0
\$\begingroup\$

Good day. The question I have is regarding a device the function of which my employer asked not to discuss in too much detail. But I believe the engineering problem itself is general enough and I will try to give as much engineering data as I can without disclosing any commercially sensitive info. The device I am designing at the moment has to switch the following waveform on or off:

enter image description here

This is a high impedance differential signal (10kOhm) with voltage of 72 Vpp, that can reach 400 Vpp. The load for this signal is also high impedance (I am using a 1 MOhm resistor as a model). I am trying to achieve a switch-like functionality in my circuit that will be controlled by a microcontroller. In one state it will pass the signal to the load, in another it will block the signal.

I was thinking about different approaches from primitive to more complex:

  1. Mechanical relays - Pros: easy to operate. Cons: bit expensive at the voltage specification, have a clicking sound that might not be suitable for the product, wear out
  2. SSR - Very expensive at the specified voltage
  3. Optocouplers - Pros: also easy to operate, cheap, solid state. Cons: usually has a low reverse voltage limit
  4. Discrete transistors (BJT, FET)- This might be a solution but most of the components I've looked at also have a low reverse voltage breakdown. Maybe I'm missing some specific type?

I have experimented a bit with optocouplers in a simulator as well as on a breadboard.

Here's an LTSpice setup with an MOC205 (datasheet):

enter image description here

According to the simulation the circuit works just as I intend it to:

Voltage across the load with 5 volts at the MOC205 input:

enter image description here

Voltage across the load with 0 volts at the MOC205 input:

enter image description here

I couldn't get my hands on MOC205 in the current state of lockdown, so I used K1010 (datasheet) that I had on hand.

The result was quite a bit different. With 5 Volts at the input the signal was identical to the first picture in the question, so it was completely passed through. With 0 volts at the input:

enter image description here

I guess this result should be expected due to the nature of the transistor in the optocoupler. Both MOC205 and K1010 list their reverse voltages to be 6V, so I'm not sure I can trust the result of the MOC205 simulation either. So the question is: are there any optocouplers, or other devices that would be able to pass/block such high voltages? Or maybe I should change the topology? I thought of connecting two optocouplers back to back but the result wasn't perfect either (still was passing some negative voltage pulses), plus it seems like there should be a more elegant solution. Or maybe I should ditch the idea of optocouplers alltogether and work with a different device? Any help would be highly appreciated!

EDIT from the comments:

The rise time of the signal is 500ns.

The output impedance of the device is about 10k, which I guess is not too high an impedance.

Load is not returned to the ground but I guess, I can ground one of the differential signals, grounding the voltage.

If 90% of the signal goes through in the on-state, I would call it acceptable for the application. Inversely of 10% goes through in the off-state, that is also acceptable.

\$\endgroup\$
  • \$\begingroup\$ How fast to operate? What maximum signal loss when on. What minimum signal loss when off? Is load returned to ground (ie could a shunt element improve isolation)? What is source impedance? Although rep rate is about 10mS, what is risetime/bandwidth of signal? \$\endgroup\$ – Neil_UK Apr 13 at 12:44
  • \$\begingroup\$ Voltage is highish, but, two MOSFETs in series opposed polarity may meet the need. Sources joined, gates joined. Drive gate to source to turn on and short gate to source to turn off. A true bipolar switch. drive need to be a floating voltage which can have its challenges. Nei;s questions should be addressed - preferably as part of the question. \$\endgroup\$ – Russell McMahon Apr 13 at 12:52
  • \$\begingroup\$ Right, should've included those in the question: 500 ns rise time. I haven't specified any specific signal loss margins, but let's say if 10% of the signal is lost when on we call it good enough. I'm not sure I understand your question about a shunt element. I will do some measurements and estimate the source impedance. \$\endgroup\$ – Anthropomorphous Dodecahedron Apr 13 at 12:55
  • \$\begingroup\$ @RussellMcMahon Thanks for the suggestion, will edit the question with new data. \$\endgroup\$ – Anthropomorphous Dodecahedron Apr 13 at 12:56
  • \$\begingroup\$ Inverse series connected optocouplers is not a terrible idea if it works for you. Has less problems with driving due to optical drive coupling. \$\endgroup\$ – Russell McMahon Apr 13 at 12:59
1
\$\begingroup\$

OptoMOS is somewhere between an optocoupler and a SSR. It is a packaged version of what Russel describes. Here is what Digi-Key has in stock that are in-stock, active part numbers rated over 100 V.

https://www.digikey.com/products/en/relays/solid-state-relays/183?FV=-8%7C183%2C41%7C307140%2C41%7C307321%2C1989%7C0%2Cyr0V+%7E+1000V%7C2133%2Cyr0V+%7E+230V%7C2133%2Cyr0V+%7E+250V%7C2133%2Cyr0V+%7E+280V%7C2133%2Cyr0V+%7E+300V%7C2133%2Cyr0V+%7E+350V%7C2133%2Cyr0V+%7E+400V%7C2133%2Cyr0V+%7E+500V%7C2133%2Cyr0V+%7E+600V%7C2133%2Cyr0V+%7E+800V%7C2133&quantity=0&ColumnSort=0&page=1&stock=1&k=optomos&pageSize=500&pkeyword=optomos

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Thanks for the suggestion! This might be the perfect solution as far as I understand. \$\endgroup\$ – Anthropomorphous Dodecahedron Apr 13 at 14:03
  • \$\begingroup\$ What I don't quite understand is that it still lists a max reverse voltage of 5V, even though the part seems to be internally symmetrical. Any suggestion as to what does it refer to? \$\endgroup\$ – Anthropomorphous Dodecahedron Apr 13 at 15:06
  • \$\begingroup\$ @AnthropomorphousDodecahedron is that max reverse spec for the input or the output? \$\endgroup\$ – brhans Apr 13 at 15:20
  • \$\begingroup\$ Sorry should've linked the datasheet. The input reverse voltage that is. ixysic.com/home/pdfs.nsf/www/CPC1025N.pdf/$file/CPC1025N.pdf \$\endgroup\$ – Anthropomorphous Dodecahedron Apr 13 at 21:57

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.