Getting Started on Preamp Design

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

First post on this forum, I've been a huge fan of all the info that's available on here. Basically I want to get started on building a pre-amp circuit. I want to integrate it into an ADC/DAC design (specifically, something similar to the XMOS reference designs, if anyone's familiar with their products). I'm looking to design a clean, low-noise amplification for instrument level inputs to line level outputs. I'm new to the audio design community, and while I've done a good amount of research on different pre-amp circuits, I was wondering if any of you had some suggestions on documentation that helped you build pre-amp designs. I'm not sure what type of op-amp to use, if I need multiple op-amps (if I need none?), if it should have an inverting or non-inverting input, etc, so any help is much appreciated! Thanks in advance,

Pat
 
- I need two inputs for instrument level signals.
- Isn't this standard across instrument level signals? I looked it up quickly, and it says 0.1V to 1.0V for passive signals, up to 1.75V for active, and I'd like the whole range, so I guess 0.1V to 1.75V?
- I don't need any features other than gain control.
- No levels for power amps, just the line level output.

I'll check out the NE5532 for now, thanks for the reply!

Pat
 
.
<< I need two inputs for instrument level signals - Isn't this standard across instrument level signals? I looked it up quickly, and it says 0.1V to 1.0V for passive signals, up to 1.75V for active, and I'd like the whole range, so I guess 0.1V to 1.75V? >>

You generally stick with the non-inverting input of an op amp because that's the high impedance side.

Broadly speaking, the input of an op amp can be within 2 volts of its power supply voltage. That is, if you power an NE5532 with +/- 15 volts, then any input up to 13 volts is OK. Data sheets give exact numbers, of course.

There is no lower limit of input if you use a buffer. Within reason, of course. Nuclear submarines have special requirements.

Back in the day "line" meant one volt. But today there are at least 3 "line" standards (if they are standards) that I know of. Still you'll be OK if you consider line to be the classic one volt.

"Instrument" can mean a lot of different things around here...or anywhere. Electret mics put out maybe 2 millivolts, piezo transducers might put out several volts, as might magnetic pickups on guitars (instruments). Might be best to specify.

I'm posting a schematic that's something of a do-all. Not that you should solder it up right away (although you could, and it would work), but as a starting point, or an illustration of what you seem to be saying.
.
 

Attachments

  • NE5532-buffer-preamp_01_b.jpg
    NE5532-buffer-preamp_01_b.jpg
    74 KB · Views: 221
Last edited:
bentsnake, thanks so much for your response, your provided schematic especially helps visualize everything you said. Firstly, I'd like to note (since I haven't flat out said it) that I'm using this preamp for an audio recording interface for musical instruments, specifically direct-in guitars and possibly expanding to microphones. I used the vague term "instrument level" because thats what the -10db level in consumer recording electronics is referred as, which specifically is the range of most passive guitar pickups. So to sum it all up, this project is for signals coming from a guitar pickup.

Before I go on, I've done quite a bit more research, and while the NE5532 seems to be a great, low cost op-amp that's been used in so many audio applications, I've begun looking into the OPA series by TI, specifically the OPA134. It seems like this op-amp and the others in the series are considered some of the highest quality for audio recording, and while the difference between the two may be small, I'd definitely prefer to build my project off of the best quality op amp off the bat, rather than do it later. Any opinions on this decision would be appreciated. But for now, I'll go off of the information you've provided so far!

.

Broadly speaking, the input of an op amp can be within 2 volts of its power supply voltage. That is, if you power an NE5532 with +/- 15 volts, then any input up to 13 volts is OK. Data sheets give exact numbers, of course.

.

So this is the first time I've noticed that it's the input of the op amp that is restricted to the supply voltage. I assumed it was the output that was restricted to the supply voltage. Does this mean that since my guitar levels will never exceed 2 volts, I could run my op amp off of a 5V supply?

.

