I am trying to heat a 1000w heating element with 120VAC. I am currently using a thermocouple and SSR to regulate the temperature, but the SSR is getting too hot at full power even with a heatsink. In open air, it's fine, but once it's in an enclosed, even with vents, there isn't enough cooling. I want to avoid moving parts so I didn't want to use a relay or a fan; is there an alternative?

Edit: Since people mentioned zero-crossing, I am switching at the zero-cross points and while I say it was an SSR, it wasn't a prebuilt one, but one I made with a BTA-16 600B triac, 4n25, and moc3021. It was based on an AC dimmer module sold by RobotDyn since I wanted to originally use PSM to control heat output, but now I simply need the AC load to switch on or off and I'm not sure if there is a better way to do it.

  • 4
    \$\begingroup\$ Please edit to include the make, model and datasheet link for the SSR and details of your SSR heatsink. \$\endgroup\$
    – Transistor
    Apr 18, 2020 at 9:44

4 Answers 4


Alternative would be a over-specced SSR, based on the load, I presume your using something like a 10 to 25A one, so try a 60 or 80A one, the device is made to run at X temperature for that rating, and half the current means quarter the heating over the switch. so a really big SSR will run with a 10A load at a much cooler temperature.

Edit: Based on his new edit, Clearer what the issue is now, change out your opto-triac for a MOC3041, that way it will not trigger until it crosses through zero again, instead of monitoring it with the micro

Edit 2: Also the above still applies, a larger triac will reduce the dissipation, e.g. 30A triac datasheet in the same package as what your currently using would have half the dissipation.

  • \$\begingroup\$ This answer probably wouldn't be true if the SSR was triac based rather than MOSFET based. Please don't be too hasty and wait for the spec of the device. \$\endgroup\$
    – Andy aka
    Apr 18, 2020 at 10:08
  • 1
    \$\begingroup\$ Yep, experience shows that SSR problems, especially if not even a type is giving, are often due to "I bought this really cheap SSR on a website that doesn't give any datasheet for it, but the product title said 'TOP: AWESOME 1000A SOLID STATE RELAY FOR ALL THINGS', and now it doesn't work", and the answer to that is: Get a proper SSR with actual ratings. \$\endgroup\$ Apr 18, 2020 at 10:24
  • \$\begingroup\$ It could indeed be that he is using a mosfet based one, switching on and off at arbitary points in the cycle, instead of a zero crossing triac based one, which would increase the total heat, When he updates the information, I will in turn update my answer, the information is still valid, \$\endgroup\$
    – Reroute
    Apr 18, 2020 at 10:27
  • \$\begingroup\$ at 8A a triac is going to dissipate about 10W \$\endgroup\$
    – Jasen
    Apr 19, 2020 at 3:39
  • \$\begingroup\$ I'll order the parts and try out your solution along with a few others, thank you for the help. \$\endgroup\$ Apr 19, 2020 at 21:09

If it is fine in the open air, and you want to avoid forced cooling, get a larger heatsink and a larger enclosure. You might want to consider a heatsink that can have its fins on the outside of the enclosure.


while I say it was an SSR, it wasn't a prebuilt one, but one I made with a BTA-16 600B triac, 4n25, and moc3021.

TRIACs have voltage drop, which is quite high no matter what their rating. This is caused by the TRIAC needing to 'steal' some voltage to keep itself turned on.

At 120 V you need ~8.3 A to produce 1000 W. A typical TRIAC drops about 1.5 V at rated load (one rated for higher current should drop less at lower current, but it will still be significant). Your TRIAC probably drops > 1 V at 8.3 A, producing > 8 W of power loss. That will need a reasonably large heat sink with good airflow to keep it cool.

I'm not sure if there is a better way to do it.

A switch using high voltage power FETs could be more efficient, but you need two FETs wired back-to-back for AC operation, and the driver circuit must ensure that the FETs don't turn partially on when there is significant voltage across them. High voltage high current FETs are also much more expensive than a TRIAC.

  • \$\begingroup\$ Thank you for your answer. I was planning on trying out this solution as I had seen it elsewhere; cost isn't much of an issue as long as it can run reliably. Do you have any recommendations for FETs or any resources where I can learn more about building one of these? \$\endgroup\$ Apr 19, 2020 at 21:34
  • \$\begingroup\$ Design:- Solid State Relay (SSR) Circuit using MOSFETs Suitable MOSFET:- STY145N65M5 nz.element14.com/stmicroelectronics/sty145n65m5/… \$\endgroup\$ Apr 19, 2020 at 22:08
  • \$\begingroup\$ Great, thank you for the help; I'll give it a try. \$\endgroup\$ Apr 20, 2020 at 0:05

SSRs (thyristor based ones) have a loss that is about 1W/A. Overspecced ones will be a bit less, but generally in that neighborhood.

So your options include using a mechanical relay (less loss and no heat sink but limited life and slower cycle time), getting a better heat sink (perhaps with fan), or using a higher mains voltage (using 240V with a corresponding heater would reduce the loss in the SSR to about half).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.