DIY RF Condenser Mics

[RuudNL]

Finally! Now let the fun begin...

 

[Gerard]

Isolating T2 secondary from ground and adding a second reversed diode with it's own CR network works OK,  but the return path for each half cycle  is then via the second diode and network ... so  losses there.  Again, no real advantage in voltage gain

Is there any advantage in looking again at the rectifier design used in the Baxandall article? I am not an electronics engineer, but I see he was using two transistors as rectifying switches, with the switching signal derived from the oscillator. But maybe I'm speaking from the wrong orifice.

 

[rogs]

Finally! Now let the fun begin...

 

Is there any advantage in looking again at the rectifier design used in the Baxandall article? I am not an electronics engineer, but I see he was using two transistors as rectifying switches, with the switching signal derived from the oscillator. But maybe I'm speaking from the wrong orifice.

There may well be an improvement by adopting Baxandall's balanced bridge option ... although it's moving 'up a level' in design  expertise..
I did wonder about  using 74HC4052 analogue multiplexers to try out  Baxandall's idea, but whether  switch noise would be a problem? ... (you can't get at the 'gates' on 4000 series CMOS, to slow the switching transients down - although they're not always too much of  a problem). And then there's working out the extra inductors that may be required for oscillator coupling ----

A more complex project I would think ....might be worth looking into though ?....if you're feeling brave! 

 

[Gerard]

A more complex project I would think ....might be worth looking into though ?....if you're feeling brave! 

I'll let you know if I'm feeling brave when the components arrive (it's so long since I made anything I think I threw out or gave away most of my 'stock' of components) and when I create some free time.

Like I said, I'm no electronics engineer - I'm a steam systems designer, professionally - and my electronics knowledge is more than a little rusty. 

 

[RuudNL]

This afternoon I constructed the first test version.
To my surprise everything worked the first time.
Output isn't bad, much higher than a dynamic microphone!
I found the best way to adjust the cores of the coils is with a test tone.
I held the microphone close to one half of a headphone with a 1 KHz tone on it, while measuring the output signal.
Noise isn't a problem. I think a lot of cheap condenser microphones are noisier than this design!
I used a K47 capsule for my tests.
The only problem I have with the design, is the current draw of the oscillator.
The voltage on the pins of the XLR connector is in the order of 22 V, so that would mean a current of ~8 mA.

I have changed some component values in my prototype.

 

[homero.leal]

This afternoon I constructed the first test version.
...
The voltage on the pins of the XLR connector is in the order of 22 V, so that would mean a current of ~8 mA.

That's really great RuudNL!Congrats! This has been an elusive design for DIYers, but now it starts to be feasible task!  Big thanks to Rogs and Gerard for their work and sharing the info with the group!

Quick question... is it really a problem having 8mA of current draw, on a P48V mic?

Thank you and regards!

HL

 

[RuudNL]

Quick question... is it really a problem having 8mA of current draw, on a P48V mic?

No, not really... It is just the idea that it is a much higher current than most condenser microphones.
Also the dissipation in the phantom power resistors is higher than with other microphones.
But the quality is surprisingly good and self noise is much lower than I had expected.

 

[rogs]

This afternoon I constructed the first test version.
To my surprise everything worked the first time.
Output isn't bad, much higher than a dynamic microphone!
I found the best way to adjust the cores of the coils is with a test tone.
I held the microphone close to one half of a headphone with a 1 KHz tone on it, while measuring the output signal.
Noise isn't a problem. I think a lot of cheap condenser microphones are noisier than this design!
I used a K47 capsule for my tests.
The only problem I have with the design, is the current draw of the oscillator.
The voltage on the pins of the XLR connector is in the order of 22 V, so that would mean a current of ~8 mA.

I have changed some component values in my prototype.

Glad your first build has gone well Ruud..... Certainly good to see a DIY RF biased microphone idea  enter the hobby domain - at last!

I have been impressed with how well such a simple circuit performs  - I had a feeling that these coils could have a spec that was close enough to be useful for this task!

I think the problem with the fairly high oscillator current  is down to the fact that the  impedance of the secondary of the Spectrum coils (used as the primary for T1) is a bit too low to be ideal for the task required.
But I found that the addition of R5 (which I note you have changed to 560R) allows for a reasonable balance between voltage and current drawn. R5 also permits the sine wave available at the emitter to be much better quality than trying to drive the emitter with a purely inductive load. At least that's what I've found.

One thing that I've not mentioned in this thread is how this whole concept relies on an unbalanced bridge. If for example the value of C4  exactly matched the capsule value, there may not be enough of a bridge imbalance to allow  the rectifier diode to conduct effectively.
Highly unlikely of course - and I've now fitted a BAT85 in place of the 1N4148 to help minimise any such effect.

