Reading Datasheets?

I'm super new to electronics and am still trying to wrap my head around how to read a datasheet. I've attached an image of a part of this datasheet that I'm trying to figure out:

https://www.fairchildsemi.com/datasheets­/2n/2n3906.pdf

If I understand correctly (and someone please tell me if I am correct or incorrect, I am only interested in transistors as switching devices even though I know they can be used as amplifiers) this PNP transistor requires 1mA on the base pin when there's 10mA on the collector pin (regardless of Voltage since Current is what determines if the transistor is saturated or cut-off, which also means there's a 10:1 ratio between collector and base) and that the .85V is the voltage drop of the transistor... Someone has been trying to explain this to me on a different forum and this is what I've come up with so far.

1 Attachment

This video lecture about digital applications of a transistors may be helpful in understanding what the measured parameters mean and with the wording about 'what is going on' around this transistor.

Knowing about your application / purpose of this component in your setup would be helpful to be a bit less theoretical...

you are on the right track. Vce(sat) indicates what voltage drop to expect across collector-emitter with (in this case) 10mA collector current and 1mA base current.
(http://diodes.com/datasheets/ZXMN2F30FH.­pdf)
if you want to control a transistor with a GPIO pin, have a look at n-channel MOSFETs such as ZXMN2F30FH. there are dozens of part names to choose between. Look for parts that have Rds(on) specified at 2.5V. this one will have maximum 0.065 ohm resistance with 2.5V gate voltage. beginners make the mistake of looking at the spec for gate threshold voltage, but that is specified at very low current level. this MOSFET will start to turn on at typically 0.9V, but to drive high current you need more gate voltage. a 3.3V GPIO can drive this MOSFET with drain currents over 4A.

The GPIOs deliver enough current, but what about the voltage to drive the gate? The CMOS gate guarantees to source up to 8mA at 2.9V (= 3.3V - 0.4V). The TTL ports - the one without the pink 3.3V box in the board reference picture - guarantees the same current but only on 2.4V. The 2.4V is stated as a minimum... nevertheless, choosing a N MOSFET with an Rds(on) <=2.4 gives full freedom in pin selection.

I wondered any way why a PNP transistor... that's why I asked for a bit more context.

@DrAzzy added a really good page to the Espruino site on MOSFETs: http://www.espruino.com/mosfets

If I'm honest I know probably as much as you about transistor datasheets... I tend to look at the graphs when I can :)

But yes, I think that's right - so assuming you've connected the transistor's collector to 5v, you're pulling the base down to 0v, and you want 50mA out, you're trying to apply 5mA when the voltage drop is 0.95v.

So 5-0.95=4.05v, and R = V/I = 4.05 / 0.005 = 810 Ohms

But maybe someone else should check that :)

Another brief note though - Espruino's IO pins are rated for 20mA, so if you're trying to get 10mA at 5v you're better off connecting directly :)

I'm new to this as well. As I've had to learn things quickly I've adopted this "workflow":

1. Make sure you can actually get the part you want to read up on. I recently almost ordered a PCB that relied on a part central to it's design (an SMD antenna) which was not available anywhere I looked.
   
2. Find reference designs with google (similar applications of a similar device to what you're trying to do)
   
3. Verify relevant electric properties (i.e what currents it can handle etc, the interesting parameters will vary from part to part) for you design
   
4. Look at the pin out explanations (for a mCU you'll want to check things like verifying it has support for the IO you need etc) to verify it supports what you need and expect
   
5. Look at the available packages it supports, and I also learned it's a good idea to verify it uses the most common pin layout for common parts (SOT-23 transistors for instance, some odd ball parts will not be in the common configuration)
   
6. Google everything you're unsure about, or ask a friend, or ask a forum - such as this one :) I'm sure I'll post many questions here.
   
   

For me, it's been really helpful to view reference designs and tutorials/explanations on youtube.

I also have more than one CAD installed in case I want to reference an open source project or something like that.

Last point I learned when doing my latest project was to try to find my exact part when I place the symbol in the schematic. For the aforementioned project, I drew the schematic first, then had to go over and figure out exact parts and values. Let me tell you - it was no fun.

Sorry it's been so long since I responded to this post, just been busy with other stuff...

