FET Mic with dedicated power supply.

After much pondering and building of mics I have drawn some of my own, granted biased, conclusions.

1. Tubes are great both for the way they sound and the ease in which they can be implemented. But they are noisy prone to  microphonics, distortion and sonic artifacts that are hard to ignore, though they are musical for a lot of uses. You also need to go through a small s**t load sometimes to find one that is microphone grade. They also fail eventually and are HOT HOT HOT!!!

2. What is our reticence when it comes to FET based solidstate mics to use an outboard PSU? I get the use of the phantom power standard when in the field and working live, sure it makes total sense. But when in the studio who cares! Lord knows I have done sessions with a bunch of tube mics and had PSU's all over the place, whats a few more if it means a qualitative advantage? I have also noticed that most consoles don't really have the current capacity on the phantom rails to get the job done.

3. As I have played around with mic circuits and looked at many more schematics I have come to the conclusion that for most microphone impedance amplifier designs FET based is the way to go. They are cleaner, smaller, cooler, stable, plentiful, cheap, have tight manufacturing quality control and are easy to design with. They also have high input impedance which as we all know is very desirable for mics. There are also a lot of choices both in op amp and discrete variety.

4. An outboard PSU allows for multiple power schemes in the same circuit without the need for things like DC to DC converters for capsule polarization and dropdown resistors. They also allow you to look at multistage active conversion to output amplifier designs. You can even go as far as output circuitry that goes to line level and even digital output. A dedicated PSU starts to allow us to look at chips VS discrete topologies or even hybrid designs. They also allow for not using interstage coupling caps as servo designs can be employed.

5. I have always seen phantom powering as a compromise for ease of use but never a reliable  robust solution for microphone powering. 

In my next post in a few days I will propose some components I have pondered over for various uses when it comes to this subject I look forward to the groups comments.

I have built an external powered solid state microphone and when tested against other commercial microphones tube and transistor  it did extremely well. 
I even thought of selling it however liability and and the cost of running a company made me decide otherwise.

Great topic Pip, looking forward to this.  I've often thought that a FET based mic based on Tim Campbell's CK12 with an outboard PSU could be a nice thing.

Gus said:

I have built an external powered solid state microphone and when tested against other commercial microphones tube and transistor  it did extremely well. 
I even thought of selling it however liability and and the cost of running a company made me decide otherwise.

Click to expand...

You should reconsider. I would love to see your design as you are one of the people I remember saying that once you really grasp the mic thing  FET's are the way to go.

Pip said:

After much pondering and building of mics I have drawn some of my own, granted biased, conclusions.

1. Tubes are great both for the way they sound and the ease in which they can be implemented. But they are noisy prone to  microphonics, distortion and sonic artifacts that are hard to ignore, though they are musical for a lot of uses. You also need to go through a small s**t load sometimes to find one that is microphone grade. They also fail eventually and are HOT HOT HOT!!!

Click to expand...

In addition, they are noisier than FET's because they don't work with 1G+ grid resistors. Noise performance of the head amp is directly related to the value of the grid/gate resistor.

2.  I have also noticed that most consoles don't really have the current capacity on the phantom rails to get the job done.

Click to expand...

This I have noted on a 1970's API console, at Barclay Studios that had 10k resistors before the 6k8's! Nobody noticed because all phantom-powered mics were Neumann's (47FET and 87), until the day I set out to make phantom-powered DI boxes; I spent sometime scratching my head why it worked with the Neve and not with the API...

3. As I have played around with mic circuits and looked at many more schematics I have come to the conclusion that for most microphone impedance amplifier designs FET based is the way to go. They are cleaner, smaller, cooler, stable, plentiful, cheap, have tight manufacturing quality control and are easy to design with. They also have high input impedance which as we all know is very desirable for mics. There are also a lot of choices both in op amp and discrete variety. 

Click to expand...

True, but still, choice of FET is an issue. The quietest are also often the ones with the highest gs capacitance.

4. An outboard PSU allows for multiple power schemes in the same circuit without the need for things like DC to DC converters for capsule polarization and dropdown resistors. They also allow you to look at multistage active conversion to output amplifier designs. You can even go as far as output circuitry that goes to line level and even digital output. A dedicated PSU starts to allow us to look at chips VS discrete topologies or even hybrid designs. They also allow for not using interstage coupling caps as servo designs can be employed.

Click to expand...

It certainly offers more options than the restricted constraints of phantom. I see at least two cases where they could be used fruitfully:
Powered ribbon mics, where the availability of more current would allow running the head amp at several mA, in turn allowing the use of low-Z secondary xfmr.
Thermostatic control, as in the Sony C800G.

