Relacement for SCR

Yes.

You do realize that you can use a small signal relay to do the same thing? The alarm trigger pulls the coil and a contact on the relay can then maintain that source to the coil, you need only to open the path momentarily, as you would with a thyristor to reset it.

Jamie

and if it needs to be solid state just look at the transistor equivalent of a thyristor - it only takes two bipolar transistors to make your own SCR.

It does work and I've used it a time or two in the past.

Tried that very idea. but for some weird reason it seems to remember it's o n state after power cycle unless you short the gate to ground Has me mysti fied to say the least. Was going for an alarm that would sound until turned off. Have almost decided to change to timed output using 555 and MOSFET to drive relay. A little more complicated but easy to implement as I already have the parts.

** Although a triac can do the same job as an SCR in some applications, the one you mention is not usually one of them.

Typical triac trigger currents are around 5mA to 100 mA - while the venerable C106 series SCRs ( 50 to 600V rated) have typical trigger currents of 35uA.

Big difference.

... Phil

Triacs typically have a minimum holding current which can be pretty high. You'd need a small "sensitive gate" triac.

The pnp-npn transistor thing is a good idea. The b-e resistors can set both the trigger and holding currents. It powers up off.

--

John Larkin         Highland Technology, Inc 
picosecond timing   laser drivers and controllers 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

As previously mentioned I tried the transistor option but it seems to alway s activate upon power up. If you put a hard ground on the base it resets. N ot sure what I am doing wrong. Think I need to buffer the gate with another transistor. Instead of on until manually reset thinking of just using 555 set for a couple minutes and a MOSFET. I have had lots of success with 555 timers. The sensors will continually restart the timer keeping it alarming until they leave which I would hope would be immediately. To disable/enable I will just set the reset pin appropriately.

Have you, by chance, forgotten the base resistors? The simplified circuit mentioned above only shows 2 transistors, that is not likely to turn off properly if there is any small current remaining. Small currents can be caused by capacitors slowly discharging, and remain "active" for some time after the device is turned off.

In order for the 2-transistor circuit to turn off, the current has to fall to a level where the transistors' beta effectively reaches unity - and that can be in the nano-amperes. A real SCR will have a holding current at least in the micro-amperes, and a Triac even much higher.

To make the 2-transistor circuit less sensitive, connect a resistor between the base and emitter of each transistor. Lower value resistors will raise the holding current. For a start, you can try 10 kOhm (each) and if the resulting "SCR" is either too sensitive (does not turn off) or too insensitive (does not latch), decrease (respectively increase) the resistance as needed.

Regards Dimitrij

It's latched by small capacitive and leakage currents during turn-on. The transistor circuit is too sensitive (nano-amperes may be enough). See my other post about base resistors for more details.

Regards Dimitrij

The circuit I tried had 2 resistors. One from npn base/PNP collector to gate and another from gate to ground. Using 2N3904/06 pair. What values would be suggested?

Think I see the problem. I started with 1k. Apparently way too low. Will try varying values. Thanks.

try varying values. Thanks.

Umm, you need resistors from the base to emitter of each transistor. The ba se of one goes to the other and vice versa, no resistor is needed there.

I think 10 K would assure that is starts unlatched wirthout screwing with y our sensitivity too much.

I don't know if it really matters for this application, but the transistor method is probably going to drop a littl more voltage than an SCR when latc hed. Probably not important to the relay or siren or whatever you are drivi ng with it but it might have dissipation then. In saturation like that, it won't be all that bad but one must think of these things to keep the smoke it. I think a pair of TO-220s would dissipate enough just with their tabs f or quite a bit of current. The 2N3904/06 maybe not so much. ;

What voltage and current are you driving ?

with no further evidence 10K is usually a good starting point.

for a better answer: what is the maximum trigger current?, minumum holding gurrent?

--
umop apisdn

Will be driving a 12vdc relay. I think maybe I have forgotten too much to g et the SCR idea to work. All references I have found do not show resistors from b to e. I never was the design type. Always repaired other peoples wor k. Guess maybe I should have studied more. Still think I am going to abando n this and just 555 the thing and be done with it. Will still play with it for my edification though.

Hmm, Gurrent? is that a new measurement term I haven't heard of yet? :)

I mean I've heard of "Grunt" being used to gage the energy needed to push a load..

Jamie

I never used resistors just two transistors.

Perhaps someone got a hold of some bad transistors.

I remember the old hobbyist "Poly Pak's" some idiot was selling back in the 60's. The ad ran something like "we are too busy to test 'em, you reap the savings" Out of a pack of 200 transistors there may be five that met spec, you'd be hard pressed to find so many duds without first testing them. In circuits they'd do some pretty strange things if they worked at all.

In the 70's working at a power supply manufacturer, we got a batch of "leaky" (would conduct slightly with an open, or high impendence base connection). We could use them in the circuit, but on the work bench they needed ~1K base to emitter to keep them turned off. These were TO3 power transistors...

The trick was the base emitter resistors. 1k on the PNP and 100 ohm on the NPN Using 2N3906/3904. Works the same as an SCR. Thanks to all who assisted.

e NPN Using 2N3906/3904. Works the same as an SCR. Thanks to all who assist ed.

/note that what you have there is not an SCR. It is called a latch. A side benefit is you actually do have the ability to turn the device off while it is passing current. Just create a short circuit between either of the B-E junctions and will turn off while a real SCR wouldn't.

I don't recall ever seeing that done in practice though, probably because t he whole Idea is to stay turned on until power is removed.

Another neato thing is that it effectively has two gates. One pull up and o ne pull down. Also, either one can turn the latch off. All you need is a tr ansistor that saturaes at a lower voltage then the B-E drop of the trnasist or(s) in the latch circuit. You can't do that with an SCR as far as I know. (never tried)

You do have to watch the base current, though. Unlike the usual uses of BJTs, in the NPN/PNP trick all the current has to go through the bases. That puts them into the programmable-unijunction class more than the real-SCR class. SCRs are usually beasts.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

When I started this thread it was more out of necessity. The time spent pla ying with this circuit has been well spent. No it isn't an SCR. Never meant to imply that. However the latching feature of an SCR was the idea behind it all and it does that quite well once you figure out the correct values. This circuit has the added benefit you mentioned of dual gate as well as ac tive turn off. Over all a great learning experience for myself. Never reali sed how much I had forgotten. Thanks for all the good answers guys.