LM5121: Disconnection Switch Issue

Dear Harry,

I have put the zener as it was indicated in datasheet and also your previous message. In case of having over current still I have 12V at the output. Then I removed the zener and the capasitor so the res pin connected directly to GND. It should cut the voltage reaching to 1.4A at the output. But what I can see is that while we increasing the current it reduces the voltage and never opens the circuit. 

Regards

Farzaneh

Hello Farzaneh,

What does the voltage signal on the RES capacitor look like over time?
Are you sure that the overcurrent is high enough to trip the hiccup mechanism?

Thanks and regards
Harry

Harry,

I tested it for different current limits. At the output I have 48V as it is a boost converter. When I apply any of current limits I can see 12V at the output while I expect to see 0V. 

Right now I have the circuit whithout zener and res capacitor.

Farzaneh

Hello Farzaneh,

So, the output voltage is basically the input voltage.
But with a simple number we cannot really make progress here.
We will need to look at Oscilloscope screenshots instead.

One important test::
Disconnect R5 or short R14 to keep the booster disabled.
What do you see on the output when you supply power to the input?

Do you already get 12V on the output, even if the contoller is disabled?
If so, something is wrong with the breaker FETs (T1/T2).

Best regards
Harry

Dear Harry, 

The first Photo is the circuit without zener, as you explained me before and by the datasheet I undertood that it should enter to hiccup mode if I put a zener in parallel with the Res capacitor. I don't have the zener and still it's in hiccup mode:

Pink probe: input voltage 12V

Yellow probe: output voltage 48V

green probe:Res capacitor 

We removed the RES Capacitor  to connect the RES pin directly to GND and this photo shows the result:

In this photo it seems that the chip regulates the voltage. As we increase the output current it decreases the output voltage and the power consumption remains constant. 

by the way I tested T1 and T2 as you have mentioned, they are OK.

Looking forward to hearing from you

Farzaneh

Hello Farzaneh,

Thanks for the Oscilloscope screenshots.
The first image shows that the fundamental hiccup procedure works.
After eight triangles of the RES voltage, the controller is starting over, resets the RES voltage and tries boosting.
But due to the overcurrent, the voltage on the RES pin exceeds the threshold and another eight triangles will start.

The one thing that does not work is that the breaker transistors do not turn off.

PLEASE DO NOT CONNECT THE RES PIN TO GND.
This will probably damage the internal current source.

Here again the test that I already asked for:
Disconnect R5 or short R14 to keep the booster disabled while you power up the board.
What do you see on the output when you supply power to the input?

Do you already get 12V on the output, even if the controller is disabled?

Then, enable the booster (with no load or light load) by connecting R5 or removing the short of R14.
Your output voltage should then go up to 48V.

When you then apply the short on the output, do you get the same hiccup cycles as in your first image?

Please take oscilloscope screenshots which also show the voltage on the DG pin in addition to the RES cap voltage and the output voltage.

Thanks and regards
Harry

Here you can see the screenshot. I tested the T1 and T2 , they work. I wanted also to ask a question, we connect mode pin to GND. could it be the problem?

The green probe is DG.

Thanks

Hello Farzaneh,

First, I want to apologize.
I had overlooked that the LM5121 does include some logic which allows you to connect the RES pin to GND to disable the hiccup feature.
So, there is no risk for a damage there.

You wrote:
> I tested the T1 and T2.  they work.
How did you test that? Can you somehow turn them on and off?

Anyway, what I would like to see in an oscilloscope screenshot is the behavior of the breaker FET control (the signal at the very bottom of Figure 27.

I must admit that only looking at DG versus GND may not be good enough.
Can you please also probe the Ds voltage?
If you have a differential probe it would be great because you could directly measure the difference.
If not, please measure DG and DS and show both of these signals on top of each other.
If possible, use the math function of the oscilloscope to calculate the difference.
(If you do that, please be aware that there may be glitches due to different timing offsets of the two oscilloscope inputs).

I would like to see when the VDG - VDG goes low to turn off that FET.

And, again, I would also like to see what happens at the initial power-up.
When does that FET get turned on?
You may need to set some trigger on your oscilloscope (e.g, when the UVLO voltage exceeds a limit) to catch that point.

Thanks and regards
Harry

Morning Harry,

I cheched the T1 and T2 as you have explained me. I removed R5 and I didn't observe any voltage at the output. So according to your explanations if it is so, it means that they are OK. Placing back R5 I observed 48V at the out put. By the way I changed the PCB and I took a new one to descart any damaged cuased by testing. 

Following you can see the screenshots of what you asked for:

starting up:

Yellow probe :Output

Green probe : UVLO

Pink probe : DG-DS (differential probe)

Connecting load:

Yellow probe :Output

Pink probe : DG-DS (differential probe)

I appreciate your time and attention

Farzaneh

Hi Farzaneh,

Due to bank holiday, please expect a reply by Friday.

Best Regards,

Feng

Morning, Harry,

I wanted to express my gratitude once again for your invaluable support and thorough explanations. Thanks to your guidance, we've finally managed to decipher what was causing the issue. It seems there was a misinterpretation of the datasheet on our end.

Upon closer inspection of the attached photos extracted from the datasheet:

We initially concluded that both overcurrent and short circuit instances would trigger T1 and T2 to open, thereby cutting off the output. However, our observations have revealed a discrepancy. While a short circuit does indeed result in an output voltage of "0," in the case of overcurrent, we're perplexed to find that the input voltage persists at the output.

It appears you were correct in your assessment that we weren't subjecting the system to sufficient current to observe the expected output voltage drop to "0."

Thank you once more for your patience and guidance in navigating this issue.

Farzaneh