LED signal meters. LM3914/5/6 still viable?

[TheJames]

I'm kind of wrestling with this.  Looks like the 3916 is "Obsolete", Mouser doesn't stock the 3915, is the 3914 going to be around long?

I know there are quantities of these devices out there, but I'm wondering if they make sense for new designs.  They're very simple and easy to use devices but I guess there's not much demand any more.

Any other simple (LED) meter options out there?  Maybe newer IC solutions?

 

[Andy Peters]

Any other simple (LED) meter options out there?  Maybe newer IC solutions?

It's very common to use a cheap microcontroller with built-in ADC to do the metering function, as well as whatever else that micro needs to do.

 

[Humner]

for 5 LED's you could try AN6884

It appears to be available from some of the Chinese suppliers I use. Its cheap as well.

Personally I haven't used it myself, but I'm thinking of trying it as a simple cheap solution.

 

[TheJames]

I didn't consider a MCU since I didn't think I could get into one for less than a normal analog solution.  I could get away with a single dual-colour LED, but I'd prefer something like a 10-segment one.  There's absolutely no need (in my application) for anything more than that.

A LM391x solution would be about $5 in parts.  (10-segment LED bargraph + LM391x+ a cap and a few resistors/trim pot).

LM324/339 quad comparators are less than 40 cents a piece.  For 10 segments you'd need three plus a few resistors of various values.  Not impossible to do....But more effort than a LM391x.

I did see that ebay is littered with Arduino Atmega328 boards for effectively $4 a pop, like this ad:

http://www.ebay.com/itm/3X-MINI-USB-Nano-V3-0-ATmega328P-CH340G-5V-16M-Micro-controller-board-Arduino-/222065258539?hash=item33b41ef02b:g:Mw8AAOSwGYVW9fYO

Unreal.  I didn't consider a MCU because I didn't think you could get into them that cheap.  You can't buy raw Atmega328's for that $$.

One advantage with the micro is that I could multiplex the digital outs, and use the analog in's for multiple meters.  I want four meters (L/R in and out) so this could potentially be more cost effective.  Another advantage is that meter scale and ballistics are totally programmable.

An Arduino is a little top-heavy perhaps, but since I have nearly zero MCU experience, it's not a bad solution.

Things that make you go Hmmmm....

Thanks Andy!

 

[JohnRoberts]

I've posted about this before. Several years ago I designed a microprocessor based master section meter for an analog console (still in production APB Prodesk series).

We did a sharp pencil analysis and the microprocessor, driving latches to keep multiplex noise down, ended up cheaper than the previous all analog approach, while adding a bunch of extra features (I implemented my simultaneous Peak/VU display, but allowed the end user to select Peak only, VU only, or both.)

I played around with the idea of making a generic microprocessor meter kit that is pre-programmed and the user can use micro pins to realize different semi-custom configurations (Linear or log, step size, etc).

I never pursued this since it looks like a modest sized market, and I already have more ideas than time (some even less attractive than this). 8)

JR

 

[Andy Peters]

I didn't consider a MCU since I didn't think I could get into one for less than a normal analog solution.  I could get away with a single dual-colour LED, but I'd prefer something like a 10-segment one.  There's absolutely no need (in my application) for anything more than that.

A LM391x solution would be about $5 in parts.  (10-segment LED bargraph + LM391x+ a cap and a few resistors/trim pot).

LM324/339 quad comparators are less than 40 cents a piece.  For 10 segments you'd need three plus a few resistors of various values.  Not impossible to do....But more effort than a LM391x.

The Silicon Labs C8051F850 costs a buck in single-piece quantities and has an ADC with a 16-channel multiplexor in front. With a few ten-cent transistors you can easily multiplex the LEDs for a meter and drive them from a handful of port I/O pins.

 

[gar381]

I'm kind of wrestling with this.  Looks like the 3916 is "Obsolete", Mouser doesn't stock the 3915, is the 3914 going to be around long?

I know there are quantities of these devices out there, but I'm wondering if they make sense for new designs.  They're very simple and easy to use devices but I guess there's not much demand any more.

Any other simple (LED) meter options out there?  Maybe newer IC solutions?

More info

http://groupdiy.com/index.php?topic=61733.0

GARY

 

[JohnRoberts]

I didn't consider a MCU since I didn't think I could get into one for less than a normal analog solution.  I could get away with a single dual-colour LED, but I'd prefer something like a 10-segment one.  There's absolutely no need (in my application) for anything more than that.

