The short version
Before you send an RFQ to six suppliers, make sure your spec sheet has three things locked down: seat and seal material (with the temperature rating), end connection type, and the applicable leak test standard (API 598 vs. ISO 5208 are not the same thing). Miss one of those and it's a $1,500–$4,000 mistake plus a one-to-two-week delay. These days I only work with globe valve suppliers who ask about those three items before I bring them up.
Everything below comes from nine years of procurement work, paid for in rework invoices and a couple of very uncomfortable Monday-morning calls.
Why I'm allowed to write this
I've been handling industrial valve orders since 2016 — globe valves, gate valves, ball valves, the whole catalogue for a mid-sized contractor. I've personally made and documented a couple of dozen significant sourcing mistakes, and I can put a dollar figure on most of them. Rough total: somewhere north of $9,000 in waste, rework, and expedited freight.
In my first year, I made the classic rookie error: I assumed "carbon steel" and "standard" meant the same thing to every globe valve manufacturer. They don't. A supplier quoted us a carbon steel globe valve that met our price target. We placed the order. The end connection was flanged per a different dimensional standard than the rest of our line, and the seat material wasn't right for our steam application. We caught it at receiving — 40 pieces — and the rework plus return freight came to about $1,100. That's when I stopped trusting the word "standard."
By 2022 I thought I knew the drill. Then in November of that year we placed a $3,400 order for carbon steel globe valves for a steam distribution project. Everything looked fine on the confirmation. It wasn't until commissioning that the installer flagged weeping at the packing gland on 12 of 30 units. Root cause: our RFQ said "graphite" packing but didn't specify the temperature rating. The supplier quoted a grade rated to 400°C. Our line runs at 425°C. No injuries, no shutdown — but we replaced the packing on all 12 units, re-flushed the line, and pushed commissioning by four days. That one cost $890 in on-site labor plus $400 in expedited replacement packing (unfortunately, my "I'll double-check the spec tomorrow" never happened).
The three checks that would have caught most of my mistakes
We've run this short list on roughly 260 valve POs over the past three years. It's caught 47 potential discrepancies — everything from wrong end-face finishes to mismatched pressure ratings. Here's the short version.
1. Seat and seal material — including the temperature rating
"13% Cr" alone is not enough. Neither is "graphite." You need the grade and the temperature range. For steam above 400°C, that generally means a specific graphite packing grade or a metal-to-metal seat design. ASME B16.34 is the baseline design standard most globe valve manufacturers work from, but it does not prescribe your packing — that's on your spec sheet. This is the one that got me in 2022.
2. End connections — type and dimensional standard
Flanged to ASME B16.5 Class 150, or to EN 1092-1 PN16? These are not interchangeable without adapters, and every adapter is a new leak path. I've seen this come up more often with suppliers working from translated spec sheets where "flanged" gets used as a generic word. It isn't. Write the standard number, not the description.
3. Leak test standard — and the specific rate or class
API 598 and ISO 5208 are both common references for valve leak testing, but they define test methods and acceptance criteria differently. If your RFQ just says "leak tested" and nothing else, expect a shrug when the valves land. Say which standard and which rate (for example, ISO 5208 Rate A vs. Rate D — big gap between those two).
The globe valve vs. ball valve question, because someone always asks
Here's what you need to know: this isn't a "which is better" comparison. It's a "which one is right for the line" comparison, and buyers get burned when they flip the question.
If your application needs fine throttling and tight shutoff on a steam or high-temperature line, a globe valve is usually the correct choice. If you need fast on/off with a compact footprint, a ball valve usually is. I once watched a client try to use ball valves for throttling on a condensate line — they choked the flow and scored two seats in six weeks. That's not a knock on ball valves. It's a knock on using any valve outside its intended service.
Bottom line: when you're sourcing, don't ask "which one is better." Ask "what is this valve actually doing on the line." The second question answers the first one every time.
Where this checklist doesn't apply
Two caveats, and I'd rather say them than pretend every situation is the same.
First, if you're sourcing for a low-pressure ambient-temperature water application — utility water, basic cooling loops — the temperature-rated packing question matters far less. Any reasonable spec will hold. The other two checks still apply most of the time.
Second, if you're buying from a globe valve supplier with pre-built stock and you genuinely need valves this week, you may not have time for a full spec conversation. In that case, ask for the datasheet of the exact item in stock and read it yourself. If they can't produce a current datasheet in twenty minutes, that's a red flag. Walk.
And one honest boundary, since I promised one: I'm not a valve engineer. I'm a buyer who's been burned enough times to know which questions matter. If your project involves high-pressure steam above 600°C, cryogenic service, or a hazardous fluid, bring in an engineer. Seriously — I do.
That's the whole thing. No clever framework, no magic supplier list. Just three items most people skip (me included, twice) and a short conversation with the supplier before the PO goes out. It's saved us real money. It'll probably save you some too.