Look for parts that have Rds(on) specified at 2.5V. this one will have maximum 0.065 ohm resistance with 2.5V gate voltage. beginners make the mistake of looking at the spec for gate threshold voltage, but that is specified at very low current level. this MOSFET will start to turn on at typically 0.9V, but to drive high current you need more gate voltage. a 3.3V GPIO can drive this MOSFET with drain currents over 4A.

I read the article linked by Gordon about MOSFET's and it's actually pointing to Vgs being the specification I should look for... Why would knowing the resistance levels with 2.5V on the gate tell me how much current can be passed through the transistor? Actually, I'm really starting to get confused now after looking at that transistor datasheet you linked, the graph from figure 3 says there would be much much more current than 4A flowing through the transistor with just 2.5 V for Vgs...

Also, for the general purpose BJT, I'm confused as to how much current needs to be applied to the base for the collector/emitter to reach the peak 40V given the saturation levels I highlighted in the OP...

Thanks :-)

The way that Vgs is specified is by telling you the Rds(on) at a given voltage on the gate (ie, Vgs)

The graphs with the crazy current numbers (fig 1-3) show the maximum peak current (measured in 20us pulses), with the x axis showing the voltage between drain and source while it's on (ie, the voltage drop across the fet) and carrying the stated amount of current, and the lines represent different gate voltages. You'll notice that at low current, IV relationship is linear; this is the ohmic region, and is where you want your devices operating. As a limiting current is approached, the current nolonger increases as Vds does. You want to stay out of that operating regime.

That is a beast of a transistor. At Vgs=10, it is spec'ed for 61 amps continuous current. Just a jaw-dropping number, and 31A higher than the package limitation (nice marketing).... At Vgs=3.3, that should be able to handle 20-30 amp loads easy.

You can get fets with 2.5v gate voltage that can handle 5+ amps in a rice-grain sized SOT-23 package now!

Uh, for the BJT, I don't understand your question "for the collector/emitter to reach the peak 40v"?

if we look at the datasheet I linked to, Vgs(th) is the threshold voltage at which the MOSFET starts to conduct current (0.00025A). This threshold is between 0.6V and 1.5V, at room temperature. This is good information in the way that if we have 1.5V on the gate we should be able to run at least 0.00025A in the MOSFET and can expect that the drain voltage will be 1.5V or below. But most of the time our loads are larger than 0.00025A, are they not?

Now, the threshold voltage is not useful information if we want to control a real life load. If the load is for example 2A we will need a lot more gate voltage if we want the transistor to turn on. If we look at figure 3 we could be fooled to expect that we only need 2V gate voltage. But don't forget that this curve is showing only typical behavior and with 5V across the MOSFET. totally useless information, because 2A * 5V = 10W. the MOSFET would expire in a puff of smoke if we tried that.
The only section in the data sheet that gives us any solid promise is the Rds(on) value at 2.5V gate voltage. If we apply 2.5V at the gate and the load current is 2A we know that the resistance will be max 0.065 ohm, if the transistor is at room temperature, because the data sheet says so. the power dissipation then is 0.065 * 2^2 = 0.26W, and we can be sure that the transistor can handle this heat dissipation. the temperature rise will be about 34 degrees.
With bipolar transistors (npn or pnp) the common mistake is to look at the DC current gain value and forget that this is specified at a relatively high voltage drop. for switching loads we have to look at the saturation voltage, and read the values of base current and collector current and the saturation voltage, to get an understanding of what the transistor can do. in most cases an ordinary bipolar transistor is not suited for driving amperes of load current, because the required base current is so high that it is unpractical to drive from a GPIO pin. it can be very difficult to figure out how much current is going into the base with a certain base resistor, because the GPIO output voltage will sag as the pin current increases. and it is many times not well specified what current will be available from the pin. and in the same microcontroller, the drive characteristics for GPIO pins can be different.

But don't forget that this curve is showing only typical behavior and with 5V across the MOSFET. totally useless information, because 2A * 5V = 10W. the MOSFET would expire in a puff of smoke if we tried that.

Ok, I'm getting conflicting information again... Dr Azzy says:

That is a beast of a transistor. At Vgs=10, it is spec'ed for 61 amps continuous current. Just a jaw-dropping number, and 31A higher than the package limitation (nice marketing)...

