DIY RF Condenser Mics

[Khron]

Duly noted re: the revision renaming.

The three-position on-off-on switch idea threw me off.
Surely, if the attenuation can be achieved by only connecting a resistor from the T2 secondary (ie. the JFET gate) to ground (ie. the other end of the T2 secondary), an SPST on-off switch should suffice, right?

On-off-on = middle does nothing, clicked in one direction does nothing, clicked in the other direction activates the pad. Doesn't make that much sense, does it?

And re: resistor value, like i said, i figured "split the difference", if 1k = -20dB and 4k7 = -10dB... 2k2 should then be somewhere in between

Most of my devices have just a single pad... Usually -10dB - (although my Sound Devices unit has a -15dB pad which is quite useful)... I must check, but I'm guessing your 2k2 resistor will probably be about -15dB?..

I'm wondering if it would be a good idea to change the board ident to 'RF.AMX9'  as well ...... otherwise we shall have 2 versions of RF.AMX8 --- and that's probably not a good idea?.... 

 

[rogs]

The three-position on-off-on switch idea threw me off.
Surely, if the attenuation can be achieved by only connecting a resistor from the T2 secondary (ie. the JFET gate) to ground (ie. the other end of the T2 secondary), an SPST on-off switch should suffice, right?

On-off-on = middle does nothing, clicked in one direction does nothing, clicked in the other direction activates the pad. Doesn't make that much sense, does it?

I was thinking of a switch like model 1MS3 on page 2 of this data sheet......  http://www.farnell.com/datasheets/1935700.pdf
So - pin 2 (common) is connected  to ground, and pins 1 and 3 are connected to the free end of each resistor (other end of both  resistors are connected to the gate).
Then the centre position connects neither resistor to ground - and both 'On' positions each connect one of the two resistors to ground....
Those switches are quite cheap too. ... (Strangely enough the double pole versions are even cheaper than the single pole??)

 

[abbey road d enfer]

The three-position on-off-on switch idea threw me off.
Surely, if the attenuation can be achieved by only connecting a resistor from the T2 secondary (ie. the JFET gate) to ground (ie. the other end of the T2 secondary), an SPST on-off switch should suffice, right?

On-off-on = middle does nothing, clicked in one direction does nothing, clicked in the other direction activates the pad. Doesn't make that much sense, does it?

An on/off/on switch would allow two different attenuations like 10 and 20dB.

 

[Khron]

I had asked for confirmation about that hunch, but i got none

An on/off/on switch would allow two different attenuations like 10 and 20dB.

Is that for a two-step pad (-10dB and -20dB)?

Most of my devices have just a single pad... Usually -10dB - (although my Sound Devices unit has a -15dB pad which is quite useful)... I must check, but I'm guessing your 2k2 resistor will probably be about -15dB?..

 

[rogs]

Ahhh .. my fault .. sorry! 

Ricardo had mentioned attenuating by both -10 dB and -20 dB. That would require 2 resistors of course -- and an 'on -off - on'  switch similar to the  type I linked to.

When you added a single resistor to the PCB layout, I then mentioned - in passing - that most of my devices  actually had only one pad --
Many others have 2 of course. I think that is where I confused the issue.

Two is more comprehensive of course -- and it's only one more resistor, and not much (if any) extra cost for the switch.
If it's easy to do -  two resistors and a 3 way connection header sounds like a good idea...

Apologies again for the confusion..

 

[Khron]

How's this, then?

 

[rogs]

Looks good! ----

 
I've just had confirmation from DHL that my RF.AMX8  PCBs from JLC should be here tomorrow.....

I'll be trying one out as soon as they get here 

 

[Khron]

Next step, figuring out how to do multipattern? ;D

 

[cyrano]

Next step, figuring out how to do multipattern? ;D

Yes, yes, yes...

