Picking a Torque Screwdriver: Three Numbers That Fight Each Other
The service manual gives one number for an M2.5 screw. A forum thread gives three. You buy a tool based on one of them, open the laptop, and find the scale will not dial to that value at all.
That is not your mistake. It is how this category is merchandised: the range is printed in the largest type, the graduation lives halfway down the spec table, and the tolerance sits in the corner of the manual. On the lower half of the range those three numbers contradict each other. I actually copied the range, main-scale graduation, handle length, overall length and weight of all five Wera 7400-series models out of the manufacturer's own datasheets, did the percentages by hand, and then converted everything across the imperial and metric unit systems.
This article contains Amazon affiliate links. If you buy through them we may earn a commission, and it does not change what you pay. Every specification below comes from manufacturer datasheets or public sources; prices and stock move constantly, so treat the listing page as the authority as of the day you read it (availability was checked once, 2026-11-11). Data source: public search and manufacturer datasheets, as of 2026-11-11.
⏳ TL;DR
Match on the smallest torque you will actually apply, not on brand:
- 0.10–0.34 N·m (phones, cameras, headphones, M1.6–M2): the Wera 7430 micro torque screwdriver is the model sitting in that band. At 0.10 N·m one graduation is 15% — it dials there, it does not dial accurately.
- 0.3–1.2 N·m (laptop shells, M2.5–M3, drive caddies): the Wera 7440. This is where most people should start. Its weakness: half a graduation at 0.3 N·m is 8.3%, wider than its own ±6% claim.
- 1.2–3.0 N·m (cases, monitor arms, cooler mounts): the Wera 7441 ESD. Across the whole 7400 family, this is the only band whose resolution stays finer than ±6% end to end.
- Everything from 0.3 to 3.0 at once: the 26-piece set. Its 24 bits contain no Phillips at all.
- Imperial scale and a conformance document in the box: Capri CP21075, 10–50 in-lb. Clockwise only, ISO 6789-1 only.
- No idea what torque you need: buy the 7440 alone. Sets are sold in range segments, and half of what you pay covers a segment you will never touch.
Only three numbers in this category actually collide
Open any torque screwdriver datasheet and you get a list: range, main-scale graduation, handle length, overall length, weight, drive, accuracy.
Only three of them decide whether you strip a thread. The lower range limit decides whether your loosest screw is covered. One graduation decides how finely you can set the value. The tolerance band decides how far the real output sits from the setting.
Accuracy — that ±6% on the box — is the number marketing leads with, and at the bench it is the least decisive of the four. The reason is below.
First, a distinction that gets blurred constantly: accuracy and resolution are not the same property. ±6% describes how far actual output can land from the value you dialled; it is set by the internal spring calibration. The graduation describes how close to your target value you can dial at all. The first is an instrument property. The second is a readout property. At low torque, the readout property fails first.
Lower range limit: the official table has two overlaps
Wera's product page lists five models in the 7400 series. Their ranges butt up against each other like this (from wera.de and the per-model datasheet PDFs, as of 2026-11-11):
| Model | Range N·m | Main graduation N·m | Handle / overall mm | Weight g |
|---|---|---|---|---|
| 7430 | 0.10–0.34 | 0.015 | 89 / 142 | 122 |
| 7431 | 0.30–1.00 | 0.05 | 89 / 142 | — |
| 7432 | 0.90–1.50 | 0.05 | 89 / 142 | — |
| 7440 | 0.3–1.2 | 0.05 | 105 / 155 | 203 |
| 7441 | 1.2–3.0 | 0.10 | 105 / 155 | 209 |
Read the column vertically and two overlaps appear: 0.30–0.34 N·m is covered by both the 7430 and the 7431, and 0.90–1.00 N·m is covered by both the 7431 and the 7432. The 7440 and 7441 touch at exactly 1.2 N·m and overlap nowhere.
How do you choose inside an overlap? By graduation, not by range. At 0.90–1.00 N·m both the 7431 and the 7432 graduate at 0.05, so there is no difference. At 0.30–0.34 N·m the 7430 graduates at 0.015 against 0.05 on the 7431 — a factor of 3.3. Work in that band belongs on the 7430.
