generic usb microphone and the other is this JBL CSUM06 which is twice as expensive compared to the cheap one. Both have okay ratings. The JBL has 3.7 on Amazon and the generic one has 3.9. What do you think?

Thanks a lot again!

You mean Matt, not Tom. 😉 You're probably thinking of Tom's Hardware.

I take your point about the diode mod, but that's up to Matt and since we've got a V2 ready now (and the original is years old) there's not a lot to be gained. Hopefully people will see this lengthy discussion and see that there have been a lot of developments.

I've only seen the Audiograbber (silly name but that's what they call it) on Amazon. It's unusually good when it comes to this sort of adaptor, most are quite poor and some are utterly dreadful.

Ones that come with internal microphones will almost certainly be using die-only chips which are covered in a little blob of epoxy. CMI and I think Realtek produce these for mass manufacturing direct to the factories. They don't come in a little epoxy case with legs sticking out as we're used to for through-hole or even for SMD soldering. I've never taken one apart because I don't have any way to remove the epoxy but I would assume the die is stuck to the PCB and then hooked up by bonding wires direct to the PCB - and that assembly is covered with some epoxy.

Such devices are complete - they have everything from the JFET bias voltage (typically a couple of volts) to the digitizer and even the USB endpoint interface. I've seen a few in the wild and the only way we can use the genericised ones is to get one with a "Mic" input. CMI and Realtek devices will be very good by there's no easy way to know what you're getting.

I've done a simple variation Varee (code-named Sarah) which is specifically designed to interface a JLI2555 or 34mm LDC with its own JFET (in several different configurations) to either the original design or the Michelle variation. I'm going to do a Mini-Michelle if I get a few hours spare which drops the "fully isolated" professional section but keeps everything else, including the diodes and other improvements to the power supplies.

Sarah (and you can do what it does at the capsule if you're really handy with a soldering iron) will drive a normal microphone input directly. While you won't quite get the performance and "click gain" convenience, it's a lot cheaper.

Gain of 100R isn't that high actually,  - x100 or 40dB because the JLI2555 is less sensitive than many similar devices. This is where devices like THAT151x start to come into their own because you might need x100 (40dB) to make the mic loud enough to be usable. There's a certain amount of manufacturing variability in all devices and mic capsules are no different.

The actual gain for the specified 1512 (the earlier, cheaper 1510 is "louder") is given as:

Gain = 0.5 + 5000/Rgain (ohms)

So that works out (the 0.5 can be ignored at this sort of magnitude)

Gain = 5000/100 = 50.5 or about 33dB

The THAT1510, however has a different gain and produces x100 for the same resistor.

Ideally, you're going to need about 40-50dB of gain to get a decent level from the JLI capsule:

Doing a quick bit of rearranging, we can see for x100 gain:

Rg = 5000/Gain

So:

Rg = 5000/100 = 50 ohms.

For some arbitrary gain in dB (you wouldn't usually do it this way but for the sake of example) let's take 50dB:

gain (decimal) = 10^ (50 /20) = 316

Now calculate the gain resistor Rg:

Rg = 5000/316 = 15R8

The resistor value seem weird because we're working on a log gain scale, not a linear one. It's usually better to go in 3 or 6dB steps because that sounds more natural to a human ear.

But why 40dB (x100) anyway?

We can see this from the spec sheet. The sensitivity for an equivalent output of 1V per pascal:

That the mic may go from -45 to -39dB for a to deliver 10mV (for 1V output at the pre-amp).

We need to be able to deliver between 39 and 45 dBs of gain to any given capsule to deliver the same output. It doesn't matter for a single, desk mic but if you're doing any sort of multi-mic mix, that can make a considerable difference.

If you haven't worked in Pascals before, this might bake your noodle a bit.

Standard air pressure (at sea level) is slightly less than 30mm of mercury or 14.7 lbs/inch^2.

Which is 101,000 pascals pushing down on every square inch of you and everything around you (including the mic's diaphragm)!

Which also means that the mic is responding to changes in air pressure in millipascals, or billionths of pounds per square inch.