But it does perhaps point to trying out Baxandall's balanced version - which Gerard mentioned above - as a natural progression of the project.

As you have now confirmed, this circuit works well for a simple project - certainly better than I expected when I first started it!

I'm waiting for some of your PCBs to arrive from JLCPCB - hopefully on Monday next. 
I'm getting tired of making stripboard and hand built coil assemblies for my prototypes! 

Couple of new observations:
 
•The addition of the second rectifier to create a balanced output actually reduces the sensitivity slightly, so I shall stick with half wave for the time being. In addition, the extra AC voltage available in that mode might prove to be useful  if the bridge does end up closely balanced !
• I have decided to stick with 10MHz for the time being. Although using 12MHz does permit a higher Q tuning, the amplitude is reduced by being further away from the designed frequency of the coils (7MHz), and the sensitivity is slightly reduced as a result.

Maybe one day we shall discover a more suitable IF can?... Trouble is there doesn't seem to be many that have appropriate characteristics, and not have an internal capacitor fitted.
In theory, it's supposed to be possible to remove those caps, but I've never had any luck with that....

 

[Gerard]

Are there any 12MHz (or higher) coils available, even with internal capacitors? Maybe we will have to see if the capacitor can be removed. However, my preliminary searches have not shown much.

I wonder if any of the other coils available from Spectrum Communications might be suitable? If you see any possible candidates, I'm willing to try doing a capacitorectomy, to see if it is feasible.

 

[Khron]

Would anything like this be even remotely suitable?

https://www.ebay.co.uk/itm/SUMIDA-10-70-MHZ-IF-TRANSFORMER-COIL-PINK-WITH-MULTIPLE-WINDINGS-2-PIECE-OFFER/173863899325?epid=1930905707&hash=item287b1890bd:g:7qEAAOSw-jhUFyee

Since it's spec'd "for" 10.7MHz and the primary(?) inductance is in the order of 3-4uH...  ???

 

[Gerard]

Rogs, Ruud, might the coil suggested by Khron work better at 12 MHz? It does have a capacitor, but I note that one end of the capacitor is not connected to any coil - see the description on the sellers web page (https://symkrilag.uk/shop/10x10mm-shielded/sumida-1070-mhz-if-transformercoil-pinkwith-multiple-windings-2-piece-offer/); it's also slightly cheaper direct from his website than from his eBay outlet.

 

[RuudNL]

As far as I know, the core material of those coils has a pretty broad frequency range.
There are even capacitor values given for frequencies between 3.8 and 14 MHz.
Q factor will slightly change, because the L-C ratio changes.
If you are not faint-hearted, you could even open the enclosure of the coil and reduce the number of turns...
(In fact I have used only the coil formers in the past, and rewound the coils to suit my needs!)

 

[rogs]

There are 3 important parameters that the inductors need to  comply with for this design to work effectively:

• The secondary winding of T1 needs to be centre tapped ( lots of IF transformers  are tapped  near one end of the coil... These won't work very well - if at all! )

• The  capacitive loading of T1 secondary  will ideally be around 30 to 35pF (when used with a capsule with a capacitive value around 60/65pF) .
That secondary load is formed by the capsule and the series capacitor C4 ....For maximum sensitivity there does not want to be another capacitor in parallel with that winding.

• There needs to be a 'step up' turns ratio of around 1:4 between the (untapped) primary and the (centre-tapped) secondary windings.

All the other parameters are to be 'fitted' around those 3....
 
My first prototype used Toko  3894 cans , which were 2.6uH loaded with an 82pF cap. It worked - but not well.

The Spectrum 5.3uH coils will resonate at around 12MHz  with a 33pF  capacitive load. Their fundamental frequency is specified as 7MHz, so 12MHz is some way from that.

If you select a 2.6uH coils, the 33pF load  would resonate at around 17MHz  - a  bit too high, IMHO!

As Ruud points out, these coils have a fairly broad  frequency range, but the Q will drop a little as you move away  from the specified centre frequency. 
Overall though, I think 5.3uH coils are to be preferred for this task over 2.6uH ones.

Another thing to consider is the complex  loading  presented by the coil assembly as a whole.

T1 secondary is also loaded by T2 and it's associated components.  T2 is in turn loaded by the rectifier - and 'looking back' into T1. ......It's a complex set up, the maths of which are way beyond me!

I have found the highest sensitivity - and hence the lowest noise - using a 10MHz  oscillator for my prototype.
Ruud used a different capsule, and preferred 12MHz.

There are many combinations that may well work OK... but you do need to  make sure that whatever you use for T1  has a centre tapped secondary winding, and will function - at a sensible frequency - with a 30pF to 35pF load, and without another capacitor in parallel.

