PWM and Relay

Even reed relays aren't going to work above 100Hz, much less 100kHz. If you want to lock the micro out of burning up your transistors, use logic

-- it can be done easily with 7400 series logic, and it'll be plenty fast enough.

I'm not sure about that particular chip, but some of them disallow inputs that'll crowbar the power to ground -- many even have delays in them for shoot-through protection. So your protection circuit may well be redundant.

I was thinking a J-K flip flop and maybe a couple FET's, depending on the load.

Yeah, flipping a mechanical relay at 100kHz would turn it into a radio station!

What do you mean by that? Relays have a short period of contact bounce, but no fixed frequency. Are you going to use the relay to pass some high current, or is it just signal current at logic level? Solid state 'analog switch' components are best for the latter.

Do you mean the relay CONTACTS (probably yes), or the relay COIL? The contacts may have some stray capacitance when open, if a few picofarads matters to you.

Regarding #1, the switching frequency of the relay will be whatever your program into the micro to switch the relay. Within limits. For example, you might program the micro to switch the relay on and off two times per second (2 Hz), or 10 times per second (10 Hz) or whatever. The maximum number of times you can switch a relay on/off in one second is really small. An electromechanical relay takes a (relatively) looooong time to transfer the contacts, and then another (relatively) looong time for the contacts to stop bouncing. As a _guess_, if you allow 50 mS for each transition (off to on or on to off) you probably don't need to worry. That's 10 Hz. As you go higher in frequency, the need to look at the datasheet and understand the particular relay's switching characteristics increases. You'll run into greater difficulty, the higher relay switching frequency you go.

You'll need to understand the effect of bouncing relay contacts on the rest of your circuit, and you'll need to understand the effect on the rest of the circuit when the relay contacts are connected to nothing during the transition time. For example, what happens to the HIP chip and/or the NMOS transistors in your circuit when the inputs to the HIP chip are floating as the contacts transfer from the closed side to the open side (or vice versa)?

I think what you are asking with #2 is whether or not the relay _contacts_ can handle a 100KHz signal. Yes, they can.

Ed

No it wouldn't. The mechanical goodies inside the relay can't move that fast. It would either switch once or not but it wouldn't toggle.

You might get one to do 100Hz. Above that you're on pushing your luck.

I think that was a 'shoosh' moment mook.