LM2576 - Voltage latches to 5V from a 12V supply

Hi,

I dont understand your feedback connection, it is not referred to ground.

Sorry, looked at it again and now I understood the connection... (edited later). Anyway, if the battery is giving 12V, you cant possibly meassure an input latchup to 5V, the fuse would blow so fast that you would meassure nothing. Can you test with another power source?

Hi Avi,

Thanks for responding to my query. I have had the same observation on multiple boards that were installed in different locations (completely separate installations). I have not seen the fuse blow up, but I am unsure at the moment as to what is causing this latch up.On one board that was returned, the LM2576 had physically ruptured. I had not anticipated such a situation, since I expected a thermal shutdown before things got out of hand (so to speak). Is there any test that you can suggest so that I can understand the possible cause of this observation? Any general improvement suggestions on the present design are also welcome.

Regards

Anadi 

Hi,

What I was trying to tell is that if the input was latched to 5V, and the bat. input is 12V, the fuse must blow. I cant understand how it could be different from that. I checked your design again and I dont see something to improve, if I were you I would try to understand the scenario of the problem.  I didnt see the values of your protection diodes. Could it be that a voltage is present in the output while the device has no input? I also didnt understand the 200Ohm coil at the output, is that DC resistance?

If you have devices burnt, you can contact your local rep or distributor to see if they can have them analyzed.

recheck your layout and see if there is anything suspicious, voltage nets must be wide and healthy.

Wish you luck,

Avi

Hi Avi,

Thanks for taking the time to look at my design. Unfortunately, I am also unable to understand what could be causing so much trouble :).

The diodes are SS36 diodes (schottky), which are rated for 3A current (my peaks are expected to be less than 2.2A).

I am unable to follow the following comment "Could it be that a voltage is present in the output while the device has no input?"

The 200 ohm coil is meant to be part of a low pass filter circuit, basically this is a 200 Ohm ferrite bead.

I was hoping some TI rep would help me get further with solving this issue :) (that's why posting on this forum)

Regards

Anadi

Hello Anadi, 

I am hoping I can help you troubleshoot this issue. Here are a few questions:

1. When you see the failures, do they occur after some time of proper operation or do you see them at power up?

2. Could you attach you board layout? 

3. Do you have any scope shots of the switch node (pin 2) voltage and also the input voltage?

Regards, 
Denislav 

Hi Denislav,

Thanks for responding to my query. Please see my responses inline below:

1. When you see the failures, do they occur after some time of proper operation or do you see them at power up?

I have observed both scenarios, one where the board starts up with a lower voltage and stays there and second where the board starts up as expected, but starts to report a smaller value during run time. It also recovers from such a situation and starts to report the normal value after a little while.

2. Could you attach you board layout? 

Is it possible to email you the board layout, since there may be some concerns about sharing the layout on a public forum?

3. Do you have any scope shots of the switch node (pin 2) voltage and also the input voltage?

Unfortunately, I do not have any scope shots of this issue, since none of the boards in the lab have shown a similar behavior as yet.

Regards

Anadi

Hello Anadi, 

Yes, you can email the layout file to denislav.petkov@ti.com.

Regards, 

Denislav

Hi,

Same issue happend when i connect to car battery (LM2576 blowup). The schematic is the same as pdf provided from ti.

Hi Denislav,

Thanks for your time and effort in solving this issue. If you feel that certain modifications / improvements should be made to the LM2576 part of the circuit, please let know also.

Br

Anadi

Hi Anadi,

I know that I'm reviving an old thread, I apologize if this is against forum policies.

I'm working on a DC-DC power supply for a car and basing it on the LM2576 as well. I was hoping you could let me know whether the issue was resolved and what was causing the problem to occur. It may help me avoid getting myself into a similar pitfall.

Thanks,

Arthur

Hi Arthur,

Sorry I forgot to post a follow up on the thread to publish the results of my investigations.

As it turns out it was not an issue with the power supply design after all. We have a low pass filter (butterworth) placed before the ADC for measuring the supply voltage (just to smooth things out). This filter had an issue leading to a voltage on the ADC pin of the uc, when the mains power supply was disconnected. The LM2576 was "not at fault here".

The issue was observed from the field as the mains power source (car battery) was disconnected through operator error and the system shifted to the internal battery backup supply. We then started observing smaller measured voltages causing all sorts of grief. 

On a side note unless you are planning to use the HV version of 2576, I would suggest that you move to the 2596 (better performance because of 150KHz).

HTH

Br

Anadi

hi every one !
i have to make power supply for gsm sim 900, and i have tried this with lm2576 regulator, but i failed...i have used the basic power supply design for lm2576
http://e2e.ti.com/support/power_management/simple_switcher/f/858/p/255499/1039477.aspx#1039477  but at the output i dont get 5 volts, i get the output voltage equal to input... i dont have  IN 5822, any schottkey diode so i am using simple diode , and inductor of some different value, so i
guess becoz of these two i dont get the desired output,, am i ri8??can i use simple diode or zener diode instead of schotkey, plz help me out in this
 

You cannot use a zener diode instead of a schottkey diode. But a regular diode should be ok. It just generates more power loss than a schottkey diode.

Another common reason for Vout going up to Vin is the feedback path has a bad connection. Please check you are using a fixed 5.0V output voltage version and Vout is connected to FEEDBACk (pin 4) correctly. Or if you are using the adjustable version, proper resistor dividers are used and connected to pin 4.

If you still have problems, please send us your schematic, layout. Waveforms of Pin1, Pin2 and Pin4 are also very helpful to identify any issue in your circuit.

Yang