Current limiter Circuit

Hello all,

I am trying to build an adjustable current limiting circuit. The supply is fixed at 5V, but can supply between 0-50A, I want to limit it to a maximum of 48A, but also have it adjustable between 0A and 48A.

So here are the specs:

Vs = 5V
Imax = 48A

If I could use BJT's, Rsense was going to be 2 resistors, one 0.015 Ohm 50W resistor, in series with a 1 Ohm 50W adjustable power resistor.

I have found an IGBT with the following specs

Vces = 600V
Vce(sat) < 2.6V
Ic (100C) = 55A, Ic (25C) = 100A
Vge = +- 20V

I have this diagram (below) but it uses BJTs and I cant seem to figure out how to modify it for using IGBTs because I have never used them before.

**broken link removed**

Normally I would start by choosing a transistor that can handle the Vs and Imax, then calculate the base current of Q1 using Imax and Hfe but there is no Hfe with IGBTs? I am at a loss since I will need to use IGBTs. I need some help!

I was also going to use some large resistors and adjustable power resistors, around 50W, for Rsense.

If you guys have any better way to have an adjustable current limit of 0A to 48A with a fixed 5V, than this circuit let me know!

The current sensing method of the circuit you show turns on Q2 when the voltage drop across Rsense reaches about ~0.65V. At 48A, Rsense is dissipating a whopping 0.65*48 = 31W. Since the drop is ~0.6V, it will screw up the regulation of the 5V supply unless your supply can remotely sense the load voltage, and you can connect the remote sensing downstream of the current limiter.

The Vbe at which Q2 turns on is strong function of temperature, so this is a pretty crummy current limiter. Besides, regardless of what you use for Q1, if a current overload is detected by Q2, Q1 limits the current by turning partially off, meaning that the dissipation in Q1 could go very high... Imagine that you short the supply. The current is limited to ~50A, and now the entire 5V from the supply appears across Q1, meaning that it is dissipating 250W!!!

A much better solution is to buy a supply that includes current limiting built in. For example, I recently bought some used MeanWell 650W 15V switchers that have pot settable output voltage and pot settable current limits on Ebay for $30.

That would be a really bad limiter. The changing temp of the transistors would shift the limit point all over the map.

Can't you alter the circuit of the 5 V 50 A supply itself? If it's a SMPS, it'll have current sense function in it all ready to protect from shorts, and there's also a voltage control loop, which is easy to find looking at the PCB. You could sense the current outside the PSU and feed back a voltage to the PSU's voltage control circuit, so it turns down the voltage.

Otherwise you'll have to have massive heatsinks, power fets (don't use BJTs) and a low dropout current sense. And even then after you do that, you'll need to get it "stable" when it's current limiting and not an oscillator!

MikeMl said:

The current sensing method of the circuit you show turns on Q2 when the voltage drop across Rsense reaches about ~0.65V. At 48A, Rsense is dissipating a whopping 0.65*48 = 31W. Since the drop is ~0.6V, it will screw up the regulation of the 5V supply unless your supply can remotely sense the load voltage, and you can connect the remote sensing downstream of the current limiter.

The Vbe at which Q2 turns on is strong function of temperature, so this is a pretty crummy current limiter. Besides, regardless of what you use for Q1, if a current overload is detected by Q2, Q1 limits the current by turning partially off, meaning that the dissipation in Q1 could go very high... Imagine that you short the supply. The current is limited to ~50A, and now the entire 5V from the supply appears across Q1, meaning that it is dissipating 250W!!!

A much better solution is to buy a supply that includes current limiting built in. For example, I recently bought some used MeanWell 650W 15V switchers that have pot settable output voltage and pot settable current limits on Ebay for $30.

Click to expand...

The meanwell, that is what I got, from eBay (350W), but it doesn't have any sort of pot for current limiting on it, only voltage, the pot says V+ ADJ above it.

I couldn't find one on eBay that has any current adjustment either.

marcbarker said:

And even then after you do that, you'll need to get it "stable" when it's current limiting and not an oscillator!

Click to expand...

Actually if you can make it oscillate at a low frequency with a low duty cycle when short circuited, then oscillation is desirable because it cuts the power dissipation.

adaminc said:

The meanwell, that is what I got, from eBay (350W), but it doesn't have any sort of pot for current limiting on it, only voltage, the pot says V+ ADJ above it.

I couldn't find one on eBay that has any current adjustment either.

Click to expand...

Have you downloaded the Spec Sheet for the supply you have? All of the MeanWell switchers I have used are intrinsically current-limited. Some have an adjustment; others just limit just above the max design load current.

Hero999 said:

Actually if you can make it oscillate at a low frequency with a low duty cycle when short circuited, then oscillation is desirable because it cuts the power dissipation.

Click to expand...

Yes, good point. Kind of 'switch-mode' current regulation. Trouble is of course if it's oscillating 'flat-out', the regulation accuracy gets a bit 'non-linear', and there may be some severe EMC issues to sort out too! Way round that is to ensure the oscillation is well defined.

I gave an example in the h-bridge thread.
https://www.electro-tech-online.com/threads/basic-problem-with-transistors.95681/#post774423

I can see a current-limiting function there. Is this a self-oscillating PWM?

Yes, the comparator acts as a Schmitt trigger oscillator when the current limit is exceeded due to the RC constant and hysteresis.