Tips for inspecting vintage oil-cooled units that nobody talks about

I dealt with a pair of 1958 Westinghouse units at a municipal substation last summer. I followed that advice about flushing the drain valve, but on those specific units, I found I had to clear closer to a gallon before the sediment settled down. The old brass fittings tend to corrode internally, and if you don't clear the line thoroughly, your DGA results often come back with false positives for copper and iron that can realy mess up your trend analysis. One thing I've noticed on those external cooling pipes is that the gaskets aren't just leaking oil; they're often drawing in air during thermal cycling. I started using an ultrasonic leak detector on the upper flanges near the conservator tank. Even if there's no visible wetness, you'd be surprised how much ambient air can bypass a brittle cork gasket when the load drops at night and the tank pressure goes negative. Regarding the cloudiness, I once saw a unit where the oil looked like dark coffee. We pulled a sample and the dielectric strength was down to 18kV. We ended up doing a full oil reclamation with a mobile rig, and that old beast is still humming along at 110% nameplate capacity during peak hours. Those vintage windings have way more copper in them than anything you'll find in a modern shop.

While the discussion regarding sediment and gasket permeability is certainly pertinent, I whould strongly advise inspecting the Buchholz relay’s float assembly if these units are equipped with a conservator, as the mercury switches in those vintage Westinghouse models frequently suffer from oxidation that leads to nuisance trips or, conversely, a failure to actuate during a geniune gas accumulation event. Furthermore, if you are observing that negative pressure during thermal cycling, you must verify the integrity of the silica gel breather's oil seal becasue an improperly maintained trap there will allow atmospheric moisture to bypass the desiccant entirely, wich effectively negates your efforts to preserve the dielectric strength of the insulation.

I'm pretty new to all this and honestly half of what you guys are saying sounds like another language. You mentioned the oil looking like dark coffee—how can a beginner tell if it’s just old or if the whole transformer is acutally about to blow? Also, where do you even get an ultrasonic leak detector, and is it hard to learn? I’d be terrified of missing an air leak just beacuse I didn't know what to listen for.

Spot on about those old Westinghouse units being over-engineered. I've seen them handle massive overloads without breaking a sweat, but you hit the nail on the head regarding the oil and seals. One thing people often overlook on those cylindrical tanks is the thermometer well. On the older models, the seal around the dial thermometer housing can dry out just as bad as the bushings, and you'll get moisture seeping directly into the top oil layer right where it’s hardest to spot. I always give the glass a quick tap; if you see any internal condensation on the lens, that's a dead giveaway your top seal is shot.

Quick update — I took that advice about the drain valves and ended up flushing closer to a gallon like suggested. You were right; that sediment was stubborn, but once it cleared, the DGA came back clean. Also checked the thermometer well while I was at it. These old beasts are running smooth again. Thanks for the tips, defintely saved me some headache on this site.