Bought many LM3914 (VU-meter style 10 LEDs driver), now what to build?!?

[RoGeorge]

Got a bag of LM3914, NOS from the '90s, made by National Semiconductor. 

LM3914 - Dot/Bar Display Driver

The LM3914 is a monolithic integrated circuit that senses
analog voltage levels and drives 10 LEDs, providing a linear
analog display. A single pin changes the display from a mov-
ing dot to a bar graph. Current drive to the LEDs is regulated
and programmable, eliminating the need for resistors. This
feature is one that allows operation of the whole system from
less than 3V.

• Drives LEDs, LCDs or vacuum fluorescents
• Bar or dot display mode externally selectable by user
• Expandable to displays of 100 steps
• Internal voltage reference from 1.2V to 12V
• Operates with single supply of less than 3V
• Inputs operate down to ground
• Output current programmable from 2 mA to 30 mA
• No multiplex switching or interaction between outputs
• Input withstands ± 35V without damage or false outputs
• LED driver outputs are current regulated, open-collectors
• Outputs can interface with TTL or CMOS logic
• The internal 10-step divider is floating and can be referenced to a wide range of voltages

(from LM3914 datasheet, https://www.alldatasheet.com/datasheet-pdf/pdf/8898/NSC/LM3914.html )



It was an impulse buy, don't ask why.  I have no regrets.  ^-^
Looking for interesting applications, any ideas are welcome.

[ledtester]

There's always this idea...The PDF contains the fill-page ad for the device.

[RoGeorge]

I'm thinking as a first test to make a Li-ion single-cell discharges (to measure the mAh of phone batteries):
    - the idea is to use the constant current in the LM3914 LEDs to make a 20-200mA adjustable constant current load
    - the LEDs will draw their current from the DUT Li-ion battery
    - the mAh can be deduced by counting the total time during which the DUT battery was able to keep the LEDs on

    - a trim-pot could set the number of LEDs always on (so to set the discharging current according to the battery capacity)
    - LM3914 can be powered with as low as 3V, so it might work to self-power the entire circuit from the DUT it discharges 
    - the on-chip Vref should be good enough to keep the current constant while discharging from 4.2V to 3V
    - should use red LEDs, for their lower V_f drop, to let some margin for the IC open collector BJTs while the DUT approaches 3V

[themadhippy]

I used one years ago to make a power up/down sequencer,simple ramp generator to the input and relays on the outputs.

[Doctorandus_P]

LM3914 is the linear version. For the "VU" meter, you would want the LM3915, which has a logarithmic scale for the resistor string

Oopsie, I forgot about the LM1916, so read on to the next comment.

[themadhippy]

For the "VU" meter, you would want the LM3915,

the lm3915 is equal 3db steps ,for proper vu metreage its the lm3916 you wants

[oPossum]

.

[UnijunctionTransistor]

If you don't find enough projects to consume all of the ICs, you can always sell them.
NOS National Semiconductor LM3914s command attractive prices.

But are you sure that they are legit?

[RoGeorge]

That should be fun, thanks. 
The date/name stamp suggests you just drawn that, or is it a copy/paste?  I'm interested what editor was used to draw the schematic.

[RoGeorge]

But are you sure that they are legit?

Bought them from a brick and mortar shop, the seller is also the shop owner.  I think those are genuine NOS parts.  Bought them to build funny things, not for reselling.  I don't even have an eBay account.

[oPossum]

That should be fun, thanks. 
The date/name stamp suggests you just drawn that, or is it a copy/paste?  I'm interested what editor was used to draw the schematic.

Yes, it is my design. Drawn in Eagle 7.

I got some LM3914 clones that I had to test. There are at least three brands of LM3914/15/16 clones. Two are available from LCSC.

https://www.lcsc.com/product-detail/C53199171.html
https://www.lcsc.com/product-detail/C55109573.html

Here is a video showing that circuit and a 555 + 393 quadrature circuit.

[RoGeorge]

Clever those 4 LEDs! 

[oPossum]

Clever those 4 LEDs! 

