SIMPLE, cheap latching circuit - one transistor?

xanatosxanatos Posts: 1,119
:smile:Feeling somewhat dumb at the moment. I need a VERY simple latching circuit that will simply be on until a signal hits it (can be either + or -, I have no preference) and the circuit will remain off until I reset it.

I have been playing with SCRs and stuff but apparently my current draw isn't sufficient to get them to stay in saturation.

So I went to a S/RS Flip Flop arrangement using an ANCIENT 7400 chip and that works well... as soon as a signal hits the input of one of the gates it flips and stays put until I put the signal to the Reset input and it goes back, no matter how many times I wiggle the input. (Note - the 7400 chip I used I have had since I was a kid - literally got that in the 1970s, and it's still working. Love those old 7400s!)

I remain convinced however that there should be a way to do this with a single transistor that feeds back onto itself, but I can't quite get it to work.

So I'm either looking for that magical single-transistor circuit, or a more up-to-date and perhaps more efficient circuit than the 7400s wired as S/RS Flip Flops. I need 5 of these circuits, that would be 3-7400 chips. Is there anything like a "hex S/RS Flip Flop" on a single chip? :-)

Thanks for the ideas, should any come forth. If worse comes to worse, I can live with the 7400 arrangement, but my gut is telling me there has to be something better.

Dave
Like Xanatronics on FB: https://www.facebook.com/xanatronics

There are only 10 kinds of people in this world - those that know binary and those that don't.
«13

Comments

  • 75 Comments sorted by Date Added Votes
  • kwinnkwinn Posts: 7,775
    edited August 2012 Vote Up0Vote Down
    I don't think this can be done with a single transistor but a dual transistor would work. You could use the '279 quad sr latch. That gets you down to 2 chips and available as ls and hc for sure. I have even seen single chip single sr latches somewhere, but can't recall where. Here is an scr with a 5mA holding current. http://www.datasheetcatalog.org/datasheet/unisonic/MCR100-4.pdf
    In science there is no authority. There is only experiment.
    Life is unpredictable. Eat dessert first.
  • Mark_TMark_T Posts: 1,653
    edited August 2012 Vote Up0Vote Down
    I believe it can be done with a single transistor - just has to be a unijunction transistor operating in its negative resistance region?
  • kwinnkwinn Posts: 7,775
    edited August 2012 Vote Up0Vote Down
    Ah, never thought of a UJT. That might work.
    In science there is no authority. There is only experiment.
    Life is unpredictable. Eat dessert first.
  • ercoerco Posts: 18,942
    edited August 2012 Vote Up0Vote Down
    I've used SCRs with success several times. They are available in a wide range of sizes, current capacities, and holding currents (down to 1 mA). What's your minimum current? Your particular SCR might not be sensitive enough. From http://www.allaboutcircuits.com/vol_3/chpt_7/5.html :

    If the SCR fails to latch, the problem may be with the load and not the SCR. A certain minimum amount of load current is required to hold the SCR latched in the "on" state. This minimum current level is called the holding current. A load with too great a resistance value may not draw enough current to keep an SCR latched when gate current ceases, thus giving the false impression of a bad (unlatchable) SCR in the test circuit. Holding current values for different SCRs should be available from the manufacturers. Typical holding current values range from 1 milliamp to 50 milliamps or more for larger units.
    "When you make a thing, a thing that is new, it is so complicated making it that it is bound to be ugly. But those that make it after you, they don’t have to worry about making it. And they can make it pretty, and so everybody can like it when others make it after you."

    - Pablo Picasso
  • xanatosxanatos Posts: 1,119
    edited August 2012 Vote Up0Vote Down
    Cool info guys, thanks. The UJT thought really brought the idea back - it was indeed a UJT I had seen used in this capacity before. If I have the time before I have to get this rolled out, I'll order up some SCRs and 279s to try out. Thanks again for the ideas & assistance - as always, the most helpful crew on the web! :-)

    Dave
    Like Xanatronics on FB: https://www.facebook.com/xanatronics

    There are only 10 kinds of people in this world - those that know binary and those that don't.
  • ercoerco Posts: 18,942
    edited August 2012 Vote Up0Vote Down
    When in doubt, you can't go wrong with a relay. :)

    You can use a sweet little latching relay like http://www.goldmine-elec-products.com/prodinfo.asp?number=G2011 which only requires a brief pulse to change states. Or you can wire up any DPDT relay into a latching configuration, latching the coil connections through one channel using the NO contact. You reset by removing power briefly (NC switch typicallly) to open the relay contacts.
    "When you make a thing, a thing that is new, it is so complicated making it that it is bound to be ugly. But those that make it after you, they don’t have to worry about making it. And they can make it pretty, and so everybody can like it when others make it after you."

