EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: ivan747 on April 12, 2012, 09:47:19 pm
-
How can I increase the gain of this transistor amplifier?
It is a dynamic microphone preamplifier that goes into an electret microphone input.
(http://www.hobby-hour.com/electronics/dynamic_microphone_amplifier1.png)
The circuit only works when I knock on the microphone. Even worse, this thing is not providing me any gain, it sounds more quiet, so something must be doing something wrong with the circuit. I made it on perfoard and I may have burned the transistor, but I am not sure. I tested it and it behaves just like two diodes when it is unpolarized, so it should be ok. I am using a 2N2222, which is a suitable replacement for the transistor shown in the schematic. I didn't have a 220k resistor so I paralleled two resistors to get around 214k.
This is all the info I have:
http://www.hobby-hour.com/electronics/computer_microphone.php (http://www.hobby-hour.com/electronics/computer_microphone.php)
I want to fix it and provide some gain (maybe 20dB from what it currently is).
Thanks for your help,
Ivan
^ An analog n00b ^
-
I'm sure someone more knowledgeable will chime-in but you might check the datasheet for your mic element and make sure it will work without a power source like that. A lot of electret microphone elements need a power source.
-
I am using a dynamic microphone, the ones with a moving coil.
-
Most of the time you shouldn't need a preamp for a dynamic microphone. If you're connecting to a PC mic input, just connect a blocking capacitor in series with the mic and it should work.
As far as the preamp goes, have you checked the voltage levels to make sure that you have power going to the circuit ?
-
I have just checked them. The tip seems to measure something between 1.1V and 1.7V. The ring is floating. Apparently somebody screwed this sound card up. It is a cheap chinese USB sound card. The line out(s) are inverted too (i.e. the left channel is assigned to the right speaker and vice versa). I will try the circuit on a different computer.
Interestingly enough, it seems like 13" MacBooks use the audio input as a line in rather than a microphone input. They use a combo port that can be used both as headphone out and line in. I think they want you to use the built in microphone or their headphones with microphone built in or buy an audio interface thing.
-
Now I am in doubt. When I plugged the microphone in directly to the USB sound card I got sound, but if my theory is right (two pins are swapped) the input is on the
sleeve ring, which is grounded because I am using a mono plug, so why did I get sound?
Teardown time?
-
If your circuit diagram above is correct, it looks like the sleeve of the plug would be GND, the middle ring would be V+ (power out to amplifier) and the tip would be the amplified signal input from the mic to the sound card. If you used a mono plug instead of the stereo plug pictured then the transistor would be unpowered and the output from the microphone element would be fed unamplified to the tip.
But I'm not familiar with microphone circuits at all, so I don't know how things are intended to be connected.
-
OK I have improvised an adapter to correct the problem with the swapped pins on my sound card. It is now louder but it is amazingly noisy, it has more noise than signal. It also sounds bassy and distorted. Also there is a DC offset that goes up and down (takes up to several hundred ms) when I record anything, probably due to the pin swap.
I plugged a different microphone to the sound card and the level is acceptable. I'm sticking to this solution for now, but I plan to do stereo at some point (the sound card has two inputs) and I need two identical microphones for some situations.
I have to pairs of microphones and a spare one. Only the spare one is giving acceptable levels without amplification.
-
The 2N2222 is not a great replacement for the BC548B. The 2N2222 is designed as a switching transistor with good gains over 100mA. The BC548B is a higher gain lower current device designed more for audio use.
The 2N2222 will still work though. Since the dynamic mic is low impedance, the current gain difference is not much of a probllem, so I would keep using the 2N2222 for now. For best performance and lowest distortion, adjust the 220K resistor so you get about 2.5VDC on the collector. Double check that there is 5V from the soundcard.
Now it is possible you are getting too much gain. The bigger the output amplitude, the more distortion you will get. Put an AC voltmeter on the output and put a continuous noise like a whistle into the mic. You probably do not want any more then 0.1V AC on the output.
If you have too much gain, adding a small resistor between the emitter and ground will reduce gain and significantly reduce distortion. Perhaps 100 ohms might be a good starting point.
Richard.
-
The 2N2222 is not a great replacement for the BC548B. The 2N2222 is designed as a switching transistor with good gains over 100mA. The BC548B is a higher gain lower current device designed more for audio use.
The 2N2222 will still work though. Since the dynamic mic is low impedance, the current gain difference is not much of a probllem, so I would keep using the 2N2222 for now. For best performance and lowest distortion, adjust the 220K resistor so you get about 2.5VDC on the collector. Double check that there is 5V from the soundcard.
Now it is possible you are getting too much gain. The bigger the output amplitude, the more distortion you will get. Put an AC voltmeter on the output and put a continuous noise like a whistle into the mic. You probably do not want any more then 0.1V AC on the output.
If you have too much gain, adding a small resistor between the emitter and ground will reduce gain and significantly reduce distortion. Perhaps 100 ohms might be a good starting point.
Richard.
I would like to do that, but, ignoring the wrong pinout problem, the sound card only outputs 1.5V approximately, for some reason. Tomorrow I am going back to the breadboard to try you suggestions.