This is also why "buy the widest range" is wrong here. The wider the range on one tool, the coarser its graduation has to be.
Main graduation: one division as a percentage is the real user experience
Divide the graduation by the value you are setting and you immediately see which part of each range works and which part you merely tolerate. Half a division is the finest position you can physically dial.
| Model | Bottom of range | One division | Half division | Top of range | One division |
|---|---|---|---|---|---|
| 7430 | 0.10 | 15.0% | 7.5% | 0.34 | 4.4% |
| 7440 | 0.30 | 16.7% | 8.3% | 1.20 | 4.2% |
| 7441 | 1.20 | 8.3% | 4.2% | 3.00 | 3.3% |
| Capri CP21075 | 10 in-lb | 10.0% | 5.0% | 50 in-lb | 2.0% |
One conclusion in that table is ugly but unavoidable: at 0.3 N·m the 7440 steps 16.7% per graduation, and half a step is 8.3% — wider than its own ±6% rating. At that end you cannot even set the value accurately, so the advertised tolerance is decoration.
The 7430 is worse: at 0.10 N·m one division is 15% and half a division is 7.5%, again above ±6%.
The 7441 is the one model in the family that holds up across its whole range: half a division at 1.2 N·m is 4.2%, under 6%, so resolution never becomes the bottleneck.
You can invert the arithmetic to find where resolution stops mattering: the point where one division equals 6%. That is 0.833 N·m on the 7440, 0.25 N·m on the 7430, and 1.667 N·m on the 7441. Below that point your finest dial position is wider than the tolerance band. Above it, the band dominates.
The boundary needs stating: this does not mean real output is 8% off at the bottom. Output deviation still comes from spring calibration. What it means is probability. Dial 0.30 and the target can sit anywhere from 0.285 to 0.315, and ±6% of output error stacks on top — a combined window approaching ±14%. Plastic clips and aluminium threads fail inside that window.
One accessory note: Wera's page says the 7430, 7431 and 7432 ship with a magnifier that clips onto the scale, which materially improves legibility. A magnifier makes the scale easier to read. It does not make the scale finer. The graduation is still 0.015.
Drive: a 1/4-inch hex decides how small a screw you can reach
Every 7400-series tool takes 1/4-inch hex bits, specified as DIN ISO 1173-C 6.3 and E 6.3 on the datasheets. Individual tools ship without bits; only the sets include them.
That creates a structural problem. Micro bits — PH000, PH0000, T4, T5 — mostly live in the 4 mm hex ecosystem. You can buy a 7430 that starts at 0.10 N·m and still be unable to drive an M1.6 self-tapper, because what limits you is the bit, not the torque.
The 26-piece set shows the contradiction plainly. Its Set contents section lists two handles (7440 plus 7441), nine TORX bits from TX6 to TX30, nine TORX PLUS bits from 6IP to 30IP, and six slotted bits at 2/2.5/3/4/5/6 mm — 24 bits in total. Not one Phillips bit, and the smallest Torx is TX6.
So if your work is MacBooks and thin laptops, that set will turn the T6 screws and will not touch the PH00 battery bay.
The decision tree
Answer four questions in order and you are done. If you cannot answer any of them, take the fallback.
One: is your smallest torque below 0.3 N·m?
Yes → the Wera 7430 is one of the few models sold in that band. If your usual target sits between 0.10 and 0.25, accept dialling by whole divisions rather than chasing accuracy, or move to a preset or digital tool.
No → question two.
Two: is your usual target within 1.2 N·m?
Yes → the Wera 7440, 0.3–1.2 N·m. It covers laptops through desktops and is the default starting point in this article. Its weaknesses are in the next section.
No → question three.
Three: is your usual target within 3.0 N·m?
Yes → the Wera 7441 ESD, 1.2–3.0 N·m. Choose it for the resolution as much as the range; its graduation holds ±6% across the whole band.
No → you need a larger tool. This comparison does not cover it; read the manufacturer's range chart and do not force a 7441 to 6 N·m.
Four: are you working on static-sensitive boards?