Oh yes, I meant Matt. No Idea where I got Tom from lol. About the diode mode, I was only talking about a disclaimer in the guide without instructions or anything so people don't damage their devices, nothing more.

I read it all but I did not understand about "Sarah" or "Michelle". I assume you are talking about new circuits that you have made but I would need to check them separately, I suppose they are there in the forum.

Thanks for the lengthy explanation on gain btw.

I will try and buy the cheap mic with the usb and see what it holds inside :p

Also, do you know if anyone has made a 3d printed version of this Mic? Since you are an active part of the community, I reckon you know more. Otherwise I shall be the first to try that out :p

Hey Marc,

An update. I tried different gain settings but the output is abysmally low. I tried 500ohms and even 2.2k Ohms and weirdly enough it sounded the exact same on all 3 resistances.

What do you think could be the issue?

I am still using the USB-C DAC Dongle. I do not think it's the issue because every other mic sounds fine. What do you think? Thanks a lot for your continued help! 

Edit: The diodes I used were not 1n4004 but instead 1n4007. Do you think that's the issue? I couldn't find 1n4004. I have the following: 1N4148 1N4007 1N5819 1N5399 FR107 FR207 1N5408 1N5822

Pretty much any GP diode will do, even a 1N4148 will be fine. In normal use they won't even operate, they're there as a type of fuse - or transient suppressor if you prefer - to stop very large voltage peaks which the THAT is capable of propagating any further and damaging the digitizer.

The output gain increases with small values for the gain resistor - so for max gain you'd use something in the region of 5 - 10 ohms.

Background noise is a possible (and won't necessarily show up) as is VHF "hash" (noise) which is another screening issue. If your screen doesn't cover the whole of the capsule and connect back to a ground, all sorts of nasty crud ends up and gets amplified, draining the amount of power the THAT has available to work with. You'll need an oscilloscope to see this though.

Even my own designs which are very heavily screened - on all four layers can't keep enough noise out unless they are screened at the head and up to the connection at the board.

It's also possible *while unlikley* that the JFET isn't self-biasing correctly. You can test for this by (temporarily) fitting the largest resistor you have to hand (at least 1Meg) between GATE and the shield. Varee and Sarah have a place for a 1G resistor or a smudge from an HB pencil (yes really) which works in the same way.

You can find ZIP files of all of my current designs (which may give you a better idea of where this ended up) here:

https://github.com/marcdraco/Sarah
https://github.com/marcdraco/Varee2
https://github.com/marcdraco/Michelle2

There's no documentation here (all I have is with Matt so he can do the video) so I can't any real support as such. There's a smaller version of Michelle in the works that doesn't have the professional level, isolated mic and balanced output.

And I think you'll be the first person to do a 3D print but remember that the shielding is essential.

I tested the gain again with a potentiometer. No changes whatsoever from 500ohms to the lowest resistance. Abysmal volume. 

As for the JFET, I used this video to test it and it seemed fine?

For your method of testing, could you please guide me a little but more? I am not too great at electronics as you might have picked up :p

Also what do you mean by the shield? The outer cover for the microphone? I have been testing without that for now… do you think that would cause the volume to be very low?

Yup, that's it. 🙂

You absolutely MUST put a screen (that's what the brass mesh and various gubbins around it) do. What you have there is (in essence) an AM radio aerial booster.

Without the screen, all sorts of radio-frequency "EMF" will get picked up by the microphone, buffered by the FET and then sent to be amplified. Which means that almost all of the audio signal from the capsule is overwhelmed by the noise.

Here's some more information -  but if you follow Matt's instructions he does detail how to make one to fit the capsule.

https://backyardbrains.com/experiments/faraday

I'm in the final stages of building the microphone! I have all the components and have started assembling them. Looking at my transistor, I'm wondering if it’s better to solder it directly onto the capsule like it was shown in the video or integrate it into my PCB design. What do you guys think?

Here's the PCB circuit

Thank you in advance!

