I want to know exactly what I have and how far I can push them

Unless they have built-in self-protection, you should be able to push them all the way to the point of letting out the magic smoke. Ahh, but then you won't have them any more.

 

I want to know exactly what I have and how far I can push them, to truly know their limits.

All you need for that is power supply and resistor.

 

The load testing that I have done used incandescent light bulbs as the loads. Certainly they are non-linear, but quite safe for dissipating a few thousand watts. Of course not everybody needs to verify a system operating correctly while delivering multiple kilowatts. For much smaller loads 12 volt headlamps and backup lamps worked.
The bad news is that with the removal of incandescent bulbs from the market this handy source of loads is difficult to use.

 

All you need for that is power supply and resistor.

allright, how morespecific?
How do you linkeverything? How do you calculateor observe or measure? What areyou thinking at?
I kind of having an idea of what you might think about, but is betteryou to expose a muchdetailedexplanationandexemplification.
Thankyou.

 

It all depends on what you want to achieve. Yes you can test a power supply with just a resistor or two, but if you want to measure load regulation, step response, etc you need to be a little more sophisticated. My active load is primarily for characterizing LFP battery packs at 28v and up to 85A (120A peak for a short while) to calibrate voltage v capacity both at constant current and across a known 'load profile' which can run for days. I also test individual LFP & LiPo cells at 1 - 10C loading to validate & match capacity before creating battery packs.. I've used it for testing/evaluating EV-bike packs at up to 96v and 20A. Finally, because it is fully programmable, I can emulate a a battery under charge to test battery chargers for correct CC/CV operation. Admittedly not something most people need/want to do!

 

betteryou to expose a muchdetailedexplanationandexemplification.

Is the space key on your computer broken?

 

allright, how morespecific?
How do you linkeverything? How do you calculateor observe or measure? What areyou thinking at?
I kind of having an idea of what you might think about, but is betteryou to expose a muchdetailedexplanationandexemplification.
Thankyou.

Lets say you have a 5v 1A 'wall wart'. 5v @ 1A = 5ohm 5W. so a 5ohm 10W resistor will load it to capacity. Measure current through resistor and measure volts across PSU output with resistor connected and disconnected. This gives you a view of load regulation (how well it copes with demand). Or use a series of 1ohm as well as a 5ohm to test with loads of, say, 10, 9, 8, 7, 6, & 5ohm to give more detail about regulation from 500mA to 1A. Use n x 1ohm in series to see how it copes with overload ie 4ohm is 20% overload...

 

allright, how morespecific?
How do you linkeverything? How do you calculateor observe or measure? What areyou thinking at?
I kind of having an idea of what you might think about, but is betteryou to expose a muchdetailedexplanationandexemplification.
Thankyou.

You have a variable power supply, right?

You want to see how far you can push a MOSFET. Connect the power supply directly across the drain to source. Supply 10V to the gate. Turn up the Vds voltage up slowly until it goes up in smoke while observing the current. Do this with about 100 of them.

How do you propose to do it with a PWM circuit?

Or you can RTFDS (read the fine datasheet) which will tell you how far you can push it without destroying any.

 

If I get a chance later I'll post a snippet of one of the 400W modules that make up my 2400W active load.

you are the only one with a schematic. Im very curious.

 

The bad news is that with the removal of incandescent bulbs from the market this handy source of loads is difficult to use.

Light bulbs are a very useful tool to troubleshoot power electronics.
Several years ago, I arrived at a friend’s home as he was replacing all the incandescent bulbs with LED ones. He had gathered all the old bulbs on a cardboard box which he was meaning to throw into the trash bin.
He agreed that I should take them instead and I am very happy that I did.

