EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: rfclown on March 27, 2022, 07:20:51 pm
-
I don't remember what made me think of the LM3909 the other day. When I Googled I was pleasantly surprised to find that my 60 year old brain correctly remembered the part number. I built two blinkers using C batteries when I was a kid. Put one in a medicine bottle, which I kept; and the other in half of a Leggs pantyhose container, LED sticking through a hold drilled in the top center, which I gave to a friend. He called it a blipping iglo. They were still blinking after a year passed. Don't remember exactly how long they lasted.
I found a link describing building the discrete circuit using the datasheet schematic:
https://hackaday.io/project/29179-disintegrated-lm3909-15v-led-flasher
I decided to build it on a breadboard, and I ordered a part online (haven't received yet). The junk box parts I used today are pretty much the same as I had way back when except for the Poly Paks LEDs. My LEDs today are way better. I first looked at it in LTspice. I didn't include the zener since I'm only interested in 1 cell operation. In my breadboard I have R2=10k, R4,R5=470 and C1=100uF. The operation is pretty simple: cap charges until the emitter of Q1 is low enough to conduct, which in turn turns on Q2B, which turns on Q3, which grounds the end of the cap that was formerly positive. Now the LED has V1+Vcap across it, and blink... but I can't figure out what Q4 is for. I deleted it from the schematic, and it still works (just changes time). I pulled Q4 from the breadboard: everything ok. Works fine at 1.1V. Works at 1.0V with noticeably dimmer LED (with or without Q4).
Any ideas on Q4? Maybe something to do with using it at a higher voltage where the zener conducts? Attached below are LTspice plots of the Vbe voltages with and without Q4.
-
Seems to limit capacitor voltage to ~1V regardless of supply.
R4,R5 voltage is relatively low, R6 voltage is ~400mV because of Q4 and the Vbe to trigger Q1 is ~600mV. Do the math ;)
OTOH, there appears to be no limit on how low Vc2 can go. More input volts, more volts across the LED and R1.
-
ok, so this limiting is a function of the base emitter diode of Q4. The collector of Q4 doesn't need to be connected for this to occur.
-
Further explanation can be found here: http://www.righto.com/2021/01/reverse-engineering-low-power-led.html (http://www.righto.com/2021/01/reverse-engineering-low-power-led.html) (including a guess on a JFET not shown on the datasheet but found when looking at the physical die). There is a link in the blog to a datasheet that includes a description of Q4's purpose.
-
Marty, start the DeLorean! I wish this IC was still in production, for the noobs and kids starting out in electronics. It was supposed to take over the world, with everyone's flashlight blinking in the dark...
LM3909 operation is explained in AN-0154 (http://www.bitsavers.org/components/national/_appNotes/AN-0154.pdf) Dec. 1975, the IC was formally announced around Sept. 1975.
-
So I rummaged though old stuff in the garage and found my original unit. It wasn't a medicine bottle, it was a vitamin bottle. The two capacitors are in series to make 147uF. The masking tape inside holding the LED to the top and insulating the back side of the board wasn't so good after more than 45 years. Sure enough, the LED is pretty crappy. You have to look right down on it to see it well. The die explanation was very interesting. Interesting App Note as well. Never knew you could make a radio out of one of these. How about a flux capacitor modulator?
-
looking at the price genuine vintage ic's are going for id get it of the workbench and in the bank vault.
-
I wonder why people want them? I bought a LM3909 at Radio Shack when I was a kid, I found it a few years ago and it was defective by then, not sure what happened. I built a few discrete versions out of transistors that worked fine. I can't think of anything that makes the IC particularly compelling.
-
For a first IC, name something as much fun for kids learning electronics. This IC blinks, beeps, boops and more... all off a battery or two. National Semiconductor also stuffed the datasheet and app note with circuit ideas.
I got one when it came out, loads of learning and made a few props with it, the long battery life was a novelty. It was when LED's were just starting to be affordable.
-
All I really did with it was blink an LED. It was the late 80s by the time I got one so I guess it was not as exciting as it was when it was new. Still about the only novel aspect is the ability to run on a single 1.5V battery.
-
Further explanation can be found here: http://www.righto.com/2021/01/reverse-engineering-low-power-led.html (http://www.righto.com/2021/01/reverse-engineering-low-power-led.html) (including a guess on a JFET not shown on the datasheet but found when looking at the physical die). There is a link in the blog to a datasheet that includes a description of Q4's purpose.
Ahem, the alleged N-ch JFET connects Q3 base with Q1 emitter rather than ground and has its alleged "gate" (P-type) pulled up to VCC by a resistor.
Am I the only one who sees that this theory makes absolutely zero sense whatsoever? :-//
edit
I get it now.
This thing is an NPN transistor with two emitters diffused into a common base layer over a common collector layer.
The collector (left connection) is grounded to the substrate nearby.
The base (center connection) is fed from VCC through ~100kΩ or similar.
The transistor operates in reverse, sinking perhaps a few µA through each "emitter" (acting as collectors).
This takes much less space than two ordinary NPNs.
Breakdown voltage of such current sink is only ~7V, not a problem here.
One current is sunk from Q3 base to prevent spurious triggering by leakage currents in Q1 and Q2 at high temperature.
The other is pulled in parallel with 6kΩ/3kΩ resistors, to ensure that even at very low supply voltage there is enough current available at Q1 emitter to activate Q3 through Q2.
-
I can sort of follow what you are saying, coverage for high temperatures where leakage current is a problem. The IC is rated 1.15V to 6V (6.4V max.) and I thought some current-limiting of base drive to Q3 would be necessary to cover that range. The schematics show a 6.5V zener across Q3 C-E.
More LM3909 trivia I found:
Pink Floyd Pulse (1995) CD box has a red blinking LED, inside is a single AA cell and LM3909. A commercial success 20 years late lol. Blink. Blink.
Had to laugh, later they resorted to using 2 AA batteries and COB, RIP LM3909. Same CD as the one I have. There's a 2022 re-issue that blinks too but needs 2 batteries.
James E. Thompson, P.E.: "The LM3909, originally called "The Miser", was designed by fellow MIT class of 1962 grad Robert A. Hirschfeld back in the '60's."
The only other low voltage IC in the 1970's NS had that I can recall is the LM10. Few manufacturers thought of low voltage (<1.5V) IC applications in the 1970's. ZN414 was another.
I made a LM3909 blinky to spoof a car alarm, it did thwart thieves paid for itself lol.
-
I have a copy of Pulse, mine is the newer version with the blob IC and 2 AA batteries. Unfortunately the IC was defective so I replaced it with a discrete circuit.
-
Have yourself a "de-integrated circuit":
http://www.circuitous.ca/LM3909.html (http://www.circuitous.ca/LM3909.html)
I built the upper schematic, and it works.
I miss the 3909, blinking LEDs had lots of uses in the late 80s.
Steve
-
My copy of pulse has no blinking LED,so much for vinyl being better.