I decided to try making a simple electret microphone preamplifier for a PC headset, and found a design on GreatScott's YouTube channel here:
I soldered it on perfboard, and it works well except I get a constant "rushing" sound, not hum, on the output. It's quite loud, not just a low background noise. Things I've tried so far are reducing the gain-setting resistor R2 and R5 values while keeping the gain similar (around 22), and increasing the power voltage to 12V since 5V is low for a 5532 / 5534 (the rushing sound increased with power voltage).
Does anyone know what could be causing this noise?
A screen grab of the circuit:
I soldered it on perfboard, and it works well except I get a constant "rushing" sound, not hum, on the output. It's quite loud, not just a low background noise. Things I've tried so far are reducing the gain-setting resistor R2 and R5 values while keeping the gain similar (around 22), and increasing the power voltage to 12V since 5V is low for a 5532 / 5534 (the rushing sound increased with power voltage).
Does anyone know what could be causing this noise?
A screen grab of the circuit:
Where to begin?
The 5532 is probably not happy at 5V supply.
The inverting connection is BAD, it hisses. Use non-inverting.
Feedback resistors for mike level should be like 100Ω/2k, not 47/1Meg.
Your opamp bias should be filtered/bypassed.
I usually see your R1 as 2k or so, not 1k. The improved output may compensate the higher hiss resistance.
You NEED an output cap... this thing leaks DC and the hiss could be the sound of the next stage's input frying.
The double duplicated caps are pointless.
Of course you know that wide-open breadboards will suck-in ALL the hum/buzz in the room. Not traditionally hiss, but in today's rooms the riot of WiFi and Cellphone may work out as a sort of random hiss.
That guy has an excellent drawing hand, and a drafting pen I have not seen in years. But IMHO his circuit is far from optimum.
The 5532 is probably not happy at 5V supply.
The inverting connection is BAD, it hisses. Use non-inverting.
Feedback resistors for mike level should be like 100Ω/2k, not 47/1Meg.
Your opamp bias should be filtered/bypassed.
I usually see your R1 as 2k or so, not 1k. The improved output may compensate the higher hiss resistance.
You NEED an output cap... this thing leaks DC and the hiss could be the sound of the next stage's input frying.
The double duplicated caps are pointless.
Of course you know that wide-open breadboards will suck-in ALL the hum/buzz in the room. Not traditionally hiss, but in today's rooms the riot of WiFi and Cellphone may work out as a sort of random hiss.
That guy has an excellent drawing hand, and a drafting pen I have not seen in years. But IMHO his circuit is far from optimum.
here is a link to ti app note for mic pre design an1534
https://www.st.com/resource/en/appl...r-microphone-amplifier-stmicroelectronics.pdf
https://www.st.com/resource/en/appl...r-microphone-amplifier-stmicroelectronics.pdf
Firstly, good on you for actually trying to build something yourself!
The trouble is that the source you picked isn't a very good one. I'll link to a better circuit at the end of this post.
To PRR's comprehensive list of problems with the current circuit, I would like to add that any noise on the +5V line is also fed directly into the input of the op-amp stage via R1.
C2 is intended to mitigate hiss on the supply, but without a series resistor in the B+ line, it can't do much.
Why use 5V, anyway? Is the plan to power this from a USB power supply? Those are switching supplies, and there will be plenty of noise on the +5V line if nothing is done about it. And, in this circuit, nothing is done about it - C2 can't really do it's job without also inserting a suitable series resistor between the +5V from the power supply, and this circuit. Unfortunately, inserting such a resistor would excacerbate the problem of having insufficient power supply voltage.
This is the kind of circuit I would be tempted to run on a pair of 9 volt flat batteries wired in series, to provide clean DC, +/- 9V rails for the op-amp, and better output headroom.
If powering from a small SMPS is the goal, then don't use a 5 volt one. 12 volts is usually no more expensive, and the op-amp will be a lot happier. You may have to increase R1 a lot, to something in the range of 10k.
