AC Solid State Relay

[kb1vyi]

I'm experiencing trouble in both my model and in my measurements of a real circuit of a solid state relay. I'm attempting to switch a 200kHz high power source. However it seems as though I cannot get the MOSFET to remain off during the zero crossing event. I'm also using a photovoltaic isolator at the input of the SSR for isolation.

My initial suspicions was the recovery diode characteristics and I attempted to optimize it which did help some. I also attempted to add some ultrafast recovery diodes across the mosfets as well to no positive results in my model. I tried a series of other small things I theorized but had no luck. Does anyone know a confirmed cause or have a solution to this issue?

 

[barry]

Maybe if you told us WHAT SSR you were using we'd be able to help. And, you do realize that most SSRs are intrinsically opto-isolated, right? And 200KHz sounds kind of high to me to expect zero-crossing. And you may be experiencing some dv/dt issue. And you don't even tell us what voltage you're switching, 10 Megavolts?

 

[kb1vyi]

My apologies. It appears my schematic and waveforms did not show up for some reason! Let me try this again. It is a custom SSR circuit utilizing N Channel MOSFETS so I could pick lower output capacitance MOSFETS so that way the impedance is higher when the SSR is off at higher frequencies. The output plot demonstrates probing at the 50 Ohm load with the SSR is in the off state.

 

[dick_freebird]

Power MOSFET gate capacitance and weak control drive
probably just let the HF signal blow through, drain-to-
source-to-drain.

You might look at signal source common mode and
amplitude, availability of any auxiliary supplies, and see if
some other switch-FET topology could serve.

And what's wrong with a regular ol' relay, if the rate of
commutation is low? Getting gate impedance low enough
to stand off 200kHz (though probably the rise / fall time
is the real problem, high enough frequency content to
blow through the Cgd to override / fight the gate drive )
might end up putting you on par with coil current in a
MOS implementation. A latching relay don't need no
steenkin' coil current, just to stay on or off).

 

[barry]

That's not much better. What are the MOSFETS? What is the "Isolator", a magical infinite current driver? dv/dt=2*pi*200000*100=125V/uS, that's pretty high.

 

[FvM]

Some info is in the schematic, e.g. 100 V sine voltage, but a lot is missing, MOSFET and photovoltaic driver type.

As a first step to understand the circuit behaviour, you can look at the gate-source voltage in your simulation.

 

[crutschow]

Need to know specific part info.
Otherwise we are just guessing.

 

[Easy peasy]

read the data sheet of the isolated driver ic, it is far too slow for 200kHz

--- Updated Jun 9, 2021 ---

also - you need a very low Z pull down to keep the mosfets off, due to dv/dt induced gate rise ....

 

[kb1vyi]

Hi all, thank you all for your feedback!

dick_freebird: I would use a standard relay but reliability is a great concern!

Barry: It happens with almost any MOSFET and most gate drivers (Some worst than others), but specifically the IPLK60R600 by Infineon. The gate driver is a VOM1271, obviously any of the photovoltaic drivers does not deliver a lot of current. So this really seems to be a design problem rather than a parts selection problem.

FVM: Yes when I looked at it you could definitely see the crossover point is being coupled onto the gate of the FET and is likely what is turning it on briefly. It also seemed to drift a bit.

EasyPeasy - I had similar thoughts and wondered if a totem-pole driver would help deliver the lower impedance I need at the gate. I will give that a shot today.

 

[FvM]

Even a good photovoltaic driver with active pulldown circuit like Vishay VOM1271 has an output impedance in the high kOhm range, effectively useless to prevent parasitic turn-on by fast rising Vds.

--- Updated Jun 9, 2021 ---

You didn't tell about intended switching speed.

 

[kb1vyi]

Even a good photovoltaic driver with active pulldown circuit like Vishay VOM1271 has an output impedance in the high kOhm range, effectively useless to prevent parasitic turn-on by fast rising Vds.

--- Updated Jun 9, 2021 ---

You didn't tell about intended switching speed.

Good news is that adding a PNP gate drive turn off circuit has seemed to remove any of the parasitic coupling. I measured the rise time of the turn on of the SSR and it is sufficient for what I need! There won't be much switching in the way of the SSR.