SSR vs MOSFET on a heaksink.

[BlackICE]

I want to control the power (on/off) to a ZVS board like this one with a micro controller output pin. The ZVS board using up to 40 volts at <=20 amps at a duty cycle of <= 6 seconds on and >10 seconds off. How would using a SSR of unknown quality and specs from China compare to just using a n-channel RFP30N06LE MOSFET on a heatsink for a low side switch? Should I add a flyback diode for protection across the MOSFET. I noticed many circuits and this for inductive loads even though the MOSFET has a protection diode built-in.

ZVS
https://www.amazon.com/KOOBOOK-Voltage-Induction-Heating-Flyback/dp/B07WD9KZ4J

SSR
https://www.amazon.com/TWTADE-SSR-40-3-32V-5-60V-Solid/dp/B079BGGVYX

[wizard69]

I want to control the power (on/off) to a ZVS board like this one with a micro controller output pin. The ZVS board using up to 40 volts at <=20 amps at a duty cycle of <= 6 seconds on and >10 seconds off. How would using a SSR of unknown quality and specs from China compare to just using a n-channel RFP30N06LE MOSFET on a heatsink for a low side switch? Should I add a flyback diode for protection across the MOSFET. I noticed many circuits and this for inductive loads even though the MOSFET has a protection diode built-in.

ZVS
https://www.amazon.com/KOOBOOK-Voltage-Induction-Heating-Flyback/dp/B07WD9KZ4J

SSR
https://www.amazon.com/TWTADE-SSR-40-3-32V-5-60V-Solid/dp/B079BGGVYX

Have you considered a mechanical relay?    Driving the relay would be a little bit more involved but that positive thing here is that you get better isolation.   

As for an SSR module, being from the industrial world I would lean towards an SSR module before trying to engineer my own solution.    Just make sure it is properly rated.   Just be aware of the heat sinking requirements, in some cases you will need a fan solution.

[BlackICE]

I may have to cycle the circuit off and on very rapidly for 0 to 2 seconds a relay wouldn't be ideal and noisy.

I'm using this drive an induction heater that will bring a part up to temperature. An IR temperature sensor will monitor the temp and once reached will vary the duty cycle like 100 times a second up to an additional 2 seconds.

The cycle would be something like on for 3 seconds (up to temp), cycle for 2 seconds off for 10+ seconds, then repeat for the next part.

Even if a relay could switch fast enough the contacts wouldn't last long. IMO relays are great when you don't switch often or quickly, reliable no wasted power. But if you have to switch often semiconductors should last longer at the cost of wasted power.

[Zero999]

The advantage of an SSR is isolation.

If it's an inductive load, a diode across the MOSFET will do nothing to protect it, unless it has a lower breakdown voltage, than the MOSFET's maximum voltage rating i.e. is a TVS, or zener diode. The diode needs to go in parallel with the load.

The RFP30N06LE has an on resistance of 0.047Ohms, with a gate voltage of 5V. Presumably the MCU is powered off 5V, otherwise it's no good.

The on resistance given on the data sheet is specified at a junction temperature of 25^oC and increases with temperature. The data sheet doesn't give the on resistance at higher temperatures, but the article linked to below suggests it will be around 50% higher at 120^oC and 80% higher at the MOSFET's maximum rating of 175^oC. If you aim for a maximum junction temperature of 120^oC, the on resistance will be 0.07Ohms and power dissipation P = I²R =  20²*0.07 = 28W.

https://toshiba.semicon-storage.com/eu/semiconductor/knowledge/faq/mosfet/is-the-on-state-resistance-of-a-mosfet-dependent-on-temperature.html

The thermal resistance of the die to package is 1.55^oC/W and if the maximum ambient is 40^oC, the permissible temperature rise is 80^oC. The maximum acceptable thermal resistance between the MOSFET junction and case is T_RMAX = T_RISE/Power = 80/28 = 2.86^oC/W.

Subtract the thermal resistance of the die to package and the total thermal resistance of the heat sink plus any thermal pads is 2.86-1.55 = 1.31^oC/W.

I advise using a MOSFET with a lower on resistance, or lower thermal resistance between the junction and case.

Another question: does the induction heater have an enable signal? If so, keep it powered and connect the enable signal to the MCU output.

[BlackICE]

Yes 5V MCU. The heater is just a coil connected to the output of the ZVS module. Couldn't I use 2 or more MOSFETs in parallel adding a gate driver? But since I'm not planning to make a PCB switching high currents from discrete wires connected to the pins of a TO-220 seem like a PITA, so maybe is best to get an SSR module to save the work and complication.

[rpiloverbd]

I would have suggested just an ordinary opto-isolated relay module. But, since you already have explained your reasons behind not using it, I think SSR will be the best option for you.

[Peabody]

The problem with cheap SSRs from the Far East is that if you take one apart you find that the parts inside are rated at 1/2 or even 1/4 of the advertized SSR rating.  Particularly "Fotek" SSRs.  You will find lots of Youtube videos about this.  The other issue is whether even good SSRs are designed to be cycled rapidly.  Assuming it uses an optocoupler, it's hard to see how you would get the really rapid switching of a DD SSR needed to keep the heat down.

[b2theory]

I would also make sure the SSR isn't implemented with and SCR. Often they are and that would only be usable in an AC application.

[Peabody]

Here's a video of a guy using a Chinese DD SSR to test a ZVS module.  It was a "100A" SSR, but failed short at 25A.  He opened it up, and found that the switching device was a BJT.