Electronics > Projects, Designs, and Technical Stuff
Storage cap. before or after regulator
JJalling:
Hello all,
I'm making a small circuit for an electricity meter. I want to connect a ESP8266 to the meter, however the meter specs states 4V+- 250mV @75mA.
So, I need to lower the voltage to 3.3V for the ESP8266, but also need enough power to actually transmit something (the datasheet states 170mA when transmitting).
Where would you put a bunch of "big" capacitors? At the input (~4V) before the regulator, or after the regulator, at the 3.3V rail?
And what about the current limit? Just a series resistor or something more "clever"?
Thanks for all your input,
BR Jonas
pardo-bsso:
Which voltage regulator are you using to power the ESP8266?
The datasheet should specify a range of values (and esr) that would give a correct operation without risk of it running into oscillation.
Most of the ams1117-3.3 clones that I have here seem to be happy with a cheap 10uF electrolytic and the esp works fine.
A cap before, sure.
A cap after, read the docs.
JJalling:
My first plan was to use a zener diode and a resistor, so I don't really see any problem with oscillation.
You mention a 10uF capacitor for powering the ESP8266 - I don't really see how that will give me enough power to transmit a full ethernet frame. Remember that my max current draw at the "high" side of the psu is 75mA.
So maybe you can elaborate a bit on your suggested solution.
BR Jonas
Siwastaja:
For wifi, the burst of current are so long it's impractical to store the energy required in capacitors.
This means, the regulator and the input power supply must be able to deal with this peak current.
In your case, it seems your power source is unable to supply the peak current you need. If the average current is OK, you can try to add LARGE capacitors. Try with some 10000uF first; the solution will be bulky. Capacitors go before the regulator so that you can maximize the voltage swing on the capacitors; the stored energy depends on the delta V. Note that the power supply may be stressed in this use.
For the same reason, use the lowest drop out regulator you can find not to waste any precious voltage. 1117 is total and utter shit from the history books, oscillates easily with wrong type of output capacitor, and the specified HUGE 1V dropout makes it a non-starter: you require max 0.7V dropout, and that has zero tolerance against input variations. Look for regulators with drop-out less than 0.2-0.3V or so. Yeah, 1V drop-out was "low drop-out" in 1980's.
mariush:
You don't need to limit the current, the linear regulator will produce as much current as the chip needs.
You will need a capacitor before and (often) after the regulator.
When your ESP chip suddenly starts transmitting and consumes all of the sudden 150+ mA of current, the linear regulator has to take a lot of energy from the input and output that desired stable voltage. If the wires on the input are long, they can behave as inductors, resisting the change of current, so the input voltage can slightly go down for a bit. The wires also have some resistance, so again the voltage drop on the wires will increase as the current increases. So a capacitor on the input can act like a buffer of energy to minimize that fluctuation of input voltage when power consumption of the chip changes suddenly.
The actual value on the input is not very important, typically as little as 10uF is recommended if the incoming voltage is already somewhat regulated. If you have an AC transformer and you rectify the AC voltage using a diode or some bridge rectifier, a bigger capacitor may be required.
If the voltage is somewhat regulated, like i said, something like 10uF is probably enough, but there's no reason why you can't use bigger values like 47uF or 100uF or 220uF ... it's just often pointless to use big capacitance, not giving you any benefit.
For output capacitance, it depends on the regulator you use. You'll have to pay attention to datasheet, because a lot of regulators will actually require that capacitor to have some technical properties.
For example, the 1117 regulators are made by lots of companies, it's a generic chip ... but lots of datasheets will mention the chip needs a capacitor on the output and says that the ESR of a capacitor on the output should be between 0.1 ohm and 1 ohm, while the actual capacity is not important.
So that's where it's important to realize that ceramic capacitors have very low esr, under 0.1 ohm, so you shouldn't use a 1uF ceramic capacitor on the output.
A lot of datasheets will use 10uF or 22uF on example circuits and don't mention that "esr must be between 0.1 ohm and 1 ohm" but say "a tantalum capacitor is recommended" - basically you're supposed to know that tantalum capacitors typically have esr higher than 0.1 ohm
Electrolytic capacitors in the 1uF .. 100 uF will typically have esr above 0.1 ohm, some super low esr high end series may have less than 0.1 ohm esr above 33-47uF.
Navigation
[0] Message Index
[#] Next page
Go to full version