Help with troubleshooting microphone preamp

Thread Starter

LizzyBoi

Joined Aug 17, 2023
7
Hi

Info:
I have recently been trying to make a microphone pre amp as a fun project. I am currently using a JLI-2555BXZ3-GP mic capsule and follewed the Texas Instruments article called "Single-Supply, Electret Microphone Pre-Amplifier Reference Design" for their pre amp design. This is my specific pre amp design where V1 represents the electrect microphone (not certain if this is a fair substitution for simulation) with a ADA4897 op amp: Mic design.pngThe amplified signal is feed into an audio grabber to get the audio out as USB.

To the issue:
Now i have gotten audio out of this setup but all of it noisy and inconsistent. It seems as if the microphone has a life of its own. For the most part the setup is completely silent. But sometimes by touching the gate of the jfet (and occasionally just by itself) a lingering static occurs where the microphone picks up my voice for a couple of second before fading into nothing again.

Question:
What causes this wierd behaviour and what can be done to remedy it.

Thanks for reading my post and i hope it was informative enough to be helpable.
 

Ian0

Joined Aug 7, 2020
9,503
What is the voltage on J1 drain?
Should there be a resistor between J1 gate and ground? Probably a very large resistor in the region of 1GΩ?

Just wondering. . . why did you substitute TI's recommended op-amp (OPA172) with the ADA4897? The OPA172 has 1.6fA/√Hz of current noise and you substituted it with the ADA4897 which has 2.8pA/√Hz, 1750 times higher?
 
Last edited:

BobTPH

Joined Jun 5, 2013
8,664
The mic is not going to produce 100mV output, more like 5mV.

And is the FET and resistor a part of your circuit, or is that supposed to model the FET in the mic capsule? In any case it is not a functional FET amplifier. What does the simulation show at the drain of the FET?
 
Last edited:

Thread Starter

LizzyBoi

Joined Aug 17, 2023
7
What is the voltage on J1 drain?
Should there be a resistor between J1 gate and ground? Probably a very large resistor in the region of 1GΩ?

Just wondering. . . why did you substitute TI's recommended op-amp (OPA172) with the ADA4897? The OPA172 has 1.6fA/√Hz of current noise and you substituted it with the ADA4897 which has 2.8pA/√Hz, 1750 times higher?
When probing the drain on J1 it reads 5 volts.
I don't currently have a 1GΩ resistor would 1MΩ do? What is the purpose of this resistor exactly

About the the op amp. It was my first time buying electronic components myself and tried to order from farnell after some research sugested they sold to private individuals. They didnt have the OPA172 in stock and i tried to find something else. Turns out i couldnt buy from farnell in the end and by this point i was so confused and tired that i just ended up ordering the same things from mouser instead not realising they have the OPA172 stock there :),

Will the noise realistacally be an issue with what im working with here? Or do i need to order a OPA172 or something similar. I think find the ADA4897 in a mouser article about some op amps good for audio amplification and took their word for it not seeing the pA instead of fA.
 

Thread Starter

LizzyBoi

Joined Aug 17, 2023
7
The mic is not going to produce 100mV output, more like 5mV.

And is the FET and resistor a part of your circuit, or is that supposed to model the FET in the mic capsule? In any case it is not a functional FET amplifier. What does the simulation show at the drain of the FET?
The jfet is apart of my circuit since the JLI-2555BXZ3-GP mic capsule doesnt have an internal FET.

This is probing the drain of the FET (with your suggested 5mV mic response)
FET Probe.png
 

Ian0

Joined Aug 7, 2020
9,503
The resistor sets the DC bias on the JFET. The time constant Rg.Cmic gives the low frequency cutoff. Unfortunately the datasheet doesn't give the microphone capacitance. If you guess at around 50pF, you will need a load resistor of >150MΩ to get to 20Hz. 1MΩ will give a low-frequency cut-off of 3kHz.

It's not a question of whether an op-amp is "good for audio" (whatever that may mean). The noise is the root-mean-square sum of the voltage noise and the product of the current noise with the source resistance.
Assuming that the output impedance of the JFET is ∞, then the input impedance of the amplifier is 13kΩ. The noise resulting from the amplifier voltage noise is a very impressive 1nV/√Hz, but the noise resulting from the current noise is 2.8pA/√Hz multiplied by 13kΩ: a rather mediocre 36nV/√Hz giving a total noise of 36.1nV/√Hz
A cheapy old NE5534 would perform better: 3.5nV/√Hz and 13k*400fA/√Hz = 6.2nV/√Hz
The OPA172 manages 7nV/√Hz and 1.5fA/√Hz*13k = 7nV/√Hz (the current noise has no effect) but the OPA172 needs a lot less current than the NE5534.

