Silicon-Controlled Switch SCS

[Aaron]

Does anybody know a source for a really inexpensive SCS chip? The only one
I can find on google is an NTE239 and BYP62, but digikey doesn't carry
either and I wouldn't know where else to look.

Thanks,

Aaron

 

[Leon Heller]

Does anybody know a source for a really inexpensive SCS chip? The only one
I can find on google is an NTE239 and BYP62, but digikey doesn't carry
either and I wouldn't know where else to look.

Must be 35 years since I saw an SCS.

Leon

 

[Aaron]

Crap! What is a better (newer) technique for achieving the same
functionality of an SCS? What I am trying to do is use a momentary switch
to turn on a system that can time itself out (power itself down).

Thanks,

Aaron

 

[R.Legg]

Crap! What is a better (newer) technique for achieving the same
functionality of an SCS? What I am trying to do is use a momentary switch
to turn on a system that can time itself out (power itself down).

Thanks,

Aaron

A real SCS or four-layer diode has been a rare animal for a while.
Part numbers that used to be seen are illustrated at:

http://www.wps.com/archives/solid-state-datasheets/Shockley/

ITT still listed the 4Exx types in the 80's. The inventory of Clevite
was distributed by 4LD Inc as late as 1992. I believe these were
hand-tested.

Cheapest replacement used to be a Diac, a bidirectional version with a
break-over voltage of ~ 32V. These were common as surplus.

ST-2, BR100, DA-4, IN5411.

Sidacs can perform the same function with breakover voltages extending
above 200V, similar to a four layer diode. They are not cheap.

http://www.onsemi.com/site/products/searchresults/0,4533,,00.html?searchString=sidac
http://www.littelfuse.com/teccor/SIDACtor.html

A PUT is much more versatile. Unless you actually need the peak
current capabilities of the older parts, I'd advise one if all you
want to do is substitute the part location in the old timing circuit.

Otherwise, there's probably another way to do what you're talking
about, if you'd care to explain.

RL

 

[Robert Baer]

Crap! What is a better (newer) technique for achieving the same
functionality of an SCS? What I am trying to do is use a momentary switch
to turn on a system that can time itself out (power itself down).

Thanks,

Aaron

As someone else mentioned, the SCS is a 4-layer device, it is really
an SCR with the anode gate also available; i think GE may still make
them.
In any event, they can easily be made with two transistors: an NPN and
a PNP; connect collector of one to base of the other (two symmetrical
connections); the NPN emitter becomes the "cathode", the PNP emitter
becomes the "anode", the NPN base becomes the cathode gate (SCR gate),
and the PNP base becomes the anode gate - which is like a PUT gate.
Note PNPs tend to have a larger B-E breakdown, allowing a larger
stand-off voltage for PUT type applications.
You can use small signal transistors, power transistors, or a mix as
you see fit and depending on tha application requirements.
If you need very sensitive gate requirements, Zetex has very high beta
transistors (300 minimum on PNP and 500 minimum on NPN if i remember
corretly) that could be used. Do not use a darlington, as thsy cannot be
saturated.

 

[Leon Heller]

Crap! What is a better (newer) technique for achieving the same
functionality of an SCS? What I am trying to do is use a momentary switch
to turn on a system that can time itself out (power itself down).

Why didn't you say that in the first place???

Devices like the SCS can probably be simulated using a pair of BJTs. 40
years ago I needed a UJT when they were very new. I couldn't get any so
I used a couple of ordinary transistors. My little circuit actually went
into production on the LEO III computer.

I'd be inclined to use one of the 8-pin MCUs like a PIC or AVR for your
application.

Leon

 

[Aaron]

OK, let me start from the start, then you guys can send me in the right
direction. My (probably mis-) understanding is that SCRs and SCSs have very
little current leakage. I am building a battery operated toy (the core
module is a PIC chip) that I want to turn on when the toy is moved (or
shaken). The idea is that the PIC chip will act as a timer and then kill
power to itself in order to save the battery life.

