Current Limit Circuit

Ref. your item 2, there is no way a 2N2906 can supply 1 amp with 10 volt drop, which is 10 watts!! Look at its data sheet, it is a "small signal" transistor with a maximum dissipation of ~300 mW!! Circuit simulators do not always include overheat simulation. You need a transistor, say 3 amp and 30W rating, mounted on an adequate heatsink, in place of 2N2906 pass transistor. Most likely your other [item 1, 3] pass transistors are inadequate or overheating!!

With ref. to item 2 - The 2N3906 is only in sim. In real I used TIP42 for the pass transistor and TIP31 as the current sink.

http://cr4.globalspec.com/thread/61092#newcomments

I want a current limit circuit, not a charging circuit.

Not fair!! Your question wrote "I am designing current limit circuit for a battery charger".

Sorry man, the title says Current limit circuit. If I just say I need a current limit circuit, then ppl as for what purpose. So I mentioned the application.

This is a good example of trusting a SIM program without understanding the real-world aspects of practical circuit design. No current limiting frying some parts and power dissipation (heat) causing unpredictable circuit behavior. I don't normally like to do other's homework for them, but ...
Try setting up an experiment.
The red LED voltage and the emitter resistor for Q1 will determine the value of a (quasi) constant current source.
(Vled-Vbe)/(R2) = Iq1
(1.6V-0.7V)/(1ohm) ~ 0.9A
The rest of the circuit should form a "simplified" battery charger that provides constant current up to a peak threshold voltage, then turns off until the battery discharges to a lower voltage threshold. Set-points (hysteresis) will need to be adjusted for the particular supply voltage, op-amp or comparator used, and the battery type. Ideally one would also add some form of temperature compensation. Good luck.

I never said I trusted the sim. I don't and never will trust it 100%. Simulation is a good way to save time and see if an idea works. I just wanted to know why there were differences.

In the circuit you mentioned, can it be implemented without the LED. What exactly happens when the current reaches or crosses 0.9A (the set limit).

Circuit simulators can be very powerful design tools but it is good that you have some reservations. Trusting simulation results takes time and experience. Eventually you will learn what works well, what doesn't, and why.

The LED is used as a voltage source. It can be replaced by 2 or 3 rectifier diodes in series, a properly biased zener diode, or a voltage source model in most simulation programs. The voltage source + transistor + emitter resistor form a simple constant current source circuit and the example equation is good starting point for design. Instead of trying to further describe operation in text, I think you will gain a much better understanding of how the circuit works if you simulate it. This is one case where simulation can provide results that are very close to actual circuit performance. Good luck.

mjb, which schematic software did you use for the battery charger a1 circuit.

gEDA: GPL Electronic Design Automation

gschem version 1.6.1.20100214

I am using Ubuntu 10.04 (Lucid Lynx OS) and this is just one of several open source engineering design programs I am "test-driving". This is the first schematic I created with this specific software. While it was very easy(intuitive) for me to use, I don't have enough time invested yet to advise for or against this particular schematic editor.

1. About your item 2 real circuit. Have you checked it is not oscillating?? And did you consider the heat sinking, which I mentioned. Have you thought what the collector leakage current in Q2 will do??
2. Are you sure "foldback" is a good thing for a battery charger??
3. About your item 3 circuit. Have you looked at a Mosfet data sheet which shows how "gate threshold voltage" changes with device temperature?? And between one MOSFET specimen "out of the box" and the next??

[quote]1. About your item 2 real circuit. Have you checked it is not oscillating?? And did you consider the heat sinking, which I mentioned. Have you thought what the collector leakage current in Q2 will do?? [/quote]

No, yes and No. I did consider heat sinking. Instead I limited the current to 0.5Amps, which reduced the power dissipation across the transistor.

[quote]2. Are you sure "foldback" is a good thing for a battery charger??[/quote]

Good question. I don't think it is necessary. But i would like to find a good foldback design.

[quote]3. About your item 3 circuit. Have you looked at a Mosfet data sheet which shows how "gate threshold voltage" changes with device temperature?? And between one MOSFET specimen "out of the box" and the next??[/quote]

No, will check now.

Ref 2. Foldback may be useful for a power supply to protect the series pass transistor from overheating. But it can lead to problems. Suppose the load behaves like a constant current ( some op-amps are close to that), or increases current with falling voltage (like a lot of switch-mode supplies, which hold their output at steady voltage and current despite falling primary supply voltage). The load line can cross the regulator output voltage versus current curve at two points. Below the lower crossing point, a combination of increasing load versus falling supply current causes a collapse to an equilibriuum point where the load current is falling with voltage faster than regulator current. Overheat protection circuits which respond only to overheating are best, since they will not act quickly. Battery chargers are designed to current limit at a little over rated current on short circuit, they have to be rugged - if they do not work, a whole unattended system could fail, instead of just automatically recovering after a power loss and battery discharge.

Well, a series resistor doesn't need much designing.

LM317 voltage regulators can be easily configured as a current source/limit. Google "LM317 current limit" and you will see a lot of examples.

Tom D.