Little electret mics are very sensitive, and you might not need very much gain before the op-amp output starts to clip if the mic is exposed to normal speech or similar sound levels. This circuit sets up the op-amp for a voltage gain of roughly 20 times, or 26 dB, and it's not adjustable. It might be too much - assuming the op-amp is working at all on the limited 5V voltage.
There are newer op-amp models that are designed to run on low supply voltages like this in contemporary mobile electronics, but the 5532 is from a different era and designed for pro-audio usage in mains-powered electronics devices. It's a superb audio op-amp, but it was never intended to run on meagre supply voltage.
The good news is that you can indeed make a sane mic preamp using most of the same bits and pieces you already have. This circuit makes much more sense than GreatScott's disaster: https://circuits-diy.com/low-noise-microphone-preamplifier-circuit-using-ne5534-ic/
Instead of 2x 68k, you can still use the two 10k resistors you're using now (R3, R4), though they will waste a bit of battery current. Note the addition of a 10uF cap across R4 to filter out any noise on the supply rail.
You can also continue to use your C2 (22uF) across the power supply, rather than the 100uF suggested in the new circuit.
If built properly, this one will work, and will not hiss excessively, unless the electret mic itself is defective.
-Gnobuddy
The trouble is that the source you picked isn't a very good one. I'll link to a better circuit at the end of this post.
To PRR's comprehensive list of problems with the current circuit, I would like to add that any noise on the +5V line is also fed directly into the input of the op-amp stage via R1.
C2 is intended to mitigate hiss on the supply, but without a series resistor in the B+ line, it can't do much.
Why use 5V, anyway? Is the plan to power this from a USB power supply? Those are switching supplies, and there will be plenty of noise on the +5V line if nothing is done about it. And, in this circuit, nothing is done about it - C2 can't really do it's job without also inserting a suitable series resistor between the +5V from the power supply, and this circuit. Unfortunately, inserting such a resistor would excacerbate the problem of having insufficient power supply voltage.
This is the kind of circuit I would be tempted to run on a pair of 9 volt flat batteries wired in series, to provide clean DC, +/- 9V rails for the op-amp, and better output headroom.
If powering from a small SMPS is the goal, then don't use a 5 volt one. 12 volts is usually no more expensive, and the op-amp will be a lot happier. You may have to increase R1 a lot, to something in the range of 10k.
Little electret mics are very sensitive, and you might not need very much gain before the op-amp output starts to clip if the mic is exposed to normal speech or similar sound levels. This circuit sets up the op-amp for a voltage gain of roughly 20 times, or 26 dB, and it's not adjustable. It might be too much - assuming the op-amp is working at all on the limited 5V voltage.
There are newer op-amp models that are designed to run on low supply voltages like this in contemporary mobile electronics, but the 5532 is from a different era and designed for pro-audio usage in mains-powered electronics devices. It's a superb audio op-amp, but it was never intended to run on meagre supply voltage.
The good news is that you can indeed make a sane mic preamp using most of the same bits and pieces you already have. This circuit makes much more sense than GreatScott's disaster: https://circuits-diy.com/low-noise-microphone-preamplifier-circuit-using-ne5534-ic/
Instead of 2x 68k, you can still use the two 10k resistors you're using now (R3, R4), though they will waste a bit of battery current. Note the addition of a 10uF cap across R4 to filter out any noise on the supply rail.
You can also continue to use your C2 (22uF) across the power supply, rather than the 100uF suggested in the new circuit.
If built properly, this one will work, and will not hiss excessively, unless the electret mic itself is defective.
-Gnobuddy
I am grateful for the advice! Thank you.
A lot of good advice here. Despite the failure of my build, I really enjoyed making it. I will attempt all 3 suggestions - the changes PRR suggested (a bad design it seems but I'd like to learn), as well at the ST app note one and the circuits-diy one.
