SG3525 PWM controller OCP cell

[ym58]

How does that cell work exactly ?
Does it produce some kind of delay (47uF/220nF ?) to the 'dis' output once a current limit has been reached on R9 (current sense) ?



I am myself triggering OCP thru a sense resistor in a Meishile 1000W/12V/87A SMPS (that I want to limit to 50A) but my OCP feedback cell is much simpler than this one (I am using an opamp comparing the voltage across the current sense resistor to a REF voltage then sending a HIGH signal to the shutdown input if Vshunt > threshold) and it works as far as current limiting is concerned.
But my cell keeps inducing an annoying 'click-click-click' sound somewhere within the SMPS !
--
Will such a cell prevent this drawback ?

[jonpaul]

the overcurrent sense is designed to shut down the 3525, and after a dealy timeout to restart.

This creates a periodic start, stop cycle upon overload, very common technique especially in the 1980s ancient 3535 SMPS controller

Jon

[ym58]

Yes, that's what I thought ...
A few questions though :
- Aside from the R9 value which is not quite convenient to source and swap (vous en conviendrez ), do you have any idea where within the cell a threshold adjustement could be implemented (I am thinking of making one of the 2.2K resistors adjustable but maybe I am mistaken) ?
- How does exactly work the delay-before-restart function, which capacitor (and resistor) are involved in the timeout process (and calculation) ?
- What's the use of the diode (not the Zener one !) ?
- Where possibly does the 'click-click-click' I mentioned in my OP come from ?
Merci.

[Njk]

- How does exactly work the delay-before-restart function, which capacitor (and resistor) are involved in the timeout process (and calculation) ?

I think you'll get the answer sooner if you change the picture adding reference to every component of the "OCP cell" circuit (like R1, C1, etc.)

[jonpaul]

click....magnetostriction of transformer and magnetscs at START><OVERLOAD><Shutdown.

Repeats about 1/sec..1/5 sec

Very typical, see  Tektronix scope PSU repairs
j

[xavier60]

For OCP or current limiting circuitry to work properly, any leading-edge voltage spike at the CS resistor needs to be filtered out. That will be the purpose of the 2.2K/220pF RC filter. A crude but stable method of current limiting I have used with SG3525 is to simply connect the RC filter to the Base of a BC548 or similar. Emitter to ground.
Connect its Collector to the Soft Start pin,8. Put a 100Ω resistor in series with the Soft Start capacitor. When current limit is reached, pulse-by-pulse limiting will first occur, then transition to stable PWM as the Soft Start capacitor discharges.
The idea can possibly be refined using a comparator and reference.

[ym58]

I think you'll get the answer sooner if you change the picture adding reference to every component of the "OCP cell" circuit (like R1, C1, etc.)

Good idea ... done !

[ym58]

click....magnetostriction of transformer and magnetscs at START><OVERLOAD><Shutdown. Repeats about 1/sec..1/5 sec  Very typical, see  Tektronix scope PSU repairs

Well ... thanks  !
And what about my other questions ?

[ym58]

@xavier

Very constructive and meaningful answer ... thanks !
I need a clarification though : you wrote

Connect its Collector to the Soft Start pin,8

I realized that my red arrowhead was masking the PWM controller pin the OCP cell is connected to (see my original post, I've changed it).
Actually it is connected to pin 10 which is a shutdown input, isn't it this input that you meant rather than the softstart input, pin 8 ?

[xavier60]

A High on pin 10 or Low on 8 for shutdown and also for pulse-by-pulse limiting.

Extra: Analog voltage on 8 can take control of PWM.
What is the nature of the load?

[ym58]

LFP4/150Ah (and higher capacity).
I need to monitor and control the output current coz such batteries may induce a high current demand when connected initially as they can accept charging currents as high as their capacity  (BMS permitting) ... I've fried two times the 2x W45NM60 mosfets, don't want to repeat it anymore !

[xavier60]

The W45NM60 are a big improvement over the originals which were being pushed close to their limits. I don't like seeing large MOSFETs being driven directly by the SG3525 but we won't worry too much about that for now.
Under what circumstances did the W45NM60  fail? That original OCP cell should provide more than adequate protection and should always be left functional.

[ym58]

should always be left functional

That's the point, shame on me ... during some testing I did with my own OCP, I left it disconnected, assuming that 3-5 seconds with high current output won't do any harm to the mosfets ... NADA  !
I think I shoud have put two fuses in the mosfets drains, at least during the testing period ...