The device I have to wake the toy up has two states: on/off. It is a
mechanical device. What I really need is the device to wake up when the
mechanical devices changes states from on-to-off or from off-to-on. When
the device is at a "steady state" it should not allow current to pass so as
to save battery life. I was thinking about using a capacitor in line with
the DC current so that when the device changes states, current can flow and
trigger an SCS ... then the PIC chip can shut it off using the fourth lead.

I could use the technique below with two transistors but I was worried about
current leakage draining the battery pretty quickly. Maybe that is an
unfounded concern?

One of the other posters has mentioned a PIC chip. Perhaps there is
something built it to do what I am doing already and I didn't even realize
it?

Thanks,

Aaron

 

[Spehro Pefhany]

OK, let me start from the start, then you guys can send me in the right
direction. My (probably mis-) understanding is that SCRs and SCSs have very
little current leakage. I am building a battery operated toy (the core
module is a PIC chip) that I want to turn on when the toy is moved (or
shaken). The idea is that the PIC chip will act as a timer and then kill
power to itself in order to save the battery life.

The device I have to wake the toy up has two states: on/off. It is a
mechanical device. What I really need is the device to wake up when the
mechanical devices changes states from on-to-off or from off-to-on. When
the device is at a "steady state" it should not allow current to pass so as
to save battery life. I was thinking about using a capacitor in line with
the DC current so that when the device changes states, current can flow and
trigger an SCS ... then the PIC chip can shut it off using the fourth lead.

I could use the technique below with two transistors but I was worried about
current leakage draining the battery pretty quickly. Maybe that is an
unfounded concern?

One of the other posters has mentioned a PIC chip. Perhaps there is
something built it to do what I am doing already and I didn't even realize
it?

Look at using sleep mode, wake on port pin change, that sort of thing.
If there's no regulator draining the battery you don't need to cut
power to the chip, just have it shut down the oscillator and other
analog circuitry on the chip. It should only draw a few uA under those
conditions, if you do everything right.

Best regards,
Spehro Pefhany

 

[Aaron]

OK I will really go over the PIC specs a lot deeper then.

Thanks,

Aaron

Look at using sleep mode, wake on port pin change, that sort of thing.
If there's no regulator draining the battery you don't need to cut
power to the chip, just have it shut down the oscillator and other
analog circuitry on the chip. It should only draw a few uA under those
conditions, if you do everything right.

Best regards,
Spehro Pefhany

http://www.speff.com

 

[Robert Baer]

OK, let me start from the start, then you guys can send me in the right
direction. My (probably mis-) understanding is that SCRs and SCSs have very
little current leakage. I am building a battery operated toy (the core
module is a PIC chip) that I want to turn on when the toy is moved (or
shaken). The idea is that the PIC chip will act as a timer and then kill
power to itself in order to save the battery life.

The device I have to wake the toy up has two states: on/off. It is a
mechanical device. What I really need is the device to wake up when the
mechanical devices changes states from on-to-off or from off-to-on. When
the device is at a "steady state" it should not allow current to pass so as
to save battery life. I was thinking about using a capacitor in line with
the DC current so that when the device changes states, current can flow and
trigger an SCS ... then the PIC chip can shut it off using the fourth lead.

I could use the technique below with two transistors but I was worried about
current leakage draining the battery pretty quickly. Maybe that is an
unfounded concern?

One of the other posters has mentioned a PIC chip. Perhaps there is
something built it to do what I am doing already and I didn't even realize
it?

Thanks,

Aaron

A SCR or other SCR-like device is not too easy to turn off with a DC
load, so i do not think you want touse that for controlling the load.
You could cross-couple two transistors or FETs as a R-S flip-flop and
one could control the load.
If bipolars were used, you may need to drive the power transistor to a
forced beta of 10 or less to guarantee saturation for a heavy load. That
would require a fair current for the driver (low power side of the
flip-flop), also hurting your concerns.
If you use FETs, there is no steady-state power for controlling them
(no gate current), which can be a factor for better battery life.
The lower power side could be set to run at low currents; for FETs one
could use microamps for the ON, and the OFF current can be picoamps.
This all assumes slow switching times are OK.