Depletion MOSFET

Hello, totally beginner here!
I’m looking for a depletion MOSFET, the idea being to put it to a solar panel, so that when it’s dark current would flow from gate to source allowing a battery to power on the Arduino, and vice-versa.
While reading a datasheet of one (attached), i’m confused: why the 0V line is associated with current, shouldn’t it be 0A as it is a depletion MOSFET? It looks like an enhancement curve to me…

Thanks a lot!

Screenshot from 2020-04-13 15-48-47.png

Current doesn't flow from the gate. Current flows between source and drain. The gate-source voltage
controls the channel width which allows current to flow, but the gate is electrically isolated from the
source and drain.

A depletion mode FET conducts when Vgs = 0V, and requires a gate voltage to reduce the drain current.

For an n-channel depletion mode device the gate voltage has to go negative to reduce the drain current.

Depletion mode MOSFETs a specialty, hard to source likely to go obsolete. All JFETs are depletion mode.

The only difference between enhancement and depletion mode is what the gate threshold voltage is - for
enhancement its above 0V (n-channel), for depletion its below 0V. Threshold voltage is the turn-off point.

Thus the curves are the same, except that the gate voltages are shifted.

Yes i meant voltage would be applied to the gate allowing current to go from drain to source.
Thanks for your recap but still i can't what you said with the graph i posted:

A depletion mode FET conducts when Vgs = 0V, and requires a gate voltage to reduce the drain current.

Then why the 5V VGS curve is above the 0V one on the graph?

Depletion mode MOSFETs a specialty, hard to source likely to go obsolete. All JFETs are depletion mode.

I guess there's an alternative then, what is it?

Thanks again

Ah ok so i checked an enhancement graph and actually numbers are negative so there's something i don't get still, but at least it's logical...

They make depletion mode MOSFETs by implanting charges into the gate oxide, which shifts the threshold directly. Naturally a MOS structure is unbiased and enhancement mode.

For power electronics depletion mode devices are useless as they are heavy short-circuits at power on, as
there is no voltage then to turn them off! Since power MOSFETs are 99.9% used in power conversion,
enhancement is the rule.

You haven't explained your scheme for auto-switching the solar panel, normally you can use a couple of diodes
to switch between solar and battery, perhaps an ideal diode is what you are looking for?

Thanks for your answer again, i have uploaded a schematic of what i have planned to do (this is my very first one :)), that’ll probably be clearer.
There would be any intervention from me, just the sun would switch on off the panel. :smiley:
This is a bit simplified of course, “load” will be an arduino and leds.

n-channel or p-channel?

You haven't labelled source or drain.

You've drawn a non-standard symbol - the dotted line suggest enhancement mode. Learn the
right symbols and use them if you want to prevent confusion. Power/discrete MOSFETs have different
symbols from VLSI MOSFETs since they are not symmetric.

You have nothing to prevent battery overcharging.

Basically with a solar panel and a battery you need a proper solar charging controller or your battery will
die through over charging or over-discharge. Battery management isn't this simple.

Hello, finally i think i have found another way (attachment) using an enhancement mosfet, i had to put a capacitor in the schematic instead of a battery but the idea should be ok.

Indeed, i have read about overcharging and battery management, but the battery i plan to buy has a PCM, isn’t that enough?

Thanks!

No, can make even less sense of that circuit.

Where's the solar panel?

Where's the Arduino?

What is the big red dot?

http://tinyurl.com/qwft4os

Maybe that'll be clearer with a direct link.

The power source represents the panel.
The switch is day/night.
The capacitor is the battery.
The red dot, here a led, is the load (Arduino + leds irl).

Ah, like this:

Well yes that's going to gradually switch off the Arduino and LEDs as the sunlight increases. Some hysteresis
might be good to have (large value resistor from fet drain to NPN base could do this).

Nothing to prevent overcharge of battery through... It will be trashed if you don't protect it.
Sizing the battery to last overnight after an overcast day means that it will be 10x overcharged on
a sunny day...

Just use a TP4056 Charging Module.

Hello, i've been reading a lot lastly to improve my understanding, here i come with questions!

MarkT:
Ah, like this:

Well yes that's going to gradually switch off the Arduino and LEDs as the sunlight increases. Some hysteresis
might be good to have (large value resistor from fet drain to NPN base could do this).

Nothing to prevent overcharge of battery through... It will be trashed if you don't protect it.
Sizing the battery to last overnight after an overcast day means that it will be 10x overcharged on
a sunny day...

Yes you got it right about the circuit!
Why do you say "gradually"? I've read that solar panels are almost on/off devices, meaning a slight light is enough for providing the design voltage, only the amount of current varies...
I've read about hysteresis and even though i think i've understood the global concept, i don't understand how it applies here, could you elaborate please?

Now about the battery (replying to Idahowalker too): i plan to buy a LiFePo4 (max 7.2V so i guess it's a 2-cell one, 3000mAh).
So the TP4056 won't work i guess. I have found that BQ24105 could be working. Circuits look pretty complicated though.
One thing which isn't clear is that sometimes i read battery stop charging when its voltage equals the incoming voltage, so why do you say it'll burn?

Also a more global question: is it a good idea to try to build stuff myself (meaning not using IC, only basic components) in order to learn better, or not?

Thanks, i keep getting myself informed!

Solar panels most definitely are not on-off devices. Most logic chips expect their power rails to come up in milliseconds to a correct value from zero. Browning out at half voltage is not a good idea.

Hysteresis uses positive feedback for a snap-on and snap-off behaviour. Feed the output voltage (MOSFET drain)
via a high value resistor to the NPN base - with the right value of resistor you'll get snap action hopefully.

Hi!

I'm just reading the electronics beginner book from Cathleen Shamieh and Gordon McComb, and it is written that "the tension produced by a photovoltaic cell is constant, whatever the quantity of light received. However, the intensity of the current produced depends on luminosity".

I have tried on the simulator to add a resistor where you said, the only effect i notice is that the current on the load is reduced, i may have misunderstood what you say...

Actually that's not the truth as any experiment with a PV cell and a voltmeter would immediate show.

A PV cell is a diode, and light generates a photo-current. The unloaded voltage is the forward voltage of
the diode for the amount of current that would flow into a short-circuit (which is a logarithmic function
given by the diode equation).

Basically the current either flows in your external circuit or just shunts back across the forward biased diode - this sets the forward voltage.

So the voltage increases with more light, but only logarithmically at about 60mV per decade of extra light
intensity. Loading the cell means less current gets shunted across the diode and thus its voltage must drop.

Typical forward voltages (for silicon) in bright sunlight, unloaded, are 0.55V, in bright sunlight loaded 0.5V, overcast
loaded, 0.4 to 0.45V.