There is no lower limit of input if you use a buffer. Within reason, of course. Nuclear submarines have special requirements.

.

Forgive my basic questions, but I'm still struggling to understand the applications for a buffer vs the gain stage. What does each do, and why cant they be done in one stage?

.

Back in the day "line" meant one volt. But today there are at least 3 "line" standards (if they are standards) that I know of. Still you'll be OK if you consider line to be the classic one volt.

.

Yes, just the standard 0db (1V) is what I'm looking to output.

Thanks again for the help!!

Pat
 
Another two questions upon further inspection of your schematic:

There isn't much resistance on the input of the circuit... Guitar signals are high impedance, so it was my understanding that the resistances should be high on the input. Most schematics have at least 100k on the input, with the norm being 1M ohm of impedance.

Second, can I have any value pot to control the volume? I have a bunch of 10k pots lying around, I feel as if they could get the job done.

Thanks again!

Pat
 
Administrator
Joined 2007
Paid Member
A fet opamp like the OPA series is better suited to a high input impedance. In the circuit above that means R3 (and R2 because its in parallel at AC) can be 1meg (or higher if needed). 10K pots are fine too.

The output of any opamp is limited to the rail voltage (less a small amount), the inputs need to be limited because if they exceed the rails they will begin to force current into the various semiconductor junctions within the IC, possibly destructively if there is enough current behind it.
 
.
<< ...struggling to understand the applications for a buffer vs the gain stage. What does each do, and why cant they be done in one stage? >>

I think I need to define "feedback" as it's used in electronic circuits. It means part or all of the output is returned (fed back) to the input. In the circuit I posted, U1a has feedback from pin 1 to pin 2. U1b has feedback from pin 7 to pin 6.

Now speaking of buffers, think "the nature of the beast," the beast being the little black chips we work with. A buffer occurs one time and one time only: when the output is shorted back to the inverting input, as is done with U1a in the schematic I posted . The term for this is "100% feedback."

It's characteristic of op amps that with 100% feedback they demonstrate unity gain, extremely high input impedance, and extremely low output impedance. This is a buffer, and it's exactly what you want as an audio input. Especially that "extremely high input impedance," which is in the 100s of megs.

Contrast the buffer U1a, with the gain stage U1b. The gain stage has two resistors (R4 and R5) in the feedback circuit, the buffer has none. This resistors-or-not is the difference. No resistance, which is to say 100% feedback, creates a buffer.

After that, maybe the best thing to do is memorize the fact that "Buffers have unity gain, extremely high input impedance, and extremely low output impedance." Because the math of it is a wee bit complex.

All of which relates to another question you asked.

<< There isn't much resistance on the input of the circuit... Guitar signals are high impedance, so it was my understanding that the resistances should be high on the input. Most schematics have at least 100k on the input, with the norm being 1M ohm of impedance. >>

You're right. But remember that the input is to our good friend the buffer. Again that "extremely high input impedance" thing. Also remember that the schematic as posted is intended just to illustrate a general approach, it's not a construction project.

As posted, the circuit has an input impedance of around 23k, which is the parallel value of R2 and R3. They're in parallel because capacitor C2 is transparent (or close enough) to the audio signal, all it sees is the two resistors.

For emphasis: The audio signal doesn't see pin 3 of U1a because of its ridiculously high impedance. As far as the audio signal is concerned pin 3 is an open circuit, nothing there. The audio signal sees only R2 and R3--or, again, close enough.

The upshot is that when you make R2 and/or R3 one meg, or ten megs, or whatever megs, the audio circuit still won't see pin 3. I'm trying to "illustrate with an explanation" that the input resistor(s), R2 and/or R3, can be whatever you need. Changing their value will not affect the rest of the circuit.

And that's as far as I'm going with that for the moment. Let me look at your other questions.
.
 
