i give up.. lm317 + 2n2222

[Arouse1973]

Actually...

The transistor option would have been a cheaper substitute, the main goal was to use a cheap pot in conjunction with transistor vs higher rated potentiometer. .

i've basically built what i need right here, but it's cheaper to use a 20c transistor and 20c pot rather than using a $3 potentiometer between out and adj on an lm317 in current regulation mode..

anyway, you show me how to use a cheap 20c pot without burning it on an lm317 cc circuit (that's the original goal)

I thought that's what I posted for you in post #16? You haven't even acknowledge if this is suitable for you.
Adam

 

[AnalogKid]

The classic problem with the classic 317 CC circuit is that all of the output current goes through the resistor. A nuisance for a fixed output, but for variability that resistor becomes a low resistance, high power pot, not a cheap part and not a good approach for long term reliability. OTOH, the circuit in post #20 seems a bit complicated for the task. It looks like 100% of the controlled current goes through the LED. If that is the only goal, there are better ways, including not using a 1.5 A power regulator where a 2N2222 will do. What is the output current range you need?

ak

 

[cjdelphi]

Looks like you guys posted while i was designing my own circuit, this is the first i've seen of these posts, i'll check them out now, thanks!

 

[Arouse1973]

Looks like you guys posted while i was designing my own circuit, this is the first i've seen of these posts, i'll check them out now, thanks!

I thought it strange you hadn't responded. I know it's very common not to get any responses from other members on a thread, why some of them even bother half the time god only knows.

Check your Alerts settings and make sure you are sending alerts out.

Cheers
Adam

 

[cjdelphi]

what happened was i got the replies while building the circuit, then when i posted the alerts vanished (and i never scrolled up to look)

anyway feeding the voltage to the base never occurred to me, duh, it's like 2 or 3c where my breadboard sits, so i'll build it in the day tomorrow.. thanks again

 

[CDRIVE]

Looks like you guys posted while i was designing my own circuit, this is the first i've seen of these posts, i'll check them out now, thanks!

It's no wonder this thread seems to be in a quagmire. Here's another design that demands very little power dissipation of Isense (R1).

Chris

 

[Arouse1973]

what happened was i got the replies while building the circuit, then when i posted the alerts vanished (and i never scrolled up to look)

anyway feeding the voltage to the base never occurred to me, duh, it's like 2 or 3c where my breadboard sits, so i'll build it in the day tomorrow.. thanks again

We have had issues with Alerts, this is something Ian is going to look at for me.

The circuit is not that obvious when you first look at it, but once you know how it works then the theory is self explanatory. You could have a switch box with different values of resistor to set the max current and have one POT to adjust it within that range.
Adam

 

[Arouse1973]

It's no wonder this thread seems to be in a quagmire. Here's another design that demands very little power dissipation of Isense (R1). View attachment 21109

Chris

Yep good one Chris. Another option there, removing the need for a large expensive resistor. However the OP did ask for a circuit that replaced the resistor with a BJT. Why I am not sure?
Thanks
Adam

 

[cjdelphi]

ok, the original (2am here now, flipping time zones) plan...

To avoid using an expensive potentiometer in regulation mode of an lm317, instead use a smaller pot (hence the idea to use a pot to drive a transistor, to allow fine tuning in current regulation. . eg delivering between 0-200ma for an LED or Laser Diode...)

CDRIVE, i like i like... thanks!

 

[AnalogKid]

The LM317 is an outstanding component, and was a minor miracle back in its day. But at its heart, it is a *voltage* regulator, and can make bending it to do what you want (other than voltage regulation) harder than a circuit grown to do the non-voltage-regulating task. If what you want is coarse maximum voltage limiting and more tightly controlled, adjustable constant current regulating, that probably can be done more easily when not having to work around the 317's personality.

Is the intended load grounded, highside, or floating somewhere in the middle (that is, with non-zero impedances to any power rails)?

ak

 

[CDRIVE]

To avoid using an expensive potentiometer in regulation mode of an lm317, instead use a smaller pot (hence the idea to use a pot to drive a transistor, to allow fine tuning in current regulation. . eg delivering between 0-200ma for an LED or Laser Diode...)

CDRIVE, i like i like... thanks!

Keep in mind that I only included Q2 to demonstrate a High Side current source. You can omit it and drive your laser diode between Vcc & the collector leg of Q1.

Chris

 

[AnalogKid]

Here's another design that demands very little power dissipation of Isense (R1).