5. I have always seen phantom powering as a compromise for ease of use but never a reliable  robust solution for microphone powering. 

Click to expand...

In fact, except the aforementioned case that made me wary (once bitten...), I have never had a bad experience with phantom-powered mics.

Pip said:

5. I have always seen phantom powering as a compromise for ease of use but never a reliable  robust solution for microphone powering. 

Click to expand...

Not reliable for microphone powering?

Sounds like hyperbole to me. People have been using phantom powered microphones for decades now, and if it really was all that bad, I'm sure someone else would have reached that conclusion by now.

Location recording, studio recording, broadcast... The only times I've had issues with phantom-supplied mics not getting proper power have been due to bad cables and -in one case- a discontinuous ground between panels. -An external power supply may indeed have passed signal through a bad cable, but possibly with a less-than-perfect noise or frequency response, due to one of the lines (either signal or shield) being discontinuous, or a short between two. -The phantom requirement actually illuminated the problem. (Likewise for the discontinuous panel ground situation.)

it really doesn't take much power to drive a very tiny signal into a 2kΩ load. The conventional phantom system has been doing this for decades now. -If a manufacturer doesn't think it's possible, or is not willing to make a circuit sufficiently efficient, then they're welcome to build a box-powered mic.

However, over the years I've also had dozens of cases of tube mics which have been brought to me for repair with a note saying "it hums". -I've plugged the mics in and found them to be often perfectly silent. -When the owner has plugged them back in again at the studio, they've discovered that they still have a hum issue, and have returned the mic to me saying that perhaps I didn't check it properly...
-In the end, it's turned out to be a ground loop caused by plugging in to an outlet a long way from where the other end of the mic line is finding its ground.

For every perceived 'benefit' of creating a mic with a remote supply, you have to consider that there may also be disadvantages.

Ha! Not the first to accuse me of being prone to hyperbole!

Yes phantom standard  works more than not. Grounding issues in studios always a problem in more ways than mic PSU's. I have also noticed a bizarre, again IMO, trend towards mic inputs on the patchbay?! The reality of breaking a patch while hooked up is a bad policy! Their is also the problem of transformer isolation and the non-passing of phantom when it comes to splits and the like. With the ever increasing "project studio" market and the consoles that have been targeted at this market not having individual phantom switches on each channel instead employing more global control is also problematic. The current draw problems and the choices made by manufacturers with-in the phantom standard as to voltage option are also in play and annoying.

It has been, over the years,  a very liberal standard to say the least with variants that are still in use today P12 P24 P48.  I recently restored a Yamaha PM2000 console and tooled it up for studio use with line-outs, it came with multiple PSU's. I noticed that some had 48V phantom and others had 41V phantom!? Well the 41V PSU was for "the U.S. Model"!? These types of things affect a mics performance.


Its not the best reference article I have ever read but it goes in the right direction and does have some good info.
https://en.wikipedia.org/wiki/Phantom_power

This in no way is meant to be a condemnation of phantom power I was merely trying to point out that we take it as the only means to an end and think of it as problem free.  I for one do not.

What I am "asking" is given the ground-rules of the phantom powering of microphones standard  are solid-state studio microphones being all they can be?

Pip said:

What I am "asking" is given the ground-rules of the phantom powering of microphones standard  are solid-state studio microphones being all they can be?

Click to expand...

It is undeniable that phantom power puts constraints on the possibilities, but in the actual state, they are not an obstacle. To my knowledge, the only mic mfgrs that can justify external PSU's with solid objective arguments are B&K, DPA et al.
Most mic that use tubes do so for "flavour", not for objectively superior performance.

abbey road d enfer said:

Pip said:

What I am "asking" is given the ground-rules of the phantom powering of microphones standard  are solid-state studio microphones being all they can be?

Click to expand...

It is undeniable that phantom power puts constraints on the possibilities, but in the actual state, they are not an obstacle. To my knowledge, the only mic mfgrs that can justify external PSU's with solid objective arguments are B&K, DPA et al.
Most mic that use tubes do so for "flavour", not for objectively superior performance.

Click to expand...

Yes! This manufacturer is one of the reasons I started to question this. The B&K 130V approach definitely results in some great sound transducer technology. The proof is in the pudding. I once was at a demonstration of the marriage of DPA 4012 with a Millenia preamp, this is a 130 Volt system. It was a stereo recording to DAT tape of a solo cello performance, granted done in a studio, best cello I have ever heard!