I now corrected the problem with the sound card. It used to have the tip and the ring signals swapped, but I made a little box that swaps them back to normal.
I will do my best to make this circuit work on a breadboard just for educational purposes. I may need to supply my own power though, so it isn't very practical anymore.
Since my sound card is not collaborating, I think I will go for another design later on that outputs a line level signal, which is more versatile. Something like this:
(http://orca.st.usm.edu/~jmneal/schematic.gif)
http://orca.st.usm.edu/~jmneal/preamp1.htm (http://orca.st.usm.edu/~jmneal/preamp1.htm)
-
I would like to do that, but, ignoring the wrong pinout problem, the sound card only outputs 1.5V approximately, for some reason. Tomorrow I am going back to the breadboard to try you suggestions.
If the supply is only 1.5V, adjust the 220K resistor for 1VDc on the collector. At 1.5V supply though, distortion will probably be noticeable. With a 100mV output, I am sure it will be over 1% distortion. If your soundcard works fine with lower signal levels, then the lower the output from the transistor amplifier, the better the distortion will be.
Richard.
-
I would like to do that, but, ignoring the wrong pinout problem, the sound card only outputs 1.5V approximately, for some reason. Tomorrow I am going back to the breadboard to try you suggestions.
If the supply is only 1.5V, adjust the 220K resistor for 1VDc on the collector. At 1.5V supply though, distortion will probably be noticeable. With a 100mV output, I am sure it will be over 1% distortion. If your soundcard works fine with lower signal levels, then the lower the output from the transistor amplifier, the better the distortion will be.
Richard.
With the actual circuit attached the voltage drops down to 730mV, useless because I can't even bias the transistor with that. The voltage on the base is 523mV. The voltage on the collector is 567mV. Useless, I assume.
-
With the actual circuit attached the voltage drops down to 730mV, useless because I can't even bias the transistor with that. The voltage on the base is 523mV. The voltage on the collector is 567mV. Useless, I assume.
Just means you need to change values. Increase the 5.6K to 27K. That should get you back to about 1.3V. Then try and do the bias again to 1V on the collector. You may find that the 220K has to be increased to over 1meg.
Richard
-
A much better choice for the transistor is a 2N5088, low noise and high gain (beta of about 500 vs 150 for 2N2222). Unit cost $ 0.08 at Mouser.
-
For the second try I will be using a 2N4124. It has higher gain under low currents (120-300), compared to the 2N2222 (~50). Its gain is lower compared to the BC548 (220-400).
-
For the second try I will be using a 2N4124. It has higher gain under low currents (120-300), compared to the 2N2222 (~50). Its gain is lower compared to the BC548 (220-400).
Probably will not make much difference. If the 2N2222 will not work well enough at 1.5V, then another transistor will not do much better. Since the supply looks like it is 1.5V from a 10K resistor, the circuit will not work well with any transistor without the changes such as the ones I suggested. Transistor current gain is not a big issue in this amp, and the noise of the 2N2222 is respectable.
Richard.
-
Followed the changes you suggested and I'M GETTING SOMETHING! ;D
Changed the 5.6k with a trimmer set to ~70k, changed the 220k resistor with a 5.1M and stuck with the 2N4124. Now the problem seems to be that a high-pass filter is interfering with the circuit. It has a very low -3dB point, I havent measured it but it must be in the 100ms - 500ms range.
Voltages are: 1050mV on the 22uF cap, 750mV on the collector 480mV on the base. I guess this low base voltage causes distortion when the input signal drops to a negative voltage in relation to the base voltage.
The first waveform is me knocking the mic with the 22uF cap in place. The second one is knocking the mic without the cap and some speech.
-
1V from the soundcard at a current of only 4 uA from the tip via maybe 100K instead of 5V via 2K2 from the center ring. That soundcard is really bad.
This is about the best you can do with a single transistor amp:
It only has a gain of about just under 10. A higher gain transistor might improve it, but the difference will only be just audible.
Trouble is at such low voltages, distortion will increase with the amplitude form the mic, from about 1% at 1mV in to about 10% at 10mV in. With 5V, you could actually get excellent performance, but 1V is really scraping along the bottom.
Richard.
-
Note: A few, recently manufactured PCs have implemented true stereo microphone inputs. High performance speech recognition and advanced noise canceling applications -- see the Andrea Superbeam Array stereo microphone -- make good use of this new feature, providing more accurate and reliable signals in noisy environments.
When the stereo mic input mode is selected, the bias voltage will be provided for both the tip and the ring. The wiring for a stereo microphone is simple -- see the schematic diagram on the left -- connect the shield of both microphones to the sleeve of the plug, the left mic to the tip and the right mic to the ring. For best performance, use unidirectional electret microphones.
(http://www.hobby-hour.com/electronics/stereo_microphone_schematic.png)
Seems like my soundc ard falls in this category (I found a blurry reference design) but someone had the brilliant idea of separating the single stereo connector into two stereo connectors with the ring left floating. The sound was possibly coupling through the circuit and finding its way into the biased input.
