| Electronics > Projects, Designs, and Technical Stuff |
| Battery voltage to 5V regulation - overcurrent shutdown protection: working |
| << < (4/4) |
| Psi:
--- Quote from: HendriXML on April 01, 2019, 11:35:24 pm ---Could be useful as a drop voltage regulator in a variable power supply. With 6 relays a drop of more than 38 V could be achieved (64x0.6V diodes). :-+ But I hope my circuit belongs to a different category... --- End quote --- Yep, it does, this thread just made me remember it. ;D |
| HendriXML:
--- Quote from: HendriXML on April 01, 2019, 11:35:24 pm --- --- Quote from: Psi on April 01, 2019, 11:12:55 pm ---A fun, but mostly pointless, voltage regulator to build is a system that switches in/out different number of diodes in series based on control from ADC+discrete digital logic (or a MCU). Use a low voltage drop diodes like 1N5817 or better. Use a binary approach where each bit in a byte adds different diode numbers (or shorts them to remove from series string). eg bit0 might add 1 diode, bit1 might add 2 diodes so if bit 1 + bit 2 are both set then 3 diodes are in series. Also include a way to short input to output in case no diode drop is needed. Then you can control the output voltage by selecting the binary number from ADC/MCU. It's a pretty pointless and slow voltage regulator but is kinda cool and it does have the advantage of no drop-out voltage. --- End quote --- Could be useful as a drop voltage regulator in a variable power supply. With 6 relays a drop of more than 38 V could be achieved (64x0.6V diodes). :-+ But I hope my circuit belongs to a different category... --- End quote --- :phew: |
| HendriXML:
I (partially) rebuild the circuit and tested the overcurrent protection and it works fine! I had to replace the voltage reference IC with a normal zener diode. The IC isn't suitable for very low (near zero) currents because it needs some current for its internal circuitry. Because the overcurrent regulation op amp sinks very low currents from the zener diode, the zener voltage is not the voltage at which the protection activates. It is a combination of the amplification (RV1) and the how the voltage over R9 drops. (Which can be seen as "an amplification" regulated by the zenerdiode). If these two combined result in an amplification of more than 1, the positive feedback collapses. Because it is dependent of the zener curve instead of just the one value spec and thus cannot be calculated easily, the protection needs to be set by a potentiometer RV1. This can be done by drawing the shutdown current and then tuning RV1 until the shutdown kicks in. C5 is added so that the circuit doesn't shutdown when a loadcapacitor of 1000 uF needs to be charged (my modules use thoose). The optimal value of it needs to be determined yet. If a shutdown occurs, the circuit needs to be disconnected from the input for about 20 seconds. This is the duration that is needed to discharge C4 properly. If not discharged, the startup current is to high and the circuit re-enters the shutdown mode right away. I'm pleased with the results. One of the design targets was to make the shutdown fast and as sensitive as possible. |
| HendriXML:
For my application the circuit is fine, but I’m a bit concerned about two things: * temperature influences on the zenercurve * when after a shutdown the voltage of the input drops to about 3V, the shutdown circuit becomes ineffective and “normal” high current operation continues. When having half full batteries as an input this is not an issue because the current is then very limited anyways. But it should be mentioned. Having a diode for the zener effect might lower the effective voltage of the shutdown state to the point that the mosfet won’t conduct anymore. The last point is also the reason the shutdown must be faster than a drop of the input voltage. C5 can be increased beyond 100 nF, but more might lower the input voltage to much in case of a shortcircuit at the output. However in case of C5=100 nF, a short with a 1000 uF capacitor still activates a shutdown. But that’s the thing when implementing a electronic fuse: capacitors draw a lot of current! So I think I will lower the buffer capacitors of the modules which will use the supply instead :-+. However the product has reached the moment that altering stuff means desoldering stacked components :o, so experimenting lost its appeal. |
| Navigation |
| Message Index |
| Previous page |