Power supply current limiting

[lti]

I am currently trying to build a bench power supply. I don't have any real controller ICs to use, so I need to use op-amps. I have no idea how to implement a current limiting circuit. I have found a few circuits that don't work at all in LTspice, and I don't know what the exact problem is. It oscillates for a while and then the output drops to zero.

Also, the use of a comparator in the circuits I saw confuses me. It seems like it is supposed to oscillate between 0V on the output and the set output voltage.

The circuits I found use a current sensing IC for high-side current sensing, but I was trying low-side current sensing using an op-amp. I don't think that would cause this problem.

I'm going to use a BA10324 op-amp and an IRF510 MOSFET (which is probably completely unsuitable for this purpose, but it's what I have) in the final power supply circuit. This will be powered by an 18V transformer, a bridge rectifier made from 1N5402 diodes, and some tiny (for this purpose) filter caps (a 2200uF 35V Nichicon VX and a 470uF 35V BH piece of crap). The voltage across the filter caps is 31V with no load, which will probably cause some problems when I try to build the circuit. I don't have very many components that can handle such a high voltage. The ripple from not having enough bulk capacitance will also severely limit the maximum output voltage and current. This won't be a very good power supply, but it will hopefully be good enough for a while until I get a real power supply.

I feel like I've suddenly turned stupid. I guess I wouldn't feel as dumb if I had suitable components instead of trying to make random crap work.

[stj]

Re: Power supply current limiting

http://tuxgraphics.org/electronics/2...upply-v3.shtml

[lti]

Re: Power supply current limiting

I don't have any microcontrollers, and I don't really want to buy one and the programming tools just for this.

The circuit I'm currently using matches the current limit circuit in the early EEVblog power supply design:
http://www.youtube.com/watch?v=8-qar5vgnbc
I'm using an op-amp and an IRF510 MOSFET instead of the LT3080 and a 0.25 ohm shunt. It still oscillates, but it is better than the circuit I had before. I don't know if that is a fault with the simulation or if it will really do that when I build it.

Also, trying to change the gain of the differential amplifier causes it to output only 6mV instead of the correct multiple of the voltage drop across the shunt. Making the gain less than 1 causes it to output really high voltage (7.2V with 0.25V drop across the shunt and gain of 0.47), and the output contains lots of ripple from the power supply.

[tom66]

Re: Power supply current limiting

Never do the regulation loops in software like tuxgraphics suggests! It is a sure way to make a very poor power supply. The control loop just ends up being so slow, you can directly short the output and it can take ages to respond, poor handing of load change/input voltage change etc.

How about sharing your schematic so far?

[stj]

Re: Power supply current limiting

i hate high frequency analog design.
you will probably find your simulation is bad.
they dont usually acount for track inductance, parasitic oscilation etc.

you will need to keep the track from the gate pin as short as posible, and probably put a ferite bead on the leg of the fet to avoid strange problems.

other than that, your on your own - the atmel is the easy way.
the chip is about $2 and you can program it fron the pc parallel port with just a few jump wires.

btw, unless i missed it you didnt say what maxium current you have in mind.

[ReverendJones]

Re: Power supply current limiting

Maybe this circuit can be adapted to your requirements: http://kd1jv.qrpradio.com/vps/voclps.HTM

[budm]

Re: Power supply current limiting

One thing we do not have is the spec of the power supply you want to build: Variable out put from 0~??? Fixed?, Max current out put with variable current limit function? Response time? etc.

[lti]

Re: Power supply current limiting

Here is the schematic. R1 and C3 are the load. The cap is there to make sure it won't oscillate with anything other than a purely resistive load. I have been changing their values to test the circuit at varying loads.

I would like to get at least 1A out of this power supply, but I don't know how well it will actually work since the op-amp's PSRR isn't simulated. It would be nice to get more current out of it than that.

[budm]

Re: Power supply current limiting

So what do you want to happen when the current limit is reached, should it go into latched shutdown so it will not oscillate when the load current goes back down to normal, manually reset when current goes back below the set point?, what is hysteresis threshold if auto reset? You need to figure what the requirements are.

[stj]

Re: Power supply current limiting

@Reverend Jones
you need to be carefull about that shematic,
the current-sense is on the ground-line.

what that means is that if your testing anything that has any other ground-path - like a video-cable or even the ground clip of a scope-probe,
your current regulation goes out of the window!

it's a VERY bad design.

[stj]

Re: Power supply current limiting

as for that spice "schematic" i cant even see where the load is supposed to go!

[tom66]

Re: Power supply current limiting

i hate high frequency analog design.
you will probably find your simulation is bad.
they dont usually acount for track inductance, parasitic oscilation etc.

you will need to keep the track from the gate pin as short as posible, and probably put a ferite bead on the leg of the fet to avoid strange problems.

other than that, your on your own - the atmel is the easy way.
the chip is about $2 and you can program it fron the pc parallel port with just a few jump wires.

btw, unless i missed it you didnt say what maxium current you have in mind.

