Have I correctly calculated the values for this current limiter?

[sebmaster]

Hi,

I want to build an array of ~60 current limiter (500mA) circuits to act as power sources. In normal operation, they will basically be electric fuses (i.e. the expected current is < the design/limited current). So, the current limiting does not have to be incredibly accurate, and cost and size are important.

I found this circuit online, but wondered if someone could check my calculations for the components, to make sure I have reasoned them correctly, and not missed anything?



I have worked it out as follows, assuming the worst case scenario (a short across the load):

Q1

The pass transistor is a 2STD1665 because it is high power, has a low Vce, is cheap and supplied in small quantities.

Thermal Dissipation

Design current is 500mA at 12v, so 6W.
+30% safety margin is 8W.

The 2STD1665 has a Junction-Ambient Thermal Resistance of 8.33 C\W and a max junction temp. of 150 deg.

At 8W that is a temp. rise of 66 degrees, giving a max. ambient. temp of 83 degrees, so no heatsink is required.

Rbias

Ib1 should be hfe*design current (0.5A) = 0.005A (actual hfe is larger, so we have headroom)

Vrbias should be Vcc - Vbe1 - Vrsense. Vbe1 is 0.85-1v, and Rsense is 0.7, so Rbias = ~10V / 0.005A = 2kohm.

Q2

Q2 only has to pass Irbias (5mA) so can be any NPN such that it (Ic2) handles this, and Vce2 < Vbe1. (e.g. BC847: Vce < 0.4)

R1

R1 protects Q2 in case the full Vcc appears across Rsense. I asked about this on another site and don't believe its essential in any conditions I can think of (that don't blow up a lot more than Q2), but since it is negligible to include, and covers conditions I can't think of, I am including it.

Vcc - Vce is ~11.9V, so R1 should be 120ohm.

Rsense

RSense actives Q2 when the current across it creates a voltage drop greater than Vce. I assume Vce is ~0.7. R1 and Q2 form the equivalent of a diode resistor pair, in parallel with Rsense. So it is strictly calculated as:
0.7 + Ib2*R1 = 0.5 * Rsense

Though at the design current when Q2 switches, Ic2 = 0.005, so Ib2 = 0.00005, so Ib2*R1 is ~0.006ohm which is <1% of Rsense, so it can effectively be ignored.

This leaves Rsense = 0.7/0.5 = 1.4ohm.

Have I gone about that the right way?

Sj

 

[KlausST]

Hi,

I recommend to use a simulation tool, like LTspice.

Klaus

 

[sebmaster]

Thanks Klaus. I have simulated it with CircuitLab, I just wanted a sanity check for things the simulator might miss (e.g. power dissipation).

(Also Vce in Rsense should be Vbe2 - I can't figure out how to edit the post to update it!)

I know I'm probably overthinking this simple circuit, but if I'm going to make 60 of them I really want to get it right!

 

[andre_luis]

I have simulated it with CircuitLab, I just wanted a sanity check for things the simulator might miss (e.g. power dissipation).

In LTspice, to get the instantaneous power you can either place the mouse over the component to be inspected and hold ALT key together with LEFT mouse button pressed. To take the average power, hold the CTRL button and click on the name of the net in the measurement window. A new window will open displaying the average power.

However, if you want to add a probe so that at every simulation its waveform is added to the graph, you can use the "voltage-dependent-voltage-source" to create a signal ( with the purpose to be measured, not to supply anything ) having other signals as its argument along with any math opperations. At the following case I used this approach for example to convert the voltage read at certain point to the equivalent current seen at uC ADC result, but to your case, you could just put at its argument the VxI product of interest.



I'm not familiar with CircuitLab, but perhaps you could do some stuff similar there.

 

[sebmaster]

Thank you andre! I have found that CircuitLab has the ability to measure power dissipiation as well: https://www.circuitlab.com/docs/expressions/

 

[Pjdd]

Q1

The pass transistor is a 2STD1665 because it is high power, has a low Vce, is cheap and supplied in small quantities.

Thermal Dissipation

Design current is 500mA at 12v, so 6W.
+30% safety margin is 8W.

The 2STD1665 has a Junction-Ambient Thermal Resistance of 8.33 C\W and a max junction temp. of 150 deg.

At 8W that is a temp. rise of 66 degrees, giving a max. ambient. temp of 83 degrees, so no heatsink is required.

I haven't had time to go through the whole post in detail but the part quoted above has a serious flaw. 8.33°C/W is the Junction-Case thermal resistance, not Junction-Ambient. No power transistor I know of has such a low J-A thermal resistance, not even the big ones with <1°C/W J-C. The TO-251/252 package is closer to 100°C/W J-A without a heatsink.

 

[sebmaster]

Thanks Pjdd. Since posting I built that circuit, and quickly discovered that I made a mistake in the power dissipation calculations! That is the figure on data sheet the data sheet though (it uses both Rtha-a as the symbol and says explicitly "junction-ambient") so it must be incorrect.