Need help with current limiting. / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Need help with current limiting.

1 2 3 4 next last

Move Thread

LAN_403

Patrick

Fri Dec 04 2015, 06:43AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

I'm etching a new circuit board. but need help with the current detection part of the UC28025 ic.

Ignore pin 11, I have the -25 version that's the -23.

I want to use an extremely low value resistor in the transistor path. 10 amps / 0.1V = 0.01 Ohms. . . . and 10A x 0.1V is 1W.

I see the datasheet pin 9 is at 1V but compared and triggered to what ? When my chosen 100mV value is reached ? is it the pin 2 NI or pin 7 ramp. there example circuit isn't clear. it shows a mysterious 1kohm resistor above the current sense one.

Example circuit. note the pins 9 and 7.

(datasheet for UC28025)

Transformer calculation link :

Mads Barnkob

Fri Dec 04 2015, 07:40AM

Registered Member #1403 Joined: Tue Mar 18 2008, 06:05PM
Location: Denmark, Odense C
Posts: 1968

It is tied up against the oscillator, could it just be a pull-down resistor?

Look at the table for over current shutdown conditions, but hands on experience is usually needed when example circuits look weird :) I had great trouble with a PFC IC where I used a too small current sense resistor and it could just not work with that.

CURRENT LIMIT/SHUTDOWN
ILIMIT Current limit bias current 0V<V(ILIM/SD) <4V Â±10 Î¼A
ILIMIT Offset voltage UC28023 15 mV
ILIMREF Common mode range(1) UC28023 1.00 1.25
Current limit threshold voltage UC28025 0.9 1.0 1.1 V
Shutdown threshold voltage 1.25 1.40 1.55

hen918

Fri Dec 04 2015, 05:33PM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

Ignore the third picture, look at the application diagram. That "mysterious" resistor is used as a low pass RC Filter to remove the leading edge noise, it will be required and will need to be valued inversely proportional to your operating frequency. The application diagram runs at 1.5MHz, so if you were running at 150KHz, for example, I would double the value of the capacitor and 5 times the value of the resistor.

Now for the over current detection:
The App. Diagram uses another 1k Resistor and a capacitor for the leading edge filtering. Otherwise, the voltage from the current sense resistor goes through to the Ilimit pin unchanged.
Looking at the datasheet, the over-current is triggered when this pin gets to one volt and shuts down when the voltage goes above 1.4V.
Use this to set your shunt resistor.

Patrick

Fri Dec 04 2015, 09:18PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Ok thank for the help, and keep it coming. this looks complicated.

hen918 wrote ...

. . . the over-current is triggered when this pin gets to one volt and shuts down when the voltage goes above 1.4V.
Use this to set your shunt resistor.

so the term "pulse by pulse current limiting" means it will throttle back the PWM pulses (via the ramp V), thus I choose the components to function at the 1 and 1.4 volt values. But using low value resistors, I don't want 1 or 2 ohm 10 watt resistors getting flaming hot. so I want the smaller values. How can I resolve this? is there a Vref that acts on one of the comparator inputs? would that compare against a 0.1V Cs V and push it above 1 and 1.4 V ?

Ill etch a prototype with extra holes and spaces.

0.01 ohm resistors. Specifically maeant for current sensing.

hen918

Fri Dec 04 2015, 10:19PM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

I had exactly the same problem with my 24 to 400V DC to DC power supply. To solve this, I used an fast op-amp, working single ended (with only a 0V to V++ output), to amplify the voltage from a 0.0001Ohm section of PCB track.

Unfortunately, EMI given off by the power components and track was inducing a comparatively massive voltage in my control PCB, so the circuit hasn't been properly tested (the current control had to be virtually bypassed for the circuit to function at all).

On the schematic, pin 8 on the IC is the current control input and pin 4 on P2 is from the current sense resistor on the main board.
Be warned: this is not tested!

Oh, and I didn't use a TL071, that op-amp would have been wayyy to slow, I used the OPA350 (also Ti)

Patrick

Fri Dec 04 2015, 11:12PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

lets try to get the uc28025 to work, it may work in both our applications. I always fear adding more analog IC's when I'm not sure how to use the base circuit itself. but you have more experience then I.

There has to be away, they do sell it, so it must work . . . TI doesn't sell crap.

0.0001 is extraordinarily low resistance.

Mads Barnkob wrote ...

ILIMREF Common mode range(1) UC28023 1.00 1.25
Current limit threshold voltage UC28025 0.9 1.0 1.1 V
Shutdown threshold voltage 1.25 1.40 1.55

with these high voltages, how are we to design circuits that don't have flaming hot resistors ? How in the world am I to get my 0.1 volts to be the current limit ?

Layout of the power side is critical. the transistors, Cs resistor, capacitors, transforemer, and analog input to the IC from the CS must all be well layed out well.

EDIT: the error output pin (op-amp) is receiving from the 5V side of the out put, and I think resistively divided ? ?

hen918

Sat Dec 05 2015, 10:55AM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

The problem is that the ICs we are using have been designed to work in low current circuits where a 1V drop isn't an efficiency disaster. (the application note has a maximum average primary current of just over one amp)
The reason for my very low resistance was that the PCB is designed to take a peak current of 320A with an average current of 80A, or a peak current of 160A with the average current being almost the same. (It's a 3000W power supply)

Patrick

Sat Dec 05 2015, 06:09PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

so maybe a current transformer. Then saturation becomes the issue, so maybe separate paths for each mosfet to drive through to ground.

Electra

Sun Dec 06 2015, 01:52PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

Yes Patrick just what I was thinking, current transformer with two opposing primary. You'll have to rectify the output and load it with a resistor. A lot of current transformers I've seen just have one turn sense winding through something like a 25mm ferrite toroid core

The designers of the SMPS ic's probably choose high thresholds because they want high noise immunity.

If you want to keep using a shunt, a differential amp circuit might work better if you take the two inputs directly to the sense resistor leads, you might still have to still have design your shunt for a few hundred millivolts, because otherwise the wanted signal will become lost in the noise, even then some filtering of the input to differential amp may be required, as switching noise I think that can be faster than the amp can cope with.

When I last tried this it wasn't very successful also just like Hen experienced, I think it could be made to work better but layout and screening are important....

Patrick

Sun Dec 06 2015, 09:10PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Alright current transformer it is. so lets do some math. does the current 1.0 volt threshold start clipping the pulses shorter and shorter, until 1.4 volts is reached when its off entirely?

1 2 3 4 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob