LM2596: Issue with ON/OFF pin of LM2596

Aakash,

Can you please also share waveforms of the ON/OFF pin and the OUT pin during this condition? VOUT shouldn't come up if OUT is not switching and OUT shouldn't be switching if the ON/OFF pin is turning the device off. So I'd like to see which part of this chain doesn't align with what you're seeing.

Thanks,

-Sam

Aakash,

Please also try again without the 10k in series. That may be causing the issue.

-Sam

Hi Sam,

The two plots are attached below:

Thus, When ON/OFF pin voltage is 3.3V, Vout= 0.6V and when the ON/OFF pin voltage is 0V, Vout= 4V.

Its OK, Right? My concern is that But, When I turn ON the power supply (12V), I have found a overshoot up-to 2V and then a decay for 4 seconds (although I have programmed a delay of 2 seconds) after which the regulator turns ON. So,What could be the reason of this overshoot ? Also, I have given a delay of 2 seconds (mentioned earlier), but practically it turns ON after 4 seconds, Why?

NOTE: 10K in a pull up resistor. By default, the regulator should be OFF. When I trun ON/OFF pin low by controller (D26 pin), the regulator should turn ON. So, 10K is pull up is necessary for this operation, Right?

Thanks.

Aakash,

The ON/OFF pin will have some current so the 10k may have a voltage drop across it. Please share waveforms showing the voltage on the ON/OFF pin as well as the OUT pin. I'd like to see the actual voltage on ON/OFF and I'd like to see if the device is switching.

-Sam

Dear Sam,

Thanks for your clarification. Currently Our Lab is closed due to COVID-19. I can share you the necessary plot after 7-10 days. I hope you have understood. In the meanwhile, you can suggest all other possible reasons so that I can test all the cases all together.

Thanks.

Aakash,

I understand the lab limitation. I'm living through the same situation :)

The issue boils down to "Why is the output coming up when we try to disable the device?" Step 1 is to confirm that the ON/OFF pin is really getting what we think it's getting, then comparing that against the datasheet to confirm that the device is operating in the way we expect. The OUT pin (switch node) will tell us if the device itself is the reason for the increased output or if the output is coming up from some other circuit somehow. Once we have this information, we should be able to come to a conclusion. I don't expect we'll need further investigation.

Send the data as soon as you can get it. Thanks!

-Sam

Dear Sam,

Please find the oscilloscope graph below:

So, my concern is that But, When I turn ON the VIN power supply (12V), I have found a overshoot up-to 3V and then a decay for 4 seconds (although I have programmed a delay of 2 seconds) after which the regulator turns ON. So,What could be the reason of this overshoot ? Also, I have given a delay of 2 seconds (mentioned earlier), but practically it turns ON after 4 seconds, Why?

Thanks.

Dear Sam,

Yes, the result is same with 10K and with shorting the 10K.

At the time of overshoot, ON/OFF pin voltage is 3.3V and the overshoot voltage is 3V. But How can I know the OFF threshold voltage at this time? Please explain.

Also, you are telling to use a pull up resistor/ capacitor from VIN (12V) to ON/OFF pin. But, I am controlling the ON/OFF pin by controller (ESP-32) pin whose voltage level is 3.3V. Will it be safe to pull up with 12V?

Rather, I have tested with pulling up this ON/OFF pin to 5V. The graph is given below:

The result is same. Right?

Thanks.

Aakash,

Yes the result looks the same. It may be that the 3.3V logic is not up when the 12V is applied. Please share a scope capture zoomed in to the overshoot so we can see the sequence of VIN and ON/OFF voltage. In addition, here are a few tests:

First test:  Disconnect from the 3.3V logic and tie ON/OFF to 12V and power up 12V. This would damage your ESP-32 but we're just doing a test for proof of concept.

Second test: Look at edge of 12V and ON/OFF at the instant of 12V turn on; look for the sequence.

Third test:  Try to slow down 12V edge.

Fourth test: Try with an EVM connected to the system.

-Sam

Dear Sam,

I can share you the required graph by tomorrow.

So before testing, please clarify the following:

1) Should I do the test by removing 10K?

2) You said, LooK at edge of 12V (VIN) and ON/OFF. Are you telling to compare the timing of 12V (VIN) and ON/OFF pin voltage?

3) Try to slow down 12V (VIN) edge, but how? please give me some idea.

4) "Try with an EVM connected to the system"- Which EVM?

Thanks.

Aakash,

1. Yes

2. Yes. Provide a scope shot that shows which voltage comes up first and the timing between them.

3. You could use a current limit on your power supply.

4. http://www.ti.com/tool/LM2596S-ADJEVM

-Sam

Dear Sam,

Please find the following graphs:

1) The below image shows the clear Overshoot when ON/OFF is pulled up to 3.3V line.

2) The below images shows the edge of 12V and 3.3V Power line (ON/OFF pin is connected to the 3.3V line). Clearly no rise time delay/edge.

3) The below images shows the edge of 12V and  ON/OFF pin when it is connected to 12V line. Clearly no rise time delay/edge.

Also, unfortunately I don't have EVM.

Thanks.

Aakash,

It's hard for me to conclude that there's "clearly no delay" at a scale of 1sec/div. Next time you're in the lab, please check again at a finer time scale so that we can see which one is really coming up first. It doesn't take long for the output to rise (<1 pixel on the scope at 1sec/div). It would be good to zoom in so the rise from 10% to 90% is over 1 or two divisions.

-Sam