To improve my design skills I like to look at the schematics of commercial stuff. In those designs I often see transistor switches connected as below.
. VCC . + . | . .-. . | | RL . | | . '-' . R1 | . ___ |/ . Vi--|___|---o----| . | |>
. .-. | . | |R2 | . | | | . '-' | . | | . '------o . | . === . GND . (created by AACircuit v1.28.6 beta 04/19/05
I usually only use R1, why would I need R2?
I calculate my R1 this way: R1 = (Vi-Vbe)/(0.1*iLoad) How would R2 affect this?
Thanks, Jenalee K.
To shut the transistor off if the driver can't. You can also vary the switch point.
greg
R2 has two purposes that I know of. It helps discharge the stored base charge and speed the turn off process (dump the charge delivered by the base to collector capacitance, as the collector voltage rises during turn off), if Vi does not go all the way to the emitter voltage and cannot be relied upon to do this. R2 also provides a path for collector to base leakage current that would otherwise be amplified by the transistor current gain and show up as high off state collector current. Again, this is more important if V1 does not go all the way to (or below) the emitter voltage.
R2 will make the transistor have a faster shut off time. Use a scope and look at the shut off time, with and without R2. I'm not sure what the preferred way of selecting R2 is, but I usually select R2 to have 10% of the current as the base current of the transistor. Brian
Two cases:
1) If the driving circuit goes from low to high but never floats, you dont need R2.2) If the driving circuit floats or is not always connected, R2 will keep the transistor completely off in case of stray signals or transistor base-collector leakage.
Luhan
Usually, you don't. Sometimes you don't need R1, either. I used to drive transistor bases from TTL gates directly just to annoy nearby engineers.
What matters is how much base current you get when the input is high, and how much when it's low, and whether these numbers are safe.
John
? ... I used to
Hmm, this can have been slippy - one misjudgement of the currnt limit and we have smoke.... But I like the idea, I am sure it has worked (on the nearby engineers, engineers, that is :-).
BTW, I almost have not used bipolar switches of this type since
1990... I just use a 2N7002 or something, logic output goes directly into the gate, works fine with 5 and 3V logic. I guess this must be comon practice all around he place, of course.Dimiter
------------------------------------------------------ Dimiter Popoff Transgalactic Instruments
------------------------------------------------------
John Lark> >
A 2N7002 is great with 5 volts on the gate, but really marginal at
3.3. They'll also switch 50 volts into 50 ohms in a ns or two, if you whack the gate hard.Is there an equivalent dirt-cheap SOT-23 low-threshold nfet out there?
2N7002's cost something like 3 cents by the reel.John
John Larkin scrobe on the papyrus:
Bipolar equivalent is BCR148.
Well strictly speaking they are marginal, but in many cases when you need not the highest current they can handle, but something within several mA they are quite adequate. They typically open to well below 50 Ohm just by the charge you give them with a DMM ohmmeter - which won't open a junction...
Another of my widely employed switches is an SST176 (P-JFET) when I have to switch a negative voltage powered load to GND and drive the switch by a +5V logic signal (4V high min.). These are good for analog signal switching, if on resistance about
100 Ohm is OK.Fast and fairly low capacitance to the driving signal. Anyway, it is again a single component switch :-).Dimiter
------------------------------------------------------ Dimiter Popoff Transgalactic Instruments
------------------------------------------------------
John Lark>
If you connect a 2N7002 drain to a 9v battery, an led, and maybe a series resistor, then tease the gate voltage to turn on the led halfway, and open the gate, it'll stay like that for days. I figure the gate leakage current must be electrons per second.
John
It might be common practice if you can get a 2N7002 for 2 cents !
I still use discrete bipolars for driving leds, relays and the like.
Graham
I saw a design once where the guy had used VN2222s:
I don't know where to get them, and I'm not sure how to read the data sheet, but it looks like 3.3V might turn it on - somebody who knows that stuff should take a look at it. :-)
Cheers! Rich
You got me interested in this one.
SOT-23 is Fairchild's NDS7002A
Is there a p channel complement ?
Ahhhh - mosfet not jfet. How does that handle ac signals ?
Graham
R2 parallels the base, so it has Vbe across it and Vbe/R2 current through it. Therefore R1 should conduct Iload/10 + Vbe/R2 making R1=(Vin-Vbe)/(0.1*Iload+Vbe/R2). There are various reasons for including R2 such as speed-up, linearization, input voltage attenuation, and shunting reverse CB current leakage. Most of the time it is included to shunt the CB leakage current around the BE junction.
Fred Bloggs scrobe on the papyrus:
Google for "digital transistor". The BCR148 is very useful.
Rohm also do a nice line in digital transsitors. Used them once when tight for space. They do various configurations of built in resistors.
Graham
Is that habitual holdover from germanium days? Transistors don't really leak any more, and R1 would slurp that up anyhow. Ironically, big power mosfets can have serious d-s leakage nowadays.
John
If you can drive the capacitance, they're fierce. I use them to switch
50 volts into 50 ohms in something like 2 ns. They have bigger capacitances than a jfet, but jfets don't have transconductances anywhere near these gadgets.Next step is a gaasfet, then GaN.
John
[...]
Its not an exact P channel complement but you may want to look at
Have something to add? Share your thoughts — no account required.
Ask the community — no account required