1
\$\begingroup\$

I came across this description of what are called "silicon controlled switches." I always look to balance n/p-channel electronics for stability, and using these silicon controlled switches seemed like the perfect way to do so. However suppliers seem few and far between, and the few products I could find were >=$50 a piece!

I could find very little information on these devices online, so I am wondering what technological niche they fill, and if anyone sells them at a reasonable price--i.e. on the order of the price of the components an equivalent circuit made out of regular transistors. If not, why are they so expensive?

I am beginning to suspect that they are more commonly known/sold under a different name.

\$\endgroup\$
  • \$\begingroup\$ The big cost decision will be made by the current you need. \$\endgroup\$ – Kortuk Jan 22 '12 at 23:28
3
\$\begingroup\$

An SCS is a specialist device that is now seldom available. The equivalent circuit is shown in the page that Feynman referenced. It's main claim to fame is that it is a TRIAC like device that can be driven off in the middle of a conduction cycle using gate drive alone, unlike a TRIAC.

  • A TRIAC or SCR can be driven off mid conduction by interposing an inductive reverse conduction inducing current spike, but that is unusual* and cheating, as conduction is stopped by stopping conduction.

  • A number of high power camera flashes use this method to terminate flash cycles.

You can buy an SCS here for $US17.50/25. Given that it is a 1400 Volt x 112 Amp device the price is "not bad". Datasheet here.

An SCS (Silicon Controlled Switch) is a device whose time has largely passed. It lingers on in very high voltage high current niche applications but would almost certainly not be used for new work.

An equivalent functionality can be provided by using 2 x MOSFETS (2 x P or 2 x N channel) connected in series with sources connected and gates connected (!!!).
You thus get a module where the current path is

---[DS]-[SD]--

Driving the 2 connected gates +ve relative to the two connected sources (for an N channel pair) turns the pair on and connecting the two connected gates to the two connected sources turns the pair off. this then is a true bilateral (bidirectional, 4 quadrant) swith - about the most flexible switching device available. Even better than an SCS if you can but an SCS :-).

To equal the above SCS you could use two of something like the IXYS MOSFET (same brand as the SCS) IXTY06N120P 1200V 90A MOSFET priced at $2.79/25 at Digikey but not in stock. It's a TO252 package which is capable but which may get "somewhat stretched" [tm] at those power levels.

In most cases a more normal topology MOSFET or TRIAC solution will do whatever job you have well enough.
There is little that you cannot do with an H-Bridge :-).

http://search.digikey.com/us/en/products/IXTY06N120P/IXTY06N120P-ND/2117410

\$\endgroup\$
  • \$\begingroup\$ Ok, I guess I will just use mosfets with an H-bridge, but it sure seems odd that SCSs died out of use. To me, SCS is to a pair of mosfets as CMOS is to NMOS (or PMOS if you use two p-channel mosfets.) So naturally, I am still confused as to why the advantages gained from balancing p and n channels were outweighed by the costs of tacking on one extra lead. \$\endgroup\$ – Alex Eftimiades Jan 23 '12 at 1:04
  • \$\begingroup\$ @Feynman - SCS and MOSFET use different base technologies. MOSFET is a resistive channel, SCS isn't. SCS is bestter compared to dual bipolars. \$\endgroup\$ – Russell McMahon Jan 23 '12 at 1:46
  • \$\begingroup\$ Ok, but that still begs the question of why dual bipolars caught on and SCSs did not. Perhaps I should start a new thread, but are there any FET equivalents to SCSs? I know I could just use transistors, but I was just curious as to whether it might be something to look into since I almost always seem to wind up arranging my transistors in this fashion anyway. \$\endgroup\$ – Alex Eftimiades Jan 23 '12 at 3:20
  • \$\begingroup\$ @Feynman - very few people arrange FETs's "back to back with opposed polarity" as I have described. It is done, but in most cases being able to arrange individual MOSFETs in topologies that suit the current need is laible to be less srestricting than having to use a necessarily more expensive part. A MOSFET lends itself to being "grown" in layers. Two MOSFETs in series with opposing polarities could be implemented but takes more steps and almost certainly needs many compromises. \$\endgroup\$ – Russell McMahon Jan 23 '12 at 5:19
  • \$\begingroup\$ In a semiconductor technology 'arms race', generally what works best at least all-up cost per performance metric that best reflects customer demand wins in this sort of market. The bubble memories, CCD memories, core memories Josephon junctions, Esaki diodes and their ilk are still around and may have niche applications but are in little use. In the power area SCS joins Germanium transistors, copper oxide and selenium rectifiers and to a large but not complete extent power bipolar transistors. Some may even make comebacks (eg Josephon junction) but most won't. \$\endgroup\$ – Russell McMahon Jan 23 '12 at 5:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.