Preamplifier with aux and mic inputs

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi everyone,

I'm projecting a simple preamplifier to use in an university project, that is the first module of an AM radio modulator, this preamp stage should have a computer microphone input, a aux input (for MP3, computer, etc.) and should be able to have volume adjustment for each input, and an output global volume adjustment, the output level should be 3-4 V. This circuit should have an output for the next stage (the frequency mixer), and a line output to connect audio devices (for example an audio amplifier, this one is lacking in our circuit).

I've made this circuit with my collegues, we don't know how to use the computer microphone, do you have any ideas?
This circuit could have problems?
:confused:

PS: Pot1 and Pot2 (mic pot), R18 and Pot4 (aux pot), R8 and R13 (main pot).

Best regards,
Daniel
 

Attachments

  • amp2_2.png
    amp2_2.png
    25.2 KB · Views: 304
Last edited:
Hi,

the circuit doesn´t do what You described.
Your decription doesn´t specify a mixing of the inputs, but a choice of a single source by switching between several sources.
If mixing of the inputs is required, rather use inverting OPAmp stages for U2 and U5.
Also, the Mic Input will probabely be single ended not balanced, hence Mic- would have to be connected to gnd.
As drawn, the Input stage is not balanced/symmetrical anyway.
Also AuxL and AuxR would be mixed together, for a single Mono-output.
Is Mono-Mode intended or is Stereo required?
Your simulation specs for the Mic show 70mVrms and 700mV of input voltage for Aux.
Seems a bit high for the MIC input and a bit low for the Aux-Inputs.
The gain for he Mic-Input is set to ~22times -->+27dB.
The Volume Poti should rather follow the Mic-Gain stage than preceeding it.
The choice of the OPamps may be thought over again.
The LT1022 is an rather old and not specifically lownoise type for lowimpedance sources.
Besides the possibilty of reducing the number count of OPAmps, it might also be sensible to use Dual or Quad OPamps to keep size and cost down.
The OPA134/2134/4134 (Single/Dual/Quad by TI) come to mind, or Duals like NE5532 (several sources), LM4562 (TI), LME49880 (TI), or Quads like the LME 49740 (TI) or similar.
It´d be helpful to precisely list the specifications/requirements and to add a Block-diagram.

jauu
Calvin
 
Hi Calvin and thank you for your help,

The mixing of the inputs is intentional, why I should use an inverting stage? The noninverting mixer doesn't work?
The circuit is intended to work in mono mode.
Why the MIC input is not balanced?
I'm using TL084 opamps, because this circuit is for a low cost application.

PS: The MIC circuit doesn't work with a computer microphone, I've tested it and I can only hear strange noises, that are not caused by the microphone.

Best regards,
Daniel
 
Hi everyone,

The circuit of figure 25 of the attached file (from Texas Instruments TL08x datasheet) can work as intended? The circuit seems to behave like a regenerative comparator. V+ and V- are swaped?
:eek:

Best regards,
Daniel
 

Attachments

  • Figure25_TL08x_datasheet.png
    Figure25_TL08x_datasheet.png
    108.1 KB · Views: 249
Hi,

attached is a circuit that could do what You require.
The different subassemblies are marked on the schematic and may be variied depending on Your specificartions (for example if different gain is needed for the Mic-channel).

Both channels Mic and Aux start the same with a Input filter with bandwidth limits of ~1.6Hz to 160kHz.
It prevents DC and HF to enter the circit.
Input impedance of both Inputs is 50kOhms.
Then linear gain stages follow.
For Mic the gain is set to A1= 1+R3/R2 = 20.6x = +26dB
For Aux the gain is set to A2= 1+R9/R10 = 2x " +6dB
Small compensation caps C3 and C6 prevent oscillation.
In both channels Preset-Potis follow, allowing to dial in on the desired individual Volume level.
If more input channels are required simply add similar stages.

Then follows the Mixer in the classic inverting architecture that adds both channels together.
The gain for each channel is -R14/R12 and -R14/R13.
This is followed by the Master Volume Poti.
A inverting Buffer guarantees sufficient drive capability and low output impedance.
A series Resistor is added to enable the OPamp to drive capacitive loads.
Finally a DC-blocking cap with bleeder resistor keeps the output free of any DC.

The OPamps could be Duals like I mentioned a few before.
One of the cheapest and still best would be the NE5532 (a bipolar type).
A JFET-Input equivalent would be the OPA2134.
Don´t use the TL0xx types.
They are simply outdated, grossly inferior and not cheaper than a NE5532.
Not shown in the schematic are the power supplies and associated decoupling caps.

jauu
Calvin
 

Attachments

  • Mic - Aux - Input stages Mixer.gif
    Mic - Aux - Input stages Mixer.gif
    50.1 KB · Views: 232
Last edited:
Hi Calvin, thank you very much for your great help,

I don't understand why the adder should be an inverter adder, why I can't use a non inverter adder?

The 100pF capacitors in parallel with the feedback resistor are really necessary?
This resistors are used to provide unity gain at higher frequencies, to decrease the 0dB point adding more phase/gain margin, but they are really necessary in the circuit with such an high value?
I've seen only values between 1pF-10pF.
Why you don't use a decoupling capacitor at the non inverting input after the gain setting resistor, this capacitor reduces the DC gain to unity, but at the expense of some added distortion, because it has to be electrolytic.
Why do you use a 2.7kohm resistor in the feedback path of the buffer?


