Apparantly I don't understand transistors/mosfets like I thought I did

[Red Squirrel]

So I was just messing around, wanted to play around with making a basic high enough current PWM to play with transformers and stuff, so I setup a bread board with 12 volts on one side from an ATX PSU and 5 volts on the other side from an arduino (purposely wanted something with a low current rating).  I then put the two grounds together to ensure that the arduino's +5v uses the same ground ref as the ATX.

I then put a mosfet, connected as follows:
Drain to the negative pin of a 12v fan and the + of the fan to the 12v rail. 
Source to ground
Gate to +5 volts. The fan barely moved. It needed help to get going. When I tested the voltage output I was getting maybe 4 volts when I should be getting 12, or at least I thought so. 

I did the same but with a transistor, fan turned better, but it was still not getting the 12v.  What could I be doing wrong? Or is 5v simply not enough to saturate?  The mosfet I was playing with is a IRF 610.    I'm trying to interpret the datasheet but really not sure where the saturation voltage is.  It might be called something else.

[Monkeh]

See, for example, figure 3. And bear in mind you're dealing with a much lower voltage. The 'saturation voltage' for a FET is basically whatever voltage they specify in the conditions for R_DS(on).

That FET isn't fully on until about 10V, and even then it's more of a resistor..

The 'output voltage' will be the voltage across the FET..

There are much better suited FETs for these applications. A little hunt for stuff advertised as 'logic level', or indicating R_DS(on) at 4.5V, will yield many good results. Many of which will be measured in single-digit milliohms.

[Red Squirrel]

Oh ok,  so with a voltage less than 10 then it basically starts to act as a resistor and only lets some of the current through, thus I get voltage drop if I draw too much?   Also forgot to mention but with the mosfet I was actually getting a full voltage reading with a digital multimeter, is that just phantom voltage, or is that how mosfets work?   

If I was actually doing a project I'd make sure to size the mosfets properly of course.  This is just a random one I pulled off an old UPS to play with.   Though even with the transistor there seemed to be some voltage drop but not as bad, and it got awfully hot. Though maybe it was the same idea, 5v was not quite enough to drive it fully.  Something I'd do more research on if actually buying for a project.

[max666]

I think even with 5 V gate voltage the fan should do more than just barely moved. I think something else is fishy here.

When you used a transistor, what transistor did you use, what base resistor did you use? Or didn't you use a base resistor? That would explain why it got hot 
And as always, a picture often paints a thousands words.

[atferrari]

So I was just messing around, wanted to play around with making a basic high enough current PWM to play with transformers and stuff, so I setup a bread board with 12 volts on one side from an ATX PSU and 5 volts on the other side from an arduino (purposely wanted something with a low current rating).  I then put the two grounds together to ensure that the arduino's +5v uses the same ground ref as the ATX.

I then put a mosfet, connected as follows:
Drain to the negative pin of a 12v fan and the + of the fan to the 12v rail. 
Source to ground
Gate to +5 volts. The fan barely moved. It needed help to get going. When I tested the voltage output I was getting maybe 4 volts when I should be getting 12, or at least I thought so. 

I did the same but with a transistor, fan turned better, but it was still not getting the 12v.  What could I be doing wrong? Or is 5v simply not enough to saturate?  The mosfet I was playing with is a IRF 610.    I'm trying to interpret the datasheet but really not sure where the saturation voltage is.  It might be called something else.

Could you show the actual circuit you are implementing?

Who is more positive? Drain or source? Is it N or P chanel?

Schematic please.

[Red Squirrel]

I still need to find a decent Linux program for making schematics, so excuse the poor drawing but this is basically what I'm doing:



Adding a resistor on base made it turn even slower, if at all.   The transistor seems to act better than the mosfet.  Using a TU13001 transistor and IRF610 Mosfet (separately, not at same time).  They are both N channel.

Edit: On the scope it's even more weird, if I give the mosfet +5v I get nothing on the scope, if I give it PWM I get a wave, but only 5v, not 12v.  If I disconnect the 12v power I still get the wave. Similar with transistor except the solid +5v also gives 5v out of the collector.   So it's almost like it's just passing through the 5v directly.  I wont rule out that this fet is damaged though, but it's probably me that's doing something wrong. 

[Alex Eisenhut]

Stop saying transistor "instead" of fet. They're both transistors.

[Red Squirrel]

Stop saying transistor "instead" of fet. They're both transistors.

I am playing with both, just specifying.

[Red Squirrel]

Oh wait, I know why I'm getting that reading on the scope, which begs the question, what is the proper way to measure a transistor's "output" if your circuit is earth ground referenced?  I can't just put the clip on the positive at the fan lead as that is before the transistor so it's always going to be active, and I can't just stick the ground lead of the scope anywhere but a ground referenced point.

[T3sl4co1l]

If the 12V remains stable, you can simply mentally subtract that from the reading.  If it's an alternating signal, you can use AC coupling (all the interesting stuff happens around the edges, anyway?).  Or you can perform a differential measurement with two probes and using the oscilloscope subtraction function.  But mind that the difference is never perfectly subtracted, and some common mode error will remain, especially at higher frequencies.

Preferred term for a "transistor" is BJT (Bipolar Junction Transistor).  There are two kinds of FETs, so that should be specified as well (JFET vs. MOSFET).

Tim

[Richard Crowley]

Do you have enough current gain in your BJT multiplied by however much current you are putting into the base?
Are you driving the BJT into saturation?  Is the BJT getting warm/hot?

Does your FET turn on when you connect the gate to 12V?
You are not using a "sensitive-gate" or "logic" FET which can be turned on with 5V
So it will take practically the same voltage that you are switching, (i.e. 12V on the gate).
If you want to control a higher voltage with a FET triggered by a lower voltage, you need a logic-level FET
For example:  http://www.allelectronics.com/make-a-store/item/rfd4n06l/rfd4n06l-4a-60v-n-channel-mosfet/1.html

The RFD4N06L, RFD4N06LSM are N-Channel enhancement mode silicon gate power field effect transistors specifically designed for use with logic level (5 volt) driving sources...
http://pdf.datasheetcatalog.com/datasheets/50/346688_DS.pdf

[Red Squirrel]

Do you have enough current gain in your BJT multiplied by however much current you are putting into the base?
Are you driving the BJT into saturation?  Is the BJT getting warm/hot?

Does your FET turn on when you connect the gate to 12V?
You are not using a "sensitive-gate" or "logic" FET which can be turned on with 5V
So it will take practically the same voltage that you are switching, (i.e. 12V on the gate).
If you want to control a higher voltage with a FET triggered by a lower voltage, you need a logic-level FET
For example:  http://www.allelectronics.com/make-a-store/item/rfd4n06l/rfd4n06l-4a-60v-n-channel-mosfet/1.html

The RFD4N06L, RFD4N06LSM are N-Channel enhancement mode silicon gate power field effect transistors specifically designed for use with logic level (5 volt) driving sources...
http://pdf.datasheetcatalog.com/datasheets/50/346688_DS.pdf

I'm not sure what you mean by the first sentence, can you explain?  Are you saying maybe I need a higher source of current and I'm not giving it enough?    I believe I am driving it to saturation but maybe that is actually my issue.  I notice I get more power if I connect base to 12v instead of 5v. The FET, as mentioned is probably just not meant for what I'm trying to do, though. Though even at 12v it will barely, if at all, move the fan. I'm also starting to think the fet may be damaged from when I pulled it out with a heat gun, I may have given it too much heat.  Even with the gate disconnected it still passes voltage that I can read with the multimeter, but it's not enough to actually do any work, even at 25 volts it wont drive the fan.  The transistor will though, and does get pretty warm even with a base resistor, but maybe it's not a high enough value.  This one I'm using is 100 ohm.  If I use any higher value then the fan barely turns.  I'm waiting for a solder sucker to come in the mail, at which point I'll be able to remove some power mosfets from an old UPS board to experiment further and they may be a better type for this kind of stuff.  Was unsuccessful otherwise due to the large heatsink assembly.

