Op Amp help needed desperately

Hey guys, I'm new here and I'd consider myself a noob to DIY audio stuff as well since I've only been working with it about a year. Anyway, I'm almost done with an amazing (if I say so myself :p) portable amplifier system however, I would love for it to incorporate some sort of tone control A graphic equalizer would be ideal but I'll settle for treble and base. So I bought a highly regarded (from what I can tell) audio op amp-- the LM833. However, I've correctly constructed a bunch of applications for the circuit with all unsuccessfull results. My knowledge of op amps is very limited and I find all this inverting non-inverting and negative supply rail stuff a little confusing. I would love it if someone could help me out with this, I'm really stuck in a dead end.

Here's how the system works so far: it uses a LM4752 22W amp that runs of a 25v li ion battery pack. I'm also adding optional kareoke inputs and an FM automatic scan radio. The radio and LM4752 stuff is all entirely finished an functioning beautifully, like I said I just want some sort of tone control and a small mic amp maybe. I would like to construct the Tone Control, Graphic Equalizer, and Balanced Mic PreAmp applications from the LM833 datasheet here:http://cache.national.com/ds/LM/LM833.pdf
but I've been having some problems doing this. So here are my questions:

-It says the op amp requires a -15v and +15v input. Can't I simply connect the negative and positive ends of a battery for this? Or do I need 2 power supplies?

-Is the input signal on the op amp supposed to come directly from the audio source? and the output can be fed right to the inputs of the LM4752, right? That's what I'm trying to do anyway...

-Also, I'm not getting any sound when connecting a speaker to ground and the output (with the audio inputs grouned as well). But when I connect a speaker to the output and the positive supply rail I get some highly distorted sound...is this simply because the output is inverted?


Sorry for such a long post about this but I desperately need some help from someone with these op amps. I'm willing to spend whatever it takes on parts to get this working, I just need someone to help guide me through it. Thanks.

-Segasonicfan
 
Hi,

You cannot just simply connect the opamp supplies to the 25V; I mean you can but need to do a bit extra to run the opamp single-supply. You connect the + supply to the +25 and the -supply to ground. Then you bias the opamp inputs to about mid supply. Then you connect up the signal through DC blocking caps.

Assuming you use the circuit on page 12 of the data sheet: Make a +12.5 V point by using two resistors of say 10K between +25 and ground, and connect the +input to the midpoint. Decouple it with an electrolytic of a few uF to ground to get rid of hum and noise.

Then use a series cap of say a few uF on the Vi point to connect up the input signal. The output from the opamp (Vo) to the power amp should likewise be AC coupled with a series cap, but that may already be present. Use non-polar caps for 50V or more for the signal, like polyester film or, if you want to go overboard any of the designer caps. The value depends on the circuit details, but you can first try it (I have the impression you are the experimenting type, great!). If they are too small you will notice a lack of low bass.

Jan Didden
 
Pretty much any OP-amp application requires a split power supply, and that includes power amplifiers as well.

For the tone control circuit you'll be able to get away with a simple voltage divider with capacitor decoupling. For the power output you'll need a large capacitor in series with the loudspeaker.

Look up single supply operation in the data sheets.

Rune
 

PRR

Member
Paid Member
2003-06-12 7:04 pm
Maine USA
www.diyaudio.com
> I'm almost done with an amazing (if I say so myself :p) portable amplifier system... I'm not getting any sound when connecting a speaker to ground and the output (with the audio inputs grouned as well). But when I connect a speaker to the output and the positive supply rail I get some highly distorted sound...

That is amazing.

The LM4752 can be used two ways. Because it is a low-cost part, they show it with a Single Supply and an Output Capacitor. But you can also use it with Dual Supply and no Output Capacitor. (The later is also how most op-amp plans are drawn.) Which way are you doing it?
 

segasonicfan

Member
2005-08-13 11:41 am
PRR--the LM4752 works perfectly...I was referring to the opamps not outputting properly. Don't mock me.

