We've Got A New "King Of The Low End" LDC...

[mshilarious]

That hasn't been a secret or anything for a very long time. Jon doesn't promote his goods other than address technical questions that users have, and I believe he's differed that to PMs. He's also given away some of his schematics for FREE.

I give away all of my schematics on request, although I don't think anyone would really want my latest; the parts count has grown with my ability to stuff more and more components into very small spaces. Also, my primary goal is minimization of power consumption, so all of my circuits now use class AB opamp outputs except for one that is just too small. I also like using subminiature transformers because I dig transformers and I want my products to be as small as possible. Most designers would find them too quirky, although I say hey, why waste an absurd amount of power hitting a large output transformer that is designed *not* to saturate, just to drive it into saturation? It takes very little power to saturate a subminiature! But anyway . . .

So the current mic circuits use about 30 components on a 0.8 square inch PCB. I also have a smaller PCB, about 0.6 square inch, with a breakout 0.75 diameter round PCB for remote coupling to a capsule. Nothing fancy, your basic cascoded FET feeding my opamp buffer output--although with current draw of 1mA, you get 44V to the capsule without requiring step-up. Just like in the old days of transformer output, but this time with no loss in sensitivity and a circuit that fits inside a Neutrik XX connector.

But I am actually not doing any new work on microphones, mostly I concentrate on battery-powered preamps now. While this debate has raged, I have been at work on a 6 square inch PCB for the newest preamp with over 200 components

It's really not anything for the home studio market at all though. At least as long as people conceive of a studio as a room with a bunch of metal filling up a rack . . .

I'm waiting for the next gen of USB CODECs to migrate to portable interfaces. Basically, my goal is to pile up technology for a niche market that will not be tempting to larger players. Because as we have seen from this thread, there is really no barrier to the Chinese manufacturers incorporating higher quality components and circuit designs; it's just a matter of being driven by demand from their customers.

It's a puzzle to me why so many Chinese microphones have circuits much more primitive than what Shure, AT, and the Europeans are doing, since the Chinese have no trouble producing technological miracles like the iPhone. And now we see the larger players pushing their production to China, using the same circuits they previously produced domestically.

The writing is on the wall; mics like the SP CS series, AT 20 series, and now I guess the 2003A are a part of it. It will be getting tougher for the aftermarket to add value as moderately sized importers will push those revisions back to the factories.

 

[moresound]

That hasn't been a secret or anything for a very long time. Jon doesn't promote his goods other than address technical questions that users have, and I believe he's differed that to PMs. He's also given away some of his schematics for FREE.

Not to mention giving away his microphones .... and I'll be damned if I couldn't get my hands on one then.

 

[chessparov]

BIG + 1 regarding the Studio Projects CS series on vocals (haven't tried the 2003a yet).

Thought the AT 2020 is pretty nice, though not in the same ballpark.

Chris

 

[mshilarious]

BIG + 1 regarding the Studio Projects CS series on vocals (haven't tried the 2003a yet).

Thought the AT 2020 is pretty nice, though not in the same ballpark.

Chris

I was thinking more of the AT2035. I mean it's an AT3035 with a black finish, right? $150. Wasn't too long ago that the AT3035 cost $300 . . .

 

[mshilarious]

Hey Mr. Joly, remember a few hundred posts back when I said that the 2003A's 6mA draw was on the borderline for problematic, and you kinda brushed that off? Well, here's a perfect example of the problem:

http://www.gearslutz.com/board/so-much-gear-so-little-time/528907-oktave-mk-319-low-output.html

As I see the specs for the MK319, it's 48V @ 8mA. So the voltage at the mic should be ~21V (it cannot be 48V under load, as you stated, because the mic's draw will cause a voltage drop across the parallel 6k8 supply resistors, according to Ohm's Law: 0.008 * ( 6800 || 6800) = 27.2V). Instead, measure the phantom on an empty channel with the MK319 plugged in another, that cannot drop below 44V or the supply is noncompliant.