Yes, but it is a terrible control loop! You cannot use a slow, 8-bit AVR to control such a loop. It's too slow to do that. You need a high speed software control loop, with dedicated on-chip comparators and DACs and high performance DSP. That's a worse nightmare than some comparatively easy analog control loop design. If you want to do it with an AVR you can - but build all the control loop in analog! Just have your AVR feed a voltage limit and current limit signal into a control circuit.

If you check it with a simple DMM, you'll see nothing wrong. But if you put it on a scope, you'll see constant overshoot as you turn the load on and off, or change the voltage, or the load current changes, or as you power it on, or as the bulk capacitor voltage varies. Unless you're only powering light bulbs this is one sure way to kill your circuit!

I bet if you short the output it will supply full current for 100+ms as it slowly hunts down, which makes a constant-current limiter useless for short circuit protection. (Real power supply will respond within 5ms.)

[tom66]

Re: Power supply current limiting

Here is a PSU I designed a while ago; I've never built it though since I found one cheap enough. 0-24V 0-3A. Has no tap switching on the transformer, you will want to add that to limit the power dissipated by the pass transistors. Analog control loop stuff is easy! One LM324 for all the logic.

[lti]

Re: Power supply current limiting

One thing we do not have is the spec of the power supply you want to build: Variable out put from 0~??? Fixed?, Max current out put with variable current limit function? Response time? etc.

I'm looking for 0-18V and 0-1A variable output. When it hits the current limit, it should decrease the output voltage like a normal bench power supply. That's the part I'm struggling with. Maybe I don't know enough to actually build this. A variable voltage power supply with no current limit is still better than what I have now.

I want it to respond within a few milliseconds like a real power supply, but I don't know if I will be able to accomplish that.

as for that spice "schematic" i cant even see where the load is supposed to go!

I didn't think it would be a very good schematic. R1 is the load, like I said earlier.

My PCB layout "skills" definitely suck. I don't know if this thing will work at all when I build it.

Currently, I don't like that the op-amp is running so close to its absolute maximum supply voltage. The unregulated power supply could hit 32V (the BA10324's absolute maximum supply voltage) with a light load if it is left powered for a while. I only have a 7805, a 78L09, and some TL431s to use, and I need to power a 12V fan. Also, decreasing the supply voltage will reduce the maximum output voltage.

This project is starting to look like a complete failure.

[tom66]

Re: Power supply current limiting

Essentially you need to create a "min" function: the power supply selects between the minimum of the two command signals. One signal will try to keep the output voltage at the set point, the other will try to keep the output current at the set point. The design linked is capable of reaching set-point current from direct short within 40us.

One way you can acheive this is using diodes, like I did in the above schematic. Another way is to use a transistor OR-ing circuit where transistors would turn on to limit voltage or current appropriately.

Try to construct a voltage-controlled supply initially. Then construct a current-controlled supply. Try to consider how to combine the two parts.

[3d0]

Re: Power supply current limiting

This PSU is 0-30VDC with current limit, you may find something useful for your project.

http://www.electronics-lab.com/proje...001/index.html

[stj]

Re: Power supply current limiting

schematic & pcb design for free.
http://www.cadsoftusa.com/

[lti]

Re: Power supply current limiting

I was thinking about this again, and I ended up building the circuit I was simulating. In real life, it's stuck in current limit mode because the difference amplifier (U2 in the crappy schematic) outputs a constant 0.77V, and I don't understand why. Connecting a battery across the inputs instead of the shunt resistor causes it to output the correct voltage. The resistors I'm using have really weird markings, but they measure 6.85K ohms. They appear to be labeled 01b with a line above the 0.

I wonder if low side current sensing (like the circuit in post 6) would work better. I'd have to make a lot of changes to the circuit board, but I'm using a Radio Shack prototyping board right now. It's always going to be a big mess with the components I have.

[lti]

Re: Power supply current limiting

I guess I should add some actual measurements.

The voltage on each end of the shunt is about 29V (output from the bridge rectifier). There is no voltage dropped across the shunt because the output voltage is zero (the current limit circuit thinks the output current is 3.08A, so it forces the output to 0V). The voltage measured directly at the op-amp's non-inverting input is exactly half of the supply voltage because the two resistors on that input form a voltage divider. The voltage on the inverting input is 0.4V lower, and the output is 0.77V.

Maybe the BA10324 (LM324 clone) isn't a good choice for that circuit, but the only other op-amp I have right now is a TL064.

Placing the shunt in series with ground will cause poor regulation, and I can't think of any good way to compensate for that. I could add the voltage drop back into the set voltage, but that doesn't work very well in the simulation. There is a lot of ripple on the output if I do that, and it would require more op-amps because I would have to add a buffer on the potentiometer output and figure out how to make a non-inverting amplifier with a gain of 0.5 (the potentiometer range is currently 0-9V, so I would have to divide the voltage drop from the shunt by 2) before adding the two signals. I'd probably be better off increasing the reference voltage to 18V instead of trying to divide a signal by 2.

With the current limit disabled, it does work and give 0-18V output with no apparent issues, so it is better than nothing. Unfortunately, I don't have the parts to build a better circuit, and I don't want to buy the parts right now.

[budm]

Re: Power supply current limiting

"With the current limit disabled" So you disable it by disconnecting the Collector of Q1 from the circuit?