As for 5V, yes I'd like it to be USB powered in the end. For testing I used a simple LM317 supply at 5V to start with and higher voltages later.
Good points were made about the voltages, so I will look into 12V DC-DC converters of the single and dual type.
I will report back with my findings.
A lot of good advice here. Despite the failure of my build, I really enjoyed making it. I will attempt all 3 suggestions - the changes PRR suggested (a bad design it seems but I'd like to learn), as well at the ST app note one and the circuits-diy one.
As for 5V, yes I'd like it to be USB powered in the end. For testing I used a simple LM317 supply at 5V to start with and higher voltages later.
Good points were made about the voltages, so I will look into 12V DC-DC converters of the single and dual type.
I will report back with my findings.
I just want to say that it's not a failure if you learn something from it. It's just a necessary stepping stone to your eventual success.
As a baby, every single one of us had to fall down many, many times before we were able to walk. But as adults, we're often taught that even a single fall is a failure, and must be avoided at all costs. If we had that attitude as babies, we would all be crawling around on all fours until the day we died!
It's no different as an adult. Every single thing I've ever learned to be good at, started with lots of failures. My early attempts at programming a computer produced lots of failures before I had any successes.
I taught myself to play guitar - every time I tried to advance a little bit beyond my then-current skill level, I would have plenty of failures, until one day I mastered that particular skill, and could move on to the next failure at a slightly higher skill level.
I've been trying to teach myself to cook for years - all my early attempts looked and tasted like canned dog-food. After a lot of those failures, I can now cook well enough to get compliments from my wife now and then.
And so it goes, with every single worthwhile new thing that I try to learn. Lots of failures, each one eventually leading to the next little step forward.
Coming back to electronics, do you by any chance have a solderless breadboard ( )? It makes it quick and easy to tinker with different circuits, until you have a circuit you like. If you intend to tinker with electronics for fun, a breadboard is an excellent tool to have. They are inexpensive, too.
In case you're unfamiliar with a breadboard: https://learn.sparkfun.com/tutorials/how-to-use-a-breadboard/all
Digikey is a excellent supplier of electronics components (in Canada and the USA), including breadboards: https://www.digikey.ca/en/products/filter/solderless-breadboards/638
That circuit is more sophisticated than it appears at first glance. If you look at the circuitry around the op-amp closely, you'll see that it is not a text-book inverting amplifier - the input resistor is missing, and the signal from the electret mic goes straight to the (-) input.
In fact, the op-amp is configured as a current-to-voltage converter. It avoids that noisy input resistor (PRR mentioned this problem), and takes advantage of the high output impedance of the signal from the internal MOSFET inside the electret microphone. The op-amps feedback resistor converts the high-impedance signal current from the microphone capsule into a voltage, which appears at the op-amps output.
Fair enough. In that case, the 5532 is the wrong op-amp for this particular job.
These days there is a lot of electronics that is designed to be powered from a single 4.2 volt lithium battery, so there are certainly op-amps out there that will do what you want. Many will only be available in surface-mount packages, but there will be something that either comes as a through-hole part, or can be bought pre-soldered to a "breakout board" that itself accepts through-hole pins.
I recently found an op-amp that works for one of my own projects, which also runs on 5 volts. But I'm using that op-amp as a Schmitt trigger, with its output switching in digital fashion between 0V and +5V. I'm not sure that this particular op-amp is also a low-noise type suitable for your circuit.
If I have time, I'll do a little online searching for an op-amp that might suit your purposes better.
-Gnobuddy
BTW its a common mistake to use series resistance in a low-noise circuit the first time you try to design one. Series resistance creates voltage noise, (called Johnson noise) at a level that generally won't matter for a line-level amp, but for a mic amp or phono amp where the signal is tiny its often a big deal. The noise increases 10dB for each ten-fold increase in resistance.