[Monkeh]

For starters you are using some terribly inappropriate parts.

The BJT you are using is entirely inappropriate and liable to have been damaged by what you've been doing to it. Throw it away. Don't use one without a base resistor again.

If the FET fails to make your fan spin properly with 12V on the gate, either the FET is entirely shot, or your fan is broken. Either way it's a 200V rated part with a ridiculous 1.5ohm R_DS(on).

Salvaging random components is a really terrible way to go about this..

[Red Squirrel]

I'm just trying to learn before I actually buy parts (well the transistors I bought, but I presume they're really low end, was like 5 bucks for 100), this is not a serious project.   Fan works fine if I just connect it directly so I know it's not the fan. 

[Richard Crowley]

I'm not sure what you mean by the first sentence, can you explain? 

In the first paragraph I was talking about BJT as I mentioned.
Being a current device, a BJT conducts (from the collector down through the emitter to ground) the amount of current determined by the base current you injecting, multiplied by the current gain of the transistor.  BJT transistors have a specified current gain depending on which type of device they are and some process variation. If you look at the published specifications for a particular BJT part number, you will see a range of values.  For example a 2N2222 can have a current gain (hFE) between 30x and 300x. 

The input (base) current can be calculated from the voltage and the base resistor using Ohm's Law.  I have an online Ohm's Law calculator here:  http://www.rcrowley.com/eirp.htm

Are you saying maybe I need a higher source of current and I'm not giving it enough?    I believe I am driving it to saturation but maybe that is actually my issue.  I notice I get more power if I connect base to 12v instead of 5v.

Are you talking about BJT or FET?
If you are talking about BJT, then if you get more power from raising the base voltage, then you are NOT driving the transistor into saturation.  Saturation means that the base current times the current gain is MUCH MORE than the amount of current you are switching.

The FET, as mentioned is probably just not meant for what I'm trying to do, though. Though even at 12v it will barely, if at all, move the fan. I'm also starting to think the fet may be damaged from when I pulled it out with a heat gun, I may have given it too much heat.  Even with the gate disconnected it still passes voltage that I can read with the multimeter, but it's not enough to actually do any work, even at 25 volts it wont drive the fan.

Measuring it with no-load (the multimeter) isn't a reliable test.  But your test under load DOES make it sound like your FET may be faulty.

The transistor will though, and does get pretty warm even with a base resistor, but maybe it's not a high enough value.

If the transistor is getting warm, it means that it is dissipating part of the power rather than conducting it. That is a clear indication that you are NOT driving it into saturation.  (And that is WHY we drive it into saturation.)

This one I'm using is 100 ohm.  If I use any higher value then the fan barely turns. 

You have so many unknowns here, the value of the resistor is a minor issue at this point.
What BJT transistor are you using?  What is it's current gain?

[Red Squirrel]

The transistor is a TU13003.  Could not find a datasheet but was told it's probably a MJE13003-E which I found a datasheet for.  I figured ordering through a retailer like Amazon I would have gotten a better documented part, but for a real project I'd probably order from Digikey so I know 100% that the datasheet I have is what goes with what I'm buying. 

This is the datasheet I'm looking at, but not sure if it's actually the right one as several different ones come up when I search for MJE13003-E:
http://html.alldatasheet.com/html-pdf/173514/UTC/MJE13001/55/1/MJE13001.html

Saturation voltage seems to be 0.5v.  There's lot of different numbers for current gain, not too sure what I'm looking at.  Also how does one know what all the symbols mean in datasheets?  Had you not told me what hfe is I would have no idea.

So the transistor will only pass a certain current based on the current going through the base, is that correct?  If so how do you force more current to go through?   I always figured as long as you reach the saturation voltage it will basically act as a switch and turn on and that it would draw a very tiny current only through the base.  Do you have a good tutorial or reading on transistors?  All the tutorials I've found never talk about this they only say it's like a switch.

