Various Transistors Tested for Ic and Vce(sat)

or skip the bias point calculation?

Depends on simulator and what unchecking that box actually does.

...Jim Thompson

In article , snipped-for-privacy@invalid.invalid mentioned...

Of course I use 'pencil and paper', or more like a calculator and computer. I see this schematic posted here,

Is that pencil and paper?

What does "forced beta" mean? Seriously! It's a rule-of-thumb... not a reality! Revert to fundamentals, stop guessing, otherwise you'll always be a hacker.

...Jim Thompson

Here is a guess, based on your hint about beta...

When charging, the rate of change of I with t is Vind/L. Vind is

When I > ß.Ib, then Vce will start to increase dramatically. That will cause Q1 to turn on, and turn off Q2, starting the 'discharge' cycle.

The discharge cycle ends when the current runs out, and the LED can no longer keep the voltage up; this causes Q1 to turn off, turning on Q1, and restarting the charge cycle.

Thats what causes the oscillations.

The period is derived as follows

Ib*ß = (Vcc - V2ce)*t/L

so

Ib*ß*L/(Vcc-V2ce) = tc

Ib is (Vcc-V2be)/R2

therefore

charge time = tc = (Vcc-V2be)*ß*L/((Vcc-V2ce)*R2)

the max current, Imax, is tc * (Vcc - V2ce)/L

discharge time = td = Imax/(dI/dt), since only when I gets near 0 will the LED drop the voltage enough to turn off Q1 and end the discharge cycle.

for discharge, di/dt is -(Vcc-Vf)/L. Thus,

td = tc * ((Vcc-V2ce)/L) / ((Vf-Vcc)/L) = tc * (Vcc-V2ce)/(Vf-Vcc)

Regards Bob Monsen

Looks like a good shot at the math.

I'll take a look late tomorrow... gotta face a design review first ;-)

...Jim Thompson

With mine, it prevents the bias point from being calculated which would put a steady current through L1, i.e., the step function wouldn't be there to make it do it's di/dt thing or (1/L).Int(v dt) That's *my* read on why a bias point calc would keep this circuit from starting.

It got my vco's going :) I usually try that before setting an .IC

Oh. Thanks for the exercises :)

Purrrfect!

AARGHHH! Finally got the damn thing to work in LT Spice. Jeez...I could have set up 10 breadboards or used an abacus to figure out the circuit in less time that it took me to simulate it. I'm sure that says much more about my lack of ability with Spice than with the program itself ;-)

cause

and

the

Just reworked the circuit a bit, turning it into your run of the mill cross coupled multivibrator with added base to ground resistors (6.8k). Everything seems to be more stable, more symmetrical, and less dependent upon individual circuit parameters such as coil resistance (up to at least 50 ohms), transistor Hfe blah blah. LED current is also more reasonable and transistor independent.

LOL! The abacus and the breadboard with 2000 hand drilled holes on tenth inch centers! It's a good thing we have people like Mike and Helmut around S.E.C. to help with LT Spice. Kevin's and others are adept at Spice, too.

With PSpice you have guys that can help, but if it's Orcad's interaction with PSpice (or anything Orcad), it becomes somewhat of an arcane ritual involving incantations, pidgeon entrails, and cloaks made of virgin hair ;) But help is still there even if you just need help getting the friggin' candles lit :)

I hear it coming... "I don't use Orcad ;-) ...Jim Thompson"

I don't use Orcad ;-)

If you are a PSpice user, choose Custom Install, and select PSpice Schematics... *much* more intuitive than Capture.

...Jim Thompson

It installs with Orcad and I haven't tried it. Let's see...

There's no pdf docs but a glance at the help topics and menus has revealed that it does pretty much the same stuff. PADs netlist, BOM. No back annotation... not too big of a deal.

Uh, I'm getting tired of kicking myself in the ass, Jim. Care to take over?

XFY34 for example - but what is this 6.3V pentode EF76 ? EF80 and 85 here in profusion!

-- Graham W

You could use three pin-driver chips as used in IC testors, and with a little bit of programming you can just plug in any transistor (the thing could even be programmed to recognize FET's, regulators and UJT's) and it would figure out polarity, what pins are E, B, and C, and give you characteristic curves. If you mechanically/themally attach the transistor to a combination Peltier cell/small high-power resistor, you can get curves involving varying temperature as well.

-----

I read in alt.binaries.schematics.electronic that Graham W wrote (in ) about 'Various Transistors Tested for Ic and Vce(sat)', on Mon, 5 Jan 2004:

It's a wire-ended valve/tube. You identified the basic spec from the type number. Look how much more helpful that 'EF76' is than 'XFY34' or '6245', for example.

However, a GK32 is not an octal-based heptode with 5 V heater. (;-)

Why shouldn't John (a.k.a. Watson) just hack it. It's his hobby, not his job. (And we're not in s.e.d, so tapping your MIT Honors ring isn't going to mean squat).

Mark Zenier snipped-for-privacy@eskimo.com Washington State resident

Like what? Is that 3 "pin driver" chips or "3-pin driver" chips? I suspect the former.

Yeah. While we're at it, we might as well make it a full blown curve tracer like the IC tester. Why stop short? And since he's testing these things in a boost-converter circuit, not a tranny tester, it could be self-testing, giving all the relevant circuit performance data.

We'll call it the TIRE Analyzer.

Tranny/IC/Reglator/EDN_Circuit[_analyzer] Analyzer. 'tis the age of the acronym, after all.

In news: snipped-for-privacy@news.east.earthlink.net (Active8):

I'd love to see the schematic/code for that beast! :)