Try to keep series resistance below a few 100 ohms in any mic circuit. This is why that inverting opamp circuit is very noisy (R2 adds 4µV rms of audio noise directly to the signal before its even amplified). Basically the microphone output should ideally go straight to the first input stage of the amp if possible (a DC blocking cap is fine, these have no Johnson noise). A 100 ohms only contributes 0.18µV rms, a whole 27dB less than 47k (and in practice the opamp's input noise will then be the dominant noise source). A 5532 input has about 0.7µV rms (across the audio band) for comparison.
I've used the AD8656 dual opamp, which is true rail-to-rail and runs from 5V, and is a bit quieter than the 5532. There are no doubt cheaper options but its notable as having good specs all round for a 5V audio opamp (can drive headphones directly for instance as its output is high current).
Your 10k/10k divider to produce the mid-rail (virtual ground) is also rather noisy - add a 10µF capacitor to ground at this node, and then
buffer it with an opamp to produce a stiff virtual ground (needed by an inverting opamp circuit). Thus a dual opamp can be used
for a low noise mic preamp, one side to generate a stiff (low noise) virtual ground, the other to amplify the signal.
Try to keep series resistance below a few 100 ohms in any mic circuit. This is why that inverting opamp circuit is very noisy (R2 adds 4µV rms of audio noise directly to the signal before its even amplified). Basically the microphone output should ideally go straight to the first input stage of the amp if possible (a DC blocking cap is fine, these have no Johnson noise). A 100 ohms only contributes 0.18µV rms, a whole 27dB less than 47k (and in practice the opamp's input noise will then be the dominant noise source). A 5532 input has about 0.7µV rms (across the audio band) for comparison.
I've used the AD8656 dual opamp, which is true rail-to-rail and runs from 5V, and is a bit quieter than the 5532. There are no doubt cheaper options but its notable as having good specs all round for a 5V audio opamp (can drive headphones directly for instance as its output is high current).
Your 10k/10k divider to produce the mid-rail (virtual ground) is also rather noisy - add a 10µF capacitor to ground at this node, and then
buffer it with an opamp to produce a stiff virtual ground (needed by an inverting opamp circuit). Thus a dual opamp can be used
for a low noise mic preamp, one side to generate a stiff (low noise) virtual ground, the other to amplify the signal.
While very good advice when dealing with dynamic mics, this isn't really applicable in this particular case. This isn't a dynamic microphone, but an electret microphone capsule with a built-in MOSFET, configured as a common-source amplifier. The drain of the MOSFET is brought out as the output pin.
This type of capsule will not work without an external drain resistor, which connects the MOSFET drain to the positive supply voltage. Since the MOSFET itself has an extremely high output impedance, the net output impedance of the (capsule + external drain resistor) becomes about the value of the external resistor itself.
In order to supply sufficient drain current for the MOSFET to bias up properly, this resistance typically has a value between 1k and 10k, depending on the DC supply voltage.
So 1k - 10k is the expected output impedance of this particular microphone. It's not down in the weeds at a few ohms or tens of ohms, like a typical dynamic microphone would have been.
In other words, with this sort of electret mic capsule, we are already dealing with the thermal noise from a resistor in the range of 1k - 10k ohms. The external amplifier doesn't have to be as quiet as a 100 ohm resistor - that gains you nothing in this case. All we need from the external circuitry is to not be noisier than a 10k resistor. That means a 5532 is more than quiet enough.
The circuit I linked to above is sufficiently quiet, as it introduces no separate input resistance at all. The op-amp is configured as a current-to-voltage converter, fed directly from the 10k output impedance of the (mic+drain resistor) combination.
It seems like a very nice op-amp, but only available in surface-mount packages, which creates a major headache for DIY. The extremely low input noise is completely wasted in this particular application, as the mic capsule external load resistor already produces far more thermal noise.
Exactly! PRR pointed this out as well.
In addition to noise from the +2.5V rail, the external MOSFET drain resistor for the mic capsule is also connected directly to the positive supply rail, feeding plenty of noise directly into the op-amp's non-inverting input.
