DIY RF Condenser Mics

[rogs]

Yes... I agree with Khron! It will be interesting to do a quick test with a small C0G ceramic cap replacing the capsule, just to check for noise floor sans ambience noise.

A quick test with the C12 capsule replaced with a  100pF mlcc C0G capacitor shows that  the noise floor is around  -84dB using the same gain set up  as previously.
This recording of the noise floors ( amplified by +75dB in my DAW) in this file :  www.jp137.com/las/RF.AM.noise.floors.75db.wav 
is in 3 parts...all made using the external mic input of my Tascam DR100Mk3, with the gain set to minimum (+20dB) and no pad. 
• The first 8 seconds is of the un-terminated mic input  (c. -87dB )
• The next 10 seconds is with  the mic circuit  'output' section active only  (D1 cathode grounded) ..(c. -89dB)
• The last 10 seconds is with the ground removed from D1 cathode.. (c. -84dB)

So yes, the noise floor is improved with the capsule removed  .... The PCB assembly was not fitted within the grounded mic body for this test - and you can hear a small amount of hum in the final section.
So it may actually be slightly better than the recording suggests, when fitted within the mic casing itself..... 

 

[rogs]

However, some C12 capsules intended for another project also have arrived; they might get sacrificed for the RF mic project.

You may need to make C4 a little larger when used with a C12 type capsule. 
My C12 capsule measured around 90pF (as opposed to around 65pF for a K67, and around 56pF for a K47) so I fitted C4 as 2x47pF in parallel.
(A single 100pF would also probably work OK) .

This should give a value for the 2 series bridge capacitors of around 47pF,  which is ideal for a Spectrum 5u3uH inductor, when used with a 10MHz oscillator.

 

[Gerard]

Those are spectacular noise figures. As good as good studio mics and approaching ultra low noise instrumentation mics (B&K say 4 dB above reference 0 dB for their very lowest noise mics).

Thanks for the comments on the value of C4 with a CK12 capsule. I had assumed as much. I haven't measured my CK12s from Wgtcenter yet, but the datasheet says 70 pF +/- 4 pF.

 

[rogs]

And also, what do you think about THD? ........
....If JFET gate to drain capacitance causes THD to increase and this circuit doesn't need a JFET, shouldn't THD be better? Just wondering.

Like many folk, I simply don't have a transducer that  has a low enough THD in its own right to allow for accurate THD figures to be measured.
What I can do is to take the best one I have  - a Sony 7506 headphone earpiece transducer - and do some comparison tests.

It's simple enough to create a low distortion source tone --I used a Neutrik MR1 minirator, in conjunction with  Hypex UCD power amp. The output of that is fed to the line inputs of my Sound Devices USBPre,  and the output of that is opened up in a software spectrum analyser. That gives a THD reading of c. 0.002% for the source.

Connecting the output of the power amp to my Sony phones instead - and connecting the several  test mics in turn to the Sound Devices mic preamp - will show that introducing the transducer - and a test mic - will derate the THD distortion by a factor of around 10.
How much of that is down to the mic and how much to the Sony transducer, I have no way of measuring?....

Nevertheless, taking comparison THD  readings  with several mics show that the new experimental RF mic actually seems to perform pretty well....

Using a 1KHz sine source of around -12dB, I placed each mic in turn close to the headphone transducer, and set the SD preamp in each case to read -12dB on the meter.
Using that set up, I obtained the following THD readings...

• Shure SM58 - 0.028%
• AKG D202-E1 - 0.025%
• Rode NT1 (new 2014 model) - 0.015%
• RF.AMX5 prototype mic fitted with C12 type capsule - 0.012%
• RF.AMX5 prototype mic fitted with K67 type capsule - 0.016%
• C12 capsule with 'standard ' Schoeps type  JFET impedance converter - 0.019%
• K67 type capsule with 'standard' Schoeps type JFET impedance converter - 0.018%

So there's not a lot in it - at least with this set up. 
Neither AKG or Shure publish THD figures in their specs, and Rode only quote the THD (around 1%) for max output, so the manufacturers data doesn't help much here..

I do think the experimental RF mic comes out pretty well for THD, on a par with the other mics....