 

[rogs]

I did try out one more version of a balanced rectifier that came pretty close to the same sensitivity as the present 'half wave' version....
More elegant I think, but not quite so good, sensitivity wise -  so no cigar! 

 

[ricardo]

One thing that I've not mentioned in this thread is how this whole concept relies on an unbalanced bridge. If for example the value of C4  exactly matched the capsule value, there may not be enough of a bridge imbalance to allow  the rectifier diode to conduct effectively.
Highly unlikely of course - and I've now fitted a BAT85 in place of the 1N4148 to help minimise any such effect.

But it does perhaps point to trying out Baxandall's balanced version - which Gerard mentioned above - as a natural progression of the project.

Ruud & rogs, I'm in awe of your progress in such a short time.

I think you guys are at about the stage of the 'first experiments' at the end of page 1 of Baxandall's article.  I'm guessing your noise level is about that of a 'good' Schoeps HiZ circuit.

This is definitely worthwhile as it comes with better tolerance of humidity

The next stage is to try GG Baxandall's circuit itself to get SOTA noise.  It's nearly 50 yrs since I played with it so my memories are quite hazy.  Then I could call on the Great Guru himself for help.

About the only useful thing I can dredge up is that his 1mH inductors in the output filters are non-critical and any convenient RF chokes will do.  Operating at 10MHz means you can use inductors 3x smaller.

Also your Schoeps double EF output stage makes this filter even less critical.

I await developments with jealous interest

 

[rogs]

I think you guys are at about the stage of the 'first experiments' at the end of page 1 of Baxandall's article.  I'm guessing your noise level is about that of a 'good' Schoeps HiZ circuit......

This is definitely worthwhile as it comes with better tolerance of humidity.....

The next stage is to try GG Baxandall's circuit itself to get SOTA noise.  It's nearly 50 yrs since I played with it so my memories are quite hazy.  Then I could call on the Great Guru himself for help......

I can now say that - noise wise - my own 'no.1 prototype' is now on a par with my  own Schoeps Hi-Z mics .... and not that much noisier than my Rode NT1 .
( I always keep  2  prototypes going when experimenting with RF circuits -- and never apply  mods  to both at the same time -- A golden rule I learned many years ago  )

Ricardo, I have to thank you for your comments in the original  Yahoo micbuilder's group thread a couple of years ago.
In among the ' flotsam' of that long thread (over 70 posts) there were 2 comments that proved inspirational to me...

Post #4 from Umashankar:  " 'It would be fun to build the Baxandall version with the 25 mm capsules. One just needs coil details (I think they were not included in the WW article) best would be if one can use IF cans, ready made"

and - most importantly - your own comments in post #36, which included the following:

"Can I suggest some of you guys actually read Baxandall's paper(s)."  (absolutely essential reading IMHO!) .. and

"PLL allow accurate very small bandwidth receivers without yucky Inductors .. but you lose the Q multiplication of genuine Reactive bits on good stuff like S/N and power output" (The underlining was  in your original comment).

I had tried  experimenting with  4046 PLL and basically  got nowhere.  It was when the original poster of the Yahoo thread  -Gerard - started this thread that my interest was re-kindled.
I initially tried some Toko IF transformers, but they weren't really suitable, and it was only when I came across  these Spectrum coils that I though there might be a chance to get some useful results.  I was immediately reminded of your  comment about reactive Q .

After I published my initial  circuit here, Ruud 'took up the mantle' as well, and we have both now produced  useful prototypes .
We only seem to have a few differences in ideas on component values - only to be expected when using different capsules.
In addition, I know Ruud is keen to reduce the oscillator current.

Whether the project will now move on to the 'next stage' as it were, time will tell? 
I have already  bought a 5/65pF variable cap  to help create a balanced bridge...but in the short term, I have to try and catch up on my other chores....
My wife can't quite see why I spend so much time making microphones......'How many do you actually need?' - she keeps asking 

P.S. I am extremely jealous of your having had  contact with Baxandall himself 

 

[RuudNL]

'How many do you actually need?'

That's I question I have been asking myself many times!
At the moment I have 80 of them...

 

[gyraf]

following from the sideline....

 

[RuudNL]

I have a feeling we should investigate the FM discriminator principle!

 

[Tubetec]

Microphone engineering handbook(M Gayford) has a section on Rf mics , written by Manfred Hibbing  who seems to have been involved with Sennhieser , Its an awkward book to find, I have a ring bound copy I can easily scan if someone wants it  for reference .

Its about thirty pages , anyone know a smart way of compressing the scans into something more manageable so I could post it here ? 

Heres a link to PJB's chapter on mic amps and transformers,

http://leonaudio.com.au/microphone.engineering.handbook..chapter.8.pdf


Heres the link to RFmics chapter of the book ,

http://www.filedropper.com/rfmics