Now, if that OP's mixer has noncompliant phantom power; that is, it cannot supply 8mA without the supply rail dropping below 44V, then he would experience the exact effect he is describing. Since he bought this used, without expert advice he will think it is defective and go back to the seller, who will receive the mic back and it will work properly on his compliant gear. I see Paypal chargebacks in their future If it was new gear, the store would get the return.

There is way too much gear out there with noncompliant phantom power, even rackmount gear running off wall AC with an internal power supply where there is absolutely no excuse for noncompliant phantom. That's appalling enough, but what is more appalling is that the mic manufacturers get blamed when it's not their fault--mic manufacturers almost universally provide a current spec, but most preamp manufacturers do not certify compliance at 10mA per microphone, and some of those that do are . . . well, exaggerating.

Thus, it behooves a mic manufacturer to not get greedy and stay at 5mA or less. Mr. Roy, take note!

I used to do a tube mic that had to be at 9mA or it wouldn't work at all. The major reason I killed it was because I had about a 10% return rate due to noncompliant preamps. I can't imagine how Earthworks can tolerate the same problem given their 10mA draw; one year of that experience was plenty for me. Although their customers are a bit more affluent than mine, I suppose . . .

 

[drBill]

Now, if that OP's mixer has noncompliant phantom power; that is, it cannot supply 8mA without the supply rail dropping below 44V, then he would experience the exact effect he is describing. Since he bought this used, without expert advice he will think it is defective and go back to the seller, who will receive the mic back and it will work properly on his compliant gear. I see Paypal chargebacks in their future If it was new gear, the store would get the return.

There are ignorant people everywhere. You can't design around that.

IMO, a designer should design for his/her VISION, and let the chips fall where they may. The world does not need any more rules that designers must conform to.

To design for the lowest common denomenator (cheapo mixers with faulty phantom) is to limit the designers creativity, and possibly a very cool product.

Now if, like you, the designer wants to design for consumer oriented crap so that they won't get returns, then go for it. It's a legit market. But I'd guess that for the most part, that product won't be of interest to high end users even if it is decent. Just as engines that run on regular octane gas are generally not in the crosshairs of those racing nascar.

 

[mshilarious]

Now if, like you, the designer wants to design for consumer oriented crap so that they won't get returns, then go for it. It's a legit market. But I'd guess that for the most part, that product won't be of interest to high end users even if it is decent. Just as engines that run on regular octane gas are generally not in the crosshairs of those racing nascar.

Pithy, but there is no correlation between microphone quality and current consumption. The U87ai uses 0.8mA (the original used 0.4mA). The Schoeps CMC6 uses 4mA, but 8mA with P12--that's the same amount of power as 2mA at P48. The Beyer MC930 uses 4.6mA, but it accepts down to 11V. The AKG C460B was <1mA, and accepted down to 9V; the current C451B of course is electret and manages the same performance. But the others are all externally biased. The current C414s are 4mA, and they have a LED and a microcontroller! Although modern MCUs are incredibly low-power . . . TI has a Youtube of one of their MCUs running on two electrodes stuck into a grape

What else . . . MT Gefell UM 900 does 4mA, and it has a tube! Whereas oddly enough, the FET UM 930 needs an extra 0.5mA . . . go figure . . .

These are some of the most highly regarded microphones in the world. I would submit that all of those circuit designs are much more like Formula 1, and the average Chinese LDC more like NASCAR. You can equate volumetric efficiency in engine design to power efficiency in microphones, if you like

Anyway, wouldn't you say that a brand like MXL is fairly likely to be used with "consumer oriented crap"? And here we have an example of EXACTLY WHAT I SAID would happen with an even thirstier mic . . .

And as I also said, it would be a great opportunity for a modder to alter the circuit to something that performs better and still uses less power. What if that guy sends the MK319 off for repair, and it comes back with no flaw found (of course, it may very well have a flaw, this is still hypothetical)? Or instead, what if it came back modified so that it worked with his existing equipment? Which service would he value more highly?

