DIY RF Condenser Mics

[Tubetec]

Thanks for the reminder about the Philips design Rogs ,I seemed to have missed it .

I noted he specifically makes reference to closed core inductors ,
I see the construction of the spectrum coils ,
Im no expert but the U shaped core doesnt look closed to me ,

I did a have a look at available pot cores, nothing quite as small as what Sennheiser use and mostly non adjustable types with no bobbins and no mounting hardware.
The smaller diameter cores also tend to have a smaller mounting hole , the pot cores in the Sennheiser mics are held together and mounted on the pcb with a thin layer of contact adhesive .

In a way were trying to take pot shots at a target in the dark unless we have more info about the Sennheiser inductor , the properties of the core material are unknown , its closed core , but is it gapped or ungapped we dont know that either ,

https://www.surplussales.com/Inductors/FerPotC/FerPotC-1.htmlThey look a lot like whats used in the Arends reference ,
Further down on the page is a ferrite cap core ,
similar in construction to the Spectrum cores .

 

[rogs]

I see the construction of the spectrum coils ,
Im no expert but the U shaped core doesnt look closed to me......

I think it's fair to say that the Spectrum style coils -- which are of course conventional IF transformers - are essentially of a U type ferrite contruction.
In this application, the transformer is enclosed, with regards to having an electrostatic shield, with the PCB ground plane underneath the external aluminium enclosing can completing the enclosure.

But the base of the transformer is still 'open' from an electromagnetic point of view. As I say, pretty standard IF transformer construction.

The Spectrum coils are fairly simple in construction - with only 2 windings - and it was a fairly straightforward task to take one apart and remove the windings, to see how it was put together.
Using the structure I described in my notes HERE it has been possible to copy the Spectrum coil pretty accurately.
It did of course require the destruction of an example of the original, to get at all the details.

With the Sennheiser inductors that 'destruction' approach is not really an option of course.
Without knowing all the details of the ferrites involved - and how the windings are actually applied - I think it would be a much more difficult task to copy the Sennheiser inductor(s).

 

[Tim Campbell]

I came a little further with this push/pull capsule design. I am really happy with the off axis response

 

[jarvis]

Hey Tim,

first I was a little confused about how you went about constructing this but now I think, I beginning to figure it out. Is the bottom half one of your CT12 capsules with a second backplate in between the usual clamping ring? That's actually genius!

I think, I'm going to start with an edge terminated K67 and try the same with my, admittedly, pretty crude tools. You just isolated the backplate from the clamping ring with glue, right? At least that's what I wanted to do with the other design I was working on in CAD. Did you machine the distance of the diaphragm to the backplate in or did you do it with a (conducting?) spacer in between?

Best
Jannis

 

[Tim Campbell]

I put a center insulator into the rear backplate, stripped the gold from the center of the membrane and used plastic insulating shims to raise the front backplate. The front and rear backplate are insulated from each other but in this photo they are both connected to ground with the membrane at 70v

 

[jarvis]

I put a center insulator into the rear backplate, stripped the gold from the center of the membrane and used plastic insulating shims to raise the front backplate. The front and rear backplate are insulated from each other but in this photo they are both connected to ground with the membrane at 70v

Thanks for the insight, Tim!

Okay, since I have a couple of MKH mics, I'm going to use their preamps to verify if my capsule works, once I have the time to build it. So the backplates have to be isolated either way, since they are going to be polarized and the diaphragm is ground. But I'm not sure about the insulator in the backplate and the stripping of the gold in the center of the diaphragm. What purpose does that serve in your case? The diaphragm is terminated at the clamping ring, correct? How is that connection made when there's an insulating spacer in between the ring and the diaphragm?

Best
Jannis

 

[Tim Campbell]

look closely. The front backplate is floating above the membrane by a center screw. In the Sennheiser the backplates are polarized differently and so must be isolated

 

[jarvis]

Oh, it's actually floating! I really thought that you glued it to the clamping ring!

Best
Jannis

 

[rogs]

As I have mentioned on several occasions in this thread (and elsewhere! ) I am no mathematician...
The calculations required to specify the inductors for this type of project are way above my pay grade.....
However now that we have located alternative coil formers it should be possible to escape from the limitations of the single 5.3uH transformer we have used to date........

Out of curiosity more than anything, I thought I'd ask the free A.I. app 'Chat CPT' about inductance calculations?.....
I have discovered using it before that - although it can be useful in pointing you in the right direction - its maths can be quite suspect.
I've posted a PDF of my exchange with the A.I. 'bot' here:

https://www.jp137.com/lts/AMX.inductor.calcs.pdf
Quite useful potentially? --- I hadn't realised that it's only primary winding turns that affect the most important inductance calculations for a transformer...
As in this project application there is essentially no current flowing in the secondary windings, the number of turns of the primary winding may well be the dominate factor, when it comes to calculating alternative inductance values?
Certainly my initial experiments with using a 6 turn rather than 4 turn primary has produced an inductor that measures something like 11uH experimentally - as predicted (finally!) by the bot calculations

Of course the A.I. calculations may not be that accurate, but they do seem to make some sense, looking at alternative sites for this type of calculation......

