Triac is heating too much

Thread Starter

mariosergio

Joined Feb 23, 2018
23
Triac is heating too much

Hello everyone, and thanks for reading this..

I'm trying to turn on a 500W load using a triac. The load is made of incandescent lamps. On 200W the temperature rises too much. I'm not sure if I just need another heat sink, or if the circuit is poor design. I try changing the 1K resistor for a 470, but it didn't change anything

Thank's very much for the help!

1581702182316.png

The temperature measurement is done by a lm35 attached to the triac with epoxy. (see picture below)
1581702634498.png


The temperature is at the table below. It was measured after 1,5 minutes, and I'm sure after more time it would have risen even more

1581702958096.png

Bellow is a picture of the setup. Not pretty, but cables are well connected
20200214_144708.jpg

ps: sorry for bad english
 

Attachments

crutschow

Joined Mar 14, 2008
34,280
The forward (conducting) drop of an SCR is about 1.2V, so for 500W @ 120V, that's around 5W dissipation in the SCR.
You thus need a heatsink that will handle that power and keep the SCR case below its maximum temperature (about 100°C) at the highest ambient temperature it will see.
So the temperatures you are seeing are within its rating, and seem reasonable for the small heatsink you have.

There's nothing you can do to reduce that dissipation, but if you don't like it running that hot, then you need a bigger heatsink.
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
That version of Triac comes in three versions, one is none insulated tab, so if mounting to aluminum hardware, ensure to use insulation if metalic structure, if you have that version.
Max.
 

Thread Starter

mariosergio

Joined Feb 23, 2018
23
Hello crutschow;
That is bad news.. I was hoping for an answer like "change that and everything will be cool" but let's face reality..
I'm ok with running it near 100ºC, can't go above that because the case is made with PSI plastic, and its melting point is near 100ºC. The temperatures is the table is fine, but I only reached 360W, and that is running outside of the case..
(and the actual case will have two of those circuits)

But if I keep the core at 100ºC, the heat sink (near the case) will be a little less right?

I guess I gonna have to use MaxHeadRoom's solution and try to expand the heat sink, perhaps shape it around the whole case. The good news is that my version is the insulated one. (BTA16-600BW), insulated and snuberless

One more question.. that 1k resistor, can I keep it like that? I tested with 470 and even 100 ohm resistor, didn't see any changes, but from 1k to 100 is quite a difference and I don't know with one to use..

Thank you both for the kind reply!
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
One more question.. that 1k resistor, can I keep it like that? I tested with 470 and even 100 ohm resistor, didn't see any changes, but from 1k to 100 is quite a difference and I don't know with one to use..
A good reference is the Fairchild AN-3006
Also AN-3003.
Max.
 

Thread Starter

mariosergio

Joined Feb 23, 2018
23
Thanks for the recommendation.

After reading the Fairchild application notes, I decide to change the 1k resistor to a 180, and the 470 to 270; It's not gonna make any difference in the temperature, but it's supposed to work longer and more reliable.

The notes don't mention anything about how to calculate the 180 resistor, but they use it in all the examples..

As for the temperature, the plan is to use a 2mm aluminium plate (or copper) that connects to the back of the heatsink, and goes all around the case.

In any case, thanks all for the help
 

Hymie

Joined Mar 30, 2018
1,277
As others have said, you need to increase the heatsink size, probably to as large a size as the plastic enclosure will allow, and preventing any wire insulation from contacting the heatsink.

Bear in mind that if there are no openings within the enclosure, the internal temperature will rise due to the limited heat dissipation offered by the enclosure.

Take care with that breadboard; I used to use the same for my prototype circuits – but never connected directly to the mains.
 

cmartinez

Joined Jan 17, 2007
8,218
For high loads such as the one you're working with, I always use heavy duty heat dissipators, and then attach a a small 40mm fan on top of it to maximize cooling. Here's a pic.


1581798615107.png



 

crutschow

Joined Mar 14, 2008
34,280
Bear in mind that if there are no openings within the enclosure, the internal temperature will rise due to the limited heat dissipation offered by the enclosure.
That's a very good point.
You may need to mount the heatsink on the outside of the case, otherwise the temperature inside will likely rise until it melts the case.
Alternately the case will need large ventilating holes.
 

Thread Starter

mariosergio

Joined Feb 23, 2018
23
Hello everyone,

Just an update in case anyone is interested..

I both and aluminum bar 1000x20x2 (mm) and repeated the test. The results are below.
This time I waited long enough so the temperature could reach equilibrium.
20200215_205520.jpg
I placed the component near the end of the bar, because I didn't want to lose too much of it.

1581957297205.png

After about 15cm away from the triac, the bar was as cool as the other end.
I could not use 6 lamps, because of a socket not working, and I almost lost my monitor that got super hot near the lamps..

The results are very good because 15cm is not that much, and 54ºC is very good. It will rise when inside the case, but..

Below is a picture of the heatsink shaped around the case. I had to use a bar that is 15mm wide, instead of the 20mm used in the tests.
20200217_133035.jpg
It took me way more than expected to bend the bar in that exact shape.

Further tests will show how much power it can dissipate. The plan is to open a lot of cuts in the case, as many as I can get. And since the microcontroller monitor the temperature, it will shut down automatically, to prevent the case to melt.


Take care with that breadboard; I used to use the same for my prototype circuits – but never connected directly to the mains.
Yep.. Accidents have happened in the past, but now I take extra care. I got a breaker nearby, and a red LED that turns on using the power from the mains. That helps to notice that the power is on.

That is all I have now. Thanks for the suggestions! Pretty confident about this, it may not get to 500W, but I will be happy with 400W.

Thanks everyone for the suggestions and ideas. It's helping me a lot!!
 

Hymie

Joined Mar 30, 2018
1,277
Make sure that the holes in the enclosure do not allow any parts at hazardous voltages to be touched by persons with small fingers (ideally so they cannot get within 3mm of such parts).
 

Hymie

Joined Mar 30, 2018
1,277
Looking at the photo of the finished item, the orientation of the opto couplers is not optimal in terms of maximising the isolation between circuits.

The opto packages are configured such that the isolation is between pins 1/2/3 to 4/5/6, so the packages should be turned through 90 degrees to that shown.

That said, at a nominal 120Vac, you only need a minimum of 3mm distance between SELV and mains circuits (for safety) – you may have this despite the above potential issue.
 
Top