How does a transistor amplify current or voltage?

[BrownOut]

Yes, I know that. The coil I was using in the simulation last night has 2 ohms of resistance. That really helped to reduce the current through the diode(s). Keep in mind, I did not design that circuit, I am only trying to analyze it.

Even at that high value, the transistor still burns up. But it's naive to think the coil would protect the transistor anyway. Trying to get a high enough value works against other circuit goals.

That is probably true. I will leave it to MrAl to confirm it.

Absolutely true. No confirmation required.

EDIT: The other "diode" is an arrow indicating the dirction of current. (Yeah, yeah I know the direction of current is already implied, but that's what it is nontheless)

Yes, I already said previously that transistor saturation is what turns the diode(s) off.

Then there should be no reason to ask why I don't consider the diode. That is the answer.

 

[tvtech]

Ratch....do us all a huge favour......keep your interpretations of junctions to yourself...........Thirty four pages of stuff that is basically a waste of time. All the knowledgeable guys here have posted here and there. They have been kind to you. They have had the patience to discuss stuff with you.

They have explained stuff to you. They have proven stuff to you. Note the word YOU.

How do You feel about it???

If I were you I would never mention the subject again.

I can help with CRT TV repairs though. Chill.

Cheers

 

[Ratchit]

Brownout,

Even at that high value, the transistor still burns up. But it's naive to think the coil would protect the transistor anyway. Trying to get a high enough value works against other circuit goals.

A lot depends on how long the transistor is in saturation and what its current is. Like I said before, the diode(s) are just as vulnerable when the transistor is in cutoff.

Absolutely true. No confirmation required.

EDIT: The other "diode" is an arrow indicating the dirction of current. (Yeah, yeah I know the direction of current is already implied, but that's what it is nontheless)

There are two listed names, LED and iLED, which might be a superfluous current symbol, or a incomplete diode symbol. Only MrAl knows for sure.

Then there should be no reason to ask why I don't consider the diode. That is the answer.

I don't understand, the diode is a integral part of the circuit. When I get the simulation going, I will remove the diode and see what happens.

Ratch

 

[MrAl]

MrAl,



Let us both quit postulating about what happens, and try to determine what is going on from a circuit analysis program. It has been years since I used one, but yesterday I downloaded LTSpice free for nothing from Linear Technology's Website. I spent a good part of the evening installing it, and getting up to speed using it. I was mostly using generic models that came with the program, and the result were not too good. I showed a startup oscillation of 4 or 6 cycles before Q2 stabilized in the cutoff state. This causes the diode(s) to fry, unless I add some resistance to the coil.

So, what I need from you is some information.

1) What resistance are you using for the coil?
2) How many diodes are there? Your schematic shows a LED and a iLED, which I infer is a infrared LED. Would you clear that up?
3) The program came with a PSpice model of a 2N4403, but not a 2N4401. So could you send me the PSpice models of the components you are using, or tell me where I can download them.

I want to duplicate as close as possible what you are doing. Then when I get to run satisfactorily, we can compare notes.

Until I hear from you.

Ratch

Hi again Ratch,


I read your post a while back and it talked about two diode drops. Where are you saying these diode drops are?

iLED simply indicates "the current through the LED" instead of say i(LED).
vLED would mean "the voltage across the LED".
It's just shorthand, and other examples are say
iR1 for the current though R1,
vC2 for the voltage across C2,
etc., it's like camelback but starts with a lower case letter to indicate the type of parameter.

The resistance of the real life coil could vary from say 1 ohm to 10 ohms, i'll try to measure the actual value
as soon as possible. It's strange that i didnt do this already, or else i did it way back when and forgot to
log it with the schematic. I am using 5 ohms with my inductor value of 100uH.
That schematic BTW is of a real life circuit in a commercial product named, "The Brinkmann Long Life LED Flashlight".
The transistors are marked as shown in parens, but i substituted those two '2N' transistors because they are more common.

