Had a weird failure of Dwyer Photohelic pressure gauge, where a lab glovebox started rapidly pulsing its solenoid pressure-supply valve. This oscillation stopped if I gave the inflated gloves a whack, or cycled AC power. Vacuum-supply solenoid valve stuck on?

No, instead the lamps in the Photohelic gauge are getting dim with age!

The dim lamps have shifted the gauge needle photointerrupter too close to switching threshold ...and when the big relays turn on, the lamp voltage drops slightly, which turns one relay off, which lets the lamp get brighter, which turns one relay on ...etc.

Perhaps all very old Photohelic gauges eventually do this? Or maybe we just got a bad one.

The cu look at the volts across one photocell or the other (on the little hex plug it's the white wire pair, or red pair,) then move the manual high/low needles on the gauge to notice the voltage threshold which audibly switches the big Potter-Brumfeld relays. On the red wires the normal voltage was 8.9VDC, but the threshold was only 9.1V!!! Fortunately there's a 4.7K resistor on the PCB in series with the red wires to the photocell. Soldered another 4.7K resistor in parallel to lower the series R and raise the photocell switching threshold from 8Kohms to 10.5K. Oscillation gone.

I notice that later model Photohelic gauges are using LEDs in their photointerrupters meter-needle thingies.

http://staff.washington.edu/wbeaty/chem_photohelic.html