basically here post on how you made your mic ig? (and how the preamp works on all resistors capacitors to chips and inputs/outputs if possible) Might be helpful to others who looked the DIY perks video but couldn't understand everything cuz they are new or debugging what is wrong with their own mic. (I'll post mine when I'm done tho I haven't researched the pre amp too much yet) *after all there are some things we want to learn 🐒🧠
So I pulled off the lamp head, sadly I can't tell you how it'll go as I am still on the stage of getting the preamp done, stuck on the USB interface and USB C breakout board - posted about it on the main thread.
But the IKEA lamp should allow one to skip a massively expensive and hard step for the average person. The problem that may arise is that their isn't enough slack with rubber bands, or o-rings or anything. but can't comment on that.
I would love to see an update on your project. If you could help with the USB interface part that would be phenomenal too
@someguy Yep I'm kinda stuck on it too, but right now I've basically got everything down except for shielding, mic connection, and the USB C breakout board. I also am designing a case that I will 3d print later, but lately, I have not had much time to work on the project because I have a lot of school and robotics stuff going on. I will try to update periodically starting maybe next week. but I'm not sure if I can keep up with it. Also, I will not know a lot of stuff too. 🍌
Update: 2/20/2023 My mic still doesn't work, everything should be fine but the processing chip might be a fake. I've made a CAD for people who have their own 3D printers if they like to print out a microphone shock mount similar to Diy Perks brass one. I'll add adapters for 3/8 and 5/8 microphone arms later (really quick).
Update: 5/2/2023 My mic has been built I have had it working after I replaced the fake THAT1512 with ones I bought on Mouser, it works great but there is a slight problem with it. When I turn up my gain all the way to max I can hear a radio station below I will attach 2 audio samples when I get the time. (1 of maxed out gain and 1 of the radio station being barely audible) *I have not shielded the preamp board yet.
Max gain:
Adjusted gain:
I am also using a potentiometer instead of how DIY perks did it in his video. Is there a reason why he chose to do many resistors instead? (pls tell me if anyone knows) A mute switch will also be added to my mic but does anyone know where I should attach it too? the 2 places I can think of are the signal to the AD converter or the power of the board, if anyone has a suggestion pls @ me. (I wanted a mute switch because I when I'm gaming if my family walks in and they say something private)
@bio-s I think we are at the same spot, I have the same things left to do. Would love to continuously update each other.
Oh and figuring out how to do the connection for the rotary switch as well. I'll post some updates on the project today, to show you what I have so far
@someguy I have done the potentiometer already but idk if the audio chip only accepts a interval input (5,10,15,etc.), probably not cuz that would be very un-convenient to use.
So I pulled off the lamp head, sadly I can't tell you how it'll go as I am still on the stage of getting the preamp done, stuck on the USB interface and USB C breakout board - posted about it on the main thread.
But the IKEA lamp should allow one to skip a massively expensive and hard step for the average person. The problem that may arise is that their isn't enough slack with rubber bands, or o-rings or anything. but can't comment on that.
I would love to see an update on your project. If you could help with the USB interface part that would be phenomenal too
The microphone is a sub-device if I'm not wrong. I found a USB c breakout board that is a sub-device here: https://www.amzn.com/dp/B096M2HQLK/
This post was modified 11 months ago 2 times by BIO.S
Check if your ground cable is touching any other wire on the output so it might be a problem with the audio out (to the A-D converter).
The audio out cable has 3 wires red = right channel, white = left channel, and black = gnd. Idk if you need to combine the red and white wire to have an equal output but if you do I think it will make it a bit louder?
Another thing is to just turn your gain up but you've probably thought of that already.
^thats what I think anyway
Also if you did, what desoldering wire did you use to make a copper cover for the thin copper wires? I don't know how to open the wires they all seem braided but not squashed.
This post was modified 11 months ago 9 times by BIO.S
That is phenomenal! So I took a look at the wiring and it seems fine, I typically use electrical tape around each wire to ensure they don’t short each other.
here are images:
Now something that is important to note is that I am using the J111 transistor on the capsule - I have ordered the recommended transistor to test if that is actually the issue.
for the case grounding does that mean I need the capsule grounding wire to be attached to the copper mesh I’m using for the case or does that mean I need the edges of the circuit board to be touching the copper mesh inside the case?
That is phenomenal! So I took a look at the wiring and it seems fine, I typically use electrical tape around each wire to ensure they don’t short each other.
here are images:
Now something that is important to note is that I am using the J111 transistor on the capsule - I have ordered the recommended transistor to test if that is actually the issue.
for the case grounding does that mean I need the capsule grounding wire to be attached to the copper mesh I’m using for the case or does that mean I need the edges of the circuit board to be touching the copper mesh inside the case?
Not for the pre-amp but for the mesh around the capsule I think so.
In this image of the video it looks like there is a bit of solder on the back of the mic joining the ground desoldering wire to the metal cage encasing the microphone capsule, does it have to be connected like that?
Also what desoldering wire do you use? I bought some and they were all just flat braids and not pre-squashed tubes.