The models required are for the white LED Nichia, the 2N4403, and the 2N4401, and i thought that those two transistors came
with LT's Switchercad, but i'll post all the models here now:

.model 2N4401   NPN(Is=26.03f Xti=3 Eg=1.11 Vaf=90.7 Bf=4.292K Ne=1.244 Ise=26.03f Ikf=.2061 Xtb=1.5 Br=1.01 Nc=2 Isc=0 Ikr=0 Rc=.5 Cjc=11.01p
+ Mjc=.3763 Vjc=.75 Fc=.5 Cje=24.07p Mje=.3641 Vje=.75 Tr=233.7n Tf=466.5p Itf=0 Vtf=0 Xtf=0 Rb=10 Vceo=40 Icrating=600m mfg=Fairchild)

.model 2N4403   PNP(Is=650.6E-18 Xti=3 Eg=1.11 Vaf=115.7 Bf=216.2 Ne=1.829 Ise=58.72f Ikf=1.079 Xtb=1.5 Br=3.578 Nc=2 Isc=0 Ikr=0 Rc=.715 Cjc=14.76p
+ Mjc=.5383 Vjc=.75 Fc=.5 Cje=19.82p Mje=.3357 Vje=.75 Tr=111.6n Tf=603.7p Itf=.65 Vtf=5 Xtf=1.7 Rb=10 Vceo=40 Icrating=600m mfg=Fairchild)


.model NICHIAWHITE20MA D(Is=5.12809n Rs=2.25 N=8.1397 Cjo=42p Iave=20m Vpk=5 mfg=Nichia type=LED)

Also, the battery had to be dropped to 2.5v in Switchercad for some reason, and i added 0.2 ohms of series resistance
just to get close to real world values.

As another handy circuit addition:
Because this circuit bangs the LED with pulses rather than a smooth DC current (to reduce parts count eliminating
a diode, cap, and whatever else) it is very handy to be able to measure the average dc current rather than the peak
pulsing current. Adding a 1 ohm resistor in series with the bottom leg of the LED, then another 1k resistor from
the LED cathode (lower leg) to a 0.1uf capacitor, other end of cap to ground, then monitoring the voltage
across the capacitor and doing a
simulation of up to 0.005 seconds. The voltage across the capacitor after a time equals the average current flowing
through the LED, with a little ripple where approximately the center (vertically) of the ripple is very close to the
true average voltage which is calibrated (because of the 1 ohm) to equal current exactly...thus 20mv equals 20ma.
If you feel that that changes the circuit too much (although really it doesnt) then you can use a 0.1 ohm resistor
from the LED cathode to ground but you'll have to scale in your head the voltage to get the current...ie 2mv now equals
20ma. It's a little easier with the 1 ohm resistor to sense current.

Note you can copy/paste those model statements right into the respective Switchercad files located under "...\lib\cmp\"
and next time Switchercad is opened those models will be available too.

I am happy to hear that you are interested in this circuit BTW. It's got educational value i believe that everyone should
look at at least once in their career. For illustrating my point about collector current though, we could have resorted to
the simpler one transistor forced base voltage circuit, but heck this boost circuit makes it more interesting anyway

 

[Ratchit]

tvtech,

Ratch....do us all a huge favour......keep your interpretations of junctions to yourself...........

Back again so soon? How can I do that when I am in the middle of a oscillator circuit discussion?

Thirty four pages of stuff that is basically a waste of time.

Thirty-five now. Just can't stay away, can you? Must be interesting. Or is someone forcing you to read this thread?

All the knowledgeable guys here have posted here and there. They have been kind to you. They have had the patience to discuss stuff with you.

And I thank them and appreciate it.

They have explained stuff to you. They have proven stuff to you.

Proof it may be, but the conclusions are still in the eye of the beholder.

Note the word YOU.

Yes, it has three letters, doesn't it?

How do You feel about it???

As I wrote before, thankful and appreciative.

If I were you I would never mention the subject again.

Which subject? We discussed so many in this thread.

