LM25037: Gate drive pulses not the same length

Daniel Smirr

Part Number: LM25037

Dear TI-team,

I am currently working with the LM25037 as a PWM controller in current mode in a push pull design, around 24V input / 15V output.

I attached the circuit of the LM25037.

Current sensing is made by a primary current sense transformer, feedback is made by an isolated error amplifier.

I recognized the problem that the pulse width of the OUTA and OUTB drive pulses is not equal in length. At lower loads the problem is increasing (as shown in the attached screenshots), with higher loads the problem gets better, but the difference doesnt disappear.

The OUTA and OUTB signals are connected to a mosfet driver and then to the mosfets.

At the moment I disconnected the isolated error amplifier from the COMP pin and connected a potentiometer to the COMP pin to adjust the output voltage manually. The problem stll persists. At the beginning I had RAMP connected to REF via 165k and to CS via 470pF (wasnt better).

Could you tell me, what could cause the LM25037 to vary the pulse widths of OUTA and OUTB?

I attached several measurements of CS, RAMP and COMP pin.

Many thanks in advance

Kind regards

Daniel

over 4 years ago

0 Aidan Davidson1 over 4 years ago

TI__Intellectual 2640 points

Hi Daniel,

Thanks for reaching out and for your interest in the LM25037! Do you have a filter on your CS signal before it's injected into the RAMP pin?

Best,

Aidan Davidson

0 Daniel Smirr over 4 years ago in reply to Aidan Davidson1

Prodigy 80 points

Hi Aidan,

yes the corresponding circuit looks like attached. All filter components are located next to the LM25037. I already tried some smaller or bigger values of the RC filter combination.

Kind regards

Daniel Smirr

0 Aidan Davidson1 over 4 years ago in reply to Daniel Smirr

TI__Intellectual 2640 points

Hi Daniel!

I'll get you an update on this Tuesday.

Thanks,

Aidan

0 Daniel Smirr over 4 years ago in reply to Aidan Davidson1

Prodigy 80 points

Hello Aidan,

do you have further informations for me?

Kind regards

Daniel

0 Aidan Davidson1 over 4 years ago in reply to Daniel Smirr

TI__Intellectual 2640 points

Hi Daniel!

Many apologies for the delay here! I'm still waiting on some feedback from a colleague on this topic, once I hear back from him I will get back to you!

Thanks so much for checking in, I look forward to talking more soon!

Aidan Davidson

0 Daniel Smirr over 4 years ago in reply to Aidan Davidson1

Prodigy 80 points

Okay, thank you very much for the update!

Kind regards

Daniel

0 Aidan Davidson1 over 4 years ago in reply to Daniel Smirr

TI__Intellectual 2640 points

Hi Daniel!

I got some feedback from my colleague on this topic, we think that this shortened pulse on your OUTA signal could be caused by the leading-edge spike on your current sense signal. Increasing the capacitance in your RC filter further should resolve this issue.

Thanks for your patience, I hope this information helps!

Aidan

0 Daniel Smirr over 4 years ago in reply to Aidan Davidson1

Prodigy 80 points

Hi Aidan,

thanks for your reply. In the meantime I already increased the RC filter capacitance to remove the spike.

But there was still another problem: The wire of the COMP signal is unfortunately relatively long and located under the signal plane of one of the power mosfets. So the signal was applied with an alternating ac voltage. Changing the path of the signal makes it a lot better.

Thank you for your help

Kind regards

Daniel

0 Aidan Davidson1 over 4 years ago in reply to Daniel Smirr

TI__Intellectual 2640 points

Hi Daniel!

I'm glad to hear that the RC filter capacitance and changing the signal path resolved your issue! Since it sounds like your problem is resolved here, I am closing this thread. If you have any further questions, please post again here or open a new thread.

Thanks for the update!

Aidan

Power management

Power management forum

LM25037: Gate drive pulses not the same length