U3 can be damaged by plugging in breakout board when power is on

[jonnew]

The PCIe controller with onix-fmc-host attached is inside a computer and powered on when the breakout board is plugged via the shrunk-delta ribbon cable, U3 can become degraded and eventually fail catastrophically. This never occurs if the breakout board is plugged when power is off.

• Is there some way to keep this chip disabled until the breakout is detected?
• Would this even help? Given that this is a reciever, we need top probe the input lines during a hot plug in event to see what they are doing.

[jonnew]

@filcarv Potential drop in replacements to try:

• https://www.ti.com/lit/ds/slls490b/slls490b.pdf
• https://www.ti.com/lit/ds/symlink/sn65lvds9637.pdf?HQS=dis-dk-null-digikeymode-dsf-pf-null-wwe&ts=1725031194981&ref_url=https%253A%252F%252Fwww.ti.com%252Fgeneral%252Fdocs%252Fsuppproductinfo.tsp%253FdistId%253D10%2526gotoUrl%253Dhttps%253A%252F%252Fwww.ti.com%252Flit%252Fgpn%252Fsn65lvds9637
• https://rocelec.widen.net/view/pdf/jfm1yhcaj1/ds90lv032a.pdf?t.download=true&u=5oefqw

[filcarv]

Could you check if these PN and cases are the right ones?

SN65LVDS33PWR 400 MBps SN65LVDS32PWR 100 Mbps DS90LV032ATMTCX/NOPB 400 Mbps

[jonnew]

Yes, here is my verification:

Original part: FIN1032MTCX

• TSSOP-16 4.4x5mm with 0.65mm pitch

Replacements:

1. SN65LVDS33PWR

   • TSSOP-16 4.4x5mm with 0.65mm pitch

2. SN65LVDS32PWR (datasheet says 150 Mbps, not 100)

   • TSSOP-16 4.4x5mm with 0.65mm pitch

3. DS90LV032ATMTCX/NOPB

   • TSSOP-16 4.4x5mm with 0.65mm pitch

[jonnew]

In addition to this, I also ordered some TVS diodes:

• RCLAMP0582B.TCTCT-ND
• RCLAMP3552TQTNTCT-ND

To try to hack into the U3's LVDS RX ports:

I want to know if this solves the issue and is a bodge we can make on the FMC units. Might be even easier on back because the termination resistors provide nice pads. The ground connection can just be made with scraped solder mask.