There are two aspects to this subject. The capsule and the head amp.
The need for higher polarization voltage came from the fact that, in order to accomodate higher SPL, the distance between diaphragm and backplate should be increased. This is not the only solution, but it's one that provides the highest dynamic range, i.e. does not impair the low-spl performance. That results in a decrease in capacitance and thus decrease of sensitivity, so, in order to compensate, the bias voltage can be increased accordingly. The result is a similar sensitivity with higher SPL capability. This in turn creates the need for higher headroom in the head amp, which is relatively easy to do by increasing its supply voltage. Again, this is not the only solution, but one of the easiest.
That was the basis of the B&K high-voltage power.
DPA have carried on with the 130V system.
But the capsules in DPA mics are prepolarized (electrets). So the benefits of high-voltage power are apparent only in the head amp. In the end, the objective difference between the 48 and 130V is a slightly lower THD and significantly higher spl capability.
Whether this difference is audible at moderate spl (>110dBspl) is debatable.

To further the discussion I will list some components that I have considered to start with. I realize this is a lot of info and little reason why.  I have borrowed from a lot of others I stand humbly on the shoulders of giants. I do put forward at the bottom two schematics that for me sort of layout the approaches. Obviously the whole battery powered thing is what I am looking to do away with.

for input the following:

TI OPA134
http://www.ti.com/lit/ds/symlink/opa134.pdf

TI OPA2107
http://www.ti.com/lit/ds/symlink/opa2107.pdf

TI OPA132
http://www.ti.com/lit/ds/symlink/opa2132.pdf

TI OPA445
http://www.ti.com/lit/ds/symlink/opa445.pdf

For servo and or output

TI TL062
http://www.ti.com/lit/ds/symlink/tl062.pdf

Output

TI TL062
see above

ON Semiconductor MC33178
http://www.onsemi.com/pub_link/Collateral/MC33178-D.PDF

Or a Schoeps output

Analog Devices SSM2220
http://www.analog.com/media/en/technical-documentation/data-sheets/SSM2220.pdf

Please contribute any and all thoughts regarding these choices.

One of the other mic designs out there that started me thinking was the CAD e-100 and original e-300.



http://cdn.recordinghacks.com/images//mic_extras/cad/E100-schematic.png
http://recordinghacks.com/microphones/CAD/Equitek-E300

There's something wrong with that schematic, I'm seeing it wrong or I'm missing something. When pad is at -20dB position the U2A and U2B non inverting inputs are connected together and to anything else, no bias current for them. I guess it should be V9, but I don't see it anywhere.

I got dizzy looking at U3B, seriously? you couldn't swap anything around to make it more readable?  Almost every opamp has it's inputs crossing over the other. The "YOU ARE HERE" sign looks useful...

I'd make use of the (already-there) opamp for the servo to mix some signal and add bootstrapping avoiding the 2G resistor and maybe getting away with a 10M one. Deal with PCB leaking and capacitance with 2G impedance network is something to be paying attention and all that adding that for the 2G resistor compared to a 10M, they are bigger, harder to get and more expensive. Bootstrap 99.9% of the signal and you get a 10G virtual impedance in a 10M network much easier to deal with. Maybe using the diodes trick used in Rode's mics may be an option there.

JS

Came up with this a while ago:  http://groupdiy.com/index.php?topic=54409.msg698224#msg698224.  Built it and it worked in pieces on the bench;  life intervened and I never got it into an actual mic to evaluate it.

joaquins said:

There's something wrong with that schematic, I'm seeing it wrong or I'm missing something. When pad is at -20dB position the U2A and U2B non inverting inputs are connected together and to anything else, no bias current for them. I guess it should be V9, but I don't see it anywhere.

I got dizzy looking at U3B, seriously? you couldn't swap anything around to make it more readable?  Almost every opamp has it's inputs crossing over the other. The "YOU ARE HERE" sign looks useful...

I'd make use of the (already-there) opamp for the servo to mix some signal and add bootstrapping avoiding the 2G resistor and maybe getting away with a 10M one. Deal with PCB leaking and capacitance with 2G impedance network is something to be paying attention and all that adding that for the 2G resistor compared to a 10M, they are bigger, harder to get and more expensive. Bootstrap 99.9% of the signal and you get a 10G virtual impedance in a 10M network much easier to deal with. Maybe using the diodes trick used in Rode's mics may be an option there.

JS

Click to expand...