But maybe you consider the MK319 the pinnacle of achievement of microphone circuit design, I don't know . . . I do know that 8mA plus no step-up equals a 21V polarization to the capsule. Is that really optimal, especially in terms of noise performance? I really have a hard time believing that.

 

[drBill]

Pithy, but there is no correlation between microphone quality and current consumption.

That is certainly your opinion, and you're welcome to it. Are you trying to save the rainforest with a few less milliamps consumed or something?

I'm for not telling designers - you or otherwise - how to do their jobs. That's the best way to come up with innovative products.

Perhaps you should quit designing mics, and go to work for the government. I'm sure you can convince someone there that this rampant mis-design of microphones need to be controlled!!!



Should we limit the voltage rails on outboard as well? +/- 24v seems so excessive. Even +/- 16v. We should be able to run on a tenth of that.......

 

[drBill]

But maybe you consider the MK319 the pinnacle of achievement of microphone circuit design, I don't know . . . I do know that 8mA plus no step-up equals a 21V polarization to the capsule. Is that really optimal, especially in terms of noise performance? I really have a hard time believing that.

I say let the market decide. My personal mic preference may or may not be in line with the general mic buying public. If they think it's a POS as you do, then it will die an untimely death. On the other hand, I'd say it looks to be becoming something of a cult classic. I've used one or two, and it is definately a mic with "personality". Would that personality be there running on 0.4mA?? I dunno. Don't really care. Now, whether or not one likes that personality is completely subjective. Perhaps if it ran on 0.4mA, the allure would not be there, but that's not for me to decide. I just stick talent in front of em and push buttons..... If I like it, it stays up. If not, I pull it down.

I don't just design em, handle em, package em up and send em off......

 

[mshilarious]

That is certainly your opinion, and you're welcome to it. Are you trying to save the rainforest with a few less milliamps consumed or something?

No, I am trying to make devices as interoperable as possible. That is a basic principle of good design. I am also trying to deal with preamp manufacturers, some of whom are less than forthcoming but have the color of authority simply because of sales volume.

I mean I've seen some silly things in preamps. One preamp regulated an 80V rail down for phantom. No problem there, except they used a 47V zener feeding a Darlington transistor as a regulator, and then used a 330 ohm resistor to feed the phantom rail. The transistor drops 1.5V, and when you plug in one mic at 10mA, the series resistance drops another 3.3V. Result? 42V on the rail, which is out of compliance. Now, that is probably not the end of the world, it's not far out enough to matter, but it would have cost the manufacturer *absolutely nothing* to use a 50V zener instead

But then there are other units, some of them not too cheap, one in particular I can think of had lots of expensive transformer-coupled inputs, a nice-looking external power supply . . . and could only put 7mA on any one mic input (both pins parallel) into a short circuit. FAIL.

So you are being purposely obtuse because you refuse to admit I am correct. I adequately demonstrated that power consumption and microphone quality are not correlated. I could go on all night with more examples. How about if I correlated the power consumption of all 90 microphones in Mike Jasper's shootout with their rankings? What do you think I would find? Unfortunately I can't do that now, I am going out to the shop momentarily (I'm a night person). But I am good enough at statistical methods to demonstrate significance of any relationship that exists--or doesn't exist. Unlike physics, I studied that in college

Bottom line, it is very likely that power consumption has caused the OP of that thread a problem. And TWO other users on that thread in just a very few posts may have experienced the same problem, because nobody should be comparing a 13mV/Pa spec mic to an SM7, which is 1.12mV/Pa. If you aren't good at volts, that's a 21dB difference. How is it possible that the MK319 has lost 21dB of sensitivity? What is your alternate explanation?

As an aside, when you put me down for only owning six microphones, did you know I was a mic designer? Honest answer, please.

Back to the MXL 2003A . . . this is *exactly* why I told them 6mA is walking the line. I think they are probably 95% safe, unlike our poor MK319 users, but if they are still reading this thread, I bet they also appreciate the friendly advice . . . cause I bet they don't want a lot of newbs dragging their new MXL mics straight back to Guitar Center because every other condenser mic they have works with their USB interface, but the 2003A didn't . . .

