You need a heat sink. The power dissipation will be around 7.5W (figure 1 of the data sheet) and the temperature will rise 50°C for every Watt of power dissipation (table 1 on the data sheet).
I am not using heat sink but the temperature should not rise that much, what is wrong with my circuit?
And why you made such conclusion? It will dissipate around 8W as said in previous post. From the datasheet:
Rth(j-a) Junction to ambient TOP3 / TOP3 insulated 50 °C/W
So 115
oC is rather cool actually.
Question:
what is wrong with my circuit?
Answer:
I am not using heat sink
The device may well be working within the specifications you are looking at - but that just means it can handle the voltage and the current the circuit operates with. They are not the only specifications you need to worry about!
What you are not taking into consideration is the fact that WHEN this device is operating, it generates heat - and at the levels it is being driven here, the amount of heat is significant. This heat needs to be removed, to keep the active region of the device from destroying itself or other fatal damage. It won't be from excess voltage .... It won't be from excess current .... It will be from excess heat.
The thermal limits will be a specification - and you need to include these considerations in your design.
Rth(j-a) Junction to ambient TOP3 / TOP3 insulated 50 °C/W
So 115oC is rather cool actually.
The figure on the datasheet is probably the worst case scenario and is for the junction to ambient. The original poster just measured the temperature of the package to be 115
oC but the actual junction temperature will be much higher.
The original poster just measured the temperature of the package to be 115oC but the actual junction temperature will be much higher.
NO, That would be just a few degrees C,
Rth(j-c) Junction to case (AC)
TOP3 insulated 0.9 °C/W
So with 8W dissipated, that would be just 7
oC. Other thing is how precisely that temperature was measured. For example, if thermocouple didn't have a good thermal coupling with the device, measured temperature can be significantly lower tan actual.
The original poster just measured the temperature of the package to be 115oC but the actual junction temperature will be much higher.
NO, That would be just a few degrees C,
Yeah,I agree that the readings were not precise, but they can be +/-5
oC not +/- 50
oc
Any way thanks for the help guys, one question..... will replacing the triac+heatsink setup with a 10A or higher relay will be a better choice? As I guess I can even save some space also with that.
Dave did a video....
That video was very helpful.....thank you
If relay need to be switched occasionally, say one time in a few minutes, then it is ok. If it needs to be switched very often, it will fail very fast. Don't forget that for inductive loads relay should be derated a lot. And inrush current should never exceed max current rating.
It also depends on whether it will be subject to mechanical vibration.
Would you not use Figure 2. to work out the Temperature. For ~8.5 A this gives a Case Temperature of about 120C which is close to the measured figure.
Also isn't the Thermal Resistance a measure of how efficient heat is transferred from point to point, it is not used to calculate actual temperatures for a given power dissipation. I did watch Dave's video and it made sense at the time.
I have been on this forum for a few months now but this is my first reply to a post. I'm very much a beginner so I may have got this totally wrong, feel free to correct.
Would you not use Figure 2. to work out the Temperature. For ~8.5 A this gives a Case Temperature of about 120C which is close to the measured figure.
Because you cannot work out case temperature from that figure. All it shows is maximum allowed RMS current at particular case temperature.
If relay need to be switched occasionally, say one time in a few minutes, then it is ok. If it needs to be switched very often, it will fail very fast. Don't forget that for inductive loads relay should be derated a lot. And inrush current should never exceed max current rating.
Yes, it only needs to be switched occasionally. Although, I am not sure about the inrush current (need to learn about that)
It also depends on whether it will be subject to mechanical vibration.
No, there is no mechanical vibration involved.
One more thing, in dave's video, even after putting such a big heat sink, the device case temperature is reaching 118oC (without air flow). Will that affect the life of the power transistor he's working on?
Heat is the number one enemy of electronics. The cooler things run, the longer you can expect them to last, whether resistors, capacitors, semiconductors - even a piece of wire.
One more thing, in dave's video, even after putting such a big heat sink, the device case temperature is reaching 118oC (without air flow). Will that affect the life of the power transistor he's working on?
Yes it does. It doesn't mean that it won't last say 50 years, however reliability is significantly reduced. For example, you produced a lot of devices containing particular transistor. If you run transistor at 40
oC, say only 0.05% of them would fail in 10 years. If you run them at 120
oC instead, 10% would fail during the same time period.