Ratch

 

[Ratchit]

MrAl,

Thanks a load of a lot for the info you sent. I am going be a little busy this weekend, so I might not get back to you right away. And I still have to take it slow due to my unfamiliarity with the program. But I will work on it, and get back to you with my results and analysis, or if I have any other problems.

Ratch

 

[BrownOut]

A lot depends on how long the transistor is in saturation and what its current is. Like I said before, the diode(s) are just as vulnerable when the transistor is in cutoff.

Running the transistor many times over it's Absolute Maximum Rating is just dumb. Nobody who has an inkling of electronics knowledge would support it. Small tranisitors aren't designed to be operated that way. Aside form all the bad stuff that will happen by disspation over the rated power, the transistor is designed to switch 500mA maximum. Why would you epxect it to be able to switch 5A? The facts overwhelm the fantasy that it can survive by staying in saturation. You need a better theory, one that supports survivablity of the component. Your's does not.

The single LED in the circuit is not as vulnerable, as I've already debunked. After you're shown the correct information, you keep repeating your fantasy false misconceptions. It doesn't make your case, just shows your lack of factual knoweldge.

There are two listed names, LED and iLED, which might be a superfluous current symbol, or a incomplete diode symbol. Only MrAl knows for sure.

"i" in iled stand for current. Good lord, this is so basic. The only thing you seem to know is to reject every simple, basic and painfully obvious fact.

I don't understand, the diode is a integral part of the circuit. When I get the simulation going, I will remove the diode and see what happens.

It's basic electronics. Just because something is part of a circuit doesn't mean it's vulnerable. You must consider how it operates.

 

[tvtech]

@ Ratchit.....gotta say you are one of the most decent people I have met on the Internet....

I don't agree with your thinking.....as many others don't....heck you have manners though.

Cheers

 

[MrAl]

Hello again,


For what it is worth, i found out the reason why the circuit would not work with a battery of 3v (but it would with 2.5v) and that is because the 2N4403 transistor in the two different environments had vastly different forward betas...as much as 4 times different. Thus for the Switchercad environment the 3.5k resistor has to be lowered and the other bias resistor raised to 15k instead of 10k. Then it works ok at 3v too.

 

[Ratchit]

MrAl,

I finally got the circuit to oscillate. You are right about dropping the voltage to 2.5 volts. I will look into the reason you say it is so. Also, nothing happened when I had the coil set to 2 ohms. But when I changed it to 5 ohms, away it went. The resistance of the voltage source seems to work OK at zero ohms, but I changed it to 0.2 ohms anyway. I get a pulse with an on time of 5 µsec with a average peak of 95 ma, and a off time of 1.5 µsec. The times that Q2 goes into saturation tracks beautifully with the off time of the LED.

Now the problems. You were right, there was a 2n4401 with Switchcad. I don't know how I missed it. The parameters match what you sent, too. I installed the diode model into standard.dio file with some difficulty. But, the program does not see it no matter how hard I try to select it. Even though standard.dio has been changed, switchcad still "remembers" the way it was when the Nichia LED was not there. I am currently using a different LED than the one you are using. If you have any idea what is wrong, let me know.

Tomorrow, I will be busy with other things again, but I will continue on as soon as I can.

Ratch

 

[MrAl]

Hi there Ratch,


You can try closing out Switchercad, then modifying the file with the new line with the new LED, saving it, then opening Switchercad up new again. It should show up if it worked right. Alternately, if you cant get this to work i can switch to another LED made by Nichia which comes with Switchercad, it's the one with the "500" suffix. That one might be ok too just for some tests. It's made by Nichia also and that's listed in the Switchercad parts list with the diodes and LEDs, and the last number digits are '500' and it's the only LED made by Nichia and has 500 in the part number.
Of course once you restart Switchercad you have to go back and manually change the LED to the new one by selecting it from the parts list for the components.

I measured the actual circuit inductor and it came out to 1.4 ohms exactly. I am still going to use 5 ohms though because we both get that to work and it's still a reasonable value. There will be some differences between the real life circuit and the spice circuit anyway.

