Triac vs MOSFETs for AC switching

know if it's better to go with a pair of FETs or with a Triac? My design ha s to be able to withstand a short circuit current of 200A for a few msec, b ut normally it's running around 0-8A. I like that the FETs can open up rea l quick, I have to wait for zero-crossing to open my Triac.

I use those triac "bricks" you can buy on ebay for $4. You could put a 75w light bulb in series with your load.

"Crude, but effective" - Old Chinese proverb

They have GTO's, besides, I think it's a good idea to wait for zero, don't you think?

Jamie..

On Tuesday, April 22, 2014 5:36:09 PM UTC-5, Maynard A. Philbrook Jr. wrote :

e know if it's better to go with a pair of FETs or with a Triac? My design has to be able to withstand a short circuit current of 200A for a few msec, but normally it's running around 0-8A. I like that the FETs can open up r eal quick, I have to wait for zero-crossing to open my Triac.

What's a GTO? I like waiting for zero to close the switch, but if I see lar ge current from a short circuit, I would like to open up really fast, and n ot wait milliseconds for the zero cross, I like that a FET can open quickly in response to something like that

Gate Turn Off SCR. But they don't make small ones anymore. Just large

If you want to turn off a SCR, you can try forced communitation, basically reversing anode and cathode voltages via a large capacitor.

Cheers

Just looking at a couple of SSRs that can switch at least 8 A on Digi- Key, the single-cycle surge current rating seems to be around 10 times the steady-state current rating. And if you *do* blow one up, it's a few minutes with a screwdriver to change it out.

They do have a maximum turn-off time of 8.3 millisec for DC control voltage. The turn-off time is longer (20 or 30 millisec) for AC control voltage; I guess maybe they have a small capacitor in there or something.

As a specific example, Panasonic AQA411VL or AQA421VL have ratings of

Standard disclaimers apply: I don't get money or other consideration from any companies mentioned.

Matt Roberds

know if it's better to go with a pair of FETs or with a Triac? My design has to be able to withstand a short circuit current of 200A for a few msec, but normally it's running around 0-8A. I like that the FETs can open up real quick, I have to wait for zero-crossing to open my Triac.

Well the thing is that SCRs and triacs CAN be commutated. It is just a bit complex for SCRs and a lot complex for triacs. That said, a lot more depends on whether or not you can control the application space of your switch. If you cannot, triacs have a big overcurrent durability advantage.

?-)

anyone know if it's better to go with a pair of FETs or with a Triac? My design has to be able to withstand a short circuit current of 200A for a few msec, but normally it's running around 0-8A. I like that the FETs can open up real quick, I have to wait for zero-crossing to open my Triac.

large current from a short circuit, I would like to open up really fast, and not wait milliseconds for the zero cross, I like that a FET can open quickly in response to something like that

A Gate Turn Off thyristor (SCR).

?-)

I've made a quick comparison just the other day and I've concluded that TRIACs are still better for higher voltages applications, at least in terms of power dissipation and robustness. But in my case I needed >200V, maybe in the 100...150V range there is something better.

A fundamental parameter reported in every TRIAC datasheet is the I2t rating, so it's pretty straightforward to choose the right fuse to protect them.

The only possible reason I could think of for using FETs rather than one of the wonderful thrysistor type devices out there is if it is an extremely c omplex, like partly highly inducive, load. Or maybe extremely non-linear to the point where you can't count on zero crossing so much.

Something like that. Even then there are ways to deal with it, it just beco mes a matter of what is easier.

Den onsdag den 23. april 2014 21.33.29 UTC+2 skrev snipped-for-privacy@gmail.com:

of the wonderful thrysistor type devices out there is if it is an extremely complex, like partly highly inducive, load. Or maybe extremely non-linear to the point where you can't count on zero crossing so much.

comes a matter of what is easier.

the advantage of a fet is that you can turn it off "instantly" so you can make it pretty much short-circuit prof

with a triac you have to handle a whole cycle

-Lasse

That is a gate turn off thyristor..

A note, when working with thyristor switches care must be taken driving highly inductive loads. The thyristor will likely not turn off at the line source when it reaches 0 but up to another 90 degrees after that. This should be fine since by the time the current near's 0 through the thyristor, you only get the remaining hold current of the thyristor to worry about in the collapse.

Jamie

The nice thing about thyristors is that they DO turn off at zero crossing and so tend not to have much problem with inductive flyback spikes. Kinda takes care of itself.

But the nice thing about FETs is their low voltage drop where you're stuck with drop and dissipation with SCRs and triacs.

For AC control, I tend to use thyristors and also already built SSRs with triacs in them. For the high surge currents, I just use larger SSRs or over-design the thing.

SSRs or MOCxxx drivers and discreet SCRs/triacs can also syncronously turn ON at zero crossing which can help in some situations, like reduced EMI.

boB

A pair of power FETs have a pair of source-drain diodes built-in. I think that solution is very complicated.

not really, FETs source to source, you'll need a isolated supply to drive the gates

~ | | ||-+ ||

Have you looked at IGBTs? They are quite cheap and can be turned off PDQ.

nope, I've never used those and they didn't come to mind, I'll look into them. thanks

what's complicated about the diodes? when the fets are off, those diodes block current, when the fets are on... they don't matter....

Shouldn't one of those be a PMOS?

Or, use a fullwave bridge around a single MOSFET (losing some efficiency). Any old relay would be easier and more efficient.

nope, two NMOS

relay will be more efficient but it won't last long if you need to toggle it at a few Hertz

-Lasse