Don't improve the microphone too much, as that will make it more difficult for the ADC to outperform the microphone's dynamic range
Why are there 15 dB...20 dB differences in A-weighted noise between the PUI and Primo capsules and most of the others?
So if one happen to need more gain maybe it would be an advantage to do the Scotts Linear Audio circuit and add gain afterwords in the reception stage rather than relying on the FET in the capsule...
Seeing Scotts answer it seems to be in the electronics - but if so, how strange to differentiate a product catalogue by seemingly downgrading performance....?
Wouldn't a pentagon diaphragm spread standing waves better?
//
Seeing Scotts answer it seems to be in the electronics - but if so, how strange to differentiate a product catalogue by seemingly downgrading performance....?
Here's a comparison chart for the Primo capsules, Comparison Table for Primo Microphone Capsules - micbooster.com
The numbers here line up closely to what I posted. The EM23/21 is my favorite for actual recording (the EM21 cardioid seems to be no longer available) reaching as low as 8-10 dBA self noise. The SNR is simply the self noise re: 94dB SPL (1 Pa). I prefer to convert everything to volts and nV/rt-Hz as soon as possible for the signal chain computations.
Marcel what do you mean by your comment on comparison to everything else, I find the self noise vs diameter to correlate quite well over many brands?
BTW do you need any parts, i have some SSM2019 mic pre-amps you can have?
For some reason I think I have decided to go omni. Don't ask me details but as I recall it it is partly due to the low end performance of omnis.
I haven't received your article i LA yet Scott even if I payed this morning - I have contacted Jan on how to proceed. Many thanks for the offer Scott - I will read the article and decide on how I would like to compose the signal path.
I don't know if you had the time and interest to read the thread but I plan not use phantom power in order to clean up and simplify the solution to a minimum.
//
but as I recall it it is partly due to the low end performance of omnis.I haven't received your article i LA yet Scott even if I payed this morning - I have contacted Jan on how to proceed. Many thanks for the offer Scott - I will read the article and decide on how I would like to compose the signal path.
I don't know if you had the time and interest to read the thread but I plan not use phantom power in order to clean up and simplify the solution to a minimum.
//
I indeed meant compared to cheap capsules. Five examples:
PUI Audio POM-2730L-HD-R: 6 mm diameter, 20 dB(A) SPL noise.
Panasonic WM-61A: 6 mm diameter, noise specified as less than 32 dB(A).
PUI Audio AOM-5024L-HD-R: 9.7 mm diameter, 14 dB(A) SPL noise.
CUI Inc CMA-4544PF-W: 9.7 mm diameter, 34 dB(A) SPL noise.
Primo EM273: 10 mm diameter, 14 dB(A) SPL noise.
PUI Audio POM-2730L-HD-R: 6 mm diameter, 20 dB(A) SPL noise.
Panasonic WM-61A: 6 mm diameter, noise specified as less than 32 dB(A).
PUI Audio AOM-5024L-HD-R: 9.7 mm diameter, 14 dB(A) SPL noise.
CUI Inc CMA-4544PF-W: 9.7 mm diameter, 34 dB(A) SPL noise.
Primo EM273: 10 mm diameter, 14 dB(A) SPL noise.
How do you figure the noise? Take the Primo 2 wire vs 3 wire connection there is a 9dB gain in the 2 wire connection but the same 14dB self noise spec. The noise at the output of the two can not be the same (for the same free field SPL one has 9dB more output, noise included). Somehow the input sensitivity has to figure into the RTO noise. My mic was designed to be 110mV/PA (-21dB) so any pre-amp noise was swamped even at low gains.
Another point, these capsule FET's are very low gm so the output impedance is >>100 Ohms.
Thats what I wonder also
So 0,014V for 94 dBA ->
120 dBA (loud!!) = 0,28 volt
30 dBA (quiet!!) = 7,5 mV
137dBA (coff..) = 2 V
Can that be correct? If it is I think I will add say 10 dB fixed gain but only if there was a real benefit for the ADC in any way or form...
+10dB gain:
0,044V for 94 dBA ->
120 dBA (loud!!) = 0,88 volt
30 dBA (quiet!!) = 22,7mV
127dBA (coff..) = 2 V
It seems easier to have a grip on the high levels... clipping etc... but how about in the "grass"... thinking resolution/distorsion... what say you? Move up - stay native (zero gain) with EM273 or perhaps damping?
