I want to pick a part for a "master switch" that closes the supply line (12V, about 7A) for a handful of regulators. This master switch should be controlled by a 5V logic pin. I don't know if I should choose a relay over a power FET.
When a power FET fails, it can short source and drain. When I use a relay plus a driving resistor for the coil, I can get the same problem if the driving transistor fails, can't I?
Hi @Christoph, I'd say either could potentially fail that way, plus the relay contacts could fuse shut. I've never done much with very high reliability design but what's the application, is it something where power not being removed has really dire consequences?
The regulators "behind" the master switch supply a number of heating resistors. If the MCU that controls the switch finds that it is receiving bad commands, it should switch them off. The MCU plus heating resistors are part of a control system in a scientific experiment, and all is hand-crafted. We simply don't want it to burn. Nobody would be harmed.
a thermal fuse would not do the trick. High temperatures are desired, but if the logic fails there should not by any heating at all even if the temperature is low
The outer control loop will be designed by students. I'm responsible for making the test rig survive that. The flow of information is Outer control loop (students) -> MCU (my realm) -> switch (my problem) -> heaters. Sensors then go back to the students' control system.
@Christoph that bit makes it more interesting, how would you tell the difference between a "bad command" and a simple logic error that sends valid commands that tell the heater to always be on? If I was designing it I might think about independant low and high side switching and maybe a small micro each end to verify valid commands were received on both channels in a timely fashion but that may not cover all situations.
A bad command would be anything that doesn't have the correct format. There's also a timeout that expires when the students' software locks up. I might have to add a temperature switch for that case that they deliberately overheat the device, but that's (to my experience) less probable than a lockup, or bad formatting.
I'm currently biased towards the relay, as it doesn't matter if it's high-side or low-side, and I can drive it with a small transistor taht I have around. a power mosfet requires more care while selecting
@Christoph Two relays or transistors working well below its rated max in series. At power up do an automatic individual test of the two devices. Then switch them on / off with a slight time in between. On the other hand it sounds like over-engineering if the devices are properly rated.
Could someone here explain to me how a mis-set impedance on an oscilloscope increases the voltage by a factor of two? I forgot the circuit, it's driving me crazy.
@Anthony With 50-ohm scope impedance and a 50-ohm source, you have a voltage divider. When you change scope impedance to 10 Megohm or whatever, you have a different voltage divider.
Proposed Q&A site for engineers, Professional or Healthcare startups in Healthcare/Medical device/instruments domain. Relevance to Diagnostics, Assistance tools, Mobile based interventions, etc.
Currently in definition.
Anonymous
@user33944 don't want to offend but I just don't think it's good judgement to have a site like that... something something liability. The only real good use for it is for HIPAA and I'm sure there are sites in the network already helping out with that (SF, for example)