As A.I. improves, I suspect answers to this type of question may get more accurate? ...... It's both quite exciting, and a little bit scary!

 

[kingkorg]

As I have mentioned on several occasions in this thread (and elsewhere! ) I am no mathematician...
The calculations required to specify the inductors for this type of project are way above my pay grade.....
However now that we have located alternative coil formers it should be possible to escape from the limitations of the single 5.3uH transformer we have used to date........

Out of curiosity more than anything, I thought I'd ask the free A.I. app 'Chat CPT' about inductance calculations?.....
I have discovered using it before that - although it can be useful in pointing you in the right direction - its maths can be quite suspect.
I've posted a PDF of my exchange with the A.I. 'bot' here:

https://www.jp137.com/lts/AMX.inductor.calcs.pdf
Quite useful potentially? --- I hadn't realised that it's only primary winding turns that affect the most important inductance calculations for a transformer...
As in this project application there is essentially no current flowing in the secondary windings, the number of turns of the primary winding may well be the dominate factor, when it comes to calculating alternative inductance values?
Certainly my initial experiments with using a 6 turn rather than 4 turn primary has produced an inductor that measures something like 11uH experimentally - as predicted (finally!) by the bot calculations

Of course the A.I. calculations may not be that accurate, but they do seem to make some sense, looking at alternative sites for this type of calculation......

As A.I. improves, I suspect answers to this type of question may get more accurate? ...... It's both quite exciting, and a little bit scary!

Wow, awesome!!!!

 

[gyraf]

It's not so much that the primary inductance has priority over secondary, but by simple naming convention "5.3uH transformer", we implicitly agree that we are quantifying by primary inductance, and that the secondary will follow naturally

/Jakob E.

 

[Tubetec]

Theres something about the A.I.s response thats a bit spooky ,
its a bit like HAL9000 .
Funny the way it needed a good kick in the butt and made re-check its answers and it comes back with a groveling appology , but this is a machine , dont let it tug on your heart strings .
I cant help but to think this A.I. is little more than a parrasitic organism , a word jumbler with the capabillity to mislead you if you allow it , if you catch it out it just goes back and 'recalculates' Sheldon style .

This article just showed up in my local newspaper ,
https://www.irishexaminer.com/business/economy/arid-41204116.html

 

[gyraf]

..no, it's just hallucinating - it dreams up content that sounds most right at a quick glance, has no interest at all in being correct or factual - that is simply against it's basic idea, being a language model..

 

[cyrano]

It might surprise you, but AI doesn't need much computing resources to apply logic. It does need a lot of CPU for things we humans do instantly, millions of times a day: recognising vision and auditory input, including typed text. Schematics are part of the visual processing and as such will never be perfect in the current instance of GPT. But it isn't very hard to build and train an AI to do electronics.

You really need to take that into account when dreaming about useful applications.

 

[kingkorg]

Hi @rogs
Today i got to play with different frequency crystals, my new oscilloscope with FFT, and my custom inductor, and MKH 50pf capsule.

When you refer to highest Q, what do you mean? Is it the shape of the RF signal curve in a typical sense like with Eq Q, attached image? Or is it just the voltage of the RF signal oscillator puts out?

When i look at FFT on the oscilloscope i get similar shapes (Q) with several crystals ranging from 4Mhz to 15Mhz as long as i adjust the inductors for max output. I do get some variation in voltage it puts out. I'm doing this for the first time, so just trying to figure out if i'm doing it right.


Also when i adjust for maximum sensitivity, with some of the inductors, with very low signal coming from the headphones exciting the capsule i can clearly hear some harmonics appearing, and im sure it's not the rest of the circuit distorting as it's quite low level. Could it be it's oscillator signal folding back to audible range? I can also see on the FFT harmonics above the crystal main frequency.

This project is so fascinating to me.

 

[rogs]

Hi @rogs
Today i got to play with different frequency crystals, my new oscilloscope with FFT, and my custom inductor, and MKH 50pf capsule.

When you refer to highest Q, what do you mean? Is it the shape of the RF signal curve in a typical sense like with Eq Q, attached image? Or is it just the voltage of the RF signal oscillator puts out?

When i look at FFT on the oscilloscope i get similar shapes (Q) with several crystals ranging from 4Mhz to 15Mhz as long as i adjust the inductors for max output. I do get some variation in voltage it puts out. I'm doing this for the first time, so just trying to figure out if i'm doing it right.


Also when i adjust for maximum sensitivity, with some of the inductors, with very low signal coming from the headphones exciting the capsule i can clearly hear some harmonics appearing, and im sure it's not the rest of the circuit distorting as it's quite low level. Could it be it's oscillator signal folding back to audible range? I can also see on the FFT harmonics above the crystal main frequency.