I did not draw this schematic it was the one released by the company and came with the mic. I don't think thats wrong isn't it just more resistance for 0db? Yes I agree the DC to DC conversion circuit is a pain in more ways than one. As to the audio point that's why I posted here. I will put that in the think tank.

EDIT:
I think now that I came back to this more awake I may have misunderstood your feedback. That "already there opamp" is the servo.? I think I know what you mean with the bootstrapping and I don't know if it really makes a difference. The PCB leakage is always to be considered but there are circuit implementation schemes that allow for this to be dealt with I think.? 

dfuruta said:

Came up with this a while ago:  http://groupdiy.com/index.php?topic=54409.msg698224#msg698224.  Built it and it worked in pieces on the bench;  life intervened and I never got it into an actual mic to evaluate it.

Click to expand...

Very nice! Now what is curious to me and why I asked this and started down the whole rabbit hole of the topic is; What if you just built a well filtered PSU that output the  15VDC for the op-amp and the polarizing voltage of +/- 60VDC without the DC to DC converters and done with it? Is there a sonic advantage or not?

About 15 years ago when DC to DC converter chip solutions really got available, thanks to the computer motherboard industry, I started to experiment with it for driving opamps in electret measurement type circuits. I never could get the current right and use phantom power without some-kind of reservoir. I still have some drawers in my bench with these things in them. Granted DC to DC converter technology has come along greatly since then. They are smaller a lot cheaper and more efficient and adjustable than they used to be.  But still feel like a compromise to me.

I guess I have to build two mics that are identical to each other in all but power scheme and see eh?

joaquins said:

There's something wrong with that schematic, I'm seeing it wrong or I'm missing something. When pad is at -20dB position the U2A and U2B non inverting inputs are connected together and to anything else, no bias current for them.

Click to expand...

I don't see that. When the switch is open, the opamps work as voltage-followers.

I'd make use of the (already-there) opamp for the servo to mix some signal and add bootstrapping avoiding the 2G resistor and maybe getting away with a 10M one. Deal with PCB leaking and capacitance with 2G impedance network is something to be paying attention and all that adding that for the 2G resistor compared to a 10M, they are bigger, harder to get and more expensive. Bootstrap 99.9% of the signal and you get a 10G virtual impedance in a 10M network much easier to deal with. 

Click to expand...

Although it works in terms of LF extension, virtual impedance doesn't work in terms of noise. Unfortunately the very existence of 1-2G resistors or the diode substitute) is unavoidable.

why use the E300 as a starting point?

bockaudio said:

why use the E300 as a starting point?

Click to expand...

Only because it made sense to me right away as a concept and approach. It inspired me. It was at the time some almost 20yrs ago a new approach to an old need and it has made think for quite some time why wasn't this concept broadened. I am not an engineer and design is not a strong suit of mine. I am the kind of person who lives by empirical results first and then tests later. I study schematics and then ask questions add a bit of research and then build something to see if it makes a difference. If there are other design approaches of this type out there I am not aware of them. I do know that a boutique company in Brooklyn did offer at one point a solid state based mic with a multiple voltage outboard PSU. I do know that using op-amps is not new and there are offerings out there but the search for schematics for them has yielded no fruit. So I started here with the e-100 and e-300 because they offered something in the direction I am looking to begin the journey. One has external capsule biasing and the other uses an electret approach. Those are my reasons.

The real question is whether you want another mains powered BOX between the mike and your preamp .. with a special CABLE.

If the answer is 'yes',  you can do loadsa stuff which is of practical benefit.

The obvious one is remote control of pattern as in AKG C12.

If you use a 'true' condensor, you can vary Vp and have noiseless switching of a gain pad.

Better still, why not raise the signal to line level and get away from the evils of cheapo 'prosumer' mike preamps?  You can ensure optimum amplification of your precious signal with each gain step.

130V or 200V powering as with B&K & DPA isn't of much use unless you use their $$$ capsules or are prepared to make your own nickel diaphragm capsules.

the e300 strikes me as way too much circuitry for your experiment. If you want to compare external power vs phantom build two identical high quality mics with minimal circuitry. THEN move on to solving other problems or introducing more complex circuits.  No one I know here in LA has an e300 let alone thinks it's a great circuit.  And as Ricardo mentions, a high quality consistent capsule will be important.  I would even venture that the difference your experiment may expose is different still for 1/2" and 1" pairs of microphones.
I think Gus published an interesting circuit for fet with separate psu, but I don't think it's adaptable to P48. Start with a km84 circuit just to keep the variables down.
Hell, we could build a km84 circuit, P48, with rotary switches to select output and source bypass capacitor and we'd hear provable differences.