Here's the datasheet for the chip. The sound card seems to follow the reference design at a first glance:
http://www.steila.com/USB4IO/CMedia/CM6206+datasheet+1.7.pdf (http://www.steila.com/USB4IO/CMedia/CM6206+datasheet+1.7.pdf)
There must be a way to get a DC signal from the sound card and at the same time inject sound back into the card without feedback or cancellation, but I don't think it is worth the effort when you have such a low voltage to work with and such high distortion figures. Unless the solution is as simple as an RC filter of some sort, I'm afraid i'll give up with this circuit.
Thanks to all, specially Richard who helped me a lot.
-
There must be a way to get a DC signal from the sound card and at the same time inject sound back into the card without feedback or cancellation
There is--the electret microphones in the schematic you showed have a built-in FET amplifier that does exactly this. If your microphone does not have such a built-in amplifier you can implement the same kind of amplifier externally using discrete components. This works because the power supply comes through a series resistor (causing the voltage to vary when the load current is changed by the transistor) and the sound input has a DC blocking capacitor to prevent the power supply voltage upsetting the input signal.
-
I will give amspire's design a try. It seems to use the same wire for both power and signal, or maybe I haven't understood the circuit.
-
The problem is the standard electret bias is 5V via a 2k2 resistor. This is about 1.5V via about 100K resistor. I don't think it is enough voltage or current for any electret FET amplifier. If you can find a way to get 5V out, then with a few modifications, you can get less then 1% distortion and more gain easily.
-
Looking at this, it seems a good solution is to use an old Germanium transistor, as it wll be happy with the low voltage supply. Either that or find a low enough voltage JFET to use in the circuit. I have use PNP Germanium transistors to replace carbon microphones, using an electret unit, where the available voltage was under 1V at around 10mA. A simple circuit, with only a single base emitter resistor to set bias current for the transistor to about 5mA. IIRC the resistor was around 1k5, and the whole unit fitted inside the carbon insert. worked well for many years, with better performance than the carbon it replaced ( that was worse than phone quality audio, the new one did much better than that).
-
If it's a electret the easiest and non fussiest way would be
There must be a way to get a DC signal from the sound card and at the same time inject sound back into the card without feedback or cancellation
There is--the electret microphones in the schematic you showed have a built-in FET amplifier that does exactly this. If your microphone does not have such a built-in amplifier you can implement the same kind of amplifier externally using discrete components. This works because the power supply comes through a series resistor (causing the voltage to vary when the load current is changed by the transistor) and the sound input has a DC blocking capacitor to prevent the power supply voltage upsetting the input signal.
-
The point is the normal FET supply is not there. There is no 5mA, or 1mA even. To just get 1V supply, the transistor had to run at about 3 uA. That is the problem.
Richard.
-
The best you can do then is to take a 5v supply to power the amplifier from a USB connection. Take a USB lead and cut it to length, then strip it and use the red lead to power the amplifier with 5V. Decouple it with a 100R resistor and a 100uF 16V capacitor, and use the shield connection of the USB lead as a supply common ( less noise from the very noisy USB supply then).
-
What if I do something like Sean's suggestion and use a 9V battery with a 5V regulator? I can add a 2.2k resistor in series and use the exact original circuit.
-
Since my sound card is not collaborating, I think I will go for another design later on that outputs a line level signal, which is more versatile. Something like this:
(http://orca.st.usm.edu/~jmneal/schematic.gif)
http://orca.st.usm.edu/~jmneal/preamp1.htm (http://orca.st.usm.edu/~jmneal/preamp1.htm)
The attached schematic is really "bizzarre" in the section that generates the bias voltage for the microphone, with that positive AND negative feedback loops around U1a,.
I really cannot understand the reason for all this complicated and unpredictable circuitry, that may be replaced with one or two resistors.
Another rather objectionable point is the choice of the op-amp: TL074 are noisy, and don't operate well at low voltages.
A single transistor can do better..
Best regards
-
The attached schematic is really "bizzarre" in the section that generates the bias voltage for the microphone, with that positive AND negative feedback loops around U1a,.
I really cannot understand the reason for all this complicated and unpredictable circuitry, that may be replaced with one or two resistors.
It seems like the designer was trying to get a regulated bias voltage, which is not necessary; maybe they just couldn't let 1 of the op-amps be idle, and wanted to use it up, because it's there in the package.
Also, D2, the LM4040 reference, is drawn backwards. The positive feedback, R1B, is not necessary (in this circuit).
I dug up his earlier design, it was even more bizzarre... R1A to VCC didn't exist. The reference diode had no connection
to VCC and had its cathode connected to the + op-amp input (which was fed-back from it's output via R1B). The thing
probably only worked because the LM4040 needs just micro-amps to get started and was likely able to siphon that
current at startup through the decoupling cap (C9). After it started, it would be self-sustaining, but the startup
conditions (of the earlier design) are clearly unpredictable. That's probably why this next guy added R1A to VCC,
but then he should have also removed R1B.
-
If you are going for a 9V battery then a regulator is unnecessary, just power it direct. Will do no harm in this case, and will make the battery last longer as it will not have to supply the regulator current.