This project is so fascinating to me.

In this context I'm using the 'Q' factor to describes the 'goodness' of the resonant circuitry around the inductor assembly..... A high 'Q' inductor configuration can effectively supply 'noise free voltage gain' (sort of! )

Actually, Wikipedia describes 'Q' much better than I can .... See here: Q factor - Wikipedia

The capsule movements we're talking about here provide tiny changes in capaitance, so we need all the help we can get, to amplify those signals (without adding too much noise if possible ).

The whole AMX inductive assembly -- with the primary of the second transformer connected across the bridge - is not easy to analyse mathematically.
There are too many unknown variables to permit a sensible model to be set up. Abbey did some simulations back in 2019 (around post #257 of this thread as I recall) but he concluded that the unknowns made his model less than ideal. As I recall, what it did show is that under certain conditions -- with the inductors tuned for maximum 'Q' within the whole assembly -- there were some pretty high gains available from the inductor assembly. Really sharp boosts (and therefore gain) at resonant frequencies.

So the RF output of the oscillator itself remains constant. What does change is the signal level presented across the bridge when T1 is 'tuned'. The 'Q' of the tuned inductor can create in significant gain of the RF carrier.
T2 has a fixed capacitor across its secondary, That is chosen to optimise the tuning of T2 in conjunction with the chosen oscillator frequency, and the value of the inductor ( This calc can be useful : Resonant Frequency Calculator )

What makes things difficult is when those 2 inductors are connected as we have them here..... It all gets very complex, mathematically.
What happens is that you can get a very good result, signal quality and noise wise.
But there are lots of variables --- capsule capacitance, oscillator frequency and amplitude, inductor values, bridge detuning etc etc -- that make it very difficult for the simple hobbyist (like me! ) be able to make definitive comments on what works best.

The Spectrum inductors were specified around 7MHz. They seem to work well at 8NhHz, even 10MHz with high 'Q' result still possible.
Above that frequency, I found the RF voltage available from the resulting lower 'Q' of the inductors started to drop off a bit. Not 'drop off a cliff', but get slightly less 'good' )

The further you get from the specified frequency for this type of indutor assembly, the lower the 'Q' -- and the lower the amplitude of the RF signals derived from the same amplitude source.

Regarding your point on RF 'foldback' --- I've not come across anythng like that?..
The time contstants used on the source and drain of the JFET are in the audio range. Way below the RF carrier frequency? ... I have to confess I have no ideas on that one, at present....

Now were starting to try out different inductor values, I suspect we'll open the flood gates to all sorts of possible variables. Some of them may prduce really good results --- other won't! .
I've made around 30 different version of this mic circuit over the last 4 + years -- and I'm still finding surprises. There is one problem I find makes things more difficult -- not being a graduate engineer!

One other aspect that makes formal calculations even more difficult is that tuning the inductors for best results can only be done with the headbasket on. The RF field - although quite small - is affected by the metalwork of the headbasket.

Just one more thing that makes this project 'experimental'..... I imagine it could of course all be calculated 'theoretically' - but I'll bet it never is!

 

[gnappoman]

My dearest microphone nerds (specifically targeting @rogs ) I am sorry having to point the obvious but at this point someone has to do it, so here I am:

Why On Our Beloved Planet Earth, after all this work of producing pcbs, testing capsules, hasn't anyone yet started a Kickstarter or Indiegogo campaign about the RF.AMX10? Do it before someone else copies the idea and reaps all the rewards of your hard work! Crowdfunding is also for people like me who would like to contribute to the project with their own money and wouldn't mind receiving a good microphone in return.
PS Please make a USB version as well, for us regular folks, and you'll have eternal glory (plus A TON of money!) I promise you! Kisses.

PPS as funny as it may sound I was brought here by this youtube video!

 

[rogs]

As I have made clear on several occasions, this is really just an experimental project.
Just my offering to Gerard's original post here, asking if anyone had done an RF mic, since Uwe Beis' project in 2015.

It's been a fascinating hobby project 'journey', with lots of input from some very experienced members here.
Lots of trial and error - Some good ideas, and some not so good ideas...

As for turning it into a commercial project? --- well I personally have no intentions in that direction.
But all the information is out there, both on my project website and on here, so if someone wants to have a go using my circuitry?...
I for one would be flattered!
I remember a comment some time ago - (I think on this thread, but it's 500+ posts and over 4 years, so I could be wrong!
Someone was surprised we hadn't seen either a Russian or a Chinese version already ...
But we haven't! .. ..( I suspect those pesky inductors are hard to source worldwide! )

 

[kingkorg]

This circuit is not an easy copy, paste cookie cutter type product for a company to just rip off. Product would require way too much QC, and manual tinkering for usual suspects to be interested. I don't think it will ever happen. If there were interest, they would have already copied Røde, Sennheiser...

 

[Coco Sloan]

Just check what this Gigachad did