 

yes, indeed, actually I find that the most use of a power load is to test power supplies of diverse fabrications.
My idea of using it is to test components, specifically the ones I have. I want to know exactly what I have and how far I can push them, to truly know their limits. Again, my stuff, what I have. I have brand new and also scrapped stuff. Thats my reason and my idea Im pointing. And now, I want to get close to that commercial 35W, and I think is possible, with what I have.
Im also learning about mosfets, how to drive them hard, and some characteristics I never thought for myself, thanks to these open discussions here.So Im shooting many rabits with 1 bullet, this project as weird as it sounds. But it is very fascinating subject and I like it.

So, you have some components from some scrapped stuff and you want to know how hard you push them?

That requires a destructive test. You push them until they fail.

But what has that gained you? You know know how hard THAT particular component could be pushed, but THAT component is now destroyed, so what use is that new-found knowledge?

If you have a bunch of those same components, then that knowledge gives you a HINT of how far you can push OTHER components of that same kind. But the one you pushed to destruction may not have been typical. Each individual component will have it's own point of failure, and you have no idea if the one you happened to test to destruction was unusually good or unusually bad. So you test a bunch of them and look at the distribution of points at which they failed and, from that, make an estimate of how hard you can push the next one without destroying it. How you come up with that estimate depends on the relative importance of a number of factors, but if it's important that NONE of them be expected to fail, then you take the point at which the worst one failed and reduce that by a reasonable safety factor and use that as your design limit.

Guess how the manufacturers of components determined what their datasheet specifications should be? By destroying a bunch of them and using the data to determine the limits that would keep the failure rate below their acceptable limits.

With this in mind -- namely that "to truly know their limits" -- you have to destroy the component you want to know the true limit of, what is it that you are hoping to accomplish?

 

what is it that you are hoping to accomplish?

very good point about testing to destruction and note the values down. I didn't think that far, I admit.
I didn't explain myself clear enough then. My target is to use these transistors and not to destroy them. You are right when saying to their limits is to actually destroy them. But in my mind I was thinking"to push" them to the maximum power they can start to strugle, without destroying them, and without overheating them. For me that limit is dictated by my usual50*C , or a bit around it.
Most certainly I will have to use AL heatsinks and a fan. To get as close to that 35W comercial load tester I put my eyes on.
- Im not that kind of dude to destroy its component just because. I am that kind of dude that likes to make stuff with what it has and ideally, durable in time and as practical as possible.
I hope my intention is a bit more clearer now.

 

very good point about testing to destruction and note the values down. I didn't think that far, I admit.

As I said in #48. Thank you @WBahn for adding detail and removing the sarcasm.

@q12x: the power in the load has nothing to do with the heating of the MOSFET. A variable load does not really help with this characterization. A MOSFET that has 1mOhm on resistance can power a 100W load at 50V while dissipating only 4mW. But the same 100W load at 1V will cause the MOSFET to dissipate 10W.

You can calculate the Rdson of the MOSFET when in the Ohmic region (driven hard as you say) by measuring the current through it and voltage across it. Do this at several currents to verify linearity and ensure it is in the Ihmic region. You can then determine the power dissipated in the MOSFET for any drain current (I*V). Next you can calculate the rise above ambient by using the thermal characteristics of the case and heatsink. And finally, if you want to know how far you can push it, you calculate the junction temperature based in the case to junction thermal resistance of the case type (TO220 eg.). Max junction temperature for ordinary semiconductors is 125C.

This is how engineering is done, not by trial and error.

 

This is how engineering is done, not by trial and error.

trial and error is where discoveries are born ! and when everything else fails, like help or good intentions.
Very well then, I agree with everything you said. I will try to do it as you say.
I believe I did it already, some time ago, exactly this exercise and we all hit the wall of calculating the power over a switching mosfet. Because it is switching, no one could realistically determine what is what. I believe I asked this exact question: -how to obtain a switchingtransistorvoltage?And the answer was just simply do theV*I=W and that is it. And I remember I did something,I extrapolated the measurement, butyou didnt agree with that solution. So I believe that was the ending point of mosfets experiment. If I remember right. I think I do. Hmmm. I think after that I concentrated on BJT power disipation, using the same cct with the opamp, because it was more easy to verify the CEvoltage and determine the Wattage over it.
But I have another idea how to push the projectas it is, this one with the mosfet here. Even if we all can not realisticallydetermine the voltage over a switching transistor like this mosfet in my case. It will be an aproximation I think, but it will be close enough to reality. We will see when we will get to it.

