Clipping/peak Indicator Circuit Help

I have built one channel of the peak indicator below, but am having difficulty getting it to work as expected.

formatting link

The output of the op amp sits at around 5V, and then drops to 1V when the upper threshold pot is adjusted with a signal applied (1KHz sine

4Vpp).

I assumed the LED would light, but it does not; and I don't understand how it would in terms of the circui diagram.

I am probably missing something simple. Can someone please advise.

Mark Newman

Reply to
mnewman
Loading thread data ...

The LEDs are backwards in the schematic.

The long lead is the cathode and should benegative when the lamp is on.

RL

Reply to
legg

I wish that circuit were a new low in published projects, but it isn't--bad as it is, there are lots worse. At least the op amp output moves, even if it can't actually light the LED.

For a start, no matter what the op amp does, the LED is reverse biased, so it will never light. This is because the op amp is powered from the

+12 rail, and it has no way to drag its output above the positive supply.

For another thing, the circuit relies on the leakage of the diodes to bring the voltage on the noninverting input of the op amp back down from the peaks, and there's no hysteresis to keep it from oscillating.

Really a bad circuit. Have a look in Horowitz & Hill for better ones.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

"Phil Hobbs"

** Agreed.

Even with the LEDs reversed, it ill still be almost useless.

The nature of audio signals is that the peaks are brief, too brief to visibly LED and of either polarity.

So, a proper design has a full wave rectifier stage, threshold detector plus a monostable acting as a pulse stretcher.

Ideally, peaks as short a 100uS that exceed the set threshold should make the LED flash for 100mS.

.... Phil

Reply to
Phil Allison

It looks to me as if the schematic has the LEDs wired up backwards.

The TLC272 op amp is being used as a comparator, as there is no feedback from the output to the inverting input. The noninverting input is being fed from a (fixed but adjustable) reference voltage.

The noninverting input is being fed from the signal source, through a diode which *tries* to keep the noninverting inpput from being pulled down below ground voltage (as this is outside the op amp's normal operating limits). I suspect that this protection is not terribly effective, since the noninverting input has an extremely high input impedance (I don't see any protective clamp diodes in the op amp equivalent schematic).

The way it's supposed to work: when the voltage on the circuit input is less than the (pot-adjusted) voltage at the "+" terminal, the op amp will be pulling its output "high", up to 12 volts (or as high as it can drive). Both terminals of the LED will be at about the same voltage and the LED will not light.

When the input value rises above the upper threshold you've set, and the "-" terminal of the op amp is higher in voltage than the "+" terminal, the op amp is going to pull its output down as low towards ground as it can (it's desperately "trying" to lower the inverting terminal voltage to match the noninverting terminal). This will pull one end of the LED down towards ground, through the 820R resistor, and cause the LED to light.

So, I see two problems with the circuit. The LEDs should be reversed (with the cathodes connected to the resistors, and the anodes connected to +12). And, I think you ought to place protective resistors (100k would probably suffice) between each op amp "-" input and ground, to give the 1N4148s something other than the op amp inputs to try to pull current through.

I do not know why you're only measuring +5 at the op amp outputs under "resting" conditions... it seems to me that the outputs should either be very near +12 or very near 0. By operating without protective resistors on the - inputs, you may have pulled the negative inputs down below the "absolute maximum rating" for the input voltage (which is -0.3 volts). This might have damaged the FETs in the input circuit and shorted something... in which case, replacing the op amp would be necessary.

Even with resistors, or a different arrangement of the 1N4148 protective diodes, you're still likely to end up pulling the - input down below -0.3 volts.

Frankly, from what I see, this circuit really needs a redesign.

Reply to
David Platt

tlc272/

As mentioned, it's a sucky circuit.

  • It's using an op-amp as a comparator, which is always a questionable practice.

  • It might work with certain comparators, that tend to hold the + and - inputs at the same voltage. The TLC272 has a FET input.

  • The audio is DC-coupled. That's great if the amplifier you're using is also DC-coupled, and varies around 0V. Put AC-coupled audio into it and your results will be questionable, at best. Put DC-coupled audio that's not centered on 0V, and your results will be questionabler.

  • And yes, the LEDs are backwards.

I suspect that some dufus scanned the circuit in from a magazine, in part because they didn't even bother to photoshop the ghost text from the back of the page out of the picture. Then they ignored whatever part number was actually called out and specified the '272

So:

Turn the LEDs around.

Put resistors between pin 2 and ground and 5 and ground of the '272. I'd start with 10k-ohm.

Give it a whirl. If you have an oscilloscope watch the voltage at the input to see if the mean value drops as current is pumped by D1 and D2. If so, then your amplifier is AC coupled. You can fix this by putting a resistor (1k-ohm?) to ground from each audio input pin.