Though I'm thinking maybe this transistor is just not up to the task for what I'm trying to drive with it.  It will drive a relay fine.  So maybe it's just the thing of me trying to make it drive something past it's current rating.  The fan is rated at 0.38a.  Not sure which number I'm suppose to look at in the datasheet but they're all higher than that though for maximum current ratings.  Come to think of it, is there a good tutorial somewhere on how to interpret datasheets, is it fairly standard across all components?

I probably should understand this stuff better before I move on to making something for real.

[Richard Crowley]

It would be great if you had a way to measure the gain. You don't happen to have a DMM that measures transistor gain, do you? 

If you are going to be playing with semiconductors, I would recommend getting one of the "$20 LCR/Transistor Tester" kits.
I have one and it is great.  I just ordered another one: http://www.ebay.com/itm/281476169022
That would tell you what kind of device you have, and the pinout, and the current gain.

The saturation voltage is pretty much the same for most silicon transistors.
But it is the hFE current gain, and the collector current rating that are critical for selecting the appropriate transistor.

You can only inject a limited amount of current into a transistor. You have to select a transistor with the ratings to handle the load.

It would be helpful to know how much current your load is drawing @ 12V so you know how that relates to your choice of transistor, etc.

[Red Squirrel]

I just ordered one of those testers actually.  The fan is rated at only 0.38 amps.  For some reason I could not get an actual current reading out of it when I was testing it out with a normal circuit, my DMM might be faulty, it's just a cheap Canadian Tire one and it's quite possible I blew the fuse or something at some point I'll have to check.  I want to eventually get myself a few more DMM's and at least one really good one like a fluke.

I ripped out some more mosfets (rfp50n06) from that same UPS, I knew these would be better as they were actual driving mosfets with a big heat sink, while the other one was a lone one without a heat sink, so it was probably being used for some kind of control and not driving circuitry.

This mosfet is driving the fan quite nicely, and will even cause sparks when I touch the wires together (can you blow a mosfet like that or will they self current limit and be fine provided decent heat sinking?), there's some real current going on now.  I was running into an odd issue where at random it would remain on, but through some quick researched I realize you should put a 10k resistor between gate and ground.  Guess it makes sense, since a very tiny amount of stray voltage will keep it on.

 So I'm getting progress here, and guess my issues were related to bad/improper components after all.  I got 3 of these same mosfets off the same UPS and a nice big heat sink too so I might keep these for if I want to build a little project SMPS, mostly just for learning.  When I was giving it a 50% duty cycle PWM I was getting interesting noise and spikes but that could be a lot of things like the breadboard and lack of any inductors or other cleaning circuitry. It's a start though. Still have tons of reading on SMPS before I do anything serious.  So much math though, that's the part I'll find hard I think but I'm eager to learn, and sometimes I find it easier to just improvise and dive in then figure out why stuff is not working from there.

[max666]

This mosfet is driving the fan quite nicely, and will even cause sparks when I touch the wires together (can you blow a mosfet like that or will they self current limit and be fine provided decent heat sinking?), there's some real current going on now.

Will they self current limit and be fine? I would say no 

[T3sl4co1l]

I would say so*.  Can't you see the transistors cleared the fault?

Here's another example...







In the course of switching, these poor specimens overheated, stayed on slightly too long, then gulped down some electrolytic capacitors, before eventually blowing the line fuse.  What you see is the impression of the textured sil-pad on the device's tin plating, with some cracked plastic around the edges.  The inside view shows two dies, one melted and cratered to bits (due to vaporization and arcing), the other shiny and rectangular, which was cracked in the process of separating the plastic (the co-pack diode for the IGBT; it wasn't under much stress, so it certainly wasn't the failure point, and it shows).

Tim

[Richard Crowley]

NO, semiconductors do not "self current limit".
An old joke is that a semiconductor will blow to protect the fuse.

[Red Squirrel]

Kind of figured considering how hot it was getting when I was quickly shorting it, just thought I'd check.  I know some more purpose built ones I read data sheets for can current limit but they are more complicated than just 1 transistor and most likely have special circuitry for that.

But yeah I seem to have things under control now.  Turned it the first mosfet was either bad or just inadequate.