PS: I've already made the PCB and I have to show my work at university, in my country TL08xx is very easy to find. Other ampops are very rare and have to be ordered.

Best regards,
Daniel
 
Last edited:
Hi,

one can use both the inverting and the noninverting Mixer.
There are distinct differences though.
With the noninverting Mixer one channel 'sees' the other channels resistor as load. its own resistor and the other channels resistor form a voltage divider.
If both resistors are of same value, the signal voltage is divided by two.
At the mixer-OPamps input the voltage is the sum of both input voltages divided by 2.
With 3 channels it woud be the sum of the three inputs divided by 3.
The loss in signal voltage requires that the mixer OPamp has a gain of 2, or 3, depending on the number of channels.

The inverting Mixer sums up the currents flowing into the inverting input node. The summed up current then flowing through the feedback resistor, thereby creating the output voltage.
Due to the current summation, each input is independent from the other.
One can even set different gain factors for each channel.
Also the voltage swing at both inputs is close to 0V.
The so called common mode voltage remains small, which keeps distortions small too.

The small caps in parallel to the feedback resistors prevent oscillation.
The lower the gain of the stage, the more prone it becomes to oscillation.
The caps also reduce the Bandwidth of the stage, which improves noise.
The caps may be omitted with, but the behaviour of the circuit should be tested for stability.
A cap betwwen the second feedback resistor to gnd reduces the DC gain to 1.
This is a good measure if DC offsets may become a problem for the circuit.
In my schematic the complete circuit is AC coupled in each input and the output.
Offsets created in between remain so low, that the additional caps are not needed.
The 2k7 resistors serve the purpose of reducing input bias current related offsets.
The input bias current of each OPamp input create a voltage drop over the associated impedance.
The two buffers see the Potis impedance at one input and the 2k7 at their other input. The poti's impedance seen from the OPamps input varies between
0Ohm (either end) and 2k5 (2x 5k in parallel). The 2k7 is close to the 2k5 which generates the largest input offset voltage, but its still high enough as minimum load of the OPamps output (quite often 2k is specified as minimum load).

The NE5532 is such a long lived common OPamp, that I doubt You can't easily and cheaply source it.
The TLxxx are by no means good OPamps after modern standards and they are not cheaper.

jauu
calvin
 
Calvin,

I was looking at your schematic since I needed something similar in my '3 channel quad mixer' thread, and I have a few questions:

1. C1 surely should be 1uF for a ~1.6Hz HPF?

2. a normal HPF has the capacitor near the input [1], your schematic has it on the other side. any reason for that?

3. C1 + R1 are the HPF and C2 + R2 are the LPF. what does R3 do in there, and won't it distrupt the LPF?

thanks,
peter


[1] http://sim.okawa-denshi.jp/en/CRhikeisan.htm
 
Last edited:
Very educational

Thanks Calvin for a great reply to a student question. I am a electronics novice / hobbyist and I followed your schematic better than I would have a year ago. It is people like yourself and others around here who are generous with your knowledge I have to thank. Cheers
 
Hi,

for the input filter frequency response the first 100k resistor doesn´t count.
Its C1/R3 resp. C4/R8 who define the Bandwidth limit.
R2 and R7 may be neglected here, as they are only 1% of R3, or R8´s value.
If a precision calculation is required they must be taken into account though.
For the input impedance R1 and R6 do count.
At DC the Input inpedance is limited to R1s resp R6s value.
Without them the Source device would see an open (infinite R).
At AC, hence in the acoustic work area, the impedance is R1||R2+R3, resp. R6||R7+R8, here 50k.
Again R2 and R7 may be neglected without the results getting off.
The splitting into two resistors of 100k against a single 47k or 50k is useful,
as the source device always sees a defined DC-path.
A open is typically not a condition to wish for for the sourcing device.
Second does the resistor bleed away any possibly occuring charges which otherwise could be stored within the caps.
Third may the Caps C1 and C4 be chosen smaller in value.
As designed here they needed to be doubled in Value if R3 and R8 would be chosen to be 50k and R1 and R6 left away.
Smaller Caps are positive with regard to size, cost and performance.

jauu
Calvin
 
Note: When used with a regular electret microphone, the mic input still needs a low-noise bias supply for things to work. When coming from something like +12V, I'd suggest 2nd-order RC filtering for simplicity, with something like 3k3 - 22µ - 3k3 - 22µ (more like 3k9 for +15V), with 2k2 going to mic-in. Calvin's R1 becomes redundant then. The input coupling cap must be able to take up to full +12V supply at the input side.

There are two different common pinouts for electret mics btw. Mics following the Sound Blaster pinout would have a TRS jack for separate audio output and bias. Others may expect bias supply and audio output on the same pin. You can read about this right here. Modern onboard audio typically has stereo mic inputs with bias supply on both channels (with dedicated 2k2 resistors), which will accept both types of mic. Signal output is on the tip in any case.

The mic input (signal and ground) may still need to be run through a common-mode choke for better RF rejection.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.