    - Pablo Picasso
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 21,772
    edited August 2012 Vote Up0Vote Down
    xanatos wrote:
    Is there anything like a "hex S/RS Flip Flop" on a single chip?
    Maybe not S/R latches, but possibly a transparent latch. Before that question can be answered, you need to specify whether the latches can be ganged with one latch-enable input or whether each has to have its own latch-enable.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 21,772
    edited August 2012 Vote Up0Vote Down
    BTW, you can get four S/R latches in one package: an HEF4044, which is 4000B-series CMOS.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited August 2012 Vote Up0Vote Down
    Erco seems have taken a sincere liking to those little latching relays. The problem with the IC latches is that they must always be powered to retain state. I have no idea about the UJT needing to do so, but the simplest BJT two transistor latch also requires power to retain state.

    At first I thought 5v latching relays were a bit of a joke, I liked the bigger more powerful ones for home automation. But these tiny ones do offer a way to retain state and conserve power in battery operated devices.

    With a 1/2 amp at 125VAC, these are a great way for a microcontroller backup battery to turn on and off a wallwart when a cyclical task needs power. When power one the battery can recharge.
    Hwang Xian Shen, Puddleby-on-the-Marsh.
    All things considered, I can live and thrive without Microsoft products. LINUX is just fine.
  • Tracy AllenTracy Allen Posts: 6,197
    edited August 2012 Vote Up0Vote Down
    There is also the simple buffer latch...

    buffer-latch.png


    With a CMOS non-inverting buffer, the output feeds back to the input and it holds its last state. But forcing the input, it changes state. It can be any non-inverting buffer, but the NC7SZ126 is a single unit in an SOT package. So, almost one transistor footprint.
    400 x 107 - 5K
  • Tracy AllenTracy Allen Posts: 6,197
    edited August 2012 Vote Up0Vote Down
    Here is the UJT circuit and load line from the old Sam's handbook, abc's of unijunction transistors. It sits at point A when off and at point B when on. The valley current Iv is on the order of 10mA, but is not well controlled in manufacturing, so the choice of resistors comes down to cut and try.
    UJT-loadline.jpg
    UJT-monostable.jpg


    A PUT (programmable unijunction transistor) has better controlled parameters, and can operate at considerably lower current. The valley current in the following circuit is on the order of 200µA, the peak point current around 1µA. A BJT or mosfet can be added to drive a heavier load.

    PUT-latch.png


    I'm a big fan of PUTs. Wish they were available in SOT23. Alas, it is a dinosaur and only seems to survive in TO92.
    720 x 510 - 71K
    720 x 333 - 44K
    440 x 290 - 14K
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited August 2012 Vote Up0Vote Down
    My current favorite Unijunction is the 2N6028. Digikey has it for $0.44us. This is modern made and should be around for quite a while. And yes its in a TO-92 package.

    Duane J
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited August 2012 Vote Up0Vote Down
    My current favorite Unijunction is the 2N6028. Digikey has it for $0.44us. This is modern made ans should be arround for quite a while. And yes its in a TO-92 package.

    Duane J
  • skylightskylight Posts: 1,915
    edited August 2012 Vote Up0Vote Down
    I'm with erco on using SCR's and all he said re holding current etc perhaps adding a dummy load may help?
  • Hey, what happened here? Somebody posted a comment to this ancient thread just now. Then I come here to respond and the post is gone.

    Anyway. Now that I'm here, here is my solution.

    Use a 2N7000 N-Channel MOSFET.
    Drive a high efficiency LED from it's drain.
    The light from that LED shines on a second similar LED connected to the gate. This second LED will generate a voltage, or allow current to pass when illuminated. LEDs are photo detectors as well as photo generators.
    With a suitable LED and resistor (or two) on the gate we have:
    When the FET is off and no input is connected no light shines and the gate is held low. Thus holding LED off.
    When the gate is driven high, overcoming the LED/resistor biasing. The FET turns on, the LED shines, the receiving LED gets power and can hold the gate up when the input is disconnected.
    Drive the gate low again and we reverse all that back to the starting point. Off.

    Now, the switch on voltage of a 2N7000 is a bit over 2 volts so this may require multiple receiving LEDs or careful biasing.

    It does require that whatever drives this to go into a high impedance state between transitions. But we can do that from a Prop pin.

    I have no idea if this can be made to work. But I have seen how MOSFETS can be latched on and off just by the charge stored on the gate and nothing connected. So there should be some chance.

    This calls for some experimenting...

  • Heater. wrote: »
    This calls for some experimenting...

    Is it time to re-open the case for LED touch-sensing?

    Is the jet on a treadmill? What color is the bear? Was an icicle in fact the murder weapon? How many were going to St. Ives?

