[ProAudio] Microphones question

Mon Jun 14 20:29:57 PDT 2021

This data sheet has become the new typical for TI ... I, too, used to respect the company ... but apparently all the analog people have departed.  My first experience was with a similar chip in the older TLV320 series.  The analog input specs and applications are like a page in a textbook that says "the solution to this problem is left to the reader as an exercise."  The data sheet for the PCM6240 is an analog joke!  Having recently done a project with electret mics an almost-identical CODEC in a Bluetooth module, I can tell you that the "mic" they talk about is an electret mic.  The "load" resistors set the DC operating point of the mic.  It's not about power transfer, it's about "bias" (the chip's "bias" output is generally the output of a very low-noise linear regulator (in the Bluetooth module my client chose, made by SiLabs.  Buried deep in its 70-page data sheet was advice to not use its internal "bias" regulator make one externally.
In my experience, the suppliers of cheap electret mics are completely clueless.  They'll tell you that supply voltage can be 1.5 to 6 V but that the load resistor should be, say 1 kΩ (with no explanation).  In fact, the pins on the mic are the source and drain of the internal FET.  In the engineering world, a FET data sheet would list Idss (drain current at zero gate voltage) and maximum drain voltage, among other things. The "recommended" load is chosen for the lowest supply voltage, which also results in the lowest sensitivity for the mic.  So I measured the drain current vs voltage for 10 samples to determine Idss = 450 µA and max drain voltage is 7 V for the part.  Knowing that, I used a 6 V supply and a load resistor 3 times the value of the "recommended" - which resulted in perfect class A operating point (drain voltage = 1/2 of supply) and a 10 dB increase in sensitivity. Apparently, the manufacturers of these $1 mics think their customers are all just hobbyists and don't want to confuse them with real specs.  And TI follows suit in their application notes, again without explanation. Apparently, when an IC has on-board digital processing, the bit-heads get to write the data sheets!  I just hope the analog ignorance doesn't leak over into the LTC products ... I have a lot of respect for the innovative engineers there.
I think, in some cases, we're dealing with the fact that analog design has fallen out of favor with universities and sold students on the idea that "the future is digital" ... but, as AI gets better won't they start writing code?  As I used to say in the 80s, "programmers will become the typists of the 21st century."
Ah, now I feel better ...Bill WhitlockVentura, CA

-----Original Message-----
From: Dan Lavry via ProAudio <proaudio at bach.pgm.com>
To: proaudio at bach.pgm.com
Sent: Mon, Jun 14, 2021 7:48 pm
Subject: Re: [ProAudio] Microphones question

 We are talking about audio fiction, so here we go: A new (year old) TI AD IC, the PCM6240-Q1, on figure 37 of the data sheet states: Figure37 and Figure38 show how to connect a DC-coupled microphone for a differential and single-ended input,respectively. The value of the external bias resistor,R1, must be appropriately chosen  based upon the microphone impedance. For a differential input,the value of the external bias resistor is recommended to be used for half of the microphone impedance,whereas for a single-ended input,the external bias resistor is recommended to be the same as the microphone impedance. They are advocating matching the  load to the mic output impedance (obviously for maximum power transfer). That optimal setup is of course only true if the mic could support such load while "behaving well" (linearly, no clipping or other issues). 
  I do respect TI. It is a good company with good reputation. So can we sort out facts from fiction? I think the answer is simple, but I will let you experts debate it if you wish. I am logging off for now, I am starting to feel like an old guy talking too much, and this "mic loading" is a whole other issue...
  Regards Dan Lavry

  On 6/14/2021 3:48 PM, Bill Whitlock via ProAudio wrote:

The following article, "11 Myths about Analog Noise Analysis," is a great read for most of the topics raised in this thread: 
 https://www.analog.com/en/technical-articles/11-myths-about-analog-noise-analysis.html# 
  Bill Whitlock 

 -----Original Message-----
 From: Dan Lavry via ProAudio <proaudio at bach.pgm.com>
 To: David Josephson <dlj at josephson.com>; proaudio at bach.pgm.com
 Sent: Mon, Jun 14, 2021 1:14 pm
 Subject: Re: [ProAudio] Microphones question

   Yes, we have to pick our battles. The 150 Ohm that is talking about has near 0 DC. But the 6.8K resistor to 48V phantom your mic is connected to has DC... that is where material may matter.  Yes, there is more to learn every day...  Regards Dan Lavry 

   Sent from Samsung Galaxy smartphone.  

    -------- Original message -------- From: David Josephson via ProAudio <proaudio at bach.pgm.com>  Date: 6/14/21 12:11 PM (GMT-08:00)  To: proaudio at bach.pgm.com  Subject: Re: [ProAudio] Microphones question  
    The microphone measurement standard recommends the use of A-weighted rms 
 noise specifications in addition to ITU-R BS.468-4, which specifies not 
 only a frequency weighting curve but a quasi-peak detector which 
 penalizes for low frequency pops and clicks, which can be a factor for 
 some poorly executed microphone designs.

 Yes, even some of the experts in the standards world talk about "low 
 noise" resistors. We have to pick our battles.