Finally, the op-amp is configured to have considerable voltage gain, which is probably too much for this type of mic capsule. These little electret mics are far more sensitive than typical dynamic mics, and output signal level is considerably bigger at typical SPL levels.
Summing up, in this particular circuit, I'm not convinced that the op-amp or the 47k resistor is the source of the white noise. I think the excessive noise is most likely coming from switching noise on the +5V rail, combined with poor circuit design that feeds most of that noise into the op-amp inputs. Add to that the fact that the 5532 is starved for voltage, and may not be operating properly, as it was never designed to run on only _5 volts.
-Gnobuddy
The Jfet is basically a 500uA current source, so why not remove R2 and set R5 to generate 2V at 500uA (4kohm)? That way you are not amplifying the opamp's input noise current, or adding a lot of extra noise with the 47k resistor.
The Jfet will also see a more constant drain voltage so will have less distortion.
EDIT: Oh, it needs 20x gain. So 80k then for R5.
The Jfet will also see a more constant drain voltage so will have less distortion.
EDIT: Oh, it needs 20x gain. So 80k then for R5.
Last edited:
At the very low voltages used in modern capsules, not far from V(threshold), the JFET may be quite far from "constant current, infinite impedance".
But certainly not 150Ω.
Extreme noise-design is often futile because the JFET is (should be) the "input stage" dominating the noise figure. We only have design control of "the second stage". Yes, the JFET is low-gain so won't wipe-out hiss in later stages. Selecting the capsule counts for a lot. With well-chosen capsules the acoustic noise level may be comparable to room ambient (HVAC, trucks outside), even in "quiet" venues.
But certainly not 150Ω.
Extreme noise-design is often futile because the JFET is (should be) the "input stage" dominating the noise figure. We only have design control of "the second stage". Yes, the JFET is low-gain so won't wipe-out hiss in later stages. Selecting the capsule counts for a lot. With well-chosen capsules the acoustic noise level may be comparable to room ambient (HVAC, trucks outside), even in "quiet" venues.
This has also been my experience. More than once I've pressed one of these cheap little electret mic capsules into service for a small, low-SPL acoustic jam, and mic noise has never been an issue. Invariably, room background noise dominates.
That could change if the P.A. is turned up to much higher SPL, but cheap little electret capsules are almost invariably omnidirectional at low frequencies, and acoustic feedback tends to become a problem long before the P.A. is loud enough to hear any hissing from the microphone itself.
In general, I've found electret mic capsules extremely easy to use with just about any sort of post-capsule amplifying device, from a single BC148 transistor, to full-blown semi-pro audio preamps. They are not at all fussy devices.
That's why I suspect Buckaroo's troubles have more to do with supply rail noise than anything else. And since the op-amp is running at voltages below its minimum spec, it's also possible the op-amp itself is misbehaving. Simply switching to a single 9V flat battery for power would quickly confirm (or refute) the hypothesis.
I've even used the tiny Sparkfun electret mic breakout board as a vocal mic in a pinch, and it worked quite well, the major limitation being maximum SPL: https://www.digikey.com/catalog/en/partgroup/sparkfun-electret-microphone-breakout/68600
-Gnobuddy
Although Scott is a very capable designer but he intended this mic amp to be used for entirely different purpose and under different circumstances. Despite him being capable, this circuit, like pointed out by several, has several flaws. Mic audio clips. Noise filtration is not as it ought to be. Selection of op amp is not as it ought to be. For single supply low voltage LME49720 by TI is extraordinary. Your resistor is too small for mic. Electret mic requires 0.5 mA of current and at 5V, it ought to be 10k-Ohm. Also, it's better if you use an FET like 2N3819 as amplifier. It's input impedance is high enough and by operating it in saturation region close to -2 volts, you'll get a very good and very stable amplifier. The rest is however, on you!