    "When you make a thing, a thing that is new, it is so complicated making it that it is bound to be ugly. But those that make it after you, they don’t have to worry about making it. And they can make it pretty, and so everybody can like it when others make it after you."

    - Pablo Picasso
  • O.T: Hope to see skylight and kuroneko back here someday soon.
    Platåberget
  • Heater.Heater. Posts: 20,554
    edited October 2016 Vote Up0Vote Down
    Never mind all the jets, treadmills, bears and icicles. And I have no idea what St. Ives is about.

    The challenge is to build a latching circuit with only one transistor. Presumably with the minimum of other components.

    Try to concentrate :)

    If you can make a line following bot with only one motor this must be a walk in the park....




  • Can't go wrong with an SCR. Simple to trigger/latch, any momentary positive voltage will do. To unlatch, you can either:

    1) momentarily break the external anode/cathode circuit or
    2) momentarily short the AK terminals together (bypass the SCR)

    13.png

    "When you make a thing, a thing that is new, it is so complicated making it that it is bound to be ugly. But those that make it after you, they don’t have to worry about making it. And they can make it pretty, and so everybody can like it when others make it after you."

    - Pablo Picasso
  • Very good.

    A normal SCR setup.

    But how do you do 1) and 2) ?

    Sounds like it needs some other transistors or whatever switches.

    Maybe I'm wrong but by interpretation of the problem is: One wire in. And some how flip the latch on and off with that. And only one transistor.
  • If we took the original 'one transistor' to mean one-sot23-package , then the modern choices can be

    * 74LVC1G175 - D-FF with /MR
    * xx2G00 or xx2G02 wired as cross-coupled gate S-R latch
    * Dual complementary transistor PNP&NPN, wired as PUT
  • Do I detect that there is more than one transistor in those packages?

    I don't seriously think a one transistor latch is a good idea. But it's a challenge right?
  • Heater. wrote: »
    I don't seriously think a one transistor latch is a good idea. But it's a challenge right?

    None of the solutions strictly meet the 'one transistor' dictate, so it is less of a challenge, more plain bad semantics.

  • My proposed solution does. Only one transistor.

    OK, with some LEDs.

    And, OK it may never be workable.

    Grief, do I actually have to build that now?



  • Heater. wrote: »
    My proposed solution does. Only one transistor.
    ....
    Well... maybe.... adding two more, more expensive parts, and adding mechanical and optical constraints, is not many's idea of a 'solution'.

    If you want to take that path, you can use a new Si8752 Isolated Gate Driver in an identical mode circuit.


  • Ah yes.

    Expensive? LEDs in my junk box?

    I have no idea what an Si8752 is. But I'm sure it's as much work to get as getting the FET I wanted.




  • Peter JakackiPeter Jakacki Posts: 6,902
    edited October 2016 Vote Up0Vote Down
    My favorite single transistor latch circuit is very simple, it doesn't need special chips that look like transistors or LEDs or anything like, just one other cheap and simple component, that is, another transistor. It might be possible to build a circuit with "one" transistor but I don't think it would be simple.

    I just had a quick google and boy are there so many circuits out there that would never work and some while trying to implement positive feedback miss the fundamentals of bjts such as Vbe drop etc. One circuit even has the base wired to the emitter thinking that would work :) The simplest circuits I've seen are two transistors, either npn+pnp scr-like latch, or a bistable latch.

    Would storing a charge on a MOSFET gate count? If the cap was good enough and leakage was very low then it would work like an EPROM memory cell. Eventually all those EEPROM/EPROM/Flash memory chips out there will "de-latch" as they lose their gate charge. But in the meantime they work.
    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    Tachyon Forth News Blog
    TACHYON DEMONSTRATOR
    Brisbane, Australia
  • Very sneaky Peter.

  • MIchael_MichalskiMIchael_Michalski Posts: 57
    edited December 2017 Vote Up0Vote Down
    It can be done without a unijunction transistor. In fact,it was apparently done in 1953 by someone at MIT.

    https://pdfs.semanticscholar.org/db1b/ff1537e3587862f23a1eabb1697161d84bcf.pdf

    They describe a "negative resistance bistable circuit". Its composed of a single transistor and some diodes and resistors.

    The UJT however is not a bad choice for such a thing. There are also devices called silicon controlled rectifiers. (SCR) When turned on they stay on until you remove the power to them. The ac equivallent turns off when the signal crosses 0.
    Particularly patient proactive practice positively predicates practically precise poly-processor Parallax Propeller programming paradigms.

    .
  • Wow, amazing paper from back in the day.

    I'm struck by how all the circuits in there are encrusted with diodes, capacitors and voltage sources.

    I get the impression that in the modern world it's cheaper and faster to just use a bunch of transistors for such latches and flip flops.


Sign In or Register to comment.