Author Options:

MOSFET Deisgn Answered

I just designed (if it works out) a super efficeint Beam Solar engine that works at low voltages. Before I post it I want to ask a few questions to see if I need to revise my design. 1) MOSFETs work on voltage, no current, right? 2) Do MOSFETs have a voltage drop, like biploar transistors have a voltage drop of about 0.6 volts? EDIT: I got a new design that I want you guys to evaluate that doesn't use mosfets. It's incredibly simple and I want to know what you guys think. How it Works: The solar cell energizes the transformer coil while turning on the transistor. The transistor shorts the circuit and then the transistor turns off, then it repeats creating pulsed dc that drives the stepup transformer. I chose this design because there's very little components and there's no voltage drop across the transistor to the transformer. Theoretically this could run on as low as 0.7 volts, which is good for BEAM robots.


i dont understand your topic.


10 years ago

1) Yes; voltage, not current.
2) If you're talking about a minimum gate voltage to turn the transistor on, then the relevant datasheet value is "VGS(thresh)", which is usually about 4.5V for regular mosfets, and maybe 2V for "logic level" mosfets. (the consequence is that mosfets tend to like relatively high voltages, even though nearly zero current is involved.)

For 2) I mean like whenever you put a voltage across it, does the voltage get lowered? like if you put 1 volt through a bipolar transistor, the other side you get .4 volts, does that happen with mosfets (assuming that the transistor is on)?

If I understand your question correctly, there is no "voltage-drop" per se. It is similar to asking what the DC voltage drop across a capacitor: It is (transiently) infinite (or, more correctly equal to the supply voltage) initially, then drops to zero (dependent on the RC time) when fully charged. The gate current is thus, initially very high (just as it is when beginning to charge a capacitor) but drops to nano Amps when in the fully on state. This gate current is actually just a leakage across the insulated junction. The question I have is: have you tried this circuit as written? Generally, MOSFETs turn on at (if I remember correctly) about 4 volts (ie above your supply voltage and do not fully turn-on until 10 volts or more (just like the current dependence of a bipolar: it isn't just on-off). Your circuit depends on creating an oscillator within the solar cell-MOSFET-transformer reactance, and on this oscillator switching your source-drain current. As written, when the gate goes positive, you short the solar cell across the source:drain of the MOSFET, resulting in heat in your MOSFET, not current through your transformer primary (as an inductive reactance recall, it opposes rapid changes in current.

One last comment: Your CONCEPT is excellent:: an efficient charge pump in the voltage/current domain of a solar cell is excellent. Charge pumps become most efficient at high frequencies (at low frequencies you need a big-ole, ie heavy, transformer core; at high frequencies, transformers become much more effecient without requiring a heavy iron core). Charge pumps can have efficiencies of 85-95% at high frequencies, good enough for what you need. The problem then would be to design them to work at the proposed voltage and current. I am looking forward to your latest work on this topic.

The 0.6V drop that bipolar transistors are famous for happens between the base (control signal) and emitter, NOT on the switched current (which goes from collector to emitter.) The Vce voltage drop is typically much lower than 0.6V on bipolar transistors (usually referred to as Vce(sa), and frequently down around 0.2V or so.) In bipolar transistors, the voltage is quantum-physics related, and stays more-or-less constant regardless of current. The similar path on FETs (including MOSFETs) is purely resistive, so the voltage drop DOES vary with current (and is very small for small currents.) This is the commonly quoted Rds(on) parameter. The lower the better, and down around a milli-ohm for good power MOSFETs. (homework assignment: if a power MOSFET has Rds(on) of 10mohms, at what current does it break-even with a bipolar transistor having Vce(sat) of 0.2V ?)


10 years ago

(now commenting on the schematic...) hmm. So it's more-or-less a "Joule-thief" circuit powered by the solar cell. You'll certainly need a resistor in the base connection for the transistor... The Joule thief isn't noted for its efficiency, but this could be an interesting way of getting higher voltages from relatively high-current low-voltage solar cells. (I think the JT requires that the power source deliver relatively substantial currents, at low voltages. In theory, that's a good match for some modern solar cells...)

hmmmm, well 2mA isn't that much current...

Does it work, or is it still all on paper?

it's still on paper, my solar panels should come on dec 28. my second choice is basically to use a minty boost circuit, but I don't want to, because it'd be using an IC.

Posted a new design, if all goes well it'll run as low as 0.7 volts

Coo. How low?

currently my version 1 design works as low as 1.5 volts, ramps up the voltage to about 16, which charges a 150uF, 10 volts capacitor super fast, but as it approaches 9 volts (about 2/3 into the charge cycle) it rapidly discharges it through the motor (or in my case, flexinol). I chose the 16 volt thing so that it charges the capacitor faster (according to a charge graph) and so that I could fit the 6mJ (see my other forum) in a small, 150 uF capacitor (oppose to something like 3300 uF or even 0.1F). My schematic is meant to work with 4 of the solarbotics 1/8" square solar panels, which in series createds 1.9 volts at 2mA. I'll post a schematic soon, I have to work on homework, and then redwraw the schematic on the computer, I love using whiteboards to design schematics, because it's easy to modify. But basically I scrapped the old MOSFET design because I wouldn't have enough voltage, instead I'm using a tiny, circuit to pulse the electricity through a voltage multiplier, and on the other end is the 150uF capacitor and basically a solar engine.

actually I think I'm going to drop the voltage cascade for a tiny handwound transformer

> MOSFETs work on voltage, no current, right?
. According to Wikipedia:
"Varying the voltage between the gate and body modulates the conductivity of this layer and makes it possible to control the current flow between drain and source."
. Didn't notice anything on voltage drop, but I was scanning fast. It may be there.

Thanks Nacho! I don't think there's a voltage drop because it's only going through 1 type of semiconductor, but someone else please correct me if I'm wrong.