.
<< ...while the NE5532 seems to be a great, low cost op-amp that's been used in so many audio applications, I've begun looking into the OPA series by TI, specifically the OPA134. It seems like this op-amp and the others in the series are considered some of the highest quality for audio recording >>

As I think you know, the NE5532 was top of the line for something like 30 years. There are tens of thousand (millions?) still in use in recording studios and broadcasting stations all over the world. Most people have never heard a recording or broadcast that wasn't passed though a dozen--or several dozen--NE5532s.

But yes, recently better chips have come along. Much better, in fact. We're talking numbers that have gone from .001 to .00001.

Now the obvious punch line: That is, they're better on paper, but I can't, myself, hear the paper difference in the real world.

But wait! I have no argument with those who prefer the newer chips. Only that I don't, myself, see the point. But a happy note is that most of the new chips plug right in to replace an NE5532, no circuit changes needed...at least as far as I know, you'd want to check this.

There is an additional consideration. If you're not pretty expert with a soldering iron I very strongly suggest that you start out with 50 cent NE5532s. and move up as you think best. Just as a precautionary budget move.
.
 
.
<< So this is the first time I've noticed that it's the input of the op amp that is restricted to the supply voltage. I assumed it was the output that was restricted to the supply voltage. Does this mean that since my guitar levels will never exceed 2 volts, I could run my op amp off of a 5V supply? >>

The only restriction on output is that at some point it starts to distort. But output voltage is not something to worry about, the question is output current. Or more correctly, output watts, as in will my 50 watt (output) amp drive my 200 watt speakers? (nope) The formula is watts = volts x amps, but again, output volts is not much worried about.

Getting to your question:

<< since my guitar levels will never exceed 2 volts, I could run my op amp off of a 5V supply? >>

Absolutely not. Especially if you're thinking of the 5 volts in a computer power supply. Computers have very noisy power supplies in audio terms, they won't work at all. I'm sure somebody will argue this point, but that's my story and I'm sticking to it.

And of course...run what op amp? The circuits we're talking about draw only milliamps of current, but you still have to connect the voltage that the op amp needs to run.

But the real concern is what's called "headroom," which is what you use compared to what you could use. The concept of headroom applies all though electronics, and you never, ever run on the edge. Always at least a 50% margin, because it's not a margin at all, you'll use it.

Didn't I post a power supply schematic? Doing that. As before it's supposed to illustrate the idea, it's not a construction project (although you could build it as is and it would work). Please don't be afraid of power supplies. As I hope my schematic illustrates, they're not as complicated as some make out. I'd say not so much complicated as trivial.

OK I think I'm caught up on your questions. If not please ask again.
.
 

Attachments

  • power_supply_dual_small_C.jpg
    power_supply_dual_small_C.jpg
    62.5 KB · Views: 108
.
Oh wait, not caught up.

<< ...can I have any value pot to control the volume? I have a bunch of 10k pots lying around, I feel as if they could get the job done. >>

Yes, 10k would be fine. In fact, any value down to 2k. No lower than 2k, because that pot is the output load for op amp U1a. Too little resistance = too much current flow = burned chip.
.
 
Note that they would have to be 10k log pots.

A fet opamp like the OPA series is better suited to a high input impedance.
Basically yes, but you may be in for a nasty surprise in terms of distortion when confronting them with an actual high source impedance.

I think the OP needs to do some research on what guitar amps and (pro) microphone inputs usually look like. While it is basically possibly to combine an unbalancd high-level input and guitar input (though that would be a "clean" input only), a balanced mic input with P48V phantom feed is a rather different animal again, and I would not combine it with the aforementioned if I can help it. Well, I guess you could integrate a switchable passive DI transformer if you really wanted to.
 
.
sgrossklass you're right, but I think in these forums the KISS rule usually applies.

It might be worth adding that the purpose of balanced lines is to minimize the noise picked up by connecting cables, speaking only of the cables themselves. Balanced lines have nothing to do with the audio directly.
.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.