 

[RuudNL]

Wouldn't it be an idea to use one or more varicaps to simulate the capsule to do distortion measurements?
I suppose you would need a varicap with a high C/U ratio, to operate the varicap in the most linear region.

 

[rogs]

I remember that Uwe Beis used a varicap construction to simulate the capsule in his design (see the bottom of this page: https://www.beis.de/Elektronik/HF-Mic/HF-Mic.html )  but discovered that it wasn't particularly linear -- which doesn't really help with THD measurements ! 

I think that the fact that the prototype holds up quite well against other mics - including commercial ones -  in the comparison tests gives some confidence that THD is not likely to be a problem.

The only possibility I can think of that might create linearity  differences between samples is if the 2 bridge capacitors (the capsule and C4 ) end up being identical in value.  That might mean that there could be insufficient 'standing' RF voltage from bridge imbalance to allow the rectifier to always conduct in a linear way. 
Extremely unlikely in my opinion, and any effect related to that is further minimised by using a BAT85 or other Schottky diode as the rectifier....

 

[homero.leal]

I do think the experimental RF mic comes out pretty well for THD, on a par with the other mics....

Thank you very much Rogs for your work, tests and info that you have shared with the group. Great work, greatly appreciated!

Will try it very son...

Kind regards!

 

[abbey road d enfer]

The only possibility I can think of that might create linearity  differences between samples is if the 2 bridge capacitors (the capsule and C4 ) end up being identical in value.  That might mean that there could be insufficient 'standing' RF voltage from bridge imbalance to allow the rectifier to always conduct in a linear way. 

This comment prompted me looking again at your schemo. I think that, with two equal caps, the resulting RF voltage would be zero; as a result, the audio output would be a full-wave rectified version of the acoustical signal. I believe you actually operate your circuit with the detector de-tuned enough to stay on one side of the V vs. F response.

 

[rogs]

This comment prompted me looking again at your schemo. I think that, with two equal caps, the resulting RF voltage would be zero; as a result, the audio output would be a full-wave rectified version of the acoustical signal. I believe you actually operate your circuit with the detector de-tuned enough to stay on one side of the V vs. F response.

Yes, the bridge always needs to be slightly out of balance  in this configuration..... Baxandall himslef makes that very point on the first page of his paper:  http://www.jp137.com/lts/Baxandall.RF.mic.pdf

Using a BAT85 diode means that there only needs to be a minimum amplitude difference of around 300mV or so  to allow the rectifier to conduct effectively, and the amplitude variations created by the modulation are only in the order of a few millivolts on top of that, so it doesn't need to be much of an imbalance in the bridge. 
Ideally the capsule capacitance needs to be slightly larger than the series capacitor for maximum sensitivity, but in all the 4 prototypes I've built so far there has never any issue with the bridge being  too close to being balanced to cause any problems.

As Ricardo has mentioned earlier, the next step would logically be to move on to a balanced bridge design - as Baxandall did himself - but in the meantime, this simple version has so far worked  better than I ever thought it would !

 

[rogs]

Thank you very much Rogs for your work, tests and info that you have shared with the group. Great work, greatly appreciated!
Will try it very soon...

It's been fun trying it all out..... Baxandall's original concept is very simple,  and I thought it might be worth trying it out as a simple unbalanced bridge first....

Using these inductors, I managed to get a much higher 'Q' from the system than I first imagined, and it has turned out to give much better results than I  originally expected. .... Definitely worth the effort, IMHO! 

.....And a big thanks to Ruud for his test PCB design (Gerber files attached to post #32) ...Using those PCBs for protoypes is a lot simpler than my original stripboard version

 

[rogs]

I've made a few notes on the inductors, and the 'tuning' options I adopted for them in this project:

www.jp137.com/lts/RF.AMX5.inductor.assembly.pdf

 

[ricardo]

rogs, do you by any chance have measured voltages, DC and hopefully RF, at various points of one of your circuits?

I'm still doodling to see if your circuit can be Baxandallised simply using the Spectrum Comms inductors.

 

[rogs]

rogs, do you by any chance have measured voltages, DC and hopefully RF, at various points of one of your circuits?

I'm still doodling to see if your circuit can be Baxandallised simply using the Spectrum Comms inductors.

