Power losses in a switch (MOSFET, Relay, Thyristor)

[beerbaron]

Hi, I'm in the process of designing a control unit that will be on the secondary of a 12V AC transformer. The control unit will control around 100W.

I am wondering what type of design should be used:

- Mosfet switch
- Thyristor switch
- Relay switch

What would be the average losts in % of each. Which one is the more economical to produce?

Thanks
Best regards
BeerBaron

 

[ScottyUK]

Is the load AC or DC? Do you need very fast-acting control? Do you need anything other than just 'on' and 'off'?

- MOSFET's are generally used in DC applications
- Thyristors are usually confined to AC applications and are often used in antiparallel pairs.
- Relays are great for plain on-off control provided you don't want very high cycle rates or the ability to control the output between some minimum and maximum values.

The relay will have the lowest losses. The choice of the others depends on the load - one is basically an AC device and the other basically a DC device. There are other options - BJT, IGBT, triac.

----------------------------------

Sometimes I only open my mouth to swap feet...

 

[itsmoked]

Triacs for 100W AC

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[beerbaron]

Thanks guys, what would be the average losses in the triac ?

Thanks
Beerbaron

 

[Skogsgurra]

Losses = Current times forward voltage drop. The latter is usually 1.0 - 1.5 volts at rated current.

So, if you have your triac conducting 100 % (no dimming), you will have around 10 watts at 8 amperes. You can reduce losses very little by selecting a triac with higher current rating. The fwd drop seldom gets below 1.0 volts.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Skogsgurra]

Loss = Current times voltage drop. Voltage drop is typically between 1.0 and 1.5 V. So, 100 W at 12 V is around 8 A. Losses will be someting like 8 - 12 watts. A little depending on rated current of the triac.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Skogsgurra]

Huh? Answered that two times?? Thought it didn't "take" first time. But it did, apparaently...

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[beerbaron]

Dam my boss now wants it good for 300W... that changes the whole thing.

Thanks guys

 

[Skogsgurra]

300 W at 12 V? 25 A. What kind of device is that? Sounds like a relay is the best choice. Very low losses.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[ScottyUK]

It would help enormously if you provided a specification for whatever you are building. You still haven't told us whether you need simple on-off control or modulation.

Why not just switch the transformer primary. If you need more than on-off control - a lighting dimmer for example - then some thyristor / triac controller designs can deal with inductive loading. Otherwise use a relay. If the transformer is a toroid then be generous on the relay current rating.


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[beerbaron]

It's to control some low voltage lamps. Relay seems out of the question since 25Amp relay cost around 50$ a piece. We are hoping to produce the whole unit for about 4$ in production.

I believe it's going to be Triac in the end, 30W of losses is bad tough.

BeerBaron

 

[beerbaron]

Sorry if I didn't tell you everything, I'll try to explain it better.

What we are building is a device that will be on the secondary of a 12V transformer. This device will make an On/Off switch with features like a timer or else. Ideally, this product would be all plastic molded because it is intended for outdoor operations. But my concern here is the heat losses, you could imagine that 30W of losses in a closed plastic box could raise the heat by a lot, making it very hard for UL approval (95°C max usually).

I'm not going to design the circuit myself but I do have to evaluate the costs of the whole thing. So knowing the heat losses will give me the chance to evaluate the plastic dimensions and the construction.

BeerBaron

 

[Skogsgurra]

A relay at 12 V AC needn't cost a lot. And, remember that there are three paths in a typical 10 A relay. Paralleling them will bring you home nicely. And impedance of transformer will probably reduce lamp starting current - which is otherwise known to kill realay contacts. What rating (VA) is the transformer?

I think that Scotty's idea is the best. Switch transformer primary. Such a switch will be very low cost and very low loss.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[beerbaron]

Yeah, that'S what I would do to, but the customer might have strange gizmos on it's transformer that he might want to keep.

I guess I'll have to think of a case with a metal face to dissipate heat.

Thanks for your inputs
BeerBaron

 

[Skogsgurra]

Like your handle. I have been seen at one or two Beer Bar - but not sitting On them.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[itsmoked]

beerbaron; The industry work horse relay is the T90. It can switch 30A and costs about a buck and a half. It would have just it's fixed coil heat dissipation which is just less than a watt, inside your box.

http://relays.tycoelectronics.com/datasheets/T90N_DS.pdf

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com