All OpAmp mic design (no FET at first stage)

[homero.leal]

In76d,

You mean..  a independent board with a plus-minus voltage (may be between 6V to 18V) DC converter, feeding from the 48V Phantom Power, to power OpAmps?

Sound's nice!

It could be useful for electrets, since there is no need for DC singal to the capsule.

For non-electrets LDC, it may be a little bit more difficult.

Best regards!

HL

 

[ln76d]

Maybe something much more usefull for many applications - let say one board with 48V converter circuit but with optional changes for different types of microphones - for electrets (there's not much to convert) for external polarised - for cardioid and for multipattern also, etc.
There's a challenge

 

[homero.leal]

Sounds great... but could it be realized?

I mean, I have read that Phantom Power is very limited on current. So, may be a special arrangement like batteries as on E300 will be needed, specialy for LDC polarization.

Regards!

 

[ricardo]

homero, can I ask what benefits you hope to get from the OPA on P48V?

I've designed commercial mikes with OPAs but you need to set out your Design Goals first.

If the only Design Goal is "mike with OPA", you might be better off, buying some dead SMD OPAs and glueing them to the outside of the mike case  8)

If the mike uses Electret capsules with built in FETs, you don't have to use a FET i/p OPA

 

[homero.leal]

Thank you Ricardo. Nice to read from you. 

A few days ago I had a look to a Mic design from Herbert Rutgers, with no FETS. It's an all OpAmp design, including first stage.  Also, it uses bootstraping to reduce distortion.  It uses a TSB-165A  (No FET) electret capsule, and several OPA134 OpAmps.

This words quoted from the Herbert's article, caught my interest and attention:

"The input impedance should be as high as possible. Therefore there is no discussion about the input device: a small, well-biased FET! But why trying to hack a discrete FET with its idiosyncrasies if there are FET op amps with high linear input impedances which promise to make much less noise than a resistor of 1 GΩ? E.g an OPA134 rustles as a resistor of 2 kΩ..... so noise is no item any more!"

The design looks very interesting, but it uses external plus and minus power sources. So, design is not suited for PhantomPower.

Also, I heard some recording samples on the article, using an TSB-165A  electret capsule, and they sound really, really great.

So this lead me to wonder, if this design could be "hacked" to work with Phantom Power.  I mean, it will eliminate the need for a power box, simplifying the build and the instalation. I know external power sources are the norm for tube mics, but I'm just wondering about the posibility.

I would like to have a mic with this goals on thedeisign:

• Could work with TSB-2555B / TSB-165A electret capsules
• All OpAmp (no FET at first stage)
• 48V PantomPower
• Bootstraping (if possible)

Thank you again, best regards!

 

[Pip]

The OPA134 is a Cascode FET input. So it is a FET!

 

[RuudNL]

I once had a Behringer C-3 microphone over here.
It used a TI TL072 OpAmp as the first stage after the capsule.
The thing was so noisy, that I could only use it to record (loud!) hardrock...
FET OpAmps have improved over the years, but I would stay with the 'classical' discrete FET design.

 

[Khron]

Not sure just how feasible OPA134's would be, with strictly phantom-power only - datasheet says 4-5mA idle current per channel (so double that for an OPA2134, for example).

But the idea is indeed sound (no pun intended )

 

[ricardo]

A few days ago I had a look to a Mic design from Herbert Rutgers, with no FETS. It's an all OpAmp design, including first stage.  Also, it uses bootstraping to reduce distortion.  It uses a TSB-165A  (No FET) electret capsule, and several OPA134 OpAmps

OPA134 is one of the FET opamps I would consider for such a beast.

Is this the circuit you mention? https://www.by-rutgers.nl/ME6211-PRO37R.html

Complicated circuits make my small brain hurt so I would use something MUCH simpler 

I would like to have a mic with this goals on thedeisign:

• Could work with TSB-2555B / TSB-165A electret capsules
• All OpAmp (no FET at first stage)
• 48V PantomPower
• Bootstraping (if possible)

You might want to start with one of Guru Scott Wurcer's Charge Amp circuits in Zapnspark's MicBuilders directory but leaving out the extra FET.

You also need to consider the current draw carefully as there is ju...ust enough current from P48 for this.

Power consumption of OPA mike circuits are slightly more complicated than eg Schoeps type circuits cos they are Class B amps.  The current you need is more than just the idle current.  You have to consider what happens if the current draw increases.

That's why I suggest using Wurcer/Zap's circuits as a starting point as these are proven and will have thought through some of this.

 

[Khron]

My point was that the OPA134 isn't necessarily the be-all-end-all of FET-input opamps And besides, with today's technology, you'd be hard-pressed to detect drastic differences between them - provided, of course, they're properly implemented and not oscillating or whatnot

Surely some half-decent ones with more frugal power requirements have been developed since the introduction of the x134

How about something like the ADA4610 series? Under 4eu/pc for 1-off quantities for the dual version at Mouser, specs look ok (fine, one less zero in the THD figure, but lower noise @ 1khz - that's nitpicking already but barely 1.5mA idle current per channel)...
http://www.analog.com/media/en/technical-documentation/data-sheets/ADA4610-1_4610-2_4610-4.pdf

 

[homero.leal]

Yes Ricardo... It is. It was referenced  by Jules, but it sparked on me a great interest on the design. It may be complex , but I think the acomplishment it's great.  And also fun... I mean, it's the first time I see actual LEDs as active diode components on a mic design. 

Schematic is not to complex until bootstraping is added. I think this is the main area where design could be simplifyed. There should be a simpler way to add bootstraping... I think.

May be design "goals"  in my last post will be too hard to reach. But anyway,  I'll follow your advice about having Wurcer design as a starting point.

Best regards!