Please, please, please  ;D

 

[rogs]

Next step, figuring out how to do multipattern? ;D

What happens if you were to fit a double sided capsule across T1 secondary, with  the centre terminal connected to  T2 primary, and perhaps with a small capacitor across one side of the capsule to unbalance the bridge a smidgeon?...

I've no idea what would happen -- just saying? 

 

[Khron]

I would imagine that should be easy enough to test out with the soon-to-arrive boards? :

What happens if you were to fit a double sided capsule across T1 secondary, with  the centre terminal connected to  T2 primary, and perhaps with a small capacitor across one side of the capsule to unbalance the bridge a smidgeon?...

I've no idea what would happen -- just saying? 

 

[rogs]

I would imagine that should be easy enough to test out with the soon-to-arrive boards? :

Would be -- if I had any double sided capsules! .. 
(I have a couple on the way from wgtcenter)

 

[Gerard]

How will you combine the inputs from the two capsules? I imagine anything other than a straightforward placing of one capsule on each half of the centretapped coil, with or without one side inverted, will affect the Q significantly.

Just a thought to get you all thinking before Rog's capsules arrive... or before I order some of the appropriate PCBs from JLCpcb

 

[abbey road d enfer]

What happens if you were to fit a double sided capsule across T1 secondary, with  the centre terminal connected to  T2 primary, and perhaps with a small capacitor across one side of the capsule to unbalance the bridge a smidgeon?...

I've no idea what would happen -- just saying? 

I think it's time to do some math instead of shooting in the dark.

 

[rogs]

I think it's time to do some math instead of shooting in the dark.

I've been wondering if someone is looking to take up that 'mantle'! ....

I understand there's lots of expertise on this board - Maybe one or two of the more expert  members might care to join in with some serious engineering advice on how  best to move the project forward to the next stage?...

There's only so far 'techs' like me can go, with just instinct and 'hands on' experimentation....

But in my own defence, it was only ever intended to be an experimental 'hobby' project - as I made clear in my early notes...

 

[abbey road d enfer]

I understand there's lots of expertise on this board - Maybe one or two of the more expert  members might care to join in with some serious engineering advice on how  best to move the project forward to the next stage?...

At the moment I'm on a trip but soon I'll be home and have some time to do simulations.

 

[Gerard]

Rogs, we are in awe of how much you have done and achieved with this project. Without your enthusiasm and skill it would still be at the same stage as the thread on the Mic Builders forum, going nowhere. It has been successful beyond my wildest imagining, even at this early stage. Thank you. I hope you keep developing this and that others can contribute the heavy lifting of mathematical theory and simulation.

And that's not to ignore the PCB designs from Ruud and Khron and the experienced suggestions from Ricardo and others. Thanks, guys. Keep it up. This project already knocks the Alice design into a cocked hat; let's make it into a really great DIY microphone project.

 

[Khron]

Oscillators and RF are way out of my league, but i do like to think i can whip up a half-decent PCB layout every now and then ;D

PS: To be fair, i believe the Alice merely was a lightly modded Schoeps circuit, so... :

That being said, i'm also looking forward to hearing from the more educated and/or experienced among us, on this 

Rogs, we are in awe of how much you have done and achieved with this project. Without your enthusiasm and skill it would still be at the same stage as the thread on the Mic Builders forum, going nowhere. It has been successful beyond my wildest imagining, even at this early stage. Thank you. I hope you keep developing this and that others can contribute the heavy lifting of mathematical theory and simulation.

And that's not to ignore the PCB designs from Ruud and Khron and the experienced suggestions from Ricardo and others. Thanks, guys. Keep it up. This project already knocks the Alice design into a cocked hat; let's make it into a really great DIY microphone project.

 

[rogs]

And that's not to ignore the PCB designs from Ruud and Khron and the experienced suggestions from Ricardo and others. ....

...Many, many  thanks to Ruud and Khron... an absolutely vital part of the project..... and to Ricardo with his expert suggestions.

