need a latching semiconductor switch..

i need a latching semiconductor switch that turns on with a high pulse but then latches closed and isn't affected by any other signals. this will be used to open/close a voltage to another IC. I need only one gate array, so preferably not a huge IC or at least as few parts as possible. any ideas? all my ideas seem to have some kind of fault.



:thumb:

you need a set/reset flipflop or something similar.

That's not much information.... Surely you want to be able to reset this latch somday? When you cycle the power? When you press a manual reset button??

What are the input trigger levels, voltages/currents? Application?

The classic albeit antiquated part that does this is a three-terminal device called a silicon-controlled rectifier, or SCR. It looks like two complementary transistors hooked up with a collector of one to the base of the other. You hit one of the bases with a forward-biasing pulse and the device goes into hard conduction, which lasts if a certain minimum current continues to flow.

hey Bcarso, Yeah i thought about using an SCR right after I posted that. I *was* going to use some kind of logic gate to interrupt a variable voltage, but have since come up with another way to use this, The IC i am using has a shutdown pin that needs to be pulled down to shut off. likely what I'll do is use a logic SCR to trigger a transistor in some fashion. I'll be checking more on this later.

I use SCRs on a daily basis, however for power switching, not logic control.. I guess it's kinda hard to switch the train of thought for specific parts.

for more information, I don't want to use a reset flip-flop because i only need one channel and all the ones I found are at least 4 in a package. How it would work is as follows: the part is latched by a pulse of 5V and allows an IC to turn on. Upon a hard poweroff condition, the IC would stay off and then have to be reset manually with a pulse to the switch. It's a safety mechanism really.

:green:

You can make a S-R latch out of a couple of transistors and some resistors, if that is less painful than buying a gate package and only using 1/4 of it. It usuallly needs a way of coming up in a preferred state, but that's going to be true of things like SR latches and stuff made out of NAND gates too. Also, there are these teeny SMD gate parts that have only a couple of gates per package that might work.

You could also roll your own low-current SCR out of an NPN and a PNP. Make sure you have a pullup resistor if the pin on your IC lacks one, and put an R in series with the two-Q SCR trigger base to limit current from your trigger source. You may need an R across one or both of the BE junctions to make the hold current something tractable.

what would you think of using a logic SCR by connecting the gate to the pulse source, the anode to my Vref(the 5v source..) and the cathode to the gate of the FET. the cathode side of the SCR would also need to have a current load and I think i'll use an LED(also use it for visual confirmation) and resistor to pull the needed holding current(Ih) to keep it conducting.

the only problem is that I use ONLY SMD parts and even logic SCRs are almost only availiable in through-hole unless you get a larger die part and that of course is defeating the purpose here..

I know i'm the blacksheep around here because I don't like through-hole stuff!

A real SCR is I think still going to be a higher-current device than you need, but if you have a datasheet link to one I'll be happy to look at it. You are talking about needing about 10mA if you are driving the LED and resistor I guess.

Note that if you put the little SCR in the circuit like you propose your threshold will be raised a bit by the LED voltage, which is probably ok.

Are you saying that your controlled device enables when the input is taken low?

i'm looking at logic level SCRs, the specs for firing current is only 12ua. the specific one needs 5ma of Ih to stay conducting.

I mistyped in my earlier posting, the shutdown pin needs to be pulled UP to shutoff, not down as I typed, sorry for the confusion!


I'll be putting a pullUP resistor on the shutdown pin. I'll be using a N-FET to do the actual pulling down of the shutdown pin on the IC when the SCR is conducting. The logic FET specifies a turn on at 2.5v so I should be within the ballpark even with the Vdrop from the SCR.

ok, drew up a quick schematic and ran spice simulations.. this should work well!

thanks for all your help!

I'll let you know how this works out.

if anyone was interested in this, the design i came up with DID indeed work well! my bosses seemed mighty pleased..

:green:

Glad to hear!

Brad

If you're not already done, you could use a PNP/NPN transistor emulation of an SCR. Unlike with an SCR, you get to put in resistors between the two and base-emitter resistors to diddle with the "SCR" characteristics, and you can turn it off by using the second base. It latches with either a pulse up on the NPN base or down on the PNP base, and you can arrange other transistors to get active high or active low inputs, inverting or not.

And it can be all SMD.

Thanks R.G. i had that in mind before, but for now the scr is working very well and is SMD and only one part.

thanks everyone! :thumb: