LM5064: Huge reflection Power Supply spikes when Board is inserted

Jordi Ferre Ferran

Part Number: LM5064
Other Parts Discussed in Thread: TPS2393

We are having issues with new hotplug used on latest board developments. Customer has complained that when they insert one board in the chassis, other boards already inserted are affected ( Board resets ). We didn’t saw that issue in “old” boards.

New boards are equipped with: LM5064PMHE/NOPB

Old boards are equipped with: TPS2393APW

We have been doing a lot of different tests past weeks but we are not able to find the issue, neither to improve results. We have noticed that a huge spike is generated when we insert the board which affects the other boards on the rack. That spike has a smaller slope in Old hotPlug

over 2 years ago

0 Rakesh over 2 years ago

TI__Guru** 105151 points

Hi Jordi,

Thanks for reaching out.

We have local holidays here till Wednesday. We will reply by Thursday

Sorry for the delay.

Best Regards,

Rakesh

0 Avishek Pal over 2 years ago

TI__Mastermind 31780 points

Hi Jordi,

Please share the schematics of both old and new designs. You may also refer to this app note, which discuss this issue.

Handling System Transients in Hot Swap Applications

0 Jordi Ferre Ferran over 2 years ago in reply to Avishek Pal

Prodigy 50 points

Hi,vDCU-hotplug_LM5064.pdf TRP25G-hotplug_TPS2393APW.pdf

Attached both schematics.

He had already seen the application note that you mentioned. Very interesting but it doesn't solve our issues. Maybe there is a difference between both Hot Plug circuits that we do not understand or maybe, you can give us some indications for new designs in order to avoid such huge reflections.

In "old designs" using TPS2393APW, the reflections are not so big and slope is also smaller. We would like to reduce those reflections in both cases, but specially with LM5064PMHE/NOPB.

Regards

0 Avishek Pal over 2 years ago in reply to Jordi Ferre Ferran

TI__Mastermind 31780 points

May I know what are you referring as "reflections"? If you are talking about any waveforms, please share those.

0 Jordi Ferre Ferran over 2 years ago in reply to Avishek Pal

Prodigy 50 points

Hi,

Attached the reflections measured in inserted boards on the rack. The mesurement is done on Supply voltage ( -48V ) of those inserted boards.

You can see that reflection amplitude is bigger in LM5064. Also, slope is bigger.

0 Avishek Pal over 2 years ago in reply to Jordi Ferre Ferran

TI__Mastermind 31780 points

What are the two nodes across which this voltage was measured?

0 Jordi Ferre Ferran over 2 years ago in reply to Avishek Pal

Prodigy 50 points

hi Pal,

We are measuring the main supply voltage of iserted boards in the rack. ( -48V )

0 Avishek Pal over 2 years ago in reply to Jordi Ferre Ferran

TI__Mastermind 31780 points

Hi Jordi,

It appears that the Inrush current profiles for "Old Board" and "New Board" are different. TPS2393APW is expected to have lower inrush current than LM5064PMHE/NOPB during the hot insertion of the board. If the inrush current is large, there would be a large voltage drop on the back plane. This might be the reason.

In case of TPS2393, there is no capacitor across RTN and −VIN pins in your schematic.

0 Jordi Ferre Ferran over 2 years ago in reply to Avishek Pal

Prodigy 50 points

Hi Pal,

Thanks for exaplanation regardng Irush current. But our goal is to reduce muximum this huge reflections in "old" and "new" desings using TPS2393APW or LM5064PMHE/NOPB.

Do you have any recomendations from our schematics in order to "reduce" those reflections when we insert a board ? Any modifications in schematics of old or new boards ?

Thanks

jordi

0 Avishek Pal over 2 years ago in reply to Jordi Ferre Ferran

TI__Mastermind 31780 points

We can use a resistance in series with the capacitor as shown below to limit the inrush current during hot-plug.

Power management

Power management forum

LM5064: Huge reflection Power Supply spikes when Board is inserted