Hello John,
Like, how many popes named Pius were there before Pope Pius the 12th?
These usually didn't even make it into my office. Credentials are very low on my radar scope when hiring. With very few execptions I found that excellent credentials are in no way an indication of true performance. Much less so than callusses.
Considering what comes out of academia these days that would almost warrant a Nobel prize nomination.
Regards, Joerg
Duh.
Silicon, germanium, or gallium arsenide transistor?
Collector and emitter are equal and trivial to a first order (Ve/1k). Base is indeterminate until you give me beta.
It ain't gonna oscillate as shown, even with the +5 at nearly zero impedance IF the +10 is zero impedance and that dotted line from +10 to collector doen't have little curlies on it.
Duh.
Silicon, germanium, or gallium arsenide transistor?
Collector and emitter are equal and trivial to a first order (Ve/1k). Base is indeterminate until you give me beta.
It ain't gonna oscillate as shown, even with the +5 at nearly zero impedance IF the +10 is zero impedance and that dotted line from +10 to collector doen't have little curlies on it.
Try it with a 2N2219 and get back to me.
John
your
Sorta like stepping on an upside-down DIP16 barefoot.
John
case
your
your
had
the
review
Socks don't make much difference.
Best regards, Spehro Pefhany
case
your
your
had
the
review
40 pins seem to be more common under my bench. That's why I won't go out to the shop barefoot. They are hard enough to pry out of my shoes, and with circulation problems, it wout take months to heal.Don't have some in the stockroom or in the engineering junkbox stock. What's special about the 2219 and will anything else exhibit this behavior?
Jim
I'm guessing lots of transistors will oscillate as emitter followers. The first time I saw this, I was doing a powerup-reset circuit for a TTL system. +5 went to an R-C, then an emitter follower, then a TTL gate. It wouldn't go high at the gate input, and we eventually figured out that the transistor, a 2N2219, was oscillating violently at about
100 MHz.The 2219 is a fast TO-5, and the relatively long wire bonds probably figure in.
Really fast transistors, like an NE68039 or whatever, usually oscillate as emitter followers or differential pairs. Adding a bit of series resistance at the base, 33 ohms maybe, generally fixes it. Collector resistance can stop it too, but usually a lot more ohms.
Power mosfets tend to do this too, and the fix is a bit of gate resistance. An opamp driving a power mosfet with an ungrounded source will often oscillate too... the opamp output doesn't look resistive enough.
I recently did a current source with a small PNP transistor, BCX71 as I recall, inside the loop of an opamp. Yup, it oscillated. Rather than kluging a resistor into a very dense circuit, I substituted one of those "digital transistors", the ones with integral base resistors.
John
You used a digital transistor in an analog circuit? I'm appalled or aghast, whichever is worse. ;-)
Best regards, Spehro Pefhany
Y'know, the reason I've probably never seen it is that my first solid state grad school class was taught by an old warhorse by the name of Bill Brown at San DIego State. He told us that the transistor case parasitics would cause an emitter follower to oscillate because of the inductance inherent in the large transistor cases at the time. He cautioned us always to use a hundred ohms or so in series with the base lead and we'd never have to worry about it. I just took his advice, used the series resistor, and never had the thing sing on me.
Just for giggles, I think I'll gin one up and use purposely long (1" or so) leads on an RF device I've come to love (2N5770) and fire up the spectrum analyzer to see if and where it goes bananas.
My original comment stands, though. Unless that +10 supply with the dotted lines going to the collector has some curlies on the lines (inductors) that sucker will NOT oscillate.
Jim
Aghast hurts a lot more.
John
So you are saying that I'm either lying or stupid. So why do you continue to use base resistors?
John
I'm saying neither. I continue to use base resistors so I don't have to worry about whether or not the collector is AC grounded to the center of the earth. I also use them for the same reason I throw a pinch of salt over my shoulder every morning -- it keeps the giraffes away.
You don't need to get testy with me.
Jim
I'm not the slightest testy, I'm just trying to understand the logic of your position. As I see it,
Your EE prof told you that emitter followers tend to oscillate.
I told you that emitter followers tend to oscillate.
You use base resistors to keep your emitter followers from oscillating.
You deny that emitter followers tend to oscillate.
Just trying to understand.
John
... unless you take pains to see that the collector is truly grounded for AC.
... unless you take pains to see that the collector is truly grounded for AC.
... unless you take pains to see that the collector is truly grounded for AC.
Just trying to say that ... unless you take pains to see that the collector is truly grounded for AC ... the sucker WILL oscillate.
Jim
Except that the better you bypass the collector, the harder it oscillates.
The reason most people don't see this is probably because it's unusual to drive an emitter follower from a very stiff signal source; just a little loss in the base will usually kill the squirrels.
Slow SOT-23s will often be stable, presumably because they have short wirebonds. But you can buy 5+ GHz npn's cheap nowadays, which more than makes up the difference.
John
Hello John,
Any favorites in that category? Under 5c? Always on the lookout for a bargain :-)
Regards, Joerg
My products are expensive, and I don't buy in huge volume, so I don't usually sweat a few cents. I like the NEC parts, because they're good and available. The NE680xx (10 GHz) parts cost roughly 50 cents at
100. I like the BFS17, because it's fast but no so fast that it's hard to tame; it's about 20 cents in 100s.There must be dirt-cheap RF transistors around. Actually, I do fast time-domain stuff, not actual RF (sinewaves are boring) so I haven't explored this terrain much.
John
Hello John,
50c is way above the pain threshold for most of my stuff. The BFS17 can be had for under 8c. I always keep lots these on hand but it's tough to net under 1nsec with them.Infineon has a lot but their distribution system, well, let's just say I am not enthused about that.
Another thing I am looking for is a cheap enhancement n-channel, like the 2N7002 but 1/5th or less of its silicon area so it has less capacitance. Cheap meaning under 5c or at least under 10c. Many of the RF FETs of the olden days are becoming hard to find because RF goes BJT these days.
Regards, Joerg
Have something to add? Share your thoughts — no account required.
Ask the community — no account required