Sending PWM through a relay

[eccerr0r]

Re: Sending PWM through a relay

When dealing with 120V 25A (3KW) circuits it's best to consider 240V across-both-phases for new design to reduce I*R losses.

Honestly after reading this thread and doing some analysis I'd stick with a mechanical or at most a hybrid relay (if it were mission critical) for 3KW applications, if I were forced to use 120V. At 240V I may dabble with SSR but possibly not even that, 480V SSR is a consideration but I don't have 480V readily available in my house.

[clearchris]

Re: Sending PWM through a relay

I have one like this:

https://www.aliexpress.com/item/32656181441.html

[clearchris]

Re: Sending PWM through a relay

In other applications, I might worry about dissipation, but I don't think it's going to affect this one much. Either I'm going to have enough power, or I won't. Right now, it's a regular toaster oven (standard resistive heaters) and I'm adding in two 500w halogen lamps. I'll probably have to put an aluminum shield in so the dark chips don't get baked, but I think it will work, and all with very common parts. I'm trying to keep it under 20a, but I do have a 30a circuit in my garage, so if I need more amps it's inconvenient but doable.

Most quality SSRs (not fotek) have two ratings. One amp rating for no heat sink, and another with a large heat sink. My carlo gavazzi SSR is rated at 5amp no sink and 25a with a large heat sink. If there's only one rating on the SSR it's usually for the large heat sink.

Incidentally, the carlo gavazzi SSR did fail in it's prior application, I opened it up, and there was a bad solder joint on a small transistor.

[eccerr0r]

Re: Sending PWM through a relay

Yeah, I was wondering about that, I suspect the SSR's I pulled from dead devices are of low current rating and thus do not have heatsinks. But a 1KW+ heater in a reflow oven, unless you're running at 240V+, at 120VAC the IR and VI losses in a SSR will be significant (though less than 1% of total system energy usage.)

TRIAC power loss is on the order of 1.1*I + gate current which is usually minimal
MOSFET power loss is on the order of 0.7*I+0.02*I² (Note there's an intercept point between mosfet and triac depending on current...)
Mechanical relay power loss is on the order of 0.02*I² plus coil power loss.
Of course these constants dependent on devices used.

[redwire]

Re: Sending PWM through a relay

A disadvantage of SSR's is they dissipate about 1W per amp, which is a lot of heat to get rid of. A mechanical relay wastes much less heat, no heatsink required.

Either way, use low frequency PWM with a carrier frequency in the seconds. Seconds on/seconds off kind of thing to control heat output.

[eccerr0r]

Re: Sending PWM through a relay

I suspect there would be some point a hybrid relay as described would make sense to use as both subunit relays need to carry full load without damage, but such device would add significant cost to the BOM and still no faster than the slowest part - the mechanical portion.

[goontron]

Re: Sending PWM through a relay

Probably because SSRs are "the thing" since it's newer and quieter, but I do wonder about losses. I probably suspect newer SSRs are possibly using MOSFET switches nowadays but even them and those using TRIACs probably have significant VI and IR losses where mechanical relays are pretty consistent until they burn out.

I would think using an SSR to make the connection, pull in the mechanical contactor, then dropping out the SSR would still prevent the losses and save the life of the contactor. Inverse for disconnection of the circuit of course.

[eccerr0r]

Re: Sending PWM through a relay

Probably because SSRs are "the thing" since it's newer and quieter, but I do wonder about losses. I probably suspect newer SSRs are possibly using MOSFET switches nowadays but even them and those using TRIACs probably have significant VI and IR losses where mechanical relays are pretty consistent until they burn out.

[clearchris]

Re: Sending PWM through a relay

redwire: Well, we do PID control at least on 3d printers with heaters.

You could be right about only needing mechanical relays to control the oven, but I haven't seen a single person implement that, so I'm hesitant to buck the trend. Everyone uses SSR for some reason.

And yeah, I don't need super long life on my reflow oven, if I use it 100 times, that's going to be a lot.

[redwire]

Re: Sending PWM through a relay

You can't use high frequency PWM to control a heater, that is kind of a notion from the Arduino crowd.
This is because the heater has a big thermal mass and slow response time, it doesn't react to say 1kHz. If a mains SSR is zero-cross it needs up to 8.67msec to turn on/off anyway which limits your PWM carrier frequency.

The common practice is (including using a mechanical relay) to turn on the heater for a few seconds, then off for a few seconds. You don't need high speed switching on/off. It can also be many AC cycles on, many AC cycles off I have seen from some PID controllers.

Long life is to have an SSR in series with mech relays, and first turn on the mech relays then say 50msec later turn on the SSR. Then turn off SSR then mech relay. This way a mech relay never switches any current and lasts a long time. But who needs over 100,000 cycle life on their reflow oven

[eccerr0r]

Re: Sending PWM through a relay

People even send RF through relay contacts...

[clearchris]

Re: Sending PWM through a relay

Good point on the contact bounce issue. Fortunately, I can configure the magnetic relays, then start up the PMW through the SSR, sending the power through the closed magnetic relay contacts.

I'm not planning on directly tracking the PWM in any manner aside from a k-type thermocouple to measure the resulting temperature of air.

[goontron]

Re: Sending PWM through a relay

Exactly, your verbiage is much better. I'm planning on sending PWM through the contacts of the three magnetic relays, not the coils. The coils couldn't handle such fast switching, we are probably talking about multiple hz.

Much clearer!


Some more notes to make:


Contact "Bounce" will mess up anything that "tracks" the PWM signal. For a modulated load, such as a heating element, this is not a concern. Usually only when timing is involved, so this doesn't apply if you are switching load(s) that are being fed PWM modulated power, however for SIGNALS it does.



The best time to switch on or off the relays is when the PWM is low/off as it will eliminate arcing of the contacts on the mechanical relay

[clearchris]

Re: Sending PWM through a relay

Exactly, your verbiage is much better. I'm planning on sending PWM through the contacts of the three magnetic relays, not the coils. The coils couldn't handle such fast switching, we are probably talking about multiple hz. I think I have a zero-crossing SSR, so that would be better for the contacts also.

[goontron]

Re: Sending PWM through a relay

Im not sure what you mean. Through the contacts or coil?

Through the contacts is fine. As a note, best practice would say to switch off at a zero crossing of the PWM circuit or AC line if you are using the PWM to modulate an AC signal.

Through the coil isn't the best idea, as it can lead to arcing of the contacts

[clearchris]

Long story, but some of my parts for a reflow oven have arrived, but I think I can get it done with my remaining parts. I know a standard relay can't directly switch a PWM signal, but if I have one SSR controlling the PWM signal, and three magnetic relay downstream of that to control three seperate heaters, there shouldn't be a problem with the PWM flowing through the magetic relays, right?

Said another way, the SSR will handle the PID/pwm and the three relays will be turned on or off based on how fast the heat needs to rise in that section of the heat profile.

To my understanding, arcing of magnetic relays is caused when the contacts are close, but not touching. Would sending a PWM signal through a closed relay cause any issues?