Original design by Rue Mohr

https://bsky.app/profile/ruenahcmohr.bsky.social/post/3mvdggliwuk2l
https://bsky.app/profile/ruenahcmohr.bsky.social/post/3mvcka7szp222
https://bsky.app/profile/ruenahcmohr.bsky.social/post/3mvcikmr3zc22

I don't like using op-amps as comparators, so I 'fixed' it.

[RoGeorge]

Wired two LM3914 for a test, to see if they work, and to try the Li-ion tester battery tester idea.

First LM3914 has its LEDs constantly ON no matter the voltage.  A trimpot is used to change how many LEDs to keep always ON.  The more LEDs are ON, the more current will draw from the Li-ion cell that is being tested.  Each LED that is ON means 25mA more drawn from the DUT (plus some 10mA quiescent current).  Like this:



The entire thing is supposed to be powered by the battery that it is testing, so the Vf on the LEDs has to be small enough to work down to 3V or so, which means red LEDs.

Curious thing, normal red LEDs from decades ago have a Vf of 1.6V or so, while recent high-efficiency red LEDs have a Vf of at least 2V. 

So I had to dig for my stash of LEDs from when I was a kid, because of their lower Vf. 



Could only find about 12 of them, and with heavily oxidized terminals, not enough for both LM3914.  For the second LM3914 on that perfboard I've used some white LEDs scrapped long ago from a wall appliance lights.  Those white LEDs (2700K) SMDs were for 65mA nominal and their Vf is higher than 3V, but they are so bright (they were meant for lighting) that they can be used at only 3-5mA and still be bright enough as indicator LEDs.  At very low current, their Vf is below 2V.

Both the red and the white LEDs were a pain to solder, the first ones because they were oxidized, the former because they were sited on a flexible ribon that was melting together with the SMD LED on it.  Eventually got all the LEDs planted, but I didn't wire yet the second LM3914, which is supposed to tell the DUT voltage, in dot mode, from 3.4 to 4.4V in steps of 0.1V, then to shut down everything if possible, will see.  So far it looks like this:



Kind of works, but meh, the idea of self-powering the circuit from the battery under test is at limit.  The discharging current is not so constant when the battery discharges under 3.5V.  Maybe the LM3914 should be powered from another voltage source (another battery), not from the DUT it discharges.  Will see about that, too.

Code: [Select]

- I vs Vbatt with 10 LEDs ON
- 4.5V -> 255mA
- 4.4V -> 255mA
- 4.3V -> 255mA
- 4.2V -> 254mA
- 4.1V -> 253mA
- 4.0V -> 252mA
- 3.9V -> 251mA
- 3.8V -> 250mA
- 3.7V -> 248mA
- 3.6V -> 245mA
- 3.5V -> 239mA
- 3.4V -> 228mA
- 3.3V -> 216mA
- 3.2V -> 203mA
- 3.1V -> 189mA
- 3.0V -> 176mA

- I vs nr of LEDs ON at 3.7V
-  0 -> 11
-  1 -> 36
-  2 -> 60
-  3 -> 84
-  4 -> 108
-  5 -> 132
-  6 -> 157
-  7 -> 182
-  8 -> 205
-  9 -> 227
- 10 -> 248mA
Anyway, that is only because I've stretched the design goals too far.
Good news is, the chips are genuine and working perfectly fine! 

[Doctorandus_P]

Recently I saw an AN that injected a small AC voltage into the input signal. As a result one of the LED's receives a PWM signal and thus varies in brightness. Nice trick to enhance the resolution.

[oPossum]

Dithering app note attached

[Perkele]

Tetris playing machine?
Hard mode - use 74xx circuits.

[Le_Bassiste]

Real Time Analyzer for audio (RTA), something in the line of a AUDIOCONTROL SA-3052.
1984 April / May editions of "elektor electronics" magazine had an interesting take on an RTA.

edit: "ELEKTOR" magazine featured that project in March/April 1984.

[Benta]

Drive it with a triangular signal and make a K.I.T.T. scanner.   

[LaserSteve]

If you can find the app notes, they scale and chain very well. There is always the 60 LEDS around the clock thing, It fascinated me during visits to VOA Studios and WWV as a kid. Dad was a government radio nerd in more ways than one.

Steve