Electronics > Projects, Designs, and Technical Stuff
Normally closed switch to discharge capacitor
aheid:
--- Quote from: SiliconWizard on January 23, 2019, 06:14:57 pm ---
--- Quote from: aheid on January 23, 2019, 04:19:12 pm ---P-MOSFET between resistor and GND, gate pulled down to GND via resistor, apply 3.3V to gate to shut it off?
--- End quote ---
The issue with this simple approach is that it will never be able to discharge the capacitor completely. Once the voltage across the capacitor gets below the pmos threshold, the transistor will stop conducting... Or did I miss anything?
--- End quote ---
Nope, just a brainfart :=\ N-MOSFET with gate pulled up, driving it to ground should work tho.
SiliconWizard:
--- Quote from: aheid on January 23, 2019, 06:32:32 pm ---
--- Quote from: SiliconWizard on January 23, 2019, 06:14:57 pm ---
--- Quote from: aheid on January 23, 2019, 04:19:12 pm ---P-MOSFET between resistor and GND, gate pulled down to GND via resistor, apply 3.3V to gate to shut it off?
--- End quote ---
The issue with this simple approach is that it will never be able to discharge the capacitor completely. Once the voltage across the capacitor gets below the pmos threshold, the transistor will stop conducting... Or did I miss anything?
--- End quote ---
Nope, just a brainfart :=\ N-MOSFET with gate pulled up, driving it to ground should work tho.
--- End quote ---
Well, it was still an interesting idea. If the OP is OK with the capacitor having a residual charge, they can select an appropriate PMOS so that the end voltage would be around 1V or lower, and would then discharge further due to leakage currents. Guess that's what we call a compromise.
I'm not sure I see the thing with an NMOS though? It would need a sufficient voltage to conduct, but I was under the impression that the OP wanted the circuit to shunt the capacitor while there was no additional supply available at all. If there is, then it's no problem indeed. Or maybe you have something in mind I didn't get, if so, you can elaborate.
A depletion-mode NMOS would work (there are not many available as discrete transistors, but you can find some). A probably adequate one would be that: https://www.digikey.com/product-detail/en/microchip-technology/LND01K1-G/LND01K1-GCT-ND/4918738
It would require a negative voltage to block it though (as mentioned above). Not that hard to do, but it's definitely not a single-component solution.
Moriambar:
Wow, as usual I oversimplified the problem. Thank you, I think that I'll try the N-FET (either depletion or enhancement) in the weekend. It's true that I do not have a negative rail, but I can engineer one.
Also yes: I have to completely discharge the cap.
Cheers!
SiliconWizard:
Below is an illustration. The part on the left is just to simulate the charging phase of the capacitor. The transistor chosen there is just because I could easily get ahold of its Spice model (it's a power transistor, not needed for your application unless you need to discharge the capacitor at a very high current). On the right there's a basic negative voltage generator that should work fine here. The input signal should be an oscillator. If it oscillates, a negative voltage is generated (enough to block the transistor, around -2V); when it stops oscillating, the voltage quickly rises back to zero. If you're using an MCU, it will just be a matter of generating a clock on the controlling IO. Otherwise you can add a small oscillator, for instance with an inverter (schmitt trigger input), a resistor and a capacitor. No need for anything accurate here.
All this requires a few components but with the benefit of drawing little current. You could have used a normally-closed relay instead, but even a small reed relay will draw a lot more current than this and most will require a 5V supply, a transistor and a freewheeling diode, so...
Navigation
[0] Message Index
[*] Previous page
Go to full version