MF NanoAuger: Vacuum Loss SEM Electronics Failure

Vacuum Loss

After bakeout and testing, ion gun turbo was on with gate valve.
Overnight analysis chamber lost vacuum. Probably turbo roughing pump stopped, then turbo lost speed and vented, causing ion pumps & gauges to stop.
SEM electronics running, software also but unable to control system.
Multiple errors on restarted SEM, Software could not communicate with hardware, did not identify boards.

In EM server got many "Vac Check" errors (~5) then Vac, EO, etc boards were not ID'ed (no revision, serial number, firmware).
This is a symptom of loss of +24 V power to vac board. Vac board runs a series of checks on power up (column valve, pressure, water flow etc). It quits on first failure, does not (usually) provide clear error messages.
Quick check is used volt meter to check for +24 V on gun column interlocks. If absent bad sign.
Next remove right side of SEM console, power board is exposed.

The power supply consists of 3 transformers in a shielded box, with a large PC board with multiple supplies (+5 V, +15 V, -15 V, +24 for optics/lenses, +- 15 for optics and scan, etc.
Each supply has a green LED to indicate function. Starting from the back, these are +5 V. + and - 15 V, then the 4th is 24 V for VAC board. In standby mode these 4 are on. In Run mode, once the vac tests are passed, the other (at least 6) LEDs come on.

Lots of fuses, On rear panel, externally accessible fuses in a long row, F1 to F15. These are mostly for external accessories including switched 208 VAC plugs. Also large breaker
Behind side panel, power board, large fuses below power board for transformer primary. Also a row of fuses on power PC board for individual supplies, (5x20 mm glass or ceramic).
Many pin fuses on vac board (flip-down board behind rear upper panel). These are brown 8x6 mm cylinders perpendicular to PC board with two pins. Can be bought from Newark, some spares in Zeiss kits. Also cartridge fuses on "piggy board" on vac board that supplies interlocks. These can blow if exposed metal interlock parts are accidentally shorted to chamber.

Each supply has a green typically 6 wire connector at the top bringing in the appropriate transformer output (each has multiple windings). The AC input is fused then routed to a diode bridge on a water cooled heat sink above the large board, and bridge output back to board.
Each supply has 2 or more 4.7 mF filter caps, usually two high-current nFET power transistors on the water-cooled brass heat sink at the bottom of the board, a TL431 shunt voltage reference with a trim-pot for adjusting the voltage, and a TVS zenner protection diode.
Failed 24 V supply had the TVS protection diode blown. When replaced, blew again.
Suspect shunt volt ref failed, leading to FETs dropping to minimum resistance, resulting in ~ 30+ volt rectifier output appearing across 28 V diode, which then failed.
Replacing both nFETs, TVS diode and TL431 ref fixed power supply. FETs might have been OK but taking board on and off is a pain.