As for the Op Amp power...I forgot to mention that I was giving it 15v from the same battery (drawing it from another cell within the battery). I just don't understand why these opamps aren't working for me...in the LM4752 datasheet it says only the non inverting inputs are available and I think most of the op amp applications I'm doing have an inverting output...but I'm not sure how to make them non-inverted and how to calculate the gain and all.

Do I need to do anything to the source audio before I input it to the op amp? And I've been tested the op amp by connected a speaker to the output pin and ground...is this an improper method? I just don't want to drive it into the LM4752 unless I know it's perfect.

-Segasonicfan
 
I don't think he was trying to mock you, I think he was just trying to ascertain what kind of configuration you're using. You might want to try posting your schematic so people can get a better idea of what you have. Also, just a tip, you better have "thick skin" especially if you want help from some some of the more knowledgeable people on this forum, and there are some good ones (ie..john curl, nelson pass, jan didden). There are also some on this forum that like to eat 'noobies' for lunch the first time you post something even slightly controversial or incorrect. Like I said, just a tip.


As far as your circuit, I would tend to agree with Jan.Bias the input to the middle of the positive rail (V+) and ground, and AC couple it (cap in series with the o/p) to the input of the amplifier stage. Your opamp choice is good (many will disagree with me) the LM833 is a strong performer and inexpensive, you might want to check out the NE5532 (also good, maybe slightly better).
 
segasonicfan said:
PRR--the LM4752 works perfectly...I was referring to the opamps not outputting properly. Don't mock me.

As for the Op Amp power...I forgot to mention that I was giving it 15v from the same battery (drawing it from another cell within the battery). I just don't understand why these opamps aren't working for me...in the LM4752 datasheet it says only the non inverting inputs are available and I think most of the op amp applications I'm doing have an inverting output...but I'm not sure how to make them non-inverted and how to calculate the gain and all.

Do I need to do anything to the source audio before I input it to the op amp? And I've been tested the op amp by connected a speaker to the output pin and ground...is this an improper method? I just don't want to drive it into the LM4752 unless I know it's perfect.

-Segasonicfan


Please be clear, are you trying to hang a speaker on the opamp ('833) output?? That won't work! If it did, you wouldn't need the power amp.;)

Do I read you right: the power amp part works, but the (pre) opamp part not? What exactly ARE you trying to do? Didnt you say you wanted to use the tone control schematic? Please draw a schematic and we all can move forward.

If you have 25V single ended, that will be fine for the opamp. If you limit the opamp to 15V and the power amp to 25V you will not get full output (the opamp will limit it). Please re-read my previous post.

Jan Didden
 

PRR

Member
Paid Member
2003-06-12 7:04 pm
Maine USA
www.diyaudio.com
> configure the LM833

Essentially the same as that power-amp, except the PA chip handles some dull details that you have to DIY with the universal op-amp.

You have +25V and Gnd. (I'm going to call it +24V for math simplicity.) An audio amp has to handle both polarities of audio. However, the audio signal reference does not have to be the DC ground. It could be almost any voltage between the +24V and Gnd supply pins. Half-way is often best: +12V.

We don't have a +12V supply. Neither does the 4752, so it makes its own with that box called "bias". If you probe pin 5, it sits at half the supply voltage. And if you probe the output pins, you will find them near half the supply voltage (because the feedback gain-set networks reference the Bias voltage). The output pins swing from +12V to +22V to +2V and back. Co is charged to 12V at start-up, and after that will cancel 12V, making the voltage at the speaker 0V, +10V and -10V. Swings both ways, even though the chip only does positive voltages.

Do the same with the op-amp. You "could" even use the same Bias supply, tapped at pin 5. In general, I hate to try small fine work while hitched to a hard-working Power stage. I'd probably use two resistors and a cap, as shown in the attachment, to make a Bias supply for (all) the opamp(s). The two resistors split the voltage; the capacitor makes it clean for audio reference.

Opamp power is still +24V and Gnd. But the DC reference for the inputs must be changed to the +12V Bias source. For this inverting tone control, not connected to anything else, if you set the "NonInverting" input to +12VDC, the output will also lay at +12VDC. So will the input to the tone network, so you can not connect it to a source directly. You need an input cap.

