A few dumb questions about building a small level meter circuit

[CurtZHP]

Came up with a Frankenstein's monster of a small level meter circuit built around an LM339 quad comparator I found. Schematic attached.




Component choices had almost entirely to do with what I have on hand.
It worked (no one was more surprised than I was!). I originally omitted the op amp section, but there wasn't enough signal to drive the meter, so I added it primarily as a way to boost the signal to a level the comparator section could work with. The trim pot allows for some "calibration." Seemed to work OK, but when I connect it to the audio circuit I want to use it on, it did seem to load things down a bit. I naturally suspected it had to do with those aforementioned component choices, so I consulted my copy of Doug Self's book, and the section on level meters. It seems I need a high impedance input to this circuit to prevent loading issues. Looking over a few of those circuits, I came up with a revision for mine. I added an input coupling cap to deal with any DC silliness as well. Would that also further isolate this from the rest of the world? Am I barking up the right tree? I haven't had a chance to try this yet, but I will once I get home from work.

 

[JohnRoberts]

Good job... LM339 are popular for such use.

There are several areas for possible improvement.

1- Rectification
1a there will be a diode drop from your simple half wave , with multiple possible remedies.
1b full wave rectification can detect both + and - signal swing, active rectification can compensate for diode drop.
2- current draw... driving all the LEDs in parallel increases total current consumption. Alternate approaches to drive the LEDs in series can save current.

This is a very mature technology so you will find tons of examples on the WWW, including many right here.

JR

 

[CurtZHP]

Good job... LM339 are popular for such use.

There are several areas for possible improvement.

1- Rectification
1a there will be a diode drop from your simple half wave , with multiple possible remedies.
1b full wave rectification can detect both + and - signal swing, active rectification can compensate for diode drop.
2- current draw... driving all the LEDs in parallel increases total current consumption. Alternate approaches to drive the LEDs in series can save current.

This is a very mature technology so you will find tons of examples on the WWW, including many right here.

JR

I did notice a few other circuits that did full wave rectification. They usually have a second diode in the feedback loop, right?

My biggest concern was impedance loading and whether I've properly addressed that.

 

[JohnRoberts]

consider a web search for "precision rectification"... good luck

JR

 

[CurtZHP]

consider a web search for "precision rectification"... good luck

JR

So far it's looking like full-wave precision rectification is a bit more complicated than I thought. Certainly more than just throwing in another diode!
I'll have to think that one over....

 

[JohnRoberts]

full wave is not very important, but tweaking out the diode drop matters for low level detection. Of course you can discharge hold cap to -V

There are many ways to skin this cat(s)...

I have avoid doing a TMI dump on you with my TS-1 rectifier....

JR

 

[CurtZHP]

Success! (I think.....)

I rebuilt the circuit as follows....



That took care of the loading issue. I can patch this in and out of my output stage without so much as a burp in the output.
What's more, I got rid of some tiny traces of noise in the positive peaks of the signal by adding that other diode in the feedback loop of the opamp. (That was another issue I was going to address once I solved the loading problem.)

By the way, ignore the part designations. (I've got two D501's and a few other typos...) That was a copy/paste job that I have to clean up.


Thanks so much for your help, John!

 

[JohnRoberts]

I suspect the cathode of the top D501 needs to be connected pin 6 (output) of U502.

R502 is superfluous because the hold cap will discharge back into the U502 VE - input, through R501, coincidentally 100k.

The second D501 looks OK..

JR

 

[CurtZHP]

Interesting. I was thinking the same thing about the top D501, but everywhere else I saw a circuit like this, it was "facing" the same way as the bottom diode.
I'll try it the other way. I'll eliminate R502. That was probably a holdover from where I found that section of the circuit.

 

[JohnRoberts]

Interesting. I was thinking the same thing about the top D501, but everywhere else I saw a circuit like this, it was "facing" the same way as the bottom diode.
I'll try it the other way. I'll eliminate R502. That was probably a holdover from where I found that section of the circuit.

it is facing the same way... the cathode is the end with a band/line

JR

 

[CurtZHP]

it is facing the same way... the cathode is the end with a band/line

JR

Yes, I know. I thought they were supposed to be facing the same way.

I think I just misread your previous comment. I presently have the cathode of the feedback diode connected to the cathode of the output diode. You're suggesting it should be connected to the point where the anode of D501 and Pin 6 of the op amp meet. Correct?

EDIT: Went back and looked things over. My updated schematic that I posted is wrong. What I breadboarded has it as you suggested. No wonder it worked!

 

[L´Andratté]

On a sidenote, are you aware of this excellent little device?
Electric Druid Audiometer

 

[Newmarket]

For the loading issue - just properly buffer the signal. Hi-Z non-inverting buffer or amplifier.

 

[ruffrecords]

For the loading issue - just properly buffer the signal. Hi-Z non-inverting buffer or amplifier.

Or just use the non inverting input of the first op amp.

Cheers

Ian

 

[ruffrecords]

For the loading issue - just properly buffer the signal. Hi-Z non-inverting buffer or amplifier.

This is a brilliant project. Despite searching all over the place for bargraph display designs I never came across this.

Cheers

Ian

 

[FIX]

You need a small amount (499K) and a small compensation cap between pins 2 and 6, as the op-amp may become unstable when the diodes aren't conducting.

 

[CurtZHP]

This is a brilliant project. Despite searching all over the place for bargraph display designs I never came across this.

Cheers

Ian

Just about every other one I saw was either a 10 segment (or more) meter, or just a single LED to show peak level.

 

[Diabolical Artificer]

I've used this circuit, attached to dive a panel meter. You could use the first two stages, IE buffer and precision rectifier. I used a TL074, it works with most op amps. Diodes are 1N4148's.

Andy.

 

[57sputnik]

Here's a VU meter I designed for and was used in the AudioScape XL-305 reverb:

A Simple Two Channel Eight Step Bar Graph Audio Level Meter For Outboard Gear Without Using An LM3914, LM3915 or LM3916 - Pro Audio Design Forum

The rectifier has an averaging "VU" time constant and uses diode pre-distortion to overcome the forward voltage drop.

An LM334 sets reference current to a resistor string to set the reference voltages.

LM324 op amps serve as comparators to provide clean switching without hysteresis.

5532s were used since the BOM was full of them - this also drives the choice of detector topology.

 

[JohnRoberts]

You need a small amount (499K) and a small compensation cap between pins 2 and 6, as the op-amp may become unstable when the diodes aren't conducting.

The op amp does not literally become unstable but will exhibit full open loop gain. A feedback cap will slow down the transition from + to - and vice versa, impacting HF performance.

For today's dose of TMI when rectifying very low level, very HF signals, the op amp's falling open loop gain at HF will be seen in diminished rectifier output frequency response. A typical op amp active rectifier will have trouble delivering flat output up to 20kHz, for signals below -50dBu. I learned this the hard way with TS-1 design

Here's a VU meter I designed for and was used in the AudioScape XL-305 reverb:

A Simple Two Channel Eight Step Bar Graph Audio Level Meter For Outboard Gear Without Using An LM3914, LM3915 or LM3916 - Pro Audio Design Forum

The rectifier has an averaging "VU" time constant and uses diode pre-distortion to overcome the forward voltage drop.

An LM334 sets reference current to a resistor string to set the reference voltages.

LM324 op amps serve as comparators to provide clean switching without hysteresis.

5532s were used since the BOM was full of them - this also drives the choice of detector topology.

IIRC Wayne has several threads full of sundry comparator/meter circuits on his forum.

JR