 

"TRIAL and ERROR" without adequate understanding is a waste of resources and time.!! And without adequate measurement capability AND UNDERSTANDING, there is no way to know when one has reached the maximum safe limit for a component!
So a much better choice is to learn how to locate the specifications of devices. AND learn how to read and understand specifications.

 

- After watching a couple of videos on youtube, and general search on internet websites, for a couple of weeks but mostly the last days, I start to shift my attention to the root of the problem. I believe, this is it the primary cause of all my failings or misunderstanding, and possibly some of you that didn't attract my attention until this point, clear enough or at all. So this is a lesson for all of us, I guess.
- So basically, I "discovered" that the "Power Load" naming is most often called "DC Power Load" and just by the look of it, it told me very little or nothing. In my mind DC meant it is not AC or that it is just a lower DC voltage than the supply transformer. And I didn't pay more attention than that.
- What it really means, is that this circuit is dealing with all -DC Power Supplies out there-, not everything else DC like testing components or the way I was thinking to use it. In other words, this circuit is -designed- to be used and test ONLY DC Power Supplies and nothing else !!!
Here is a list of All possible Power Supplies that can test:
-variable or fixed SMPS,
-steel core transformers,
-batteries or/and Accumulators of any type and chemistry inside,
-DC generators like the ones used with diesel motors
-and probably dynamos as well.
Everyone is pointing a bit too fast and evasive towards these enumerated supplies for test and they are calling them "LOAD" Power supply, or rarely "under test PSU" which is the same thing. Maybe this is a common thing in some circles, but not so common in my very small circle. And this is very important detail I was not completely aware, I was like 30% aware but not putting all my attention into it, believing I can adapt it to my specific and exotic needs. I still think I can adapt it, but, now I see it with better eyes because of this revelation. Now I can think about it better than before. I hope. Haha, or maybe I'm still an idiot. But this project is a very interesting lesson about a LOT of circuit problems to solve and to find answers others have already solved or comercialized or just DIY for personal use.
I am still fascinated about this project, and my intention to be very clear, from the very beginnings, was NOT to test my PSU as a LOAD, but to test components as a LOAD, more specifically my new or scrapped components like mosfets and BJT transistors, some diodes, some resistors and a couple of other types of components. I'm repeating this for the 10'nth time now, because I can still read some new and dizzy answers of what all this means or is intended for.
Hmmm...still fascinating, haha.

 

Yes, the complement of a variable load is a variable power supply.

Power supply — produces power
Load — consumes power

 

Most often, testing the majority of components involves much more than the application of a DC voltage. ALWAYS there is a need to measure current. The possible exception is testing indicator devices or actuator devices, and destructive testing.

 

Yes, the complement of a variable load is a variable power supply.
Power supply — produces power
Load — consumes power

Well, thats the thing I mentioned already, some call it " -LOAD Power supply- " !!!
See? In this particular case the Power supply is both the producing power and consuming power as a LOAD. LOAD, because is under test and not used as normal, to provide power? Is my guess. A very peculiar detail, at least for me. So the circuit or the designer is seeing this "LOAD" as a simple power resistor to test the maximum of heat dissipated over it, but also as a capacitor that you suck power from it, and how much, in terms of Amps. Is my best guess! Weird shi_tuff.

 

What you allude to is what's known as a 4-quadrent active load, which both a supply and a load depending on the direction and polarity of the voltage/current and the external circuit.

What we've been discussing to date as an active load is purely passive, in the sense that it doesn't generate, merely consumes power, either as a constant current irrespective of the applied voltage, or constant power, where the current is adjusted depending on the applied voltage (ie a constant I x V).