@g3nts Yes it should be connected like that since it is the same as connecting the ground wires to 2 flat pins on the back of the mic capsule, and if you connect your voltmeter to the mic shell and the ground tab they are connected.
V2 of this design is just about to go into production - Matt and I have been at it feverishly for a couple of months and the result is beautiful and mostly pre-assembled at the factory.
@BIO.S, to answer your question about the THAT, internally it's a three-amp instrumentation amplifier. The resistor can be variable, the reason Matt used stepping was to fix specific gains based on the calculated values and also to lock in that gain so you don't inadvertently move it during or between sessions. The design has a discrete version of the THAT (pre-soldered) and space for a THAT too. But this is what the amplifier section looks like with the calculations for voltage gain. Remember that you need to convert this into dB if you prefer that measure but to keep it easy, 20dB = x10, 40dB = x100 and 60dB is x1000.
One of the gotchas with discrete design is the GBW of the amp. A 10MHz amp like a 5532 which I've used here can't reach anywhere near the full audio range if you push it to 1000x in a single stage.
Also desolder braids are always flattened tubes. You have to open them up with a knitting needle or something like that but some (I have a few different ones) are easier than others.
Take everything I say with a pinch of salt, I might be wrong!
@marcdraco Thanks for the reply, some cheap desoldering braids I bought off amazon that were made in china didn't open up, I even asked one of my family members if they could help open it up at the company they work at, they brang home an example of one that the technician opened which was made in the USA and he was unable to open the one I had purchased off amazon. I went to a local electronics store to find a desoldering braid that was made in the USA and that one could open up.
As for electronics, I honestly do not know what im doing other than connecting wires, I have a few questions about how the Gain works.
1. If I increase the resistance does it make the volume louder or quieter?
2. If increasing the resistance will lower the volume of the microphone can I attach a mute switch to the potentiometer?
3. If I cant connect the mute switch to the gain where else would be a good place to connect it?
I think the problem with braids is the way they are made - the cheaper ones are especially bad in this regard. When I was using them back in the 1970s they even came with a flux! Those were the days.
Electronics is largely about mathematics - and some physics and a bit of logic. So it's no wonder a lot of us struggle. Most people are still (to this day) told that current travels through the wires. It does at DC but as soon as the frequency increases it moves through the insulation ... it makes it hard to design decent boards for high-speed logic unless you're very careful with the routing.
To your question.
In the circuit above, gain is set by a resistor between the terminals in J9 as (R10 + R11)/R (the one in the jumper).
Op amps are fairly predictable - they were originally designed for math operations before we had digital computers. In fact, analog computers are still used to this day for some extremely difficult computations that work in the analog domain almost instantly (due to the fact that electrons move through the copper at 2/3 light speed).
Just to run that calculation for you
R10 = R11 = 1K = 1000 ohms -
So R10 + R11 = 2000.
Divide that by gain resistor (this is for my circuit, not the THAT) and you get the specific multiplication factor.
Let's say that was 100 ohms that gives us:
2000/100 = 20.
So we have a gain of 20 times.
In decibels that works out like this: 20 x LOG10(20) = 26 decibels.
TL;DR: - in essence the larger the resistor is, the softer the sound is.
What's important to know here is that the microphone capsule only generates a relatively small signal. The new impedance matching heads I've designed with Matt (we have a new prototype going to the board shop soon) are hugely efficient but still only deliver a few 1000ths of a volt to the pre-amp.
Line input for consumer products expects about 1V for maximum "volume" (0.71 RMS) but that means we need to make the signal as much as 1000x LOUDER just to reach the signal level expected by the digitiser.
In my experiments we needed 40dB (1000x) to get this design to operate effectively at normal speech levels and deliver a good quality output to the ADC - that was expected based on the specifications from JLI but I'll not fill you on those details as we're into sound pressure and that makes most of us glaze over a bit!
TL;DR2: A 20dB (which is the usual "pad" or attenuator on professional microphones) cuts the sensitivity by a factor of 10. This is the difference between speech and a rock singer. You could "cut" the signal from the microphone by dropping it across a resistive load but in reality you only need to set the application factor to around 20dB (10x) in total to effectively mute the mic short of sound pressure that would deafen most people.
For the THAT1512 which has a gain equation of 0.5 + 5000/R-gain. So a gain resistor of 5000 (5K1 is the closest easily available value) is more than enough to reduce the gain to the level where it's effectively muted.
I'll look to put a mute switch or mutable value into this design though. I've used a couple of tricks to extend the frequency response at very high gains with a popular op-amp. The board also supports the THAT1512 though so we have an upgrade path. I don't want to pull the covers off just yet as Matt hasn't seen the final design and since his name is on it, I'm reserving the right for him to make changes.
To whet your appetite these are some of the most recent prototypes. This particular one uses a pair of 2SK170 low-noise JFETs but the final one (I've done two different "final" designs from a bunch of experimental prototypes, but more of that later). I thought the noise performance on these was "me'h" but that's largely due to the pre-amp I was using at the time and could have been affected by the general purpose transistor I used for the phase splitter. Like Matt, I'm something of a perfectionist and we both recognise that this beautiful microphone can compete with some of the best sub-$1000 mics.
