Help with LM3916 display driver pls.

[Othello]

In the data sheet for the LM3916 it says the range and/or the resolution are adjustable and in the application notes there are indeed different designs which seem to prove this point.

Unfortunately I do not understand by which mechanism they change the resolution.

I would like to build a peak program meter with a higher resolution in the 0 to 12 dB range but I first have to figure out how to adjust the range and resolution.

Maybe someone here can help.

Uwe

 

[Analog]

Its quite easy. The ic provides a 1.2V reference. So with one resistor, you can set the full scale to 1.2V. Or, with a resistor divider, you can set the full scale to 12V. See the PDF data sheet's formula for VREF. HTH

 

[Othello]

Yes, I saw this formula. And I understand (I think) that with this mechanism you can "float" the 10 LED display between any two voltages.

What interested me was that apparently you can change the resolution. In their examples they show 3, 2 and 1dB steps displayed by the driver, but unfortunately I have no idea how they accomplish this feat.

The way I understand this is that the steps are fixed by the internal resistors in the driver, but if that was true you could not easily modify the step width-or do I get all this wrong??

Uwe

 

[Nigel Goodwin]

Yes, I saw this formula. And I understand (I think) that with this mechanism you can "float" the 10 LED display between any two voltages.

What interested me was that apparently you can change the resolution. In their examples they show 3, 2 and 1dB steps displayed by the driver, but unfortunately I have no idea how they accomplish this feat.

The way I understand this is that the steps are fixed by the internal resistors in the driver, but if that was true you could not easily modify the step width-or do I get all this wrong??

You've already said you can 'float' the display between two voltages, the width of the steps is proportional to the maximum and minimum voltages (ten steps between them).

 

[audioguru]

The number of decibels between the steps is fixed by the built-in resistor ladder, when the bottom is zero. If the bottom of the ladder is higher than zero then the steps are crowded together.

This LM3916 is a VU meter with a different number of decibels between each step. An LM3915 has 3dB between each step. An LM3914 is linear and has 10% between each step.

 

[Othello]

I finally figured out what the step width is on a 3916!

The next thing which I am struggling with is the repeatability and the accuracy of my display.

I guess the repeatability comes about by building very stable amplifiers, properly nulled.

But still the diodes come on and fade out over a span of about 1dB, making it next to impossible to read any closer than that.

Is this it or is there a way to fire the diodes more precisely??

Uwe

 

[audioguru]

The LM3916 is very repeatable and accurate and the level indications are determined by resistor ratios.

Nearly any properly designed amplifier is stable. Their gain is determined by the ratio of two resistors.

Amplifiers aren't nulled. Are you trying to set them to have the same gain?

You can't hear a tiny change in level of only 1dB so why worry about it?

 

[Othello]

I think you are saying if you don't ask this question than you don't have to worry about answering it.

I guess you are right about that.

But I guess the fundamental problems remains, is there a way to switch on an LED in a more uprupt fashion, as opposed to what looks like part of a bell curve, is there a way to fire them like a square wave?

And to my knowledge the op amps I am using need to be nulled for better performance, in fact I just did it-works great, you should try it.

Burning the midnight oil in Maine

Uwe

 

[audioguru]

The LM39xx is a bar-graph display driver IC. In a bar-graph the number of lighted LEDs add. Then you can hardly see the outputs switch another LED.
You are probably using the IC in the "dot" mode where only a single LED is lighted and it fades from one LED to the next.

An opamp can have its very small input offset voltage nulled with an adjustment trimpot, maybe that is what you are doing. It doesn't affect the operation of the opamp, it just makes the output DC volts the same as the input DC volts. If the input is capacitor-coupled then it makes no measurable difference.
Attach your schematic for us to see which opamp you are using and why you need to null the small offset voltage.

 

[Othello]

I am not working off a plan of a finished device, rather I am working thru (and learning) the data sheet of the LM 3914/15/16.

As I mentioned I am attempting to build a peak meter with a fairly large range of 50 to 60 dB. It is in this application, a precision rectifier followed by amplification stages feeding into a number of display circuits, that I came across these fairly simple op amp amplifiers.
There is really nothing too difficult about these stages other than that they should be very stable and be able to deal with very small signals (in the low mV range).

This is where nulling entered the picture and if I do understand this properly then in this application it does make sense and is needed.
Some of my anticipated signals are smaller than the typical offset voltage.

And as you noticed already I am trying to build a fairly accurate device, which is why I was inquiring about the diode threshholds.

There might be a possibility to use the device for dynamic measurements (as a level meter to record on a tape machine) using the LEDs, because you are right, I won't hear 0.5 dB difference anyways. But when using the device for general audio measurements I could switch the display to a regular DC voltmeter.


Thanks for your help and your input

Uwe

 

[audioguru]

As I mentioned I am attempting to build a peak meter with a fairly large range of 50 to 60 dB. It is in this application, a precision rectifier followed by amplification stages feeding into a number of display circuits, that I came across these fairly simple op amp amplifiers.
There is really nothing too difficult about these stages other than that they should be very stable and be able to deal with very small signals (in the low mV range).

I built a sound level indicator with two LM3915's for a 60dB range. 60dB is too high a range. I had to adjust the input offset voltage of the opamp.

And as you noticed already I am trying to build a fairly accurate device, which is why I was inquiring about the diode threshholds.

An active peak rectifier has the rectifier diode inside the negative feedback loop of an opamp that has an open loop gain of 3,000 at 1kHz, so the diode's 0.6V is reduced to 0.6V/3,000 which equals nearly nothing.

I recently made another Sound Level Indicator project that uses a single LM3915 for a 30dB range plus an automatic gain control for an additional 20dB of range. It operates from a rechargable 9V (7.2V actually) battery or an adapter.
https://www.electronics-lab.com/projects/motor_light/009/index.html