Am I missing something, or is this a saturated switch, and Iload is not related to the 317 output current. What I see is the 317 providing 2.5 mA of base current to the first of two saturated switches in series. Did the op requirements change?

ak

 

[CDRIVE]

You're correct. You are missing something. Look at the plot again.

Chris

 

[AnalogKid]

Chris, are you proposing the circuit in post #26 as a solution to the question in post #1?

ak

 

[CDRIVE]

AK, no I'm not. The circuit simply demonstrates how Pass Transistors can be used to extend the limitations of a Current Source. CJ will have to modify it to suit his requirements. He will have to take into account the beta of the transistor or transistors used.

As I see it his issue was Rsense and the fact that Isense = Iload, which equates to appreciable power dissipation of Rsense. Since Q1 & Q2 are current amplifiers Isense is quite low, which results in low power dissipation of Rsense. Which in turn (I believe) solves his problem of necessitating an expensive high power wire wound pot.

Also, and to reiterate my earlier post. CJ doesn't need Q2. I used it to demonstrate a High Side current source. His laser diode can be driven from the low side (Q1) which is a Current Sink.

Now that I've answered your question do you still contend that my circuit is nothing more than a "switch"?

Chris

 

[AnalogKid]

No. I did look at it again, ran a calc on the 317, saw that you had it throttled back to only 2.5 mA output, and figured out that you were relying on the current gains of both transistors to set the output current. Way back when this was called "dangle biasing". I wanted to correct my post but was away from computers for a while, and you beat me to it.

But - with two transistors (or even just one) running wide open with no feedback, the output current will vary wildly with temperature, Ic, Vce, and particularly with Ib. For example, doubling the 317's Rsense (R1) will decrease Q1's collector current, but almost guaranteed not by 50%. The decrease will depend on a combination that transistor's gain curves for temperature and collector current. And by "that" transistor I mean that individual device, not just that part number.

Tone-of-voice disclaimer: No way am I attacking you or your circuit; I think this is an excellent discussion.

ak

 

[Arouse1973]

"Tone-of-voice disclaimer: No way am I attacking you or your circuit; I think this is an excellent discussion"

I am glad you said that ak, This was beginning to go pair shaped. It's so difficult to communicate via text alone, people can come across so rude and arrogant, but they are not really, I know I get it all the time at work with emails... worst thing for communication. The person can quite innocently type something the other person takes the wrong way.

It's very difficult to get right. That's why the like button is so important, it tells someone you like their post and makes it so much more easier to ask probing questions about their design without them taking offence. A few kind words of encouragement costs nothing. I think everyone could learn something from your statement. After all I hope that's what we all here for to help others and maybe pickup something along the way that we didn't know.

Thanks
Adam

 

[CDRIVE]

AK, the first thing you have to do is tell us your name. You've been here a very short time but it's becoming quite evident that you're going to be a very active and knowledgeable member.

I think I can arguably claim the we are the friendliest tech forum on the face of this earth. Many of us have become friends separated by geography. So how about spilling it?

Now as per your post... No arguments on ANY point you made regarding that circuit. The temperature stability would be about.. NADA!

Sometimes I post concept circuits not intended to be fully functional or applicable 'as is'. I would love to claim that my circuit was intended to be just that but I would be bullsh!tting you. In my defense, I never claimed this was the best solution either.

So, the ball is in your court. Gonna give CJ what he needs?

Oh, one more thing... You didn't retract your "Switch" statement.

Chris

 

[AnalogKid]

Oh, one more thing... You didn't retract your "Switch" statement.

Post #36:

quote
I wanted to correct my post but was away from computers for a while, and you beat me to it.
enquote

ak

 

[Arouse1973]

AK, the first thing you have to do is tell us your name. You've been here a very short time but it's becoming quite evident that you're going to be a very active and knowledgeable member.

I think I can arguably claim the we are the friendliest tech forum on the face of this earth. Many of us have become friends separated by geography. So how about spilling it?

Now as per your post... No arguments on ANY point you made regarding that circuit. The temperature stability would be about.. NADA!

Sometimes I post concept circuits not intended to be fully functional or applicable 'as is'. I would love to claim that my circuit was intended to be just that but I would be bullsh!tting you. In my defense, I never claimed this was the best solution either.

So, the ball is in your court. Gonna give CJ what he needs?

Oh, one more thing... You didn't retract your "Switch" statement.

Chris

Yeah I agree otherwise others might think ak stands for absolute Knob.