 

[drBill]

So you are being purposely obtuse because you refuse to admit I am correct.

No, I'm not being purposely obtuse, and I never said you weren't technically "correct". I suspect what you say is in fact industry standard "correct". However, I find incredible beauty, creativity, and "outside the box" ideas in "screw up's". And by putting your industry standard into force, I suspect we might miss out on some otherwise very cool stuff. (A couple that come instantly to mind are "front end overdrive" in guitar amps. Technically wrong from a design standpoint 70 years ago, but now developed into an art.... Or how about digital aliasing distortion - once avoided like the plague, now being brought into mainstream in some very cool and innovative sound libraries.) That's all I'm saying.

Now, as a designer, all these minutae might be VERY important to you. As an end user, I could care less. If a mic doesn't work on one phantom powered pre, I'll switch mics or pre's. But I'd prefer a mic that didn't work with one pre in the chance opportunity to let designers push the outer limits and make wrong, incorrect, "stupid" decisions in their designs instead of being hand tied to your choice of "industry conformity". That's all I'm saying..... There's room in your sandbox for that cantancorous, ADD Aspergers child don't you think?? He's got something cool up his sleeve....as long as you don't make him sit quiet in the corner.

 

[drBill]

PS - In case you haven't noticed, I (like the most creative people I've run across) REBEL against rules.

As soon as you tell me :

I can't put verb on a Kick Drum, can't have a bone dry vocal, can't put distortion on the violin section, can't put drums in the L channel and Vox in the R, can't split cellos into antiphonal L/R group's with a full orch playing, can't put autopan on the BGV's......

That's where I'm headed....


If MY MICROPHONES were endowed with consciousness and you asked them where they would end up tomorrow.....they would cower in fear. heh heh

 

[mshilarious]

No, I'm not being purposely obtuse, and I never said you weren't technically "correct". I suspect what you say is in fact industry standard "correct".

Well there is an industry standard, it's 10mA per microphone. I am not aware of a single microphone that violates that standard. If you want more power, you have to provide an external power supply. Marik is a proponent of that, but if you want a P48 compliant mic, you are limited to 10mA.

But the preamp manufacturers do not adhere to that standard, nor do they properly disclose their noncompliance. You see the problem? MXL is technically compliant, so is Oktava, so is Earthworks, so was I. But customers don't see it that way; read that thread for yourself.

So you could say we beautiful and perfect mic manufacturers are forced to compromise our creative principles because of the peccadillos of preamps, and you'd be right. Still doesn't get me paid . . .

Or how about digital aliasing distortion - once avoided like the plague, now being brought into mainstream in some very cool and innovative sound libraries.) That's all I'm saying.

Have to be honest, I don't know why anybody would want aliasing distortion. It's not unlike intermod distortion when playing a chord on a heavily distorted guitar amp (which I don't mind so much), but worse. Now, I enjoy playing very dissonant music myself, but even I have a hard time seeing the masses going for that.

Also, if we talk about aliasing distortion in VSTs, I think the trend is very much the other way. Many VSTs of five years ago, including some of the most highly regarded, had aliasing and it did not make them sound better. Quite the contrary, I think it was a major reason why people didn't like VSTs compared with hardware (the other reason was a lack of proper harmonic distortion modeling). Analog hardware doesn't alias. With the available increase in CPU power, I would not even consider any compressor VST that didn't oversample. I don't care how good it sounds, the exact same algorithm will sound better oversampled, period.

That's why I wrote my own compressor VST . . . not that there aren't a lot of good freebies out there these days, but mine looks and works the way I want it to . . . and it doesn't alias. Well, down -100dB, anyway. I have an old PC, so I only did 4x oversampling

 

[drBill]

Have to be honest, I don't know why anybody would want aliasing distortion. It's not unlike intermod distortion when playing a chord on a heavily distorted guitar amp (which I don't mind so much), but worse. Now, I enjoy playing very dissonant music myself, but even I have a hard time seeing the masses going for that.