Get it working first, then worry about what op-amp might be the best to use.

Your other problem with the J113 is that biassing it at Vgs=0 will give a lot of drain current.Screenshot from 2023-10-28 22-40-20.png
maybe 15mA, but you have only 5V across a 13k resistor, which is 400uA.
To use a J113, instead of a specialist electret-microphone JFET will need some bias to reduce the drain current.
 
Last edited:

Thread Starter

LizzyBoi

Joined Aug 17, 2023
7
The resistor sets the DC bias on the JFET. The time constant Rg.Cmic gives the low frequency cutoff. Unfortunately the datasheet doesn't give the microphone capacitance. If you guess at around 50pF, you will need a load resistor of >150MΩ to get to 20Hz. 1MΩ will give a low-frequency cut-off of 3kHz.

It's not a question of whether an op-amp is "good for audio" (whatever that may mean). The noise is the root-mean-square sum of the voltage noise and the product of the current noise with the source resistance.
Assuming that the output impedance of the JFET is ∞, then the input impedance of the amplifier is 13kΩ. The noise resulting from the amplifier voltage noise is a very impressive 1nV/√Hz, but the noise resulting from the current noise is 2.8pA/√Hz multiplied by 13kΩ: a rather mediocre 36nV/√Hz giving a total noise of 36.1nV/√Hz
A cheapy old NE5534 would perform better: 3.5nV/√Hz and 13k*400fA/√Hz = 6.2nV/√Hz
The OPA172 manages 7nV/√Hz and 1.5fA/√Hz*13k = 7nV/√Hz (the current noise has no effect) but the OPA172 needs a lot less current than the NE5534.

Get it working first, then worry about what op-amp might be the best to use.

Your other problem with the J113 is that biassing it at Vgs=0 will give a lot of drain current.View attachment 306130
maybe 15mA, but you have only 5V across a 13k resistor, which is 400uA.
To use a J113, instead of a specialist electret-microphone JFET will need some bias to reduce the drain current.
Thank you so much for the response i will try to read up on some FET biasing.
 

BobTPH

Joined Jun 5, 2013
8,664
You also need a source resistor to bias the FET. The are typically fully on with 0V on the gate. the source resistor raises the voltage on the source allowing the gate to be negative, which is the operating region of most N-channel JFETs.
 

Ian0

Joined Aug 7, 2020
9,503
The TS's original circuit was common source. I don't see any particularly good reason to prefer either common drain or common source.
Common source gives more gain at the first stage at the expense of a higher impedance into the second stage. Commonn drain gives a low impedance to drive the second stage. Generally more gain in the first stages results in lower overall noise, whilst a common drain buffer would likely give lower overall distortion. The TS's original circuit is the circuit most commonly used with a mic insert which has a built-in JFET (although I have seem mic inserts with common drain buffers)
 

Audioguru again

Joined Oct 21, 2019
6,610
The TS's original circuit was common source. I don't see any particularly good reason to prefer either common drain or common source.
Linkwitz invented the modification to a cheap electret mic that has a Jfet impedance converter inside (Linkwitz Electret Mic Mod). The mod changes the wiring so that the original Common Source for the Jfet is changed to Common Drain to reduce its gain and reduce its distortion when used inside a drum or inside a piano. There is a You Tube video showing the modification.

The TS used a HUGE and sensitive electret mic without an internal Jfet and the mic datasheet shows the common drain wiring for an external Jfet.
 

Audioguru again

Joined Oct 21, 2019
6,610
A common electret mic with a Jfet inside should be used with a common audio opamp.
The opamp selected is a dual (use a single opamp) and oscillates easily at very high frequencies.
The circuit shown is missing a very important supply bypass capacitor mounted very close to the opamp.
 

amile123

Joined Dec 29, 2023
1
The initial circuit of the TS employed a common source configuration. There doesn't appear to be a compelling reason to favor either common drain or common source. Common source configuration provides higher gain in the initial stage but at the cost of a higher impedance into the second stage. On the other hand, common drain configuration offers a low impedance to drive the second stage. Typically, more gain in the initial stages leads to lower overall noise, while a common drain buffer is likely to result in lower overall distortion. The original TS circuit is the most commonly used with a microphone insert that incorporates a built-in JFET, although I have encountered microphone inserts featuring common drain buffers as well.
 

Audioguru again

Joined Oct 21, 2019
6,610
Your Jfet does not amplify.
With a supply that is only 5V, you want the drain DC voltage of the Jfet to be half, at about 2.5V. Your drain voltage is only 0.017V because the Jfet conducts much too much current. The J113 conducts 2mA when its gate bias voltage is 0V. But it must conduct only 0.2mA (2.5V/12k= 0.2mA). Select a Jfet that has a much lower current, or reduce The J113 drain resistor to 1.2k ohms.
 