Take your time no need to rush this

 

[BrownOut]

Again, the sequence of events is as follows:
1. Q2 turns on.
2. Current rises in inductor
3. Current exceeds that which Q2 can hold with it's current bias
4. Vce voltage starts to rise
5. Vcb gets higher, transistor comes out of sat.
6. Voltage rises Vce
7. Rising voltage couples through cap to Q1 base, turning it off quick
8. Q1 turning off turns off Q2
9. Q2 turning off gives rise to inductive kick back and high boosted voltage Vce

If you dont agree with that sequence of events then list your own idea of how you think it happens. You really should do a circuit simulation first though because it looks like you havent done that yet.

I have a slight issue with #6. I believe as collector current increases, the bjt comes out of saturation, and so the collector current changes from a linear rising one to one which is a constant, controlled by base current. The change in current is reflected in the inductor voltage: VL = L*dIL/DT. The rapid change in inductor voltage is fed through the cap, and the rest is as stated. The curcuit might oscillate without the rapid change in inductor voltage, but this makes the transition much more rapid.

What do you think? Mr Al? Jony130?

PS This happens too fast to be captured by the simulation. If we tried with and without an inductor, it might shed some light. I just got back from a sun-burning weekend at the lake and too tired to try it. Maybe later this week...

 

[Jony130]

Hmm,
When Q2 is "ON" his base current remains constant.
And at the beginning, Q2 is in saturation and almost all supply voltage is apply to choke. So current in choke start to rise up progressively but Ib is remains constant.
But Ic increases as time increases. The operation point of a Q2 moves up the curve from point A to B. So Q2 comes out of saturation and in the same time Vce voltage rise to.

The voltage on C1 is (Vcc - Vbe - Vce_sat) so increasing Vce will cause Q1 base voltage to increase. This reduce the base current. So Q2 will enter into active region even faster, becaues Vce increasing (positive feedback).
So I think is all up to BJT and IB when transistor start comes out of sat.
And i think we have very similar situation in blocking oscillator.

http://www.covingtoninnovations.com/michael/blog/0502/index.html#050203

 

[Ratchit]

MrAl,

I tried everything to make SwitchCad to accept the model diode you sent, but to no avail. And of course SwitchCad has been restarted many times. I am going to have to send the problem to support and see what they say. So I used the diode NSWP500BS. Now the circuit oscillates fine at 3.0 volts and 1.4 ohms in the coil, so I went with that.

Here is my analysis. Starting with Q2 being saturated and ON, the current through the coil increases, and the north side of the coil is positive with respect to the south side. The capacitor is also being energized through the base circuit of Q1, so that its west side is more positive than its east side. When the 3 volts can not longer sustain a rising current in the coil, the back voltage drops to zero and the 3 volts plus the voltage of the capacitor are applied to the base of Q1. As you can see, it peaks at 5.5 volts and turns Q1 and Q2 OFF. The two transistors remain OFF until the base voltage of Q1 drops to 2.2 volts, which forward biases Q1 and runs Q2 into saturation again. The reason for this drop is because the capacitor is being discharged through R1.

I think that the Ic of Q2 does not influence its coming out of saturation, because the max current (130 ma) of Q2 is well within the saturation range of a 2N4401, and its Vce change is negligible. See the curve of https://www.electro-tech-online.com/custompdfs/2010/08/2N2F2N4401.pdf .

Ratch

 

[Jony130]

MrAl,
When the 3 volts can not longer sustain a rising current in the coil, the back voltage drops to zero.
Ratch

But how it is possible that ideal coil connect to the voltage source can not longer sustain a rising current?


PS look at this simulation (3 to 5 and gen2)

For Ib=340uA--->Ic stop rise at 34mA
And in the oscillator Ic stop rise at 33mA , very strange coincidence?

 

[Ratchit]

Jony130,

But how it is possible that ideal coil connect to the voltage source can not longer sustain a rising current?