Yeah, all of V2 designs have the transistor mounted at the capsule. The reason is that the output is very high-impedance and that means it's highly sensitive to external EMI (electromagnetic interference) through magnetic coupling. Any wires we add between the transistor and the capsule act like radio antennas so they shorter they are the better. The whole thing MUST be wrapped in a Faraday shield as Matt demonstrates in his video.

Just one question, you've got two caps in parallel at the output there. DC blocking is important (just in case, although there shouldn't be much).

But also note what I said earlier about GBW. With x470 gain you're going to lose all of your higher frequencies regardless of the setting of your volume pot.

To revise. You take the GAIN of your amplifier: 470 approximately and divide the GBW by that amount.

For the OPA2134 (considerably better chip than the little "jellybean" like a TL06/7/8x but it will still start to flatten out and fall off at just 16K. Now that's outside of out range of hearing but the loss of those harmonics does affect the sound. I'll be honest here, that part of the maths (Fourier) is above my abilities so I'm parroting what I've been told by people who do understand it.

For voice work that should be OK though. However you only need about x100 gain to bring the capsule to line level. x200 should be more than enough and you won't lose the high-end harmonics.

I can't read the value of those two paralleled caps so I'm not quite sure what you're trying to do other than a DC block.

Cheers,

Marc

@marcdraco I could not find brass/copper mesh anywhere and when I did it was at exorbitant rates. Stainless steel mesh was readily available but I had to find acid-based flux and silver tin to solder to it so I instead ordered a cheap BM800 condenser mic which I plan to gut out and place our capsule inside.

I suppose that should work..... but let me know what you think. Thanks!

Works a treat yes. The only issue I've found with the BM800 is that the XLR connector can be a bit flaky; the cable that comes with the kit is not balanced (even though it's XLR)!

However, for a cheap and sturdy Faraday cage which is where Matt's design excels is arguably THE most important part. If you don't enclose the capsule (and the JLI2555 was the worst offender, not necessarily a bad thing thing) all you'll end up doing is amplifying a lot of RF and you'll be lucky to get anything more than a nasty hum out of the the thing. I did a very simple PCB to make it easier to mount a JFET called "Sarah" which is at my GitHub. https://github.com/marcdraco and this can be connected to Matt's original design too.

For compatibility with the original layout, you'd replace C1 and C2 with "zero ohm" resistors (links) and take taps from the drain and source terminals per the diagram. A 1G resistor (or some pencil graphite) will be needed to make this work because it's unlikely there will be enough leakage on the board to bias the JFET. Sarah also has a "guard" trace around the gate terminal to reduce interference which can cause distortion. How effective it is... I don't know, but it's better to have it than not. It's more useful on Varee where it acts as a barrier for any neighboring fields cause by the 20dB gain stage.

For the brave (!) here's Dorit - (from "Little Dorit" by Dickens) which is a scaled down, simplified version of Michell that eschews the "isolated" power supply from that design and omits the balanced output for professional applications.

Additionally, there's no switch to change the power over. This version has through-hole resistors to configure the board either for Matt's original capsule design (or similar) -OR- for P12 which uses +ve/+ve phantom supplies. If polarized capacitors are used, it's up to the builder to ensure the correct orientation or use an unpolarized one for C5.

To keep costs lower, this one is only double-sided - some of the filtering can be omitted, but the actual cost of these components is so low that it doesn't reflect much of a saving.

KiCAD 8.x project here.

https://github.com/marcdraco/Dorit

Hi @marcdraco, I hope you are doing well.

I was FINALLY able to get the mic working at expected volume but it was not because of the faraday cage/mic body.

After adding the capsule with a 3d printed bracket in the BM800 body the volume was even lower.

The diode mod was the cause of the issue. As soon as I removed the diode mod, the volume issue was instantly fixed. The reason it didn't work before without the diode mod before was that I was supplying the wrong voltage (3.3v instead of 5v)

Now without the diode mod and 5V, it works as expected but I am still scared of damaging even more inputs :p

Could you help me with the diode mod, such that it offers protection but it doesn't mute the volume? I would really appreciate that! Thanks a lot!

/-That's weird.

