Base Emitter Voltage Drop over Temperature.

[rentner]

Just wondering: I divided a voltage of 40V with 56k and 1k to a voltage, where a BC547C turns on and gives a signal, to stop giving more voltage through a Buck Converter. Nothing realy interesting happening here.

Code: [Select]


5V
|
|
40V      Collector
| |
| |
| 10k
56k |_______some other circuits
| |
|_____100k____Base
| |
| |
1k |
| |
| |
GND____________Emitter





I hope, it does show correctly.
The Problem is, that I don't know, what exactly will happen, outside of the simulation. There is the problem, that I have to replace the 56k resistor with a 100k poti, because the diode BE neveris identical, right? Also, what happens, if the Temperature changes? Will that change the circuit much aswell? If so, does it mean, that I have to replace that part of the circuit with some kind of discrete OpAmp, to make the voltage control stable?

[Rufus]

If you are simulating with spice sweep the temperature and see what happens.

Vbe will drop about 2mV/C. The 2mV depends on the collector current. It changes with the log of collector current at about 200uV per decade.

[rentner]

I am using LTSpice. Simulating -20 0 20 40 60 80 100 120.
10 ms of Simulation time.
I guess I can wait at least 1 or 2 hours now, to run through all.
I hope, that my PC will not crash, when showing the values.

[edavid]

The Problem is, that I don't know, what exactly will happen, outside of the simulation. There is the problem, that I have to replace the 56k resistor with a 100k poti, because the diode BE neveris identical, right? Also, what happens, if the Temperature changes? Will that change the circuit much aswell? If so, does it mean, that I have to replace that part of the circuit with some kind of discrete OpAmp, to make the voltage control stable?

You are on the right track.  If you want an accurate trip point, you need a comparator and a voltage reference.  This is why people use ICs.

If it doesn't need to be that accurate, you can use a zener diode to increase the trip point.  This reduces the dependence on VBE.

[rentner]

Simulated that now. Have variations of the voltage between 28 to 45V. That is not close to acceptable. 40V+/-2V is acceptable.

Well. A comparator with precicion reference will do it. Do you have one , that is cheap, as solid as possible to not die from ESD, Through hole?

It doesn't need to be as precise as 0.001%, because the Buck Converter will have a much higher ripple anyway. Did not find cheap Comparators for my purpouse, since I need 4 IC's, I don't want 1€ per IC. The problem is, it cannot use a Quad one, since all Comparators need to be isolated to more than 100V. Quite impossible, since 2 of them are floating quite a bit below ground.


I don't know, what do you think of LM358? They are available for low. They work with 5V single supply. Give me over 5mA Output contious which is quite enough for my purpouse. I can drive the input to at least 3 Volts, which also is enought. Gain is enought... Would you just recomment me using that?

[edavid]

Simulated that now. Have variations of the voltage between 28 to 45V. That is not close to acceptable. 40V+/-2V is acceptable.

Well. A comparator with precicion reference will do it. Do you have one , that is cheap, as solid as possible to not die from ESD, Through hole?

It doesn't need to be as precise as 0.001%, because the Buck Converter will have a much higher ripple anyway. Did not find cheap Comparators for my purpouse, since I need 4 IC's, I don't want 1€ per IC. The problem is, it cannot use a Quad one, since all Comparators need to be isolated to more than 100V. Quite impossible, since 2 of them are floating quite a bit below ground.


I don't know, what do you think of LM358? They are available for low. They work with 5V single supply. Give me over 5mA Output contious which is quite enough for my purpouse. I can drive the input to at least 3 Volts, which also is enought. Gain is enought... Would you just recomment me using that?

It's common to use a TL431 as reference and comparator in one... very cheap.

LM358 is not a good comparator... the cheap comparators are LM311 (single), LM319 (dual), LM339 (quad).

But, a zener diode would probably be good enough.

[Rufus]

Simulated that now. Have variations of the voltage between 28 to 45V. That is not close to acceptable. 40V+/-2V is acceptable.

Use a 39v zener diode instead of the 56k resistor. That will produce more feedback gain which might produce stability issues. The 100k in series with the base seems high - just makes it soggy, but, with a zener you will need something there.

Simulate that

[Paul Price]

LM358 is an op amp
LM393 LM2903  are  both 8-pin dual comparators, low power and cheapest as well

[rentner]

I simulated it with LM358 in LTSpice an just resisors. Why? Well, the Zener diodes changed over Temperature aswell and gave me huge differences of many Volts which is deadly for my application, since 40V is ok and >46 is deadly.


If the resisors heat up a bit, they both still have the same ratio, but if a Zener diode warms up, it is dramatically. A precision Zener is very expensive and inacceptable for my purpouse.