Yes → take the ESD version. The Wera 7440 ESD matches the standard model on range, graduation and length exactly. The extra money buys a discharge path, not accuracy.
Want 0.3–3.0 in one purchase → the Wera Kraftform 26-piece set, at the cost of having no Phillips bits in the box.
Fallback, if none of the four questions has an answer: buy the 7440 alone, estimate torque from screws you already own, then decide whether to add a 7430 or a 7441. Do not buy a set out of fear of running short. Every set contains one tool you will not touch for three years.
Six models on sale, and the real shortcoming of each
Wera 7430 (0.10–0.34 N·m) — listing. Only 0.24 N·m wide, the narrowest in the family; graduation 0.015, which is 15% at 0.10 N·m, so the bottom end is dial-by-division only; sold without bits; and the 0.34 ceiling will not reach a typical 0.4 N·m laptop screw. It complements a 7440; it does not replace one.
Wera 7440 (0.3–1.2 N·m) — listing. Three weaknesses: half a graduation at 0.3 N·m is 8.3%, above the ±6% rating; the 1/4-inch drive will not take micro bits; no bits included. Its datasheet also carries the line "Unlimited torque for loosening seized screws", which reads like permission to use it as a breaker bar — metrology providers argue against exactly that, see the calibration section below.
Wera 7440 ESD (0.3–1.2 N·m) — listing. Identical range, graduation, length and drive to the standard model; the price difference buys static dissipation only. The datasheet does not publish a surface-resistance figure, so if you need it for line compliance, ask the manufacturer for a test report rather than treating the three letters as certification.
Wera 7441 ESD (1.2–3.0 N·m) — listing. The best-matched scale in the family, but the 1.2 N·m floor leaves the common M2.5 band uncovered; sold without bits; the ESD premium is wasted on home builds.
Wera 26-piece set (0.3–3.0 N·m) — listing. Two handles and 24 bits, case listed at 255×210×62 mm and 950 g. The shortcoming is specific: no Phillips bits, smallest Torx is TX6 with no T4 or T5, and the case eats drawer space.
Capri CP21075 (10–50 in-lb) — listing. The manufacturer's page states 1/4-inch hex, 1 in-lb increments, ±6% clockwise, a calibration document included, conforming to ASME B107.300-2010 and ISO 6789-1:2017. Four limitations: clockwise only; imperial scale with no newton-metre ring; the 10 in-lb floor converts to 1.13 N·m, so nothing below 1 N·m is reachable; and "certified" here means an ISO 6789-1 declaration of conformance, not an ISO 6789-2 calibration certificate. Its strength is the bit set — PH0 through PH3 and T10 through T40 — which fills the gap Wera's set leaves.
(A three-handle 27-piece set reaching 6 N·m is also listed, but the datasheet for its third handle, the 7442, was not obtainable for this article. I am not guessing its range or graduation; check the listing's spec panel if you need it.)
Units and certificates: the two places people buy wrong
Units. One newton-metre equals 8.851 in-lbf, 100 cN·m and 10.197 kgf·cm. All three systems sit on the same shelf: Wera specs in N·m, Capri in in-lb, and some STAHLWILLE models are published in both cN·m and in-lb. Convert and an awkward fact appears: Capri's floor of 10 in-lb is 1.13 N·m, while Wera's 7440 stops at 1.2 N·m. The overlap between those two tools is a narrow band from 1.13 to 1.2 N·m. "Buy two and you are covered" does not survive the unit conversion — that band is either bought twice or left empty.
Run the Capri resolution the same way: 1 in-lb steps mean one division is 10% at 10 in-lb and half a division is 5%, which is inside ±6%. Its imperial scale is actually steadier at the bottom than Wera's metric scale. That is not a brand effect. It falls out of division-divided-by-range.
Certificates. ISO 6789 split in 2017. Part 1 covers design and quality conformance and produces a Declaration of Conformance. Part 2 covers calibration method and measurement uncertainty and produces an actual calibration certificate. Calibration providers are blunt about the difference: most manufacturing audits accept Part 2 certificates, and a Part 1 declaration is not a calibration certificate.