Thread Starter

LizzyBoi

Joined Aug 17, 2023
7
The resistor sets the DC bias on the JFET. The time constant Rg.Cmic gives the low frequency cutoff. Unfortunately the datasheet doesn't give the microphone capacitance. If you guess at around 50pF, you will need a load resistor of >150MΩ to get to 20Hz. 1MΩ will give a low-frequency cut-off of 3kHz.

It's not a question of whether an op-amp is "good for audio" (whatever that may mean). The noise is the root-mean-square sum of the voltage noise and the product of the current noise with the source resistance.
Assuming that the output impedance of the JFET is ∞, then the input impedance of the amplifier is 13kΩ. The noise resulting from the amplifier voltage noise is a very impressive 1nV/√Hz, but the noise resulting from the current noise is 2.8pA/√Hz multiplied by 13kΩ: a rather mediocre 36nV/√Hz giving a total noise of 36.1nV/√Hz
A cheapy old NE5534 would perform better: 3.5nV/√Hz and 13k*400fA/√Hz = 6.2nV/√Hz
The OPA172 manages 7nV/√Hz and 1.5fA/√Hz*13k = 7nV/√Hz (the current noise has no effect) but the OPA172 needs a lot less current than the NE5534.

Get it working first, then worry about what op-amp might be the best to use.

Your other problem with the J113 is that biassing it at Vgs=0 will give a lot of drain current.View attachment 306130
maybe 15mA, but you have only 5V across a 13k resistor, which is 400uA.
To use a J113, instead of a specialist electret-microphone JFET will need some bias to reduce the drain current.
Hello again, been a while. Im sorry i didn't respond with any updates after your great advice. Life and exams got in the way :). I recently got back to this project and i am really stuck. I tried biasing the j113 in every way i could read and i simply didn't get any regular behaviour. Now i have purchased a 2n4338 transistor cause it had way better Idds charecteristics with zero or near zero Vgs (from my understanding) and it did well in ltspice.
1708524364901.png
But still i get no signal. My computer has picked up audio from it once with alot of noise but i have failed to reproduce this "state". I tried biasing the transistor as shown below but still no audio output:
1708524710990.png
I have a strong feeling either something is broken or i am doing something very wrong.
 

Thread Starter

LizzyBoi

Joined Aug 17, 2023
7
This thread is a continuation of this previous thread:
https://forum.allaboutcircuits.com/...ooting-microphone-preamp.196874/#post-1860334

For newcomers:
I am trying to build a preamp for a electrect microphone feed through a ADC for a computer microphone (and failing miserably) :) , see circuit below.

For the kind persons that responded to the previous thread
Hello again, been a while. Im sorry i didn't respond with any updates after your great advice. Life and exams got in the way.

I recently got back to this project and i am really stuck. I tried biasing the j113 in every way i could read and i simply didn't get any regular behaviour. Now i have purchased a 2n4338 transistor because it had way better Idds charecteristics with zero or near zero Vgs (from my understanding) and it did well in ltspice. Assuming i didn't miss some critical characteristic that makes it totally unfit.

1708524364901.png


But still i get no signal. My computer has picked up audio from it once with alot of noise but i have failed to reproduce this "state". I tried biasing the transistor as shown below but still no audio output:


1708524710990.png


I have a strong feeling either something is broken or i am doing something very wrong.

I also have a OPA1652AIDR opamp that shoud have a better SNR but i wanted to get the circuit working first before i resolder the IC
 
Hi

Info:
I have recently been trying to make a microphone pre amp as a fun project. I am currently using a JLI-2555BXZ3-GP mic capsule and follewed the Texas Instruments article called "Single-Supply, Electret Microphone Pre-Amplifier Reference Design" for their pre amp design. This is my specific pre amp design where V1 represents the electrect microphone (not certain if this is a fair substitution for simulation) with a ADA4897 op amp: View attachment 306120The amplified signal is feed into an audio grabber to get the audio out as USB.

To the issue:
Now i have gotten audio out of this setup but all of it noisy and inconsistent. It seems as if the microphone has a life of its own. For the most part the setup is completely silent. But sometimes by touching the gate of the jfet (and occasionally just by itself) a lingering static occurs where the microphone picks up my voice for a couple of second before fading into nothing again.

Question:
What causes this wierd behaviour and what can be done to remedy it.

Thanks for reading my post and i hope it was informative enough to be helpable.
Should be a source follower for the mic circuit instead of a common source fet amp.
Also, op amp should be a non-inverting connected afterwards.

rgrr5trtrt.jpg
 

Attachments

Last edited:
Top