There has been some discussion here about using a high-voltage power MOSFET like the IRF520 in a logic mode, and it spurred my curiosity. I don't have any of this particular device on hand, but I decided to see what I could learn from a spice simulation.
Figure1 shows an IRF520 being used in a typical logic configuration. I've arbitrarily chosen to put a 1K resistor in series with the drain, and a 100K resistor from the gate to ground. Power is provided from the 5V source on the right of the diagram. The values are really not critical to this discussion though, within reasonable limits.
Figure 2 shows the output voltage VDS (vertical axis) of the circuit as the gate voltage VGS (horizontal axis) is varied from 1 to 5V. As you can see, the MOSFET switches abruptly from OFF to ON when the gate voltage reaches a threshold level of about 3V.
Figure 3 shows the IRF520 transfer curve when R1 is reduced to 50 Ohms and the power supply voltage is increase to 12V. Note that the device still switches but that the threshold is less abrupt and has shifted to a higher (but still probably OK) level. The MOSFET current in this case is about 250mA.
Physical device characteristics will vary from my spice model of course, and the switching threshold will vary with things like temperature and individual device. But in general, it appears that the IRF520 will work just fine when its gate is driven by something like an Arduino output pin over a wide range of operating conditions.
A lower-power mosfet will have a lower switching threshold voltage, which would give more noise margin with a 5V supply voltage.
Perhaps someone here could verify these results on the bench with an actual device, but 'til then I hope this provides some insight.
(The images are easier to see if you right-click on them and then "view image". At least here at my end using Ubuntu.)
The figures don't show up very well do they? Still, thanks for the info.
I don't think anyone does any serious work anymore without simulation. Especially since so much real design is done on silicon, or with FPGA-like devices, and since things like PCB trace matching have become so important. But we all end up on the bench eventually don't we.
I expect that a lot of posters here want to use a device like the IRF520 in logic (not power) mode because it's what they've bought on eBay. But I suppose the best course is to ask before being too critical, no?
CurtCarpenter:
I expect that a lot of posters here want to use a device like the IRF520 in logic (not power) mode because it's what they've bought on eBay.
I do believe it is included in some of the Arduino "beginner kits". But both the inappropriate and the appropriate ones are pretty cheap on eBay,
IRF520 - advertised for Raspberry Pi!
... though admittedly the IRF520 is much cheaper.
CurtCarpenter:
But I suppose the best course is to ask before being too critical, no?
I'm truly sorry that you took offense srnet, but I have to wonder what motivated it. Perhaps you'd like to explain?
Letting that go though, I'd encourage you to download a SPICE program and invest the time to learn to use it. There are several freely available on the internet I believe, and they're invaluable not only in serious design work (I'll leave it to you to characterize that as you wish), but also as a learning tool. Simulation forms an important teaching element, for example, in many introductory EE courses at places like MIT. Beats doing the math by hand
And you, no doubt jrremington, are one of those very special workers, right?
But listen: whether you know it or not (and it seems that you may not), the majority of the integrated circuits you use -- no doubt with stellar expertise -- begin life as models in somebody's simulator. Even those single MOSFETS begin with a process model. They don't start with somebody sitting in a corner using "an oscilloscope and real ccts."
Some of you here sure are defensive. Or perhaps "close minded" would be a better term? I wonder why. I've been doing this stuff since 1960 and I'm still learning things (and even, God forbid, making mistakes).
When did you stop? And why? It's too bad in any event. No spirit of adventure left, I guess.
Look: if you want to ignore the simulation results, that's fine with me. They don't represent my life's work.
If you want to take some measurements with an "oscilloscope and actual ccts" and share the data, that would be terrific too. As I said, we all end up at the bench eventually. And who knows -- the SPICE model I have could be wrong. If so -- good to know.
If one or two decide to learn something about simulation and add it to their toolbox as a result of my effort, it would make my day. But if not, that's OK too.
The fact is that an "oscilloscope and acutal ccts" just aren't sufficient across a broad array of serious design work anymore -- at the component level, at the board level or the system level.
Jim Williams and Bob Pease are both rolling over in their graves as the seduction of simulation has claimed one more victim.
Your use of simulation to justify the misapplication of a part is a perfect example of how it can be used to prove things that really aren’t true. GIGO rules.
My WattsThat but aren't you a purist of the first rank
If "a part" is the only one you have, and it gets the job done, "misapplication" can sometimes serve quite well.
My impression is that a good number of people here are playing with power MOSFETS and trying to make them work. My purpose was to be of help in that, Why insist that they be burned as heritics because they don't adhere to your somewhat puffed-up standards of "appropriateness"?
As to your views of the value of simulation, save them for the designers of the LLC. They're wasted on me.
Sure larryd. But the "toy" news will be a surprise to Synopsis, Cadence, Mentor Graphics and a great many others.
You and several others here have a very... uninformed shall I say?... understanding of simulation and the role it can play across the full spectrum of engineering design activities.
PS -- nice diagram. Can I ask what schematic capture program you use?
The transistors are fine and certainly don't hurt, but they may not be needed and that's the point. In several of the applications I've seen here, the load appears to be so small that the level shifters just aren't necessary. Until somebody shows some actual measurement data to the contrary, all of the "its not a logic MOSFET!" comments are just nonsense. Sometimes, a MOSFET is just a MOSFET.