It is not an ideal coil. I modeled 0.2 ohms for the voltage, 1.4 ohms for the coil, and the transistor has resistance too. That limits the current. So the back voltage of the coil will be higher at first and less later. Therefore a positive voltage will be periodically applied to the base of Q1 via the capacitor to shut it off and bring Q2 out of saturation.

Ratch

 

[MrAl]

Hi again,


Brownout:
Yes that's right i think. I wrote that list up a bit too quick perhaps. I'll take another look once we have this other issue (about the collector current influence) is settled.

Jony:
Yes that's right too i believe, as the transistor collector current isnt the only variable, the base current is too.

Ratch:
Let me quote you just so we know what we are talking about here:

"I think that the Ic of Q2 does not influence its coming out of saturation,
because the max current (130 ma) of Q2 is well within the saturation range
of a 2N4401, and its Vce change is negligible.
See the curve of https://www.fairchildsemi.com/ds/2N%2F2N4401.pdf ."

You have to realize though that you can not make a statement like that when you talk about saturation, ie collector current 'only'. Yes, 100ma is within the 'range of saturation' but then so is 10ma, and 1ma, but without the proper drive we still cant obtain (nor maintain) saturation.
In other words, we have to specify collector current and also base current to tell if the device is in sat.
Example:
100ma collector current and 10ma base current, transistor is (most likely) in saturation.
100ma collector current and 1ma base current, transistor might be in sat.
100ma collector current and 100ua base current, transistor probably is not in sat.
100ma collector current and 10ua base current, transistor is most definitely not in sat.
100ma collector current and 0.00 base current, transistor is not in sat.
So it's a matter of BOTH Ic and Ib, not just Ic.

I think what Jony was saying is that for an ideal coil (as in simulation) the coil can support any voltage you throw at it so there's no reason for that coil to force a zero voltage (across it) condition. Also, you have to keep in mind that even without that feedback circuit (Q1) the transistor comes out of sat, and that's with no feedback whatsoever.

Is this starting to make sense now?

I realized that another way to think about this is a simple circuit where we are driving a relay with a transistor...a very common thing. We want to provide enough base current to *keep* the transistor in saturation even with some temperature variation. The key word there being "keep". What is there to keep? It's saturation. We have to 'keep' it in sat because if we dont it will pull out if the beta (due to temperature change) drops lower than we thought could occur. As example, 1ma base current and 100ma collector current might work for a while when the beta is high, but as the transistor either cools off (due to ambient change) or heats up (ambient or self heating) the beta could drop (upside down skewed parabola of beta vs temperature).

 

[BrownOut]

transistor has resistance too.

Some resistance? It would need around 90 ohm for the current to stop at 33mA. That would be quite impossible, so the whole analysis is bogus. Here is your hint, from Jony130:

For Ib=340uA--->Ic stop rise at 34mA
And in the oscillator Ic stop rise at 33mA , very strange coincidence?

Of course it's not a coincidence. This is how it really works.

 

[BrownOut]

Jony130: I think for the resistive load you'll find that Q1 comes out of saturation first. Q2 really has no reason to come out, since the load is resistive, ICQ2 won't rise with 'on' time. The feedback current that keeps Q1 in saturation eventually dies down, and Q1 enters active mode, Q2 will quickly follow. The built up change in the capacitor added to Q2's sudden rise in collector voltage then turns off Q1. The coil makes the situation a little different. At least that's the way I think it works. If course it depends on the coil's charge time vs capacitor change time, and a few other factors. If Iwasn't already tired of the subject, I'd try to present a more complete analysis.

 

[MrAl]

Hi Brownout,


Sorry to hear you're getting tired of the subject. I havent looked at the resistor only circuit for the oscillator, but i do know that any transistor will pull out of sat if the collector current rises too high. I guess we'll have to see if that happens with a pure resistance too. I think it would because with no feedback if the resistor is of a low enough value the transistor will come out of sat. So it may depend on the value of the resistor.