I'm a project engineer handling electrical equipment sourcing orders for 8 years. I've personally made (and documented) 6 significant mistakes on transformer orders, totaling roughly $47,000 in wasted budget. The $12,400 one (specifying the wrong containment setup, winter 2022) still makes me wince when I think about it. Now I keep our team's pre-order checklist so nobody else repeats my errors.
What follows are the seven questions I get asked most often — plus one that people almost never ask, which is honestly the one that cost us the most money.
An oil immersed transformer is exactly what it sounds like: core and windings sit inside a tank filled with insulating fluid. The fluid does two jobs — cooling and dielectric insulation. That's basically it. The variations (sealed tank, conservator tank, nitrogen-blanketed) only change how the fluid handles thermal expansion.
Use one when you're outdoors, when ambient conditions are rough, or when cost per kVA matters. Dry-type units make sense indoors where fire load and oil containment are concerns — but they cost more at the same rating, and they don't handle hot, dusty environments nearly as well. According to IEC 60076-1, the international standard covering power transformers, liquid-filled units are generally rated for higher continuous loading in the same footprint, which is usually the deciding factor.
My $12,400 mistake: in my first year (2017), I specified a fluid-filled unit because it was cheaper on paper. It was — until the inspector required a containment basin and a spill response plan our budget didn't include. The transformer was never the problem. Everything around it was.
In practice, 500 kVA almost always means three phase. Standard single phase distribution units — the kind covered by ANSI C57.12.20 — typically top out in the 167 kVA range for pole-mounted service. If a vendor quotes you a "single phase 500 kVA transformer," ask for a nameplate photo before you sign anything. It's either a custom build or a mislabeled quote, and you want to know which one before money changes hands.
The second half of the question matters more: who picks the kVA rating? Your load study does. Not your gut, and definitely not the previous project's spec sheet. I said "500 kVA" on a bid once because it was the number we'd used before. The vendor heard "500 kVA minimum" and quoted exactly that. Our actual peak load was around 240 kVA. We paid for a bigger tank, a heavier pad, and higher no-load losses running 24/7.
Here's the checklist that would have saved me $8,900 and 11 days back in 2019:
On that bushing point: I ordered 8 pad-mount units and the secondary bushings landed on the wrong side. Nothing on the spec sheet was technically wrong. It just didn't match the pad we'd already poured. We ate the rework (ugh).
A secondary distribution unit is basically a transformer plus low-voltage switching and protection bundled into one enclosure, serving loads at secondary voltage — typically 208Y/120V or 480Y/277V — close to the point of use rather than back at a main switchboard.
The key difference is scope: switchgear is switchgear. A distribution unit is packaged, meaning the transformer and the switching are designed to work as one assembly, usually built to a project's specific configuration.
The easy mistake is treating those two categories as interchangeable. We once issued an RFQ for a "distribution unit" while actually picturing a plain switchboard. The vendor delivered an enclosure with space for the transformer — and no transformer. Honestly, that one was on both of us. We weren't clear, they didn't push back, and three weeks disappeared.
Not really, and I'd push back on anyone who markets it that way. Modular or pre-assembled substations do cut site labor — that part is real — but foundations, cable terminations, protection settings, grounding, and commissioning don't vanish. They move from the field into the factory.
On a 2022 project we used a modular setup for a medium-voltage to low-voltage distribution scope. Field time dropped by roughly 4–6 weeks. Sounds great, right? The catch is that your design has to freeze about 12 weeks earlier than you'd expect. Late changes don't get solved on site — they go back to the factory, and if modules are already built, that's expensive.
Choose modular when your design is stable and schedule certainty beats unit cost. If the drawings are still moving, slow down. I've only made this mistake once, but it cost about $9,000.
Six things come up again and again, roughly in order of how badly they hurt:
I don't have hard industry-wide data on shipping damage rates. Based on our own receiving records over five years, my rough sense is that around 5% of deliveries arrive with some issue. Small sample, so treat that number loosely.
Yeah. And honestly, that's a dumb way to run a business.
Early on, we needed a single 75 kVA single phase pole-mount replacement. Urgent spec, three RFQs sent out, three vendors never replied. I don't need to guess why: one unit, replacement quantity, small dollar value — not worth a salesperson's afternoon.
A fourth vendor answered. Quote came back same afternoon, we asked two follow-up questions, and nobody made us feel stupid for asking. That order was roughly $3,800. Since then, we've routed about $400,000 in transformer purchases through them. I'm not claiming every small buyer becomes a big one — there's no guarantee there. But if I only allocated attention by order size, that $400,000 wouldn't exist.
If you're starting out or just need one unit in a hurry, filter vendors this way: check how fast they respond to a quote request, and whether they'll take a small order without adding conditions. The ones who don't reply have saved you time. As for suppliers with formal minimum order quantities — I get it, tooling and production scheduling have real costs, and that's a different thing from disrespecting small customers. Radio silence is the part I have zero patience for.