Electronics > Projects, Designs, and Technical Stuff
Battery voltage to 5V regulation - overcurrent shutdown protection: working
HendriXML:
For my playable modules:
playable-electronic-modules
I've designed a voltage regulator schematic, which should drop 5 x AA 1.25V rechargeable batteries to 5V.
I've set following goals:
* No regulation IC -> fun factor
* Less than 10 mA loss
* Can regulate even with a low voltage drop
* Signals voltage level too low (LED -> off)
* Should be an improvement of this one: only-1x-gain-op-amp-regulated-power-supply-very-low-noise. Faster handling of sudden 1 amp draw. More stable, maybe even without an endcap.
To reach the last goal, I've ordered a high(er) current op amp (LM7332), which can drive capacitive loads and has (no?) less input capacitance. Also will be using a Mosfet with less gate capacitance.
The reference voltage is also lower and more stable. The output offsetting of the error amplification is done in a different way (thx to xavier60).
Feel free to comment on the schematic. I'm waiting for the components to arrive and will probably start with redoing some characterization of the new Mosfet when they arrive (Automatic characterization via power supply).
David Hess:
U1B is not doing anything useful.
Current limiting would be a good feature to add.
I like the concept of floating a negative regulator so that an n-channel instead of p-channel MOSFET can be used.
For some inspiration, check out Figure 6 on page 14 of the LT1366 datasheet.
HendriXML:
--- Quote from: David Hess on March 19, 2019, 04:01:07 pm ---U1B is not doing anything useful.
Current limiting would be a good feature to add.
I like the concept of floating a negative regulator so that an n-channel instead of p-channel MOSFET can be used.
For some inspiration, check out Figure 6 on page 14 of the LT1366 datasheet.
--- End quote ---
Thanks for the schematic, nice to see the similarities!
I used the second op amp for buffering because I did not want the negative fb to affect the reference voltage. Or having the need for more current through the “zener”. The LT1366 schematic has even stronger neg. fb / less gain than mine. That is what I hyphothized about in my previous experiment, but couldn’t make to perform well. I’ll certainly experiment on different values in my circuit later on. My schematics uses lower decoupling and fb resistor values and higher fb capacitance. (And will be sourcing/sinking more from the ref voltage). But the opamp also can drive with more current.
I originally wanted to include a “electronic fuse” functionality, and thus shutting the supply down. But after experimenting with it I’ve let the idea go. Having capacitive loads (or an end cap) would trigger it.
For now I’ll try the “more controlled version” of the circuit and if this one turns to to be functional I can try to see if there’re possibilities to skip the buffered vref. (But I probably also need to create some RC V vs time graphs to get a more trained “sense” of the buffering capabilities of a capacitor)
HendriXML:
I’ve been thinking a little more about the usefulness of U1B. If I enlarge R5 and R6 to 2K, R1 should be able to sustain a stable ref in the most extreme situation without any buffering. In that case there will be less current through the zener IC, but that IC has a better stability than a normal zenerdiode. Having a capacitor will improve stability even more.
I’ve also been thinking about some current control, but that will somehow intervene with the voltage regulation. Making that slightly less optimal, mostly in respect to (expected) regulating speed. Because I want to see how this circuit compares to the previous one, I’d like it to be in the most optimal (and simple) form.
HendriXML:
I updated the schematic without the Vref buffering. I will try to show various examples how the ref voltage will act in time depending on the values of R5 and R6 and C1.
Because there's a complex interaction between:
R1, R5, R6 and C4, C1
I'll revert to simulating with a self made script. It will be about developing techniques to tackle these calculations without using complex math.
Navigation
[0] Message Index
[#] Next page
Go to full version