You measured carefully, you mixed thoroughly, and the epoxy still cured soft and tacky in one corner. Nine times out of ten the culprit is not the resin — it is the ratio, and specifically how you metered the two parts. The two cheapest, most common tools for hitting a 1:1 mix are a calibrated dispensing pump (pull a stroke of resin, pull a stroke of hardener) and a graduated measuring cup (pour each part to a printed line). Both are sold as accurate. Both fail in ways that are easy to miss until a batch refuses to cure. This guide pits them against each other, with the real numbers, and tells you which to reach for at which batch size.
Here is the short version. For repeat batches at volume, a calibrated pump is faster and more repeatable because it removes the human reading error of a cup — but it is only as accurate as your full-stroke discipline and your last calibration check. A stuck valve or a half-pulled stroke silently corrupts the ratio with no visual warning. A graduated cup lets you see the ratio, but volume-by-eye carries meniscus, wall-film, and printed-graduation error that grows fast as batches shrink. And the honest answer that overrides both: for anything under roughly 2 oz, a 0.1 g digital scale beats every volume method. The comparison table and the specs below lay all six options side by side; this guide explains the why.
Why the 1:1 ratio is non-negotiable
Epoxy is not paint, where “about right” is fine. It is a two-part chemical reaction, and the cure depends on a precise stoichiometric balance between resin and hardener. Get the ratio wrong and you do not get a slightly weaker part — you get a part that cures soft, tacky, or never fully hardens, because there is leftover unreacted resin or hardener with nothing to bond to. That tacky surface can stay chemically active for days.
The good news is that epoxies have some tolerance. Most systems shrug off roughly a 5% ratio error, and 1:1 (50:50) systems are frequently forgiving out to about plus or minus 10%. The bad news is that “5% of a small batch” is a much tighter target than it sounds, which is the entire reason measuring method matters. The looser the chemistry’s tolerance, the more your tool choice forgives you; but you rarely know the exact tolerance, so the safe move is to meter as accurately as the batch size demands.
How metered pumps work — and the partial-stroke trap
A calibrated metering pump bakes the ratio into the mechanism instead of into your attention. The gold-standard example of discipline-based metering is the WEST SYSTEM 300 Mini Pumps: one full stroke of the resin pump plus one full stroke of the hardener pump delivers the correct ratio — about 0.8 fl oz of resin per full stroke on the 105/205-206 system (0.9 fl oz on 105/207-209) — without you ever measuring a volume. You do not count milliliters; you count strokes.
That is also the trap. The ratio is only correct on a full stroke — you must bottom the pump out and let it return completely. A partial stroke gives the wrong ratio, full stop. Worse, a mechanical pump can stick or backflow into the reservoir, under- or over-dosing one side with zero visual warning. The pricier end of the category tries to engineer the discipline away: the TotalBoat MakerPoxy 1:1 pump ties both pistons together so one physically cannot move without the other, holding volumetric error under about 5%, dispensing 10 cc per stroke, and metering up to a quart of unmixed epoxy per minute with an integrated viscosity heater so cold resin does not thicken and skew the draw. The budget generic 1 oz/stroke screw-on pumps sit in between — they screw onto a 38-400 thread gallon or 96 oz jug and drop a fixed 1 fl oz (~30 mL) per stroke, but they are not piston-linked, so a sticky one quietly over- or under-doses, and that 1 oz minimum increment is far too coarse for sub-ounce work.
The takeaway: pumps remove reading error but introduce mechanical error. They are fast and repeatable, and they are only as trustworthy as your stroke discipline and your last verification.
How graduated cups introduce error
A graduated cup takes the opposite bet: you read the volume yourself off a printed scale. It is cheap, disposable, and lets you eyeball both components — but it stacks up three distinct error sources that pumps avoid.
First, the meniscus: resin curves up the cup wall, so the “line” you read is subjective and shifts with your eye height. Second, manufacturing tolerance: generic disposable graduated cups are sometimes mis-rolled at the factory, so the very first graduation sits too high or too low — your 25 mL line might really be 22 or 28. Third, wall film: viscous resin clings to the cup, so the volume you “measured” is not all the volume that pours out into the mix. Stack those, and a plain graduated 10 oz cup carries the highest reading error of anything on this page — and some of these cups only print marks up to the 8 oz line, leaving the top of the cup unguided.
This is why the cup category splits in two. Plain disposable graduated cups are the cheap baseline ($8-15 per 50-pack) and the least accurate. Calibrated stepped cups are the smarter middle ground, covered below.
