Why a flask loses its vacuum
Between the inner and outer wall of every vacuum flask there is — ideally — nothing at all. That emptiness is the insulation. When a flask "stops working", the vacuum has been compromised: a pinhole in the tail weld, a seal that didn't seat, a body knocked hard enough in handling to crack a weld line. Steel conducts heat perfectly well; it's the vacuum doing all the work.
The frustrating part for buyers is that a failed flask looks identical to a good one. It polishes the same, paints the same, packs the same. If a factory only finds out at the hot-water sample test — or worse, when your customer's coffee is cold by ten o'clock — the defect has already travelled through the entire production line and possibly across an ocean.
Our answer: test 100% of bodies, half-finished
The photo above is the actual bench, not a brochure shot — the blue crates are semi-finished bodies straight from the vacuum process. The principle is simple physics: a body that has lost its vacuum lets heat reach the outer wall, and the station senses that difference immediately. A good flask stays cold on the outside; a leaker gives itself away in seconds.
The key decision is when we test. This happens at the semi-finished stage — before painting, printing and assembly. Catching a leaker here costs us one bare body. Catching it after decoration would cost the paint, the print run and the labour too; not catching it at all would cost you a claim and us a customer.
What this means for your order
Full-batch insulation testing is the main reason our finished-goods defect rate stays below 0.5%, and why we can put retention numbers next to every model on the thermos flask category page with a straight face — 12+ hours hot for compact models, up to 24 hours for the large ones. Those figures assume the vacuum is real; this bench is what makes sure it is.
The standard behind the hours. When a supplier says “12 hours hot”, ask which test produced that number. The international reference is ISO 20342:2020 (test methods for the thermal performance of vacuum flasks), and premium and private-label buyers increasingly ask for it by name. Its benchmarks are specific: contents still above 60 °C after 6 hours when filled hot, still below 13 °C after 12 hours when filled cold, plus a vacuum-integrity check.
Two things worth knowing before you compare numbers between suppliers. First, the test is run on a full, sealed, pre-warmed flask — a half-empty one opened twice behaves very differently, and that is physics rather than a defect. Second, “ISO 20342 tested” and “ISO 20342 certified” are not the same claim; it is a test method, not a certification scheme. Ask for the report, not a logo. We can arrange third-party testing to ISO 20342 for your specific model when a programme calls for it.
It's also why we invite scrutiny rather than avoid it. Our lines have passed Hallmark's DG-Lab function-reliability testing — age and humidity cycles, thermal function, 25-cycle use tests — and third-party inspectors are welcome to stand at this bench and watch bodies pass or fail in real time.
Auditing any supplier's insulation QC — three questions
Whether you buy from us or anyone else, these three questions separate factories that test from factories that hope:
1. "Do you test every body, or a sample?" Sampling catches systemic problems; only 100% testing catches the random pinhole. For vacuum ware, random is the failure mode that matters.
2. "At what stage do you test?" Semi-finished is the honest answer. Testing only finished goods means defects were decorated and assembled first — a sign the factory optimises for line speed, not your claim rate.
3. "Can my inspector watch?" Any factory genuinely running the test will say yes without hesitation. Hesitation is data.
Frequently asked questions
Written by the Jupeng Drinkware team — Yongkang, Zhejiang, China. Manufacturing drinkware since 1998. Contact Beyond: [email protected] | WhatsApp +86 156 5791 8881