How loud will I encounter??? :-D
//
So 0,014V for 94 dBA ->
120 dBA (loud!!) = 0,28 volt
30 dBA (quiet!!) = 7,5 mV
137dBA (coff..) = 2 V
Can that be correct? If it is I think I will add say 10 dB fixed gain but only if there was a real benefit for the ADC in any way or form...
+10dB gain:
0,044V for 94 dBA ->
120 dBA (loud!!) = 0,88 volt
30 dBA (quiet!!) = 22,7mV
127dBA (coff..) = 2 V
It seems easier to have a grip on the high levels... clipping etc... but how about in the "grass"... thinking resolution/distorsion... what say you? Move up - stay native (zero gain) with EM273 or perhaps damping?
How loud will I encounter??? :-D
//
A decent source follower won't effect the EM23's output. BTW a DPA electret capsule goes for ~ $1800.
Reading a lot. Seeing that the PCM4222 has an increasing distorsion when using the last 20 dB high level input. Still its not very high... say + 10 dB...
The differential input stage of the EVM board has a 6 db attenuation. 60 dB down there is -60 dB THD so below the noise of any room I suppose with some 20 dB margin at least?
The differential EVM board input may be used single ended by grounding the - pin. There goes another 6 dB - thinking a direct connection of the capsule... But there is an impedance situation requiring a buffer!? Primo seem to indicate 5,6k as load and 5V bias in some basic configuration.
So what could be a very first trial?
Alt 0) Mod the EVM to a higher input impedance, hook it up to the EVM with -pin to earth and feed 5V to the capsule. This would I assume give some sound
Alt 1) Build a mic input stage ala' Wurcher with differential out to drive the stock EVM interface. But which one - there is quite a number of variants in the articles - all involving P48? Scott - any recommendations for this particular situation? No requirement for the standard phantom feed. I can make any local power source and feed it discreetly with own wires. What is this bootstrapping?
What I dont get is if "Capsule" in part 2 / Figure 4: Input Stage – Phase Splitter is equal to a EM273 or is it a really bare capsule without built in FET that is shown?
This "Vocm" seem to be a way to have DC offset in control and leave DC blocking caps out - nice! Is that a TI special? Something for the output stage of the mic stage to
- - - - - -
Ideally I would want a stage (F(x)) that can on the input side make an ideal situation for an EM273 and on the output side interface the EVMs differential 570 ohm input - without a trace of "P48" (Basic version without gain but optionally 10-20 dB perhaps...)
Cables between capsule and input stage to be fixed/soldered (no connectors!).
Help!? pretty please...
//
The differential input stage of the EVM board has a 6 db attenuation. 60 dB down there is -60 dB THD so below the noise of any room I suppose with some 20 dB margin at least?
The differential EVM board input may be used single ended by grounding the - pin. There goes another 6 dB - thinking a direct connection of the capsule... But there is an impedance situation requiring a buffer!? Primo seem to indicate 5,6k as load and 5V bias in some basic configuration.
So what could be a very first trial?
Alt 0) Mod the EVM to a higher input impedance, hook it up to the EVM with -pin to earth and feed 5V to the capsule. This would I assume give some sound
Alt 1) Build a mic input stage ala' Wurcher with differential out to drive the stock EVM interface. But which one - there is quite a number of variants in the articles - all involving P48? Scott - any recommendations for this particular situation? No requirement for the standard phantom feed. I can make any local power source and feed it discreetly with own wires. What is this bootstrapping?
What I dont get is if "Capsule" in part 2 / Figure 4: Input Stage – Phase Splitter is equal to a EM273 or is it a really bare capsule without built in FET that is shown?
This "Vocm" seem to be a way to have DC offset in control and leave DC blocking caps out - nice! Is that a TI special? Something for the output stage of the mic stage to
- - - - - -
Ideally I would want a stage (F(x)) that can on the input side make an ideal situation for an EM273 and on the output side interface the EVMs differential 570 ohm input - without a trace of "P48"
(Basic version without gain but optionally 10-20 dB perhaps...)Cables between capsule and input stage to be fixed/soldered (no connectors!).
Help!?
pretty please...//
Attachments
Last edited:
In the text it describes that it can be either. For the three wire capsule you only need R1 and R2, the values picked to get the most dynamic range. You will need to have some kind of buffer because the two outputs have very different source impedances (if your A/D has very high input impedance you should be OK with a shortish cable only, no buffer).
Figure 10 is very easy to build if you want to try single ended, or add a dual op-amp diff out circuit. There are some rail to rail low noise op-amps around that are a good fit.