AC MOSFET Switch

I'm trying to use MOSFETs to allow or block 450kHz sine wave. Here's the test circuit:

At low frequency, MOSFETS are totally able to switch on and off. But when the frequency becomes higher at 450kHz, it seems MOSFETS are not able to block sine wave anymore, no matter if apply or not 5V-signal at OPT_SIG: oscilloscope always shows sine wave on load (R1).

I tried with other parts such as FDP020N06B with very low rds but I get the same result.

I've also found older MOSFETs STP5NB40FP and using them, RMS voltage on load is almost 0, when signal is not applied. I was thinking the high frequency AC to be passing through MOSFET capacitance Cds, but both FDP020N06B (Coss = 3840pF) and IPD60R180P7S (Coss = 19pF) behave the same way.

STP5NB40FP MOSFET has lower continuous drain current and much higher Rds resistance compared to the previous ones. I'd like to choose MOSFET with Rds as low as possible but work as well. So it'll be another characteristic I'm not taking into account. Could you please help me?

I got some screens from my oscilloscope as a clear evidence of that. Blue track is the input sine, yellow track is the signal on the load. You can see the big differences between them. R9 was replaced with a short. SCREEN

Datasheets: IPD60R180P7S, FDP020N06B, STP5NB40FP

• Look at the voltage across R9. I think you're going to find a lot of 450 kHz there, due to Cgd. You're going to need to actively clamp the gates to ground when the transistors are supposed to be off. – Dave Tweed Mar 25 '20 at 11:09
• What happens, if you short circuit the R9 (with disabled 12V_iso, of course)? – Marko Buršič Mar 25 '20 at 11:21
• Please provide a link to the pdf datasheet on the STM website for the STP5NB40FP – Andy aka Mar 25 '20 at 11:36
• I did very similar almost 20 years ago. I used a floating supply for the gate drive and then added a shorting FET at the output to further increase isolation. I was switching maybe 100 Watts in an inductive power transfer circuit and applying on/off control to create signalling slots for comms over the loop. The power levels were such that leakage through the MOSFETs was too high. The extra clamp fixed it nicely. – Russell McMahon Mar 25 '20 at 12:21
• @DaveTweed Shorting R9 to ground makes no difference. Sine wave is always on the load. – kristian_ Mar 25 '20 at 12:28

Well, it's the right characteristic you are considering but you need to dig into the detail of the what you say here: -

IPD60R180P7S (Coss = 19pF)

The truth is, if you compare apples with apples, the $$\C_{OSS}\$$ of this device is more like 250 pF. The apples are the drain source voltage. See graph below: -

At an applied drain source voltage of 25 volts, the IPD60R180P7S has a capacitance of 250 pF (not 19 pF as said on the salesman designed front sheet).

As for the STP5NB40FP (data sheet link kindly supplied by @Hearth), the $$\C_{OSS}\$$ is 72 pF when $$\V_{DS}\$$ is 25 volts: -

In other words, the STM device is better hence it produces less residual output voltage when the SSR (solid state relay) is off.

• Ok, the explanation is clear. But in these graphs, VDS is assumed to be continuous and not time-varying. Am I wrong? What could it be the real Coss with a 100V RMS sine wave on the drain? And as well, I'm not applying sine wave directly to its VDS. I made two tests with both STP and IPD part: 1.applying 87V RMS sine wave, 2. applying 19V RMS. Blue track is input sine, yellow track is output signal on the load. test. Having to choose the right MOSFETs, shoud I find a part with Coss as low and constant over VDS as possible? Thanks – kristian_ Mar 25 '20 at 22:00
• VDS is constant but modulated with a small test AC signal in order to measure capacitance. A big AC signal means a significant change in capacitance over each cycle but, the message is clear. Choose the MOSFET with the lowest capacitance in the same signal conditions but, choose a MOSFET that can handle the maximum AC voltage you wish to supply. – Andy aka Mar 25 '20 at 22:20
• I’m not going to analyse remote screenshots of waveforms. I’ve made my answer. If you don’t understand it raise a comment. If you want to add a whole pile of other stuff, raise a brand new question. – Andy aka Mar 25 '20 at 22:22
• I'll try to find other MOSFETs with the highlighted characteristics. The screen test I uploaded was just a clear evidence of the results i found. I'll add it in the main question – kristian_ Mar 25 '20 at 23:43
• Please don’t move the goalposts by changing the focus of the original question. If something hasn’t been addressed from the original question in my answer, please focus on that. This is a question and answer site and not a forum or training school. Hope you understand. – Andy aka Mar 25 '20 at 23:50