A LM391x solution would be about $5 in parts.  (10-segment LED bargraph + LM391x+ a cap and a few resistors/trim pot).

LM324/339 quad comparators are less than 40 cents a piece.  For 10 segments you'd need three plus a few resistors of various values.  Not impossible to do....But more effort than a LM391x.

The Silicon Labs C8051F850 costs a buck in single-piece quantities and has an ADC with a 16-channel multiplexor in front. With a few ten-cent transistors you can easily multiplex the LEDs for a meter and drive them from a handful of port I/O pins.

I multiplexed LEDs in both generations of my drum tuner, but for the master section of an analog console I decided the one latch line per LED would be a better mimic of the analog circuitry it replaced.  LEDs only changed state when audio levels changed.

If the meter is off in it's own world where noise floor is not a concern multiplex away. I do a 12x3 multiplex in my current drum tuner to light  36+ LEDs.  A latch for 12 pull downs, with 3 high side (mosfet) switches to alternate between the 3 banks of 12. 

FWIW I found that the cheap low voltage (3.3V or less)  micros don't drive LEDs directly very well, while the older 5V stuff could easily. 

JR

 

[Rochey]

I''m back on development of the Expat Audio VU Meter, as I discovered a bug that I thought was a projectkiller was a software issue and not a HW one!

Those with interest can catch up with Expat Audio on facebook. I may even try my hand a livestreaming the debugging process on facebook, but don't expect miracles.

/R

 

[bruno2000]

I''m back on development of the Expat Audio VU Meter, as I discovered a bug that I thought was a projectkiller was a software issue and not a HW one!

Those with interest can catch up with Expat Audio on facebook. I may even try my hand a livestreaming the debugging process on facebook, but don't expect miracles.

/R

Looking forward to it!
Best,
Bruno2000

 

[analogguru]

I'm kind of wrestling with this.  Looks like the 3916 is "Obsolete", Mouser doesn't stock the 3915, is the 3914 going to be around long?

So you know evilbay..... why don´t you do a search there ?:

LM3915
http://www.ebay.com/sch/i.html?_from=R40&_trksid=m570.l1313&_nkw=LM3915&_sacat=0

LM3916
http://www.ebay.com/sch/i.html?_from=R40&_trksid=m570.l1313&_nkw=LM3916&_sacat=0

Any other simple (LED) meter options out there?  Maybe newer IC solutions?

The LM391x-series was outdated for VU-Meter use even in the mid 80´s.  Reason: excessive current consumption in bargraph-mode.

Skilled engineers used BA682A instead.  Benefit:  12 segments (instead of 10) and only 25% current consumption with same brightness.
datasheet:
http://media.digikey.com/pdf/Data Sheets/Rohm PDFs/BA682A.pdf

digikey:
http://www.digikey.com/product-detail/en/rohm-semiconductor/BA682A/BA682A-ND/4003430

evilbay:
http://www.ebay.com/sch/i.html?_from=R40&_trksid=m570.l1313&_nkw=BA682A&_sacat=0

 

[bruno2000]

But you CAN run the LM39XX in constant current mode, drawing only about 15mA in bar mode.......
Best,
Bruno2000

 

[fum]

I use a Microchip 12F675 for the metering in my Si mic pre ( 2008? wow, that went quick). They are about a $1, and give me 10 segment with over, peak hold, peak +over, and knight rider metering modes

I'm using a time sliced/multiplexed current sharing like John states above.  Line per light is easier, if you have the real estate.

On a more recent meter project that has hit production,  I stepped up to a larger pic, 21 segment.

Of course, the PIC stuff requires some programming skills, as opposed to ninja or nunchuk .

Yay blinky lights!

$.02

 

[Rochey]

I did a live broadcast on youtube last night discussing the EA design to do this.
The project isn't finished yet, but it's more about sharing knowledge and allowing folks to see what they will (hopefully) be buying.

https://www.youtube.com/watch?v=KRmITtitYnA

/R

 

[Rob Flinn]

I did a live broadcast on youtube last night discussing the EA design to do this.
The project isn't finished yet, but it's more about sharing knowledge and allowing folks to see what they will (hopefully) be buying.

https://www.youtube.com/watch?v=KRmITtitYnA

/R

Interesting.  Thanks.

 

[dbonin]

http://howtomechatronics.com/projects/infinity-mirror-music-vu-meter-electronics-project-using-lm3915-ic

Thought  this may help someone.
Cheers!