LMR38020-Q1: Issue with simulation results

Jim Brinkhurst1

Part Number: LMR38020-Q1

Hi experts

I used PSPICE to simulate a 42V to 2.5V, 2A converter at 500kHz and 24V enable using LMR38020-Q1DDA_TRANS. I ran into frequency foldback which makes sense then I reduced the switching frequency to 200kHz so that the on time was increased from 119ns to 298ns. This is now above the Ton-min of 131ns from the datasheet, granted this was given at 24V. With the 298ns on time, the frequency still folded back after some millisecond (see Fig 1). I even increased the output voltage to 3.3V at 200kHz and still saw some foldback with the new 393ns on time. This is different from what was said in the datasheet.

1) Have you heard of any complaint about this before and is there a fix?

Fig 1: Unpleasant foldback transition (inductor current, Load current, Switch node, output voltage)

2) The datasheet says the foldback frequency could go as low as 133kHz. Does this mode have a fixed frequency?

3) Do you have the ton/toff min at other voltages?

4) The startup from sim is way different from the result in the datasheet. I can see the 4ms soft start and the 18ms overcurrent blanking time prevents a shutdown but the 3.8A I_HS_Limit is not clamping well at startup. Have you encountered this before? Do you think this will occur in real life?

5) The output voltage also goes negative and this should not be per the footnote on pg 4 of the datasheet. Do you know why this has happened? How well is the sim correlated to real life?

Fig 2: startup transition (Load current, inductor current, Switch node, output voltage)

WEBENCH sim looked smooth so I am glad I used PSPICE to check what I was seeing.

What are your thoughts?

Thanks for your help!

Best regards,

Jim B

over 2 years ago

0 Colin Chen over 2 years ago

TI__Mastermind 21361 points

Hi Jim,

I also have simulated psipce model on your condition and don't found the waveforms as you showed,

Fig 1: startup transition (Vin, Switch node, inductor current, I_load, output voltage)-There is no negative output voltage and no overcurrent.

Fig 2: steady mode (Vin, Switch node, inductor current, I_load, output voltage)-There is no frequency foldback.

This is my schematic:

Whatever input voltage and output voltage, the Ton_min shouldn't be less than 80ns.

Colin

0 Jim Brinkhurst1 over 2 years ago in reply to Colin Chen

TI__Mastermind 44680 points

Hi Colin

I saw your response and I found the issue. Replacing the current sink with a resistor seems to remove the weird overshoot at startup. If you change the resistor to a 2A current sink you should be able to reproduce the issue.

Moreover, the premature frequency foldback still occurs. My sim parameters are: 24VIN, 2.5VOUT, 2A, 300KHz operation. From the zoomed figures below you can see that the frequency dropped from about 295kHz at the start to around 239kHz. The ON_TIME is well over 131ns from the datasheet and the duty cycle of 10% is well above the Dmin of 3.93% (i.e. 131ns*300kHz). Not having a well-defined or having a wide frequency range of operation makes it difficult to optimize input filters.

Kindly try it out and let me know if you can reproduce the same issue at your end. What do you think is the solution?
Any differences in die between LMR38020SDDAR/LMR38020FDDAR/LMR38020FSDDAR and their respective Automotive versions? What are the differences?

Thanks for your help.

Jim B

0 Colin Chen over 2 years ago in reply to Jim Brinkhurst1

TI__Mastermind 21361 points

Hi Jim,

No problem, I will reproduce and update soon.

Auto version converter has AEC_Q100 qualification, and Non Auto version don't have this one. Besides , they are no different.

Thanks

Colin

0 Colin Chen over 2 years ago in reply to Colin Chen

TI__Mastermind 21361 points

Hi Jim,

I have reproduced and didn't see your simulation result. The frequency don't change as below shown:

Fsw=297kHz at 10ms:

B R

Colin

0 Jim Brinkhurst1 over 2 years ago in reply to Colin Chen

TI__Mastermind 44680 points

Hi Colin

Can you share your simulation file? I would like to compare the differences to my simulation setup.

Thanks!

Jim B

0 Colin Chen over 2 years ago in reply to Jim Brinkhurst1

TI__Mastermind 21361 points

Hi JIm,

The simulation file I used is downloaded in https://www.ti.com/lit/zip/snvmc93.

B R

Colin

0 Jim Brinkhurst1 over 2 years ago in reply to Colin Chen

TI__Mastermind 44680 points

Thanks Colin

I will email you the design I am working for comparison. It is from a design created in WEBENCH.

Regards,

Jim B

0 Colin Chen over 2 years ago in reply to Jim Brinkhurst1

TI__Mastermind 21361 points

Thanks Jim, got it.

B R

Colin

0 Jim Brinkhurst1 over 2 years ago in reply to Colin Chen

TI__Mastermind 44680 points

Hi Colin

Please let me know if you see anything in my design that would cause the frequency foldback behavior seen.

Thanks,

Jim B

0 Colin Chen over 2 years ago in reply to Jim Brinkhurst1

TI__Mastermind 21361 points

Hi Jim,

Got it, will feedback soon.

Colin

0 Jim Brinkhurst1 over 2 years ago in reply to Colin Chen

TI__Mastermind 44680 points

Hi Colin

I don't think I've heard anything by email. Do you have any updates on this request?

Thanks,

Jim B

0 Colin Chen over 1 year ago in reply to Jim Brinkhurst1

TI__Mastermind 21361 points

Hi Jim,

Sorry for responding late due to CN holiday.

I reviewed your schematic and it looks not bad except the inductor is a bit big. It is better to reduce to 15uH.

You can use the model I sent and try again to see it.

Thanks

Colin

0 Jim Brinkhurst1 over 1 year ago in reply to Colin Chen

TI__Mastermind 44680 points

Hi Colin

I will check out the sim you sent but I have a feeling we may be slightly off track. The last thing was, I sent the sim I am using, and you were going to check to see what could be causing the foldback. This was about 18days ago. Were you able to do duplicate the foldback in your simulation?

It is nice that sch was reviewed but the inductance value needs to be larger to allow operating from 24-42V without changing the design.

Moreover, we might have a bigger issue as I have reproduced the foldback in the lab. Results are attached below.

What do suggest?

Thanks for your help.

Lab Results

As feared, the premature frequency foldback showed up when converting 42Vin to 2.5Vout, 2A. Find attached (the traces are Vin, Vout, PGood/Switch node and inductor current respectively). I used LMR38020FS automotive. I plan to increase the output voltage to the point where I don’t see this issue anymore and that will dictate the actual output range of the controller.

Jim B

0 Colin Chen over 1 year ago in reply to Jim Brinkhurst1

TI__Mastermind 21361 points

HI Jim,

Sorry for responding late, 48V to 2.5V can meet Ton-min spec (131ns max) that Fsw will foldback.

As you said it need increase the output voltage or decrease the Fsw you set to keep Fsw same.

Thanks

Colin

Power management

Power management forum

LMR38020-Q1: Issue with simulation results