Also, when I asked

for the collector/emitter to reach the peak 40v

I mean, the bjt's maximum voltage rating from collector/emitter is suppose to be 40V, but what amount of current would I need to apply to the base in order to allow 40V from collector/emitter given saturation ratings...

I was talking about the IRF3708 (linked from mosfet's doc) - I belive that's the only MOSFET that's had a datasheet linked here.

Now, the threshold voltage is not useful information if we want to control a real life load. If the load is for example 2A we will need a lot more gate voltage if we want the transistor to turn on. If we look at figure 3 we could be fooled to expect that we only need 2V gate voltage. But don't forget that this curve is showing only typical behavior and with 5V across the MOSFET. totally useless information, because 2A * 5V = 10W. the MOSFET would expire in a puff of smoke if we tried that.>

Figure 3 where? We are clearly not looking at the same device, as figure 3 of the datasheet I thought we were talking about (again, I see no datasheets for any other MOSFETs referenced, and only mosfets have a gate voltage) says nothing about behavior at 2v on the gate - the graph doesn't start until 2.7v; the IRF3708 doesn't provide any specs for Vgs lower than 2.7, other than that Vgs(th) is between 0.6 and 2.0 volts.

Re: BJT - So you have positive voltage on emitter, and want to know how much current to apply to the base to make it saturate? It looks like somewhere around 5mA with a load of 50mA, leaving a voltage drop of around 0.4v? The voltage doesn't matter.

in my post 23 days ago I was referring to ZXMN2F30FH, I don't know about any other link to a MOSFET data sheet. perhaps there is where the confusion is.

Okay, I think that explains the confusion. I had somehow missed that link, and thought you were talking about the one linked in the mosfet page.

Sorry for the confusion there.

Re: BJT - So you have positive voltage on emitter, and want to know how much current to apply to the base to make it saturate? It looks like somewhere around 5mA with a load of 50mA, leaving a voltage drop of around 0.4v? The voltage doesn't matter.

OH! lol, I always thought the .4V was the allowable voltage across the collector-emitter... not the drop out voltage... where do you see that .4V being the drop out voltage? So just to be clear, regardless of what voltage is being applied to the emitter (with the exception of voltages above 40V) it only takes 5mA on the base (which inserts 50mA to the load) to have it saturate? Say there was 10V being applied to the emitter and you only apply 4mA to the base, would that mean only 8V is allowed through the bjt?

if we look at the datasheet I linked to, Vgs(th) is the threshold voltage at which the MOSFET starts to conduct current (0.00025A). This threshold is between 0.6V and 1.5V, at room temperature. This is good information in the way that if we have 1.5V on the gate we should be able to run at least 0.00025A in the MOSFET and can expect that the drain voltage will be 1.5V or below. But most of the time our loads are larger than 0.00025A, are they not?
Now, the threshold voltage is not useful information if we want to control a real life load. If the load is for example 2A we will need a lot more gate voltage if we want the transistor to turn on. If we look at figure 3 we could be fooled to expect that we only need 2V gate voltage. But don't forget that this curve is showing only typical behavior and with 5V across the MOSFET. totally useless information, because 2A * 5V = 10W. the MOSFET would expire in a puff of smoke if we tried that. The only section in the data sheet that gives us any solid promise is the Rds(on) value at 2.5V gate voltage. If we apply 2.5V at the gate and the load current is 2A we know that the resistance will be max 0.065 ohm, if the transistor is at room temperature, because the data sheet says so. the power dissipation then is 0.065 * 2^2 = 0.26W, and we can be sure that the transistor can handle this heat dissipation. the temperature rise will be about 34 degrees.

So, just to be clear, this .045 and .065 for the different voltages is meant to give us a ratio? Does this ratio only specify the required voltages applied to the gate in relation to the current on the load? As in, assuming you had .00000000000001 amps on the load, but wanted to apply the maximum 20V on the Drain-Source, it would only take at a minimum .6V on the Gate since that's the beginning threshold for the transistor to turn on?

That is a beast of a transistor. At Vgs=10, it is spec'ed for 61 amps continuous current.