Easy there msh, you're showing your age. I'm 100% confidant my grandfather said the same thing about overdriven guitar amps. The anti-aliasing thing is a trend, and it's growing very quickly. I know what you're saying, but if there's a way to break, manipulate, fry, or otherwise mutate the audio waveform, someone is going to figure it out and do it. Right now, anti-aliasing in sound design and sound libraries is getting big. Electronica music mostly, although it's definately invading ambient and film score stuff as well. Rock is probably next. And I have to say, as much as it rubbed me wrong at first, there's definitely some beauty there. At least for people without closed minds...... If you think inside the box, you're going to HATE it.

I just turned in a CD. Got a call from the company a few days later. They said their mastering engineer rejected some of the mixes because of distortion. I laughed. They sounded puzzled. I said, "I know. I meant that." They figured it out pretty quick after I clued them in, but their mastering engineer is an old skooler..... I had to open their minds up a little, (might have been painful on their end) but I think it's going to do very well for them...... I doubt my mom would like it if she were still around, but you never know, she surprised me fairly often.

 

[chessparov]

AFAIK the capsules of the AT 2035 and the AT 3035 are similar. The 3035 was made in Japan, however, vs. China for the 2035. Haven't tried the AT2035 or AT3035 yet.

Besides the Studio Projects CS series, along with the T3, particularly like the B1 (flatter MKII) that's in the same price range as these two AT's.

BTW much of the "tech talk" here is beyond me, but wonder how handy the Triton FETHead
would be to help prosumer (underpowered) mic pre's?

Chris

 

[drBill]

BTW much of the "tech talk" here is beyond me, but wonder how handy the Triton FETHead
would be to help prosumer (underpowered) mic pre's?

Chris

Ironically, based on aforementioned tech talk, I have found the fethead doesn't like certain phantom power, even when delivering a full 48v. (Yes, I grabbed my multimeter & checked at the mic.) My tech told me why once. I forgot. When they work, they work great. I love mine. They don't work with API 512B's though. Even though they work with other pre's in the same lunchbox.

 

[mshilarious]

I wrote something about the FEThead the other day. Basically, it appears to be very simple amp with its signal pins linked directly to a FET pair's drains. As such, the impedance of the circuit will be equal to the impedance of the preamp. That does two things; it makes the gain dependent upon input impedance (see the chart in their manual), and it makes the device more subject to cable capacitance. In other words, the more gain you have (if you have a higher input impedance amp), the worse your HF loss due to cable capacitance will be. It won't be bad unless you use a long cable (say 100 ft), but that means the device should be used at the preamp end and NOT at the mic end, as they say in their manual.

Also, they imply the device has poor power supply rejection, which means it also likely has poor common mode rejection. That would come from unmatched FETs not a long-tailed pair configuration.

I don't see where simply spending the money on a very quiet EIN preamp would not be a better idea . . . gain is cheap.

I'll right a bit more on aliasing later, including a little VST I thought of. It works OK on pure tones, basically another form of synthesis. On a mix, it gives you that '80s video game sound. Like most circuit-bent stuff, a novelty that wears off kinda quickly.

Oh, and Dr. Bill, I really like dissonant music. I could listen to Xenakis' "Pleaides" all day long But there is a difference between well-written dissonance and/or atonality and tinkering with toys.

 

[mshilarious]

OK, on aliasing and synthesis. Basically, what aliasing does is create sum & difference tones between the sample frequency and the signal. For example, if we choose to sample at 44.1kHz and our signal is a 1kHz tone, there will sideband distortion at 43.1kHz and 45.1kHz. Or where we are more concerned is with higher tones, say 10kHz. That will have sideband distortion at 34.1kHz and 54.1kHz. Let's go really high, 18kHz. OK, now we have 26.1kHz and 62.1kHz.

That 26.1kHz is getting too close for comfort, but it's still ultrasonic, so maybe we feel OK. But wait, this is only what happens if we've applied a brickwall filter at 1/2 sample rate before sampling. Our sample, strictly speaking, has a lot of aliased content, but it's all above 1/2 our sample rate. So when we remove such content on playback with a reconstruction filter, all is OK.