HL

 

[homero.leal]

Thank you Khron. Will take a look to this ADA4610. Looks promising.

Best regards!

HL

 

[Khron]

I just did like on TI's site - filtered'em down to FET-inputs only, and dip8/soic8, then took a peek at Mouser's prices

Fortunately, there aren't THAT many (good) opamp manufacturers, so browsing through their sites shouldn't take too long

Thank you Khron. Will take a look to this ADA4610. Looks promising.

Best regards!

HL

 

[ricardo]

My point was that the OPA134 isn't necessarily the be-all-end-all of FET-input opamps   And besides, with today's technology, you'd be hard-pressed to detect drastic differences between them - provided, of course, they're properly implemented and not oscillating or whatnot.

Surely some half-decent ones with more frugal power requirements have been developed since the introduction of the x134

How about something like the ADA4610 series? Under 4eu/pc for 1-off quantities for the dual version at Mouser, specs look ok (fine, one less zero in the THD figure, but lower noise @ 1khz - that's nitpicking already but barely 1.5mA idle current per channel)...
http://www.analog.com/media/en/technical-documentation/data-sheets/ADA4610-1_4610-2_4610-4.pdf

I did say OPA134 was

one of the FET opamps I would consider

There are other good reasons for OPA134. 

Of the uber OPAs, it is the least sensitive to poor decoupling & layout.  Very often, this means  in real life its THD etc outperforms more highly spec'd OPAs.  see eg Kingston's thread http://www.groupdiy.com/index.php?topic=37307.80 from about #41.

Next, and important for mikes, it is VERY immune to RFI.  Dunno if other FET OPAs are as good but this type of thing takes a long time and much experience to get a good reputation with an old fogey like me.

The less than 0.8dB difference in Env between OPA134 & ADA4610 makes NO difference for a small capacitive source like a mike capsule.  The important parameter is Eni (the current noise) and this is poorly documented or even non-existent in the datasheets.

For power consumption, if the quiescent current isn't greater than what P48V will supply, it often means a greater Class A region which avoids the Class B problems I mention earlier.

But the main point against FET OPAs for me is that Guru Wurcer doesn't show an example in his Linear Audio articles.  I may be wrong.  Dis beach bum don't get to read these excellent dead tree books.  Anyone got them all?

Rutgers accuses Guru Wurcer of terrible sins but it's clear he puts words in Guru W's mouth and also doesn't understand the issues.

BTW, Guru Wurcer designed some of the quietest OPAs in the known universe.

Homero,  the reason for the bootstrapping is varying Cin on the OPA inputs.  But Guru Wurcer's charge amp circuits DON'T HAVE VARYING Cin!  No bootstrapping is needed.  In fact it would introduce more THD.

Some people are impressed by complex circuits.  With my small brain, I prefer simple circuits which outperform complex circuits  8)

Of course you can still have LEDs etc shining out through holes in your mike body.  They don't have to 'do' anything as long as they don't degrade performance.

 

[homero.leal]

Ok Ricardo... I had the idea that varying Cin with Bootstraping had an effect in reducing THD. :S

I already have built several Occam-3B from Zap's schematic, which are based on Scott Wurcer's charge amplifier design. They work great using TSB-2555B capsules. Also, they have a great ouput level and detailed sound. Don't know about distortion, since don't know how to measure it yet.

I think that OPA134 won't be an option for a 48V Phantom Powered mic design. Quiescent current is 4-5mA per amplifier. Having a power box with positive and negative 12V source, and feeding the mic with a 7 pin XLR cable similar to the way tube mics work, could be an option for using OPA134 on a mic design.

But Anyway, I will try to modify Zap/Wurcer's design, to use a low power (<1.5mA) FET input OPA, removing original FET and using a charge amplifier for first stage as you suggested. Let's see ho it goes!

Regards!

HL

 

[ricardo]

Ok Ricardo... I had the idea that varying Cin with Bootstraping had an effect in reducing THD. :S

Actually varying Cin introduces distortion.  In a voltage follower like Rutger, the input goes up & down wrt the rails so Cin varies.  Bootstrapping the rails keeps the inputs sorta constant wrt the rails and reduces THD due to this.

But so does a charge amp.  So no varying Cin and no THD.

If you have a OPA134, it is easy to modify one of your Zap/Wurcer charge amps to use it.  As only one OPA is used, there is more than enough current from P48V.

If you find another FET OPA which works well in this, let us know.

 

[homero.leal]

Well Ricardo, I was thinking on using two OpAmps... something more or less like this...

It's clipping... need to know why

I removed filter caps just to show the general idea. They will be added later on the design.

Function generator is feeding a 5Vp sinusoidal wave.

What do you think?

Best regards!

HL

 

[Khron]

Are you sure the other end of that 1G resistor is supposed to go to the output of the top opamp?

Also, since it's a single-supply arrangement, a series coupling cap between the capsule/1G node and the opamp input might be in order. But i stand to be corrected

 

[homero.leal]

Hi Khron,

I took this as the base design:

http://s114.photobucket.com/user/zapnspark/media/OCCAM-3b-1.png.html


It's a Charge Amplifier design from Zapnspark (Occam-3b), based on a  Scott Wurcer's article from Linear Audio Magazine.

This design is available on yahoo  MicBuilders  group, on Zap's folder.

I removed the FET and 5.6k resistor,  replaced OpAmps for FET-input OpAmp (OPA134), an connected input directly to first OPA.

Cheers!

 

[Khron]

How about putting a (virtual) scope probe on the positive supply rail for the opamps?

They're gonna suck about 10ma at idle(!) together. With those 1.5k resistors added there, you're left with a short-circuit current of about 11mA, so that rail's gonna sag considerably...

EDIT: But then again, it also depends just how accurate the models are, so...