My volt-notes in purple are (roughly) the voltages you should observe. Not critical: it "should be" +12.5V for 25V battry, but +/-10% deviation for resistors makes no difference. Measuring some of those points with a typical voltmeter will load-down the bias a tenth-volt or so.

> does the signal have to be inverted again?

No. Audio works with any number, odd or even, of inversions. Switch the speaker leads, sounds the same. (For stereo, flip both speakers or you get a hole in the middle and lame bass.)

Or "I say no", with a weasel. On a few recordings, for some ears, for some speakers, there is an "Absolute Phase": a big bass drum's initial transient is (often) positive air pressure at the audience, and in the right situation an inversion makes a difference. Audio producers don't obsess about it: boxes are usually non-inverting but there is a Phase Invert switch on the input of every mike channel, and every patchcord is potentially wrong-phase.

> Your opamp choice is good (many will disagree with me) the LM833 is a strong performer and inexpensive

Compared to most any chip power amp, it is a jewel. National must have been really lazy to get nearly-no market penetration with the 833. Here, its low noise is a strong point: that power amp has gain=50, quite a bit.

5532 is another go-to, but with these tone-pot values you get a lot of DC error (more than 833, which is hardly DC-error-free). I suppose that may be moot now that we jacked the whole bias up 12V.

The other go-to, TL072, can drive the 4752 well, and has zero input current so zero DC error in a 100K pot. Often used for EQ knobs in a certain class of Pro gear. But 2uV input noise times a gain of 50 in the output is marginal; the BJT chips will be a whisper quieter.
 

Attachments

  • segasonicfan.gif
    segasonicfan.gif
    27 KB · Views: 500
Hey, thank you SO much for all the help. You even put in the schematic for me =) I tried everything you said and tested all the voltage places. I got all the correct volt readouts on everything with the op amp and the LM4752 is functioning perfectly. However, the output (when going through the opamp) is very very soft and you can barely hear it. The bass knob is working perfectly but the treble only sounds right when it's at full or off. In between it gets very fuzzy and sometimes intercepts the radio..heh. The only thing I changed from your schem was I used 3k resistors instead of 2.7k on the bias input. I don't think this makes any diff though,

Maybe my op amp is messed up though from previous testing. I'm going to try another one tomorrow and see what happens!

-Segasonicfan
 
> sometimes intercepts the radio..heh.

The last time I intercepted radio, I seem to recall the problem was an overly-long unshielded floating input to an amplifier running at very high-gain. The funny part about that episode was that I was using a radio tuned to that same station to test the amplifier. I just didn't realize it wasn't actually hooked up, and was trying to figure out why it sounded like crap. Actually plugging the radio into the amplifier made it sound *much* better.

Incidentally, I know you have the amp working properly, but I thought I would point out that you could achieve 12.5-0-12.5 voltage for this system by choosing your ground from halfway through the battery pack. Then the negative end of the battery pack would be your negative rail, and the positive end of the battery would you be your positive rail.

Not that it makes much of a difference, adding the voltage divider and some AC coupling at this stage of the game seems like the way to go.

Wes
 
hrmm I guess I could try dividing the voltage on my battery for the bias. I tried the mod with another (brand new) op amp btw and there was no differnce. Still a really soft sound and only the base control worked. =/ Maybe the LM4752 just wasn't designed to work with op amps at all? Just a thought cause this is still bugging the hell out of me. =/

-Segasonicfan
 
I've checked many times. The pot is ok but it doesn't even need to be the proper resistance. A diff resistence in the pot would only alter the equation that determines the crossover frequency. I think the problem is feeding it 25v, these circuits are only made to run off 15v with an 18v maximum.

-Segasonicfan
 
Back to basics

Sega: I am suggesting you abandon your attempt at single supply operation for now.Try using two 9 volt batteries as a bipolar supply, tie one + terminal of the battery to the negative of the other battery and ground this.Now just connect the + battery terminal to pin 8 of the chip and the - battery terminal to pin 4.