In order to do that (which I can't with the JLI2555) it meant designing the whole thing from the ground up. We threw away our design books and ran hundreds of simulations to get an affordable solution that sounds as absolutely beautiful as it looks. The one thing I don't control is the JLI2555 which is a pre-charged (electret) condenser and they tend to be a little noisy. Phantom powered capsules are a lot quieter (>30dB) but right now I've got something that supports the original components.
The main takeaway that I've fixed for the production designs is a jumper block so you don't have to solder the wires to the board which is, to put this mildly a bit tricky. As are the pads for the 2SK170s - another thing I've fixed in the new one which features one of the industry's lowest noise JFETs, the LSK389 (or it's brother, the LSK489) https://www.linearsystems.com/_files/ugd/7e8069_580dfb20b43a409abf5194e309102e14.pdf . Now these aren't the quietest FETs you can get but the next rung up the ladder comes from InterFET and their prices *start* at over $50 a piece and you have to buy, I believe, 10 at a time. They're intended for military use where the signal is amplified by factors that make my brain hurt. The other option comes from Texas (TI): https://www.ti.com/lit/ds/symlink/jfe150.pdf but it's only available as SMD (surface mounted) and we wanted to make these boards with as little soldering as was practical and make it all "through hole" because the components are easier for the average dude/dudess to handle. I've worked at this (on and off) for many decades and even soldered under a microscope but no one needs that harassment when you can have it done at a factory!
It's been a few months in development but I'm fairly confident everyone will be delighted with the result. I expect they will be sold in kit form, mostly pre-assembled but won't cost the earth. A couple of parts are optional and interchangeable so (for instance) you can use a "quieter" op amp than the NE5532 in the high-pass filter and boost circuits. The 5532 is already a professional grade amp, but TI (particularly) do a bunch of others which outperform it in most key areas.
It all comes down to the signal from the capsule though - which is why we've expended a lot of effort (Matt sourcing, me designing) in some designs that really squeeze every last ounce of signal out of the capsule while adding as little electron noise (Johnson-Nyquist, popcorn, etc.) as possible. My gut says the v2 Cerberus (it has three transistors so named for the three headed dog) will be the quieter one but the Hydra should have lower distortion. Both are pretty quiet though - on paper at least.
I'll write all of this up, in "less testicle terms" (so it's not such a ball-ache 😉 ) so everyone can understand our design choices. Matt is a master at finding modern, high quality components - whereas I'm still stuck in the Stone Age. I still think a MOSFET is a pretty neat idea and have (honestly) made circuits with germanium transistors in the days before silicon was a thing and MOS/IGFETS were an upcoming technology - for home experimenters at any rate.
Together we've come up with something quite beautiful if I do say so myself.
Take everything I say with a pinch of salt, I might be wrong!
Your post about your mic creation process is like giving others a guided tour of a complex city using a taxi. Sharing insights about the mic's construction, including details about the preamp and the roles of components like resistors, capacitors, and chips, can be a beacon for newcomers who might have found DIY perks videos challenging to grasp fully. By shedding light on these intricacies, you're like a taxi driver navigating through the maze of electronics, helping others troubleshoot their mic issues or grasp the basics. We're all eager to learn, and your contribution can be a valuable map for fellow enthusiasts.
Looking forward to your post and the collective knowledge it will add to our DIY journey!
Thank you for your kind words @gossi123 it sometimes does feel like navigating a corn maze by moonlight though.
So everyone knows this is a developing project (much already works) there are some more developments on the "head" end which are notable. This 3D rendering of a near final board(s) gives you a better idea. All through-hole (DIP, TO-) devices are home-solderable or can be fitted into sockets. Many benefit from the socket since it allows a better "breather" space for the rear of the capsule.
There are still some changes I need to make here as this is WIP.
The schematics are not that special (although I will include everything when it's settled) but what I've done for this batch is standardised the header positions for most of them so that all but the very special ones can be run from 48V phantom power using a stacked board with the power regulator and high-voltage capacitors.
I've used this set to include some BiFET amps like the OPA2134 which has a low-noise FET front end, low-noise and low distortion and a FET conversion for a standard bipolar amp. This can be any standard 8-pin dual audio op amp and delivers balanced output.
I've really had to "jump on the suitcase" and gone a bit crazy with the last two which mate two different FETs (2SK208 and LSK170) with a THAT1512... the very device that Matt used on his design only this time, the whole thing is on the head!
The THAT needs a split supply to operate and the usual trick of biasing one of the inputs doesn't work. Which means it means it needs a bias voltage on a pin specifically designed to alter the ground reference. We'd usually do this with voltage divider impedance of 100K or more so it doesn't drain current. But the ground ref pin on the THAT (and other instrumentation amps) requires an impedance of (ideally) less than 1 ohm.
To do that you need a reference and the easiest (most stable) way to do that is to drive it with an op amp. I can't stress enough that this is still experimental and I haven't built these yet - although they do work on paper and in simulation. This is the
