LM5175: output voltage issues - Overvoltage

Hi Bodo,

I am not if I fully understood the description above and the issue you see.

So my understand is that you have some units (returned from customers) which are still working as expected with only one exception when decrease the input voltage from high (what is high?) to 20V.

In this case the voltage increases first to 13.5V and then the controller turns off for 10ms and then restarts with normal operation.

Please confirm or correct if necessary.

To your questions:

• Are there known topics which have the same behavior?
  Never seen this behavior before
• Do you recommend some additional measurements?
  YEs, please provide measurements with (and zoom in to the area where the increase and shutdown happens
  - COMP
  - SS
  - VIN
  - VOUT
  Provide Schematic
• Is it good to x-ray the defect IC. Is it possible that we see some defects?
  Do not think you will see anything with X-Ray.

Best regards,

 Stefan

Hey Stefan,

yes thats the case with the customer. Our internal µC detects an 12V5 abnormal behaviour and sends an software error. Due to this error we got our devices back from the customer to investigate. The device is still working in all other aspects. High in our case is 20V ... 32VDC and worst-case transients to 58V. 

One difference is the power consumption. The malfunction IC has 8,9W power consumption than a good IC 7W (base load) at 24V input. At other input voltages the consumption varies to.  

The 13,5V controller switch off is correct. We measured the time more precisely, it is 9ms instead of 10ms. In this time the high side input mosfet Ugs=0V. 

One inconsistency in my mind is that the bode-plot shows stability. On the other hand a passive probe on Pin 27 vs 0V results in no more occuring problem. For me this means bad layout and or noise coupling into the IC. Is this a possibility? 

I will make the measurements. They will take some time to complete. 

Can i send you the schematic in private? Dont think i am allowed to share it here.

Hi Bodo,

to send the schematic on a private chat just hover with the mouse over my name and select send private message.

Please drop a note here to let me know to check for that private message.

Best regards,

 Stefan

Hey,

send you a private message with the pdf file. 

Hi Bodo,

i just entered the data from the schematic into the quickstart calculator. 

Here it looks like the design can easily run into instabilities.

Have you checked the design with a Bode analyzer?

Also the slope compensation does not follow the recommendation.

Best regards,

 Stefan

Hey Stefan,

yes we did check the design with a bode analyzer from 9V ... 32V input voltage. In my opinion it is stable with enough phase margin etc.

At some points our gain and phase has small fluctuations in the curves. Is this due to possible instability? I will also check the design with quickstarter calculator.

Do you have some recommendation for slope and comp, should i use the data sheet values?

Hi Bodo,

slope compensation and Compensation parameters need to be calculated based on the power stage parameters.

The math is given in the datasheet - you can also use the quickstart calculator to make it easier.

Best regards,

 Stefan

Update to the schematic i send you. I forgot all output capacitors 2,2mF who are placed on another schematic page as buffer. Sorry for that.

i calculated and simulated our design per hand and with the quickstart guide. The system is stable after standard stability criteria. 

i looked at our available bode-analyzer data. i found one measurement what was declared as "wrong measurement" with 32V input. There was a phase drop to around -150° at a gain near 0dB. More measurements with less critical behaviour is available. 

here the picture of stated bode-measurement. im am not sure how to interpret it on stability criteria. can you give a hint?