You may have ground the case of your Pot,this might reduce the RFI coupling into your circuit.Please post a picture of your wiring as this would be helpful. Bob
 
wes-ninja, you really confused me. In the schematic you showed 25v connected directly to the op amp and you also said "Opamp power is still +24V and Gnd." That's why I connected it to the full power supply. So are you saying to just do a 12v input to the bias and a 12v input to pin 8? That definitely makes more sense.

As for the 25v li-ion pack, I'm using 6 li-ion cells in series. I'm charging it by using a swithc to split the + and - terminals of the pack and input 12v. After half the pack is charged you just switch it back and it charges the other half. The whole thing charges in 2 hours which I'm fine with. I could build a 25v li-ion charger but I rather keep it as a 12v for my other li-ion packs.

So for the opamp, 12v input on pin 8 instead of 25v and keep everything the same from the previous schematic? Should I use the same + and - as the bias inputs or draw the positive 25v line and ground from the middle to keep it at 12v?
Thanks for your help guys =)

-Segasonicfan
 
Hey, SSF!

I didn't post that schematic, PRR did, and he's definately smarter than me. :)

***** I am not an expert -- so the thinking below could very well be wrong *****

Okay. What the posted schematic does is bias the op-amp at 12+ in your 0-25V system. This should effectively make the 0-point 12V as far as your op-amp is concerned, giving it the same effect as if you had a 0-12-25V system.

But, remember, voltage is merely electrical potential difference. That's another way of saying "it's all relative" -- so as far as the op-amp is concerned, biasing it up at 12V is the same as running a -12-0-12V system. You're lucky that op-amp has a bias pin, many (LM324, TL072, etc) don't and use an internal voltage divider to figure out the where the zero point is.

Have you created a voltage divider to give you your 12V bias? That would be two resistors of equal value from rail-to-rail, with the mid point yielding you your 12V. Lower valued resistors will give you higher current (and put more load on your PS), but you shouldn't need any real current for the bias pin of an amplifier. So I'd use matched resistors of at least 10K ohms as this is a battery-powered circuit.

-- OH, I just re-read PRR comments, you should be able to get the 12V bias from the 4752, eliminating the need for the voltage divider.

So let's see, you should be connecting pin 5 of the 4752 to the bias pin of your op-amp. Next, pin 8 of the op-amp goes to 25V, pin 4 goes to ground. Measure the purple voltages that PRR outlined for you, and see how it goes.

Are you able to put your circuit on a 'scope with a signal generator? Seeing the waveform with a bias reference (dual trace 'scope) would be extremely enlightening, I'm sure.

Are you *also* using the 2.7k-resistor voltage divider circuit shown in the schematic? I would think that you could one or the other, but not both.

Do you measure 12V on the "ground" of your audio input? I think that you should.

Me? I find audio circuits confusing enough to think about without trying to monkey with raising the bias. I would run +/- rails. But that's just me. This should work, too, since it's really the same thing if you connect the black lead of your meter to the bias pin..

Oh, one final thing. Is it possible that you have too much gain on the op-amp and are over-driving the inputs to your amplifier? Have you tried testing the op-amp circuit separately (i.e. on a breadboard) with a pair of amplified PC speakers? Have you measured the output voltage swings of the op-amp with a 'scope? Are you driving the input with something other than a line-in?

Good luck. If I had this circuit in front of me with appropriate diagnostic tools, I could probably fix it -- unfortunately, my theoretical knowledge in this area is quite lacking, so I can only take stabs in the dark without seeing the waveforms and following them through the circuit.

Wes
 
wow, I really hate op amps.... >_<

Well, I don't have a scope unfortunately. You said I DO connect 25v to pin 8? But I shouldn't also connect the voltage divider (12v) to the bias input? That's what I've been doing...and it's easier to draw 12v from the battery. Like I said I've been using 3k resistors instead of 2.7k though.

Well, I gave up on the tone control for this project. I've finished my portable speaker and it's nice, even though it lacks the tone control. I'm not too concerned about it though since my MP3 player has a 5 band graphic equalizer. I'd still like to learn how to do this right for future projects though...

-Segasonicfan