Heat in a triac

[kybert]

Hi,

Is there a rule of thumb calculation for calculating the self heating of a triac?

Im looking to switch 2000W (dimmer), 220VAC, so ~10A.

Im looking to use a rated at approx 1.5 times the max current.

BTA216-600B (16A) or BT139-600 (15A).



Thanks.

 

[MikeMl]

The basis of the self heating calculation is on the TRIAC's data sheet.

 

[kybert]

Hi Mike, thanks foer the reply.

https://www.electro-tech-online.com/custompdfs/2010/09/0900766b800ab66c.pdf

Could you please point out where and how to use it? im at a loss with this.

 

[MrAl]

Hello,

You can look at the devices "on state" voltage. For that triac it looks like it is 1.5v max at 20 amps, so with your load of 10 amps that could mean peak power dissipation as high as 15 watts, but average likely closer to 10 watts or even less because that 10 amps is a sine wave. Could be as low as 8 watts, but if you get lucky with a low on state voltage it could be even lower than that.

So the rule of thumb would be:

Pmax=Vmax*I/1.4

 

[kybert]

Ah, yes, that makes sense. I've looked at the datasheet, and it doesnt say how many °C per watt the device heats up by?
Most FETs do give this value, so how am i suposed to work that out?

Then i can start thinking about what heatsink °C/W to use.

 

[MrAl]

Ah, yes, that makes sense. I've looked at the datasheet, and it doesnt say how many °C per watt the device heats up by?
Most FETs do give this value, so how am i suposed to work that out?

Then i can start thinking about what heatsink °C/W to use.

It gives the thermal resistance junction to case as 1.2K/W so you can almost ignore that and go with whatever spec the heat sink has. For example, if your heat sink has 5 deg C per watt, then at 10 watts that would be a 50 deg C temperature rise, so expect around 60 deg C temperature rise.
Something around a 3 deg C per watt heat sink would be even better.

 

[kybert]

I calclated the heatsink, and heres my working, please feel free to comment/correct.

From datasheet:
Thermal resistance Junction to ambient (RthJa) = 60K/W
Max junction Temp (Tj) = 125C

Assumed Ambient Temp (Ta) = 40C

Therefore;
Max power disapation of the part = (Tj-Ta) / RthJa = (125-40)/60 = 1.416W

As this is less than 15W we calculated eariler (1.5V * 10A = 15W), a heatsink will be required.

---

Total thermal resstance (Rt)

Rt = (Tj-Ta) / Power = (125-40)/15 = 5.6C/W

Thermal resistance of Junction to base (Rjb) = 1.7K/W (from triac datasheet)

Thermal rsistance of Heat Sink Compound (Rbh) = 0.5C/W (from datasheet for compound)
This is known as the base to heatsink resistance, hence 'Rbh'



Thermal resistance of heatsink (Rhs) =

Rhs = Rt - Rjb - Rbh

Rhs = 5.6 - 1.7 - 0.5 = 3.4C/W


So, i need a heatsink with 3.4C/W or lower.




Rt: is this the total thermal resistance required to ensure the part does not exceed 125C ?

 

[MrAl]

Hello again,


That sounds right but the thermal resistance is not 1.7K/W it is 1.2K/W unless you really want to operate with only one half cycle all the time.
Also, the power comes from a sine wave and a 'somewhat' constant voltage drop (it's not really constant but we'll approximate it that way) so the peak power might be 15 watts but the average power is the instantaneous power averaged over time which would come out closer to 10 watts so we'll use that figure instead.
Also, it's not a good idea to allow the device to get as hot as 125 deg C, and a 60 deg C rise is probably a bit much too but we'll assume that for now.

Doing it that way we have:

RTreq=(100-40)/10=6.0C/W (this is total required for a 60 deg C rise from ambient to cap at 100 deg C)

Thermal Junction to Case Rjb=1.2K/W from datasheet

Thermal resistance of Heat Sink Compound Rbh=0.5C/W

Total device thermal resistance RT=1.2+0.5=1.7C/W

Heat Sink: RTreq-RT=6.0-1.7=4.3C/W

So the max spec for the heat sink comes out to around 4.3 deg C per watt, but again 3 deg C per watt would be even better. With a 3 deg C per watt heat sink we would end up with around 4.7 deg C/W total which would mean only around 50 deg C rise over ambient. My guess is that the triac will do even better than 1.2K/W though but we have to do it right to make sure.

 

[kybert]

Many thanks for your reply.

I used 1.7 because its a dimmer, so it uses between 10 and 100% of the sine wave.

 

[MrAl]

Hi again,


I think if you are switching on both halves of the cycle you can use 1.2 even if it is less than a full half cycle for each half cycle, but you can use 1.7 if you like as these are only estimates anyway and if you use 1.7 you'll end up with a cooler temperature design overall.