Even without at O-scope you can measure the input voltage with a volt meter. It'll probably average out the audio and thus will probably show you the DC bias -- if it doesn't stay at 0V then you need the input resistor. If it doesn't _start_ at 0V, then you need the input resistor to ground, but you _also_ need a blocking cap (1uF would work down to about 250Hz or so, more is better).

Or, get yourself a copy of H&H and go from there.

--
Tim Wescott 
Wescott Design Services 
 Click to see the full signature
Reply to
Tim Wescott

I agree with the others, it is not a good design. I also consider it useles s.

For one, if you want to detect clipping in an amp there are too many variab les for such a circuit with preset thresholds to do any dman good. It might tell you when a certain power level is exceeded but other than that it tel ls you nothing.

To detect clipping on an amplifier, you must detect a discrepancy between t he input ad output. First of all you need the right point for one of the in puts. That will be the input to the power amp stage after all volume and to ne controls and ANY other signal processing. Usually it is maybe 1 volt P-P for rated output.

The power amp circuit has a specific gain after this point and you have to null it. Here's a qwikndirty of it :

formatting link

The box on the left is whatever, the preamp and everything like equalizers, dynamic range expanders and whatnot before the power amp. If you are deali ng with an integrated amp or a reciever without pre out/main in jacks you h ave to go inside and find the inputs to the power ampp itself. This is not difficult usually, except in newer junk where they try to force the space s huttle electronics into three square inches.

R1 and R2 simply take the level of the input side down to the point where i t is easy for the OP AMP to handle. On the output side, R3 and R4 match tha t level so that it nulls, that is to apply the same level to the + and - in puts of the OP AMP. Any difference is rectified by the diodes and fed to Q1 , which lights the LED. That will not happen until and unless it clips.

MOST power amp stages are non inverting, but if you happen to get one that inverts then R1 an R3 are chosen to null right there and the OP AMP only ne eds serve as a buffer.

The REASON you need to do this to detect clipping in an amp is because you never know what the power supply voltages are going to be as they are usual ly not regulated and they vary with line voltage, and how the varying speak er impedance reacts to the input signal.

Another way to do it, more hillbillyish, is to ust use the power supply rai ls in the amp as a reference and when the speaker voltage gets within about a volt of that then trigger something to light the LED. This also works we ll enough for some. The circuit on the back of the proverbial napkin in the picture has an added advantage of detecting if the amp is not working well with an uncooperative load, like certain amps with peizo or electrostatic tweeters, or if there is actually a fault in the amp that is not apparent t o the ears, or if the current limiting is kicking in for some reason.

The ciruit that is (somewhat) depicted does not really detect distortion in the conventional sense, it does but it is simply detecting the differecne between the input and output signal. Yes it is distortion but this is not, well, quite, the IHF method and any measurements you can take with this are not admissible in the official spec sheet.

Reply to
jurb6006

Thanks to everyone for their helpful suggestions. I will try them all out and see if I can get the circuit to work.

I will also substitute another type of op amp before totally giving up. The simplicty appeals to me.

This looks like a slightly more sophisticated approach. It's for a single purpose. I don't really need a hi-fi type solution.

formatting link

Mark Newman

Reply to
mnewman

From the style of the schematic, I'd guess that the magazine is Elektuur / Elektor (originally Dutch).

--

Tauno Voipio
Reply to
Tauno Voipio

The diodes on the inputs really bothers me. Depending on leakage is well put, but in my mind, the inputs are floating. I try to limit the number of high impedance nodes in a circuit. If a node is high impedance, there better be a good reason.

Reply to
miso

It may look more sophisticated but its not much better. I especially hate C5 which seems intended to give Q1 as hard a time as possible - I would remove it but the circuit still won't work very well.

Michael Kellett

Reply to
MK

On a sunny day (Thu, 13 Feb 2014 09:38:43 +0200) it happened Tauno Voipio wrote in :

Yes I noticed that too, and the 'lek van elektuur' could have the errata..

Reply to
Jan Panteltje

I'm wondering if it was initially designed for an op-amp with an appreciable bias current?

I would also recommend the OP finds better circuits although they're likely to be more complex.

--
Mike Perkins 
Video Solutions Ltd 
 Click to see the full signature
Reply to
Mike Perkins

You got that one right, especially in a built-up circuit. Of course, all it's doing is blinking a LED. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

--
Version 4 
SHEET 1 1620 1624 
 Click to see the full signature
Reply to
John Fields

(Snip)

Thanks John. This looks rock solid. I'll order in the LT part.

One day I hope to be able to design like this.

Mark Newman

Reply to
mnewman

--
Thanks for the compliment. :-) 

I used the Linear part because it was convenient, but an LM393 and a 
 Click to see the full signature
Reply to
John Fields

--
Looks like the transistor can go away: 

Version 4 
 Click to see the full signature
Reply to
John Fields

Thanks John. I will think of you every time I see that lttle clear plastic thing flash ;-)

Mark

Reply to
mnewman

--
:-) 

OK, here ya go... 
 Click to see the full signature
Reply to
John Fields

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.