 On 6/14/2021 12:01 PM, Dan Lavry via ProAudio wrote:
 > Excess noise is important in some cases, and it is material dependent. I 
 > agree that the 150 for the noise measurement can be "any: 150 Ohm.
 > 
 > Regarding "A weighting", I just publish both figures. In most cases it 
 > is 2dB difference.
 > 
 > Regards
 > 
 > Dan Lavry
 > 
 > On 6/14/2021 11:03 AM, Bill Whitlock via ProAudio wrote:
 >> Dan, I'm glad you caught that one!   Noise is a "stand alone" that 
 >> needs no reference except a unit of measure - and, as frequently 
 >> omitted by marketing types, a stated bandwidth.  And I have no problem 
 >> in using weighted figures as long as it's clearly stated.
 >>
 >> In my mind, the biggest barrier to understanding noise and its 
 >> implications is the lies, distortions, and half-truths perpetrated by 
 >> marketing folks!  A measurement, with test conditions and references 
 >> fully disclosed, is not subject to interpretation.  Sadly, many folks 
 >> are "educated" by such misleading information and will believe, for 
 >> example, that the 150 Ω resistor used for testing preamp noise must be 
 >> a special "low-noise" resistor.  In fact, thermal noise has no 
 >> dependence on the resistive material at all.  So-called "low-noise" 
 >> resistors are low in "excess noise" that occurs when DC current flows 
 >> in it - and there's a huge difference among those!  Anyone whose ever 
 >> heard me lecture is well aware that no love is lost between me and 
 >> marketing people.
 >>
 >> Bill Whitlock
 >>
 >>
 >> -----Original Message-----
 >> From: Dan Lavry via ProAudio <proaudio at bach.pgm.com>
 >> To: Mike Rivers <mm1100 at yahoo.com>; proaudio at bach.pgm.com
 >> Sent: Mon, Jun 14, 2021 9:27 am
 >> Subject: Re: [ProAudio] Microphones question
 >>
 >> I think you are confusing things. EIN does not need to be measured at 
 >> full scale, and it is a real representative of the noise contributions.
 >>
 >> Micpre noise is low and not easy to measure directly. Say you test 
 >> system can measure down to 1uV, but your signal is 0.1uV. It would be 
 >> difficult to measure the noise directly.  Say you set the gain at 
 >> 60dB, the noise will become 100uV, and a 1uV system can measure that. 
 >> At 40dB gain you get 10uV, still good enough to measure. In the first 
 >> case Ein is 100uV/1000, in the second case 10UV/100, both yield 0.1uV 
 >> which is the real input noise.
 >>
 >> Of course that will not hold if the design is poor. Any reasonable 
 >> design will yield very similar ein at 60dB and 40db.
 >>
 >> Regards
 >> Dan Lavry
 >>
 >>
 >>
 >>
 >> Sent from Samsung Galaxy smartphone.
 >>
 >>
 >> -------- Original message --------
 >> From: Mike Rivers via ProAudio <proaudio at bach.pgm.com>
 >> Date: 6/14/21 8:07 AM (GMT-08:00)
 >> To: proaudio at bach.pgm.com
 >> Subject: Re: [ProAudio] Microphones question
 >>
 >> How about a little discussion on the value of knowing the EIN of a 
 >> preamp and how the figure is useful to the designer?
 >>
 >> Here's why I ask:
 >>
 >> When I was writing reviews regularly, EIN was (and still is) often 
 >> quoted in the product's specs. The numbers ranged from -125 dBu (just 
 >> a couple) to a majority being either -127 or -128 dBu. Manufacturers 
 >> liked to tout it because it was a nice low number with "noise" in its 
 >> name. And it was always measured at a level within a dB or so of 
 >> clipping since that's it looked the best.
 >>
 >> But unless I missed the appropriate chapter, EIN is a calculated value 
 >> - the measured noise level with the gain subtracted out. So a preamp 
 >> with 60 dB of gain that advertised EIN=-128 dBu could be expected to 
 >> put out -68 dBu of noise, measured, of course to the advantage of the 
 >> marketing department with the appropriate input termination (0, 100, 
 >> or 150 ohms usually) and output load.
 >>
 >> So, among design engineers, what's the big deal about EIN? Is there a 
 >> better way of measuring it that's more meaningful? And if you can 
 >> squeeze another dB of EIN out of a design, how significant will this 
 >> be to the user?  I
 >>
 >> , know "it depends."
 >>
 >> -- 
 >> For a good time callhttp://mikeriversaudio.wordpress.com  <http://mikeriversaudio.wordpress.com/>
 >> _______________________________________________
 >> ProAudio mailing list
 >> ProAudio at bach.pgm.com <mailto:ProAudio at bach.pgm.com>
 >> http://bach.pgm.com/mailman/listinfo/proaudio 
 >> <http://bach.pgm.com/mailman/listinfo/proaudio>
 >>
 >> _______________________________________________
 >> ProAudio mailing list
 >> ProAudio at bach.pgm.com
 >> http://bach.pgm.com/mailman/listinfo/proaudio
 > 
 > 
 > ------------------------------------------------------------------------
 > Avast logo <https://www.avast.com/antivirus> 
 > 
 > This email has been checked for viruses by Avast antivirus software.
 > www.avast.com <https://www.avast.com/antivirus>
 > 
 > 
 > <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
 > 
 > _______________________________________________
 > ProAudio mailing list
 > ProAudio at bach.pgm.com
 > http://bach.pgm.com/mailman/listinfo/proaudio
 > 
 _______________________________________________
 ProAudio mailing list
 ProAudio at bach.pgm.com
 http://bach.pgm.com/mailman/listinfo/proaudio
    _______________________________________________
 ProAudio mailing list
 ProAudio at bach.pgm.com
 http://bach.pgm.com/mailman/listinfo/proaudio

  _______________________________________________
ProAudio mailing list
ProAudio at bach.pgm.com
http://bach.pgm.com/mailman/listinfo/proaudio
 _______________________________________________
ProAudio mailing list
ProAudio at bach.pgm.com
http://bach.pgm.com/mailman/listinfo/proaudio
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://bach.pgm.com/pipermail/proaudio/attachments/20210615/2ae62243/attachment-0001.html>