The batch-size math: why 3 mL is 1% on 300 mL but 10% on 30 mL
This is the single most important number on the page, because it explains why the same tool can be fine and disastrous on the same day. A measuring error is roughly fixed in absolute terms — a sloppy pour or a mis-read line is off by some small volume regardless of batch size. Call it 3 mL.
On a 300 mL river-table pour, 3 mL of error is 1% — comfortably inside any epoxy’s tolerance. On a 30 mL jewelry batch, that same 3 mL is 10% — right at the edge of even a forgiving 1:1 system’s limit. The error did not change; the batch shrank, so the percentage exploded. This is precisely where cheap volume methods fail: small batches. A pump’s 1 oz coarse increment and a cup’s meniscus error both become proportionally enormous as you scale down, which is the entire argument for reaching for a scale on small, critical mixes.
Calibrated stepped cups: the middle ground
Between trust-the-stroke pumps and read-the-line plain cups sit calibrated stepped cups, and for low-cost accuracy they are the most foolproof path. Instead of a number to read, they give you a fill-to-the-line target. The WiseBond WiseMix is a reusable 32 oz non-stick cup with numbered stepped lines 1-7: you fill resin (A) and hardener (B) to matching marks, supporting both 1:1 and 2:1, and when it cures you flex the cup to crack the epoxy out and reuse it. The GlassCast A-to-B cup is even simpler — fill resin to the lower A mark, then top up with hardener to the upper B mark on the same step, a 2:1 by volume pour with no math and square corners that scrape clean for thorough mixing, at pennies per unit in bulk.
The win is that you see the ratio — instant visual confirmation, no stroke-counting trust and no meniscus-number guesswork. The catch is that they are still volume tools tuned to a specific chemistry: GlassCast explicitly states its cup is 2:1 by volume and must NOT be used for parts-by-weight resins. Always confirm whether your resin’s ratio is specified by weight or volume before trusting any cup.
When a $500 pump is worth it — versus a $1-$3 cup versus a scale
The price spread here is wild: a TotalBoat MakerPoxy pump runs around $500, a WiseMix calibrated cup about $1-$2, and a 50-pack of disposable graduated cups $8-15. That ~300x gap should drive your decision as hard as accuracy does.
The expensive pump only pays back at throughput. If you pour 1:1 river tables, furniture, or production craft runs all day, metering a quart a minute with linked pistons earns its keep fast. For weekly coasters or jewelry, it is a $500 paperweight — a $1-$3 calibrated stepped cup gives you adequate accuracy, a $16 set of screw-on pumps speeds whole-ounce batches, and a digital scale covers the small and critical mixes neither volume method handles well. For the full equipment picture see our best resin equipment buyers’ guide; if your real fork is pumps against a scale rather than against cups, the dispensing pump vs measuring by weight comparison goes deeper on production consistency.
Verify any method against a 0.1 g scale
Whatever you choose, treat a digital scale with 0.1 g resolution as the referee. It is generally faster and more accurate than any volume method and is immune to the temperature and viscosity effects that quietly change a liquid’s density — and therefore the volume-to-mass relationship every pump and cup depends on. For small or critical batches, skip volume entirely and weigh.
Use the scale to audit your volume tools, too: periodically dispense a known number of pump strokes, or fill a cup to its line, then weigh the output and confirm it matches the manufacturer’s by-weight ratio. This is how you catch a pump that has started to stick or backflow into the reservoir before it ruins a production run, and how you catch a mis-rolled cup whose first graduation lies. One caveat that trips people up: a volume-calibrated cup or pump is not valid for a weight-ratio resin, and a weight target is not valid for a volume tool — match your number to your method. If you are choosing between cups and a scale specifically, the digital scale vs measuring cups breakdown covers that decision, and you can browse every head-to-head on the comparisons hub.
Safety note: epoxy resin and hardener are skin and respiratory sensitizers — mix in a ventilated space, wear nitrile gloves and eye protection, and follow the manufacturer’s safety data sheet. An off-ratio batch that never fully cures can stay tacky and chemically active for days, so when a batch matters, confirm a small verification mix cures hard before committing.
The verdict
Read the verdict block and specs table below for the at-a-glance call. In short: reach for the pump when you pour the same 1:1 batch over and over and throughput is the bottleneck — just keep your strokes full and verify it hasn’t drifted. Reach for a calibrated stepped cup (WiseMix or GlassCast A-to-B) when you want cheap, reusable, visually-confirmed accuracy on modest batches. And reach for a 0.1 g scale the moment the batch drops under a couple of ounces or the part is critical — because at that size, 3 mL of slop is 10%, and no amount of careful pouring fixes a method that cannot resolve the difference.