Antiparallel diodes can only prevent an over-voltage to an input - in this case no more than about +/-0.7V  by shunting any voltage greater than that to ground. We use two like this with an AC signal as that can swing both ways. It's a protection system and shouldn't normally be conducting unless they voltage is getting to the point where it might damage the input transistors.

You can see this work on a scope (if you have one) or on a multimeter at a pinch (but to do that you really need a signal injector). Or you can do it in simulation with LTSpice, KiCad or even Paul Falstad's online sim. like this.

https://tinyurl.com/2dc48mxh which in this example the cutoff is about 600mV. YMMV depending on the exact diode you use. But it shouldn't affect the volume as such - it should work up to a certain level and then start to distort quite badly when the diodes start conducting - shorting the excess signal to ground.

May sound like an odd question by what value have you got the gain resistor set to? And are you using the digitizer Matt used (the Audiograbber)? The other ones I've tried don't seem anywhere near as sensitive. It's entirely possible I've missed something of course, the newest iteration of the head adaptor hits this week so I can try a whole slew of different devices with direct microphone ports.

One thing that springs to mind although I can't see why this would work, would be to use two pairs of diodes in series - two going in each direction so the clamp voltage is raised to roughly 1.2 - 1.4V peak which is more than enough to drive a line input and I would expect not enough to destroy it.

EDIT: does the THAT get warm when your mic has been running for a couple of minutes? It shouldn’t even feel warm to a fingertip. 

Hello all,

I am new to this but I want to get into content creation and I saw this video a while back. I finally have the money to do this and a decent skill set, but I am not exactly sure where to start. For starters, I am in Arkansas, US. And I have no idea where to get anything for this, but I want to do it soon. If anyone could help me at all, I would be very thankful. All I really need to know is where to get all the items for now. Please help!

- Nick Law

Hi Nick, welcome to the forum.

eBay is your friend (locally) or AliExpress.

The capsule is the part most people have trouble with, so I've experimented with a few different types. ANY "jfet-less" electret capsule will work but that's crucial to these designs.

There are three major types of condenser mic capsule:

Electret with FET - these are the most common ones and cost about the same as a thimble full of gasoline.

Electret without FET - this tends to be LDCs in the main that's the ones in the 25mm (1") and 34mm (1.5") types and this is the type Matt picked for the original picked for the original - JLI2555 - the number gives you the dimensions in millimetres - 25mm diameter, 5mm high.

Unbiased: these are the most expensive ones and REQUIRE a biasing voltage to work so they won't work with this project.

I know a lot of the audio perfectionists swear by the unbiased ones but for simple vocal work, a 25mm back electret LDC will perform very well indeed. The mass produced Chinese ones which sell for about £8-£10 in the UK are nasty looking things but are more sensitive than the (otherwise superior) JLI capsule. I've used them extensively testing the various adapters I've discussed in this rather long thread. 😉

The latest one (Jaime) just arrived from China and I'll be testing it as soon as I can get my hairy white ... out of bed. It's just bitter cold here and my soldering iron has little icicles hanging off it. This one is the result of a lot of suggestions from the community and might be the best solution so far since it outputs directly to a USB sound card (the type that come with microphone inputs on a TRS or TRRS connector). I think Sabrent do the cheapest one but I've also got a UGreen and a few others to try. This one is really intended for use in a donor body although a short run of screened cable should also work. It's just not as nice as the balanced output and Matt's rather gorgeous case.

I've heard some people are having trouble sourcing the THAT1512 IC - but you can use several others including the THAT1510 or any audio-grade Instrumentation amp (INA) of which there are a few. The gotcha is that they all have different gain configurations so you'll need to work out your own resistor chain if you went that route.

Pretty much everything electronic you need (including those) but not the capsule is available at DigiKey. The NMA0515 is a standard part so you can replace that with a couple of others also at DigiKey provided they deliver at least 1W and have + and - 15V outputs. These work in Matt's stripboard or my PCBs (Michelle/Dorit) without issue.

Finding the metal might be more tricky - I don't know what the supply of brass sheet etc. is like where you live but I expect Amazon will have something.