I've attached two screenshots of the datasheet that confuse me with what I'm being told about Rds(on) and this maximum peak current... The Rds says the transistor can only handle 15A at 10V but the maximum specs say it can handle 61A at 10V... what does it mean! LOL

2 Attachments

So, just to be clear, this .045 and .065 for the different voltages is meant to give us a ratio? Does this ratio only specify the required voltages applied to the gate in relation to the current on the load? As in, assuming you had .00000000000001 amps on the load, but wanted to apply the maximum 20V on the Drain-Source, it would only take at a minimum .6V on the Gate since that's the beginning threshold for the transistor to turn on?>

the 0.065 ohm Rds (on) is the resistance between drain and source, when the MOSFET has the specified voltage on the gate. This doesn't depend on source to drain voltage - as soon as the MOSFET is turned on, Vds is becomes tiny, since the drain and source are practically connected.

For even very small loads, yes, you'd need to get the gate up to around Vgs(th) for it to start to conduct. However, note that leakage current - current that will flow even when the transistor is off - is not quite zero; it's specified in the datasheet, and is generally very, very small, but probably larger than your example very small load.

I've attached two screenshots of the datasheet that confuse me with what I'm being told about Rds(on) and this maximum peak current... The Rds says the transistor can only handle 15A at 10V but the maximum specs say it can handle 61A at 10V... what does it mean! LOL>

First one specifies Rds(on) at the specified gate voltage and current.

Second one specifies maximum continuous current at that gate voltage - at the higher current, the resistance may be a bit higher (see the charts)

But don't forget that this curve is showing only typical behavior and with 5V across the MOSFET. totally useless information, because 2A * 5V = 10W. the MOSFET would expire in a puff of smoke if we tried that.

So, I just want to be sure that I'm understanding this correctly since when I look at the datasheet, it says the transistor you linked to can handle a continuous current of 4.9A at 4.5V on the gate which is definitely a lot more than 10W... am I reading this and understanding you correctly?

1 Attachment

With 4.5v on the gate, it is rated for that 4-4.9A continuous current. Rds(on) typically 0.065ohm @ Vgs=4.5, sat 4.0A, ohm's law gives us 0.26 volts dropped over the fet. 4A*0.26V=1 W

His first statement was in reference to figure 3 (note - this is in the second set of graphs, not the ones on the second page w/no figure numbers), which shows a graph of current vs gate voltage - that at first looks to be saying that at 2v you can get 2A of drain current. But the graph says that that's at Vds=5V, that is with a 5v drop across the fet - that is to say, this isn't a graph depicting useful operating conditions. As you can see, the lines fly off the top of that graph when Vgs enters typical operating conditions :-)

I think DrAzzy is doing a good job of explaining how to read the data sheet. And there is also lots of good detailed information if you google "how to read MOSFET data sheet". if we just look at this particular MOSFET that I mentioned (ZXMN2F30FH), I can show how I go about finding out how much current it can handle:
first, find out how much power dissipation it can handle. to do this, I have to decide how hot I want to run the transistor. I usually set +125C as maximum. then I decide on the hottest ambient temperature. let us say +40C. 125-40=85C. then I find the figure for thermal resistance from junction to ambient, 131K/W. so the maximum power dissipation it can handle is 40/131=0.3W. then I estimate the resistance Rds(on) at 3.3V. for this transistor I have values for 2.5V and 4.5V, and I decide that the resistance at 3.3V will probably be close to the average of these two values, so I say that at room temp and 3.3V gate voltage the resistance is (0.065+0.045)/2=0.055 ohm. but at 125C the resistance will be higher. Fig 4 tells me that it is about 1.3 times higher at 125C than at 25C, so I will use 0.072 ohm for the resistance value. Now that I know R and P it is possible to calculate the maximum current I that the transistor can handle. it is sqrt(P/R) or 2.0A.

this is the procedure that I normally use for quickly figuring out how much steady state load current I can safely run through a specific MOSFET.
2A is a lot less than 4.9A, is it not?
if the MOSFET is switching on and off there will be additional power loss especially is it is driven from a GPIO pin, so it may be necessary to use a gate driver between the GPIO pin and the MOSFET gate - especially if the MOSFET is larger than this SOT23 size device we have been discussing.