LM5045 hiccup problems with LM5045 in current mode

Jeff Ransdell

Other Parts Discussed in Thread: LM5045

Hello TI-

I've got a serious problem with my DC/DC converter using the LM5045. The converter operates in the current mode with a current sense resistor and small RC filter. It has a small amount of slope compensation added. Both the CS and RAMP waveforms look good to me, well below the 0.75 restart current limit, and besides the RES cap is huge anyway since I don't wish to restart. My problem is that the controller periodically stops switching after a cycle and pauses for a couple cycles. When pausing (hiccuping?) the slope compensation appears to be adding but rolls off mysteriously.

This is my second layout with the same problem. I've paid very close attention to bypassing, filtering, etc, and whats more I'm using augmented gate drivers near the switches so the 5045 is only delivering a hundred mA or so. The noise on all the other SS, RES, UVLO, etc lines looks tolerable.

Any ideas? Please help I'm dying here.

FYI its a 28/600VDC hard switched full bridge with about 20Arms primary current.

over 9 years ago

0 Jeff Ransdell over 9 years ago

Prodigy 40 points

Update:

Today I shorted both the RES and CS pins, attempting to let the 5045 run open loop with max duty cycle. The problem persists.

Note: I have 50K delay time resistors.

0 Kevin LaRosa over 9 years ago in reply to Jeff Ransdell

TI__Genius 17380 points

Hi Jeff,

I've moved this issue to AC/DC forum to get the attention of the proper experts.

0 Jeff Ransdell over 9 years ago in reply to Kevin LaRosa

Prodigy 40 points

thanks. It looks like the oscillator is not stable or something. The double pulse is above the nominal switching frequency. Perhaps its noise but I don't know how to do any better with this layout:

0 Wangxin Huang over 9 years ago in reply to Jeff Ransdell

TI__Intellectual 1550 points

Hi Jeff,

The waveform look weird to me. Can you post your schematic pdf here or send it to my email w-huang3@ti.com?

Regards,
Wangxin

0 Jeff Ransdell over 9 years ago in reply to Wangxin Huang

Prodigy 40 points

Here is the controller schematic page:

p.s. Today I fixed the problem by putting a 1000pF capacitor across the RT resistor, directly on the IC pins.
I still don't understand why this was necessary.

0 Wangxin Huang over 9 years ago in reply to Jeff Ransdell

TI__Intellectual 1550 points

Jeff,

Thanks for providing the schematic. Based on the waveform and schematic, it seems the slope compensation ramp is not reset properly at the end of one of the oscillator cycles. This can be verified by measuring the slope compensation ramp current flowing through R137. If it is really not reset properly, I would agree that the part does not behave the way it should which may be caused by many reasons such as layout, bypassing, decoupling, etc. Since you said the problem was fixed by adding a 1000pF cap on the RT pin, it would be useful to compare the voltage on the pin before and after you add the cap. It is stated in the datasheet that the RT pin should be nominally regulated at 2V if the oscillator is free running.

Regards,
Wangxin

Power management

Power management forum

LM5045 hiccup problems with LM5045 in current mode