But what if don't have an anti-aliasing filter on input? Hmmm, now we have ultrasonic content that we can't hear, but will create sideband distortion in the audible range. Say, for example, you have a 35kHz tone from crickets chirping. That will alias as a 9.1kHz tone. Hmmm.

Now, let's talk about compressors. Compressors distort, especially really fast ones. This is true of both analog and digital compressors. This is harmonic distortion, and if the compressor operates symmetrically (we hope), it will be all odd-order.

In analogland, that's OK, that's the "flavor" of the compressor. In the digital world however we are creating ultrasonic content with those odd-order overtones. The resulting sideband distortions are not harmonically related to the signal. That is normally thought of as bad.

Here's a little picture of the UAD 1176SE plug taking on a 4kHz tone. Yellow is slow attack, blue is fast attack (both fast release). Note how yellow is pretty much just third-order distortion. This is probably somewhat like the analog version would do.

Blue, however, what up with dat? Suddenly we have second-order, OK, but why? This is a symmetrical compressor, and even-order distortion is always an asymmetrical distortion.

Well, it's because that distortion is actually an alias of the 4kHz fundamental and the 12kHz third-order distortion (really higher overtones, but you get the idea). But wait, what are those little peaks around 8kHz, offset at intervals of 100Hz? More aliases, due to our 44.1kHz (note the .1 part of that sample rate--100Hz).

The result is a hazy mazy crazy kind of fuzz scraped all over our audio. Some might call that "digititis", I dunno. I just call it a plug that doesn't oversample, or if it does, then the filter isn't good enough.

OK, on to creative use of aliasing. Well, actually that goes way back; it's perfectly possible to create sum and difference tones in analogland too, we call that device a ring modulator, just feed a high-frequency tone into one of its inputs and you'll get the same effect.

There are plenty of modulation plugs out there so I will let you play with that. Basically, when you have a relatively pure tone that you feed into such a device, you get a lot of overtones that are not harmonically related to your signal. This is described as "bell-like", because bells do that too. If the signal is more than 1/2 the control tone, there will be sub-inharmonic content generated. These usually aren't used on complex signals because there are too many tones generated to be considered musical useful, but hey, knock yourself out . . .

This faceless VST simply undersamples by 8x (by replacing every 8 samples with the first sample value in that set), for a 5.5kHz sample rate (at 44.1kHz). That's a pretty deep cut into the audible spectrum and will produce many many aliases. It would work OK on a guitar solo, not quite as well on drum overheads . . . put it on program material and I think it sounds like one of those greeting cards that play music. Those are really embarassing by the way, as they've packed more technology into a greeting card than I've ever put in one of my circuits

http://www.naiant.com/vst/Alias.dll

 

[Marik]

Well there is an industry standard, it's 10mA per microphone. I am not aware of a single microphone that violates that standard. If you want more power, you have to provide an external power supply. Marik is a proponent of that, but if you want a P48 compliant mic, you are limited to 10mA.

I am actually, not the only one who is "current hungry" . I remember reading Josephson proposing a new standard for phantom power of total 43ma available, with 2.2K resistors, each (of course, in this case all the Schoeps circuits will need to be re-worked).

Gus also posted about external PSU. While (I think) his ideas are somewhat different from mine, the bottom line, the external PSU would give quite a bit more flexibility in the design and open room for new topologies and design solutions, impossible for the current P48 standard.

I wrote something about the FEThead the other day. Basically, it appears to be very simple amp with its signal pins linked directly to a FET pair's drains.

Well, at least they saved about $0.40 for an additional pair of BJTs, pair condensers, and couple resistors, which would take care of all the drawbacks of such topology. What a marvel of engineering approach !!!

Which FETs do they use?

Best, M

 

[chessparov]

So...

Guys, would it be reasonable to deduce that a "slower" response microphone tranducer,
ala a dynamic (or ribbon/subset), would help to mitigate this distortion vs. a condenser?

Hey you're all rubbing off on me!

Thanks,
Chris