And to you of course ... for having another try at getting this project going, after the  not very positive response you got from the  earlier Micbuilders thread!

I'm hoping to receive some of Khron's version 8 PCBs tomorrow, and I'm intending to build a couple up straight away...
Hopefully, we shall be one step nearer to an actual repeatable, cost effective project build!

I do look forward to seeing what the experts come up with, maths and 'simulation' wise....
I would think that simulation software has improved since I last tried using it some 10 years ago.
I had a go at some audio circuits  with something called 'LT Spice'  -  which was hopeless!...
It produced 'designs' that may well have been correct mathematically, but were often hopeless from a practical build point of view.....too many things like layout, grounding and sensible component values it couldn't cope with well at all.....
I gave up on it at that time, and went back to experimenting with actual circuitry !

I'm sure things will have improved by now... it'll be interesting to see what improvements the simulations  suggest...

 

[ricardo]

This is my analysis of the files that rogs posted in #196.

They were done by loading the file(s) into Audacity and splitting the stereo file into Left (Rode) and Right (AMX8) files.

You select the part of the file in Audacity and then Analyse 'Plot Spectrum' with 'Hanning window', 16384 & 'Log frequency'.  Then Export saves the 8192 point spectrum as a text file.

I read this and plot it out with various bells & whistles .. in this case as though it is a constant bandwidth analyser eg 1/3 8ve though this is more like 1/20 8ve or finer.

With this, pink noise appears flat, white noise rises at 3dB/8ve with frequency .. and '1G' noise which is 'reddish' falls at 3dB/8ve

Alas, my programs are DOS which will soon be extinct but this beach bum is too lazy to move into the 21st century and learn a modern programming language.

In each case Blu is the Rode NT1 and Red is AMX8.

The 2 top curves are the Pink noise part of the files in an attempt to see whatis  the difference in response as a guide to matching sensitivities.

You can see the danger of trying to match sensitivity with a single frequency.  I prefer to match sensitivities using music and voice and then noting what my figures give when measured.  In this case, AMX8 is about 15 or 16dB louder than NT1.  It has more bass below 100Hz and less between 400 & 1k5 Hz.  The plots use 16dB matching.

I don't believe the response above 3k5Hz at all as this depends on mounting, room bla bla.  I have my own tools for 'anechoic' response above 200Hz but alas, they are also DOS programmes. 

You can clearly see the effect of rogs 100Hz filter on both response & noise.

rogs, how are the mikes mounted for the Pink Noise measurements?  Is this your 30cm measurement?

The interesting bit is the lower curves where I've increased the gain by 40dB.  Remember, this is with the 16dB matching so represents the noise when the 2 mikes are adjusted for the same sensitivity.  I've overlaid 3 noise curves out of the 4 files to check we are not looking at (very quiet) ambient noise.  The 10kHz files had more LF noise than the other 3 so I've discarded it as probably ambient.

NT1 has classic '1G' noise merging into white 'resistive' noise at about 2kHz.

I'll need to do more work (eg listen to the noise) to see if my guess that RF mikes 'don't have 1G noise' is valid ... but ... AMX8 has between 3 & 4dB less noise than NT1 below 1kHz.

Their noise spectrums crossover at 3k1Hz and above that AMX8 is noisier.  The noise is 'whiter than white' so not something I can identify immediate.  It might be the result of some response peaking due to the RF gubbins which we haven't soused out. 

This is sorta borne out by AMX8's greater HF response compared to NT1 but I'd like to see a proper 'anechoic' response before pontificating further.  There's a great temptation to EQ the extra HF which would then make AMX8 quieter all round  8)

So I think AMX8 is at least as quiet as NT1 though not as much as originally thought.  Still a great achievement rogs 

Gotta go but I'll write more later

AMX8 has less noise below 3k1 Hz.  As it also has more bass below 100Hz, it is clearly quieter than NT1 at mid & low frequencies.