Compare the wording. Wera says measurement accuracy is ±6% in accordance with DIN EN ISO 6789. Capri says certified in accordance with ASME B107.300-2010 and ISO 6789-1:2017. Both are conformance claims. Neither is a Part 2 certificate in the box. For an audit, both get sent out for calibration.
One rule also changed in 2017: the calibrated range moved from 20%–100% of capacity to "from the lowest marked value to 100%". The folk advice to stay in the middle 60% is therefore no longer a compliance exemption. It is a habit.
On intervals, the providers converge: 12 months or roughly 5,000 cycles, whichever comes first, and immediately after a drop, an overload or a repair.
Which brings back that 7440 line. The manufacturer is saying the clutch does not cap loosening torque. The metrology side is saying that cracking a seized fastener with a torque tool applies shock loads that shift the reading permanently. Both are true. Read together, though, the first one invites you to use a calibrated instrument as a breaker bar. If you do, think about whether you still trust it next time.
A ten-minute checklist
Four steps, faster than reading reviews:
1. Find the lower range limit, write down your smallest torque, confirm it sits inside the range and above one third of the floor.
2. Divide the graduation by your target value. Above 6%, that end is dial-by-division only.
3. Check the drive is 1/4-inch, then check whether your smallest bit belongs to the 4 mm system.
4. Check whether the paperwork says Part 1 or Part 2. "Compliant with" is Part 1.
For bit-material and tolerance comparisons across screwdriver kits, the site's four checks on precision screwdriver kits pair with this one; if you are also doing post-solder rework, the bench-vise clamping piece and the solder-wire spec piece are the upstream half of the same bench.
FAQ
Q: Can a torque screwdriver be used as an ordinary screwdriver to loosen screws?
A: Not advised. Calibration providers describe what happens when you crack a seized fastener: ratchet teeth, pawl mechanisms and springs take shock loads outside their design envelope, and the reading can drift permanently. Most warranties exclude that use. Loosen with a ratchet or a plain driver and reserve the torque tool for the final tightening pass. Wera's "unlimited torque for loosening" refers to the clutch having no cap, not to a recommendation.
Q: Why are screwdrivers rated ±6% when wrenches reach ±4%?
A: ISO 6789 bands tools by type and class. Among Type II setting tools, capacities below 10 N·m are generally rated ±6%, and only above 10 N·m is there room for ±4%. Screwdrivers all sit in the small-capacity band, so ±6% is the category norm rather than a sign of a bad tool. When you see ±10% on an adjustable model — Wera's own 1430 carries it — that tool is declared to a wider band.
Q: Does a torque screwdriver need annual calibration?
A: The baseline is 12 months or about 5,000 tightening cycles, whichever comes first, and sooner after a drop, an overload or a repair. A home user doing dozens of fasteners a year follows the calendar; a production line follows the cycle count. Only a Part 2 certificate carries audit weight.
Q: Will a 1/4-inch drive take a PH000 micro bit?
A: It needs an adapter, and the result is not worth it. Micro bits mostly belong to the 4 mm hex system; the smallest common sizes in 1/4-inch (6.35 mm) run out around PH0 and TX6. For self-tappers below M1.6, switch to a 4 mm handle instead of stacking an adapter on a 1/4-inch drive.
Q: Digital readouts resolve finer. Why not recommend one?
A: Resolution and data logging are genuinely better on digital tools, but no official statement of tolerance and conformance class for current digital models was obtainable for this article. Without it there is no way to say which part of the range the ±x% figure covers, or which edition of the standard it is declared against. That is not a parameter to infer from listing copy; check the manufacturer's spec sheet.
Closing
What you are buying here is not accuracy. It is the position of your target value on a scale. If the range covers your work, one division divided by your target stays under 6%, and the drive accepts the bits you actually own, the tool is sufficient. Miss one of the three and no advertised tolerance will save the job.
For most people the answer is one 7440 plus a short 1/4-inch-to-4 mm bit holder.
👉 Join Xiaomi MiMo Platform: Leading AI model platform with cost-effective inference
👉 Join Aliyun AI: Top AI products with exclusive coupons for business innovation
📌 This article was AI-assisted generated and human-reviewed | TechPassive