View source for Agilent E3633A/E3634A

{{Device page
|Manufacturer=Agilent
|Device type=Other
}}
{{stub}}


== Guides ==
{{List Guides}}

== Device pictures ==
<gallery showthumbnails="1">
File:Example device pictures.jpg
</gallery>

== PCB pictures ==
<gallery showthumbnails="1">
File:Example pcb pictures.jpg
</gallery>

== Reference measurements (also schematics if available) ==
<gallery showthumbnails="1">
File:Example measurement pictures.jpg
</gallery>

== More Information/External Sources ==

Both Agilent E3633A and E3634A are covered by the same service manual<ref>https://ridl.cfd.rit.edu/products/manuals/agilent/power%20supplies/cd1/service/E3634ser.pdf</ref>.

=== Risk of easy damage while debugging. ===
The '''GND/VSS''' pins of the logic on the display board is powered by the -17.4V power rail, while the '''Vcc''' supply is actually an internally generated -12.4V power rail. But at the same time, all the metal work around the VFD, the metal parts of the encoder, and also lots of exposed copper traces around the contacts of the membrane keyboard is connected to 0V, which is marked as "down-arrow A" on the schematic. This 0V, actually looks like a +17.4V compared to the GND pin of the logic devices. You must be VERY CAREFUL while probing anything around that board, to not short any logic pins to this exposed metal work, or you can burn out the chips on that board. If the MCU (with firmware in mask rom) has died on the display board, it can be replaced with a programmable replacement, the firmware has been dumped, and there are multiple pin-compatible replacements available, see <ref>https://www.youtube.com/watch?v=wpOJ_3-evMw</ref> and <ref>https://www.ianjohnston.com/images/stories/IanJ/Agilent/n80c51bh_E3633A_Display_Board.zip</ref> for more information.

Furthermore, the membrane must at all times be absolutely perfectly aligned above the front panel PCB. If there is even the smallest misalignment, then pressing a button could create a path to the 0V power rail (which really looks like +17.4V to the CPU), and this can destroy the CPU. (or sometimes it will only reset it if the current sent into the CPU pin is small enough)

Probing around the filament pins is also dangerous, if you accidentally short one of the filament pins to the metalwork around the VFD, then you will likely burn the "VR1" 6.2V zener diode on the mainboard. This is because the filament power is generated by a center-tapped transformer winding. The AC that is generated is around 5V, however the center tap voltage is at -11.2V which is generated using VR1 zener diode. If you short one of the filaments to the metal work, then the zener diode will be overloaded and it will pop.

=== Typical fault (possible design flaw) ===
The 2.4V zener diode "CR9" has been observed as blown on multiple examples of this device. (Observed as blown by the author of this article, but also see <ref>https://www.eevblog.com/forum/repair/hpagilent-3634a-repair/</ref> and <ref>https://www.youtube.com/watch?v=3EjmCWmarwE</ref>.

People suggested that it may be a design flaw, and recommended to replace it with a Zener diode in a bigger package that is able to dissipate more current.