LME49720NA help!!

[bassfrontier]

I am not an engineer and have little experience with preamp circuits. So I was browsing the National Semi. site and came across the "high performance audio" section, took a look inside and learned about the LME series op amps. Although I am relatively new to chip amps, I decided to give it a shot since the data sheet looked so good! I purchased a LME49720NA last week and got to work.

- I am running it with a proven +/- 15V power supply with super low ripple.
- Im keeping the layout as clean as possible (though its still on a breadboard)
- Im using relatively good quality external passives


Its set up as non inverting and my signal path is:

- input signal (from high impedence source) through a 1K resistor into inverting input
- 10K resistor from inverting input to ground (to match source impedence)
- Output straight to power amplifier with negative feedback through 10K resistor (Rf) into non-inverting input.
- 1K resistor (R1) from non-inverting input to ground

which should result in a gain of 11dB

Im experiencing a significant hiss and a buzz problem and even a little RF pickup.

I tried a couple of puffs between both inputs and a .01uF from inverting input to ground as well as a couple of puffs as negative feedback in parallel with the feedback resistor.

these helped, but Im not sure whats going on (since I dont have a scope) so I cant directly address the problem(s).

Any ideas? circuit recommendations?

Im running a (high impedence) guitar and a cd player for signal.

Thank you for your time.

[theAnonymous1]

It must be a layout issue. I've built plenty of non-inverting circuits with this and other LME IC's without any of the issues you're describing.

I normally bypass with 0.1uF X7R + >47uF electrolytic VERY close to each supply pin to ground.

Why are you using 1k from the non-inverting input to ground (or did you mean inverting)? I use 50-100k and normally leave out any series resistance on this input, though it's probably good practice to have it there.

If you would like one of these SMD boards I made let me know.

[AndrewT]

Hi,
10k feedback resistor and 1k0 inverting to signal ground is a gain of 10/1+1 = 11times = +20.8dB

The impedance of your source will not be 10k. It is more likely to be <=2k and with luck <500r.
The 10k is maybe the recommended load impedance for the source.

The NFB has no DC blocking caps. That makes the DC gain the same as the AC gain.

The input has no DC blocking caps.

The result of omitting both DC blocking caps is that the pre-amp output offset ~=11 * source output offset.

You have not mentioned any input filters.
The DC blocking would create a high pass filter on the input.
You also need a low pass filter. Adding a small capacitor to signal ground after that 1k0 series input resistor will attenuate RF at the input.
Try somewhere between 100pF and 470pF for a pre-amp input.
Add another 47pF directly across the input socket from pin to signal ground.


I don't know your Power amp sensitivity but I doubt your CD output needs 20dB of gain to help it along. I suspect all you need is a buffered volume control.

The high impedance guitar output needs a dedicated guitar pre-amp, preferably located inside the guitar and phantom powered via the output cable..

[bassfrontier]

theAnonymous1,

Thank you. Your opinion is greatly appreciated. I take solace in your good experiences with the LME series amps.

I have capped both rails to ground in the way you specified, but not as close as you have on your boards.

My layout, was, cause for concern (alligator clips holding most of the circuit together, long leaded resistors all over the place, possible ground loops, etc...).

The 1K resistor your asking about is there for gain set. It is the (Rg). And it does go from non-inverting input to ground.. causing a gain of 11 in conjuction with the 10K feedback resistor (Rf). Please correct me if Im wrong, but it would result in a gain of 1.2 or less if I use a 50K-100K resistor from non-inverting input to ground in conjuction with the 10K feedback resistor given (gain=1+Rf/Rg)

But I have since made a solid board with relatively better layout.
This change improved a lot of things but I still have a problem or two.

- It produces a constant "buzz" on the output when Im not touching any grounded part (like the power supply chassis) and stops when I touch it.

- The "buzz" increases in amplitude when I extend my fingers towards the op amp, or get within 2" of the thing.

- The longer the input cable is (with an open end; unplugged), the louder the "buzz" gets in amplitude.

Any ideas?

Thank you for helping me out and for letting me know about your boards.

[bassfrontier]

AndrewT,

Thank you very much. Please tell me more about mandatory input filters for this amp. I wasnt aware that it needed any.







theAnonymous, Im sorry.. you were right.. I meant non-inverting= inverting and vise-versa.

[AndrewT]

Hi,
the input has not got a filter on it.
Worse, you have left no route for the input offset current to pass.
Finally by leaving no route for input offset current you have create a very high input impedance.

The unfiltered input is acting like an RF receiver. The buzzes are it's attempt to amplify anything it can grab a hold of.

[bwaslo]

- input signal (from high impedence source) through a 1K resistor into inverting input - 10K resistor from inverting input to ground (to match source impedence) - Output straight to power amplifier with negative feedback through 10K resistor (Rf) into non-inverting input. - 1K resistor (R1) from non-inverting input to ground

Um.. are you mixing the terms "inverting" and "non-inverting" in your connection description?

Because if you are running "negative feedback through 10K resistor (Rf) into non-inverting input" you aren't running negative feedback, you're running positive feedback and it has no chance of working that way. Feedback resistor (and shunt resistor to ground to set feedback level) go to the INVERTING input.

[bassfrontier]

bwaslo,

Your right. Thanks. I meant to say "10K feedback into inverting input.




AndrewT,

I see. Would you recommend some schems for input filtering?


Thanks a lot!

[AndrewT]

Hi,
expletive deleted (self censorship).

First, have you inadvertently swapped +IN & -IN?

Second, Post3 lists typical component values for a low pass filter.

Third, of course I recommend input filtering. It's what I have been saying consistently from my first reply onwards.

RF attenuation on both the input socket and at the PCB.
DC blocking on both the input and on the NFB route.
You can omit the DC blocking ONLY if you put in place other systems to ensure your pre-amp and subsequent stages can tolerate/survive with DC contamination.

[bassfrontier]

AndrewT,

- Yes, I had accidentally mixed the two input names in post #1.