[xavier60]

With the original OCP cell left functional, hopefully any extra current limiting circuitry should not be able to cause damage.
You could try something simple at first. The added transistor I explained earlier with a 1K/470pF RC filter on the Base fed from the top of the CS resistor.   Put the 100Ω resistor from Collector to pin 8. The current limiting transistor needs to act at a lower current than the OCP cell. Add some positive bias to the Base by connecting a resistor from the Ref 5V to the Base. Try a 22K first up.
Even if the OCP cell trips due to a sudden overload, the idea is for current limiting transistor to take control as the Soft Start gradually ramps up the PWM during the proceeding restart.

[ym58]

@xavier60
Well, actually it's a bit more complicated than that ... let me explain real quick.
The schematic that I posted in my OP is something that I've found here and there.
Starting from there, I just wondered if I could adapt this type of cell to the current PSU I am working on !
The PSU I am working on is based on a half-bridge topology and, actually, there is no RS (shunt for sensing current) in the source of the bottom Mosfet, nor in the ouput DC stage of the PSU !
However, there is a small transformer T3 "before" the AC/DC final transformer (T4) and IMHO this T3 senses the current by ways of a rectifier that feeds the shutdown pin of the SG3525 through a simple transistor Q1 (no RC filtering).
That's something I am not really familiar with ... needless to say
I am currently doing some reverse engineering on this PSU (not completed yet, but  I think it will be done by t'row !) but so far here's what I have been able to find ... I think that the circle named 9 is the key issue and, as you can see, it only involves pin 10 (shutdown) of the PWM controller :

[xavier60]

What's connected to the Soft Start pin 8?

[xavier60]

Does this PSU current limit rather than shut down? It looks like it should. Bridging the diode connected to R13 should cause a drop in limited current.

Or increase the 33Ω resistor.

[ym58]

@xavier60
No diode connected to R13, that was just an arrow !
R13 (which value is finally 10R and not 100R) is directly connected to the rectifier though (D4/D5/D6/D7) and also to R7 which value is 3.3R
The Soft Start pin#8 is connected to a simple C6 100nF capacitor ... reason why I keep thinking Meishile (manufacturer) handles the current limiting function through the Shutdown pin#10, but I don't quite understand 'HOW' it works !
I've redrawn the thing for you :

[xavier60]

The end result is kind of the same. Try increasing R7 to about 5Ω
 to check for effect. Double check that R7 is likely to be 3.3Ω by measuring it.
Im assuming that the PSU is already capable of smooth current limiting? no rattles or squealing.
What bothers me a bit is that primary snubber current is also being sensed, adding to leading-edge-spikes.

From the Osemi's data PDF, https://www.onsemi.com/pdf/datasheet/sg3525a-d.pdf

"An alternate approach is the use of the shutdown circuitry
of Pin 10 which has been improved to enhance the available
shutdown options. Activating this circuit by applying a
positive signal on Pin 10 performs two functions: the PWM
latch is immediately set providing the fastest turn−off signal
to the outputs; and a 150 A current sink begins to discharge
the external soft−start capacitor. If the shutdown command
is short, the PWM signal is terminated without significant
discharge of the soft−start capacitor, thus, allowing, for
example, a convenient implementation of pulse−by−pulse
current limiting. Holding Pin 10 high for a longer duration,
however, will ultimately discharge this external capacitor,
recycling slow turn−on upon release."

[ym58]

"primary snubber current" ... ? 

"a 150 A current sink begins to discharge the external soft−start capacitor" ... actually C6 is a micro-SMD decoupling-type cap, in no ways could it stand a 150A discharge current !

[xavier60]

"primary snubber current" ... ? 

"a 150 A current sink begins to discharge the external soft−start capacitor" ... actually C6 is a micro-SMD decoupling-type cap, in no ways could it stand a 150A discharge current !

The "150µA" didnt copy and pasted properly. It's on page 3 of the data PDF.

[ym58]

Can you believe that ?
Now I understand why I kept blowing MOSFETs !!!

[xavier60]

There is no standard value for CT load resistors. It all depends on the CT turns ratio. 3.3Ω seems unusually low. Any idea what it originally had?

[ym58]

3.7R means the resistor is FAULTY for it should be a 33R (orange-orange-black) !
It means that the current measurement was definitely underestimated ... I will replace it with a valid 33R first then increase it step by step until I get my targeted lower current limit ...
I will re-post to report the testings.

[xavier60]

3.7R means the resistor is FAULTY for it should be a 33R (orange-orange-black) !
It means that the current measurement were definitely underestimated ... I will replace it with a valid 33R first then increase it step by step until I get my targeted lower current limit ...
I will re-post to report the testings.

There appears to be a band after the black? The brown at the end is likely the 1% tolerance. It's very unlikely that a faulty resistor would drop from 33Ω to 3.3Ω.