Using the component values shown in the RF.AMX5 schematic ( copy here: http://www.jp137.com/lts/RF.AMX5.pdf ) --- fitted with a K67 type capsule -  I measure the following voltages at different points:

DC volts (with respect to system gnd) ...
• 16.4V at the junction of R6 and C7. 
•    9.3V at the junction of R6 and C6.....So the oscillator and associated circuitry draws around 7mA.

AC volts at 10MHz  (with respect to system gnd)

• 13V p-p at the emitter of Q1
• 12V p-p at the T1 primary end of R5
• c. 15V p-p at each end of T1 secondary ( slightly different values at each end)
• 1.5V p-p at T2 primary  (junction of C4 and the capsule)
• 2.5Vp-p at D1 anode.

These voltages are approximate ... I'm using a x10 10MOhm/15pF oscilloscope lead,  which  loads the measuring points by a small amount at 10MHz.
(The actual calibration of the inductors is carried out audibly - by tuning the cores for maximum audio level - with a tone applied to a  test transducer next to the capsule )

The measured voltages for my C12 type prototype -  which presents a bigger capacitive load to the bridge - were somewhat lower, even though that version is  actually slightly more sensitive....

 

[rogs]

A couple of points that have been mentioned throughout the thread -- the high oscillator current and the lack of a balanced audio output...

It suddenly occurred to me that there may be a single change that could help solve both those problems ....

That is to fit a JFET 'infinite impedance' detector in place of the diode rectifier.....

Advantages include:
• Much higher Q from T2  - load is now only a JFET gate
• Balanced output by using the JFET as an audio phase splitter as well.

Fitting that idea to the existing circuit , and the voltage swing now available at T2 secondary when the system is tuned can be way too high for the FET (15 -20V p-p) .

One obvious solution- increase R6 to reduce the oscillator DC supply.

Result  is  the oscillator current is now cut by around 50% -- and the sensitivity is increased by about another 6dB

Provisional schematic  is attached... (It may be that some of the new values still need to be 'tweaked'...)

The only disadvantage so far is that it's now a bit more tricky to get it all to fit onto Ruud's original prototype PCB ....

 

[Khron]

Well, i was sorta-kinda close... ;D

Not that i know anything about oscillators, but i was starting to wonder - would it help anything, switching (no pun intended) to a JFET or a MOSFET for the oscillator itself (Q1)?

Jokes aside though, great progress / evolution! 

 

[homero.leal]

The only disadvantage so far is that it's now a bit more tricky to get it all to fit onto Ruud's original prototype PCB ....

But now, as there is a JFET in the signal path, THD may increase specially if jFET has a high gate-to-drain  capacitance. :-\

Also, no biasing requred for JFET Gate?

Regards!

HL

 

[rogs]

But now, as there is a JFET in the signal path, THD may increase specially if jFET has a high gate-to-drain  capacitance. :-\

Also, no biasing requred for JFET Gate?

Regards!

HL

I haven't measured any changes in THD yet -- although as I mentioned earlier, I'm not sure how much of the distortion in my set up is from the audio transducer anyway?

The FET gate is effectively grounded, and I can measure c. 5V across both the drain and source resistors - so there's around 1mA of current  flowing in  the FET. 

With the reduced oscillator current, the lower modulated 10MHZ carrier presented to the gate from T2 secondary should allow the FET to operate within it's linear region......

Another advantage is that there are now fewer volts across the capsule itself.....

If this does turn out to be a backward step - distortion or noise wise --we can always go back to the diode! ....That's the beauty of an experimental project !   

 

[rogs]

Well, i was sorta-kinda close... ;D

Jokes aside though, great progress / evolution! 

Thanks -- yes, it's turning out to be an interesting project ...it certainly now performs way better than I first thought it would..

I'm looking forward to seeing how the results from others  experimenting with the same idea turn out.....

 

[RuudNL]

The 'FET detector' may be a big step in the right direction!
Have to try this as well! (And maybe design a new PCB for it...)
I think you have almost reached the maximum performance from such a simple circuit.

 

[abbey road d enfer]

But now, as there is a JFET in the signal path, THD may increase specially if jFET has a high gate-to-drain  capacitance. :-\

Cgd is a source of THD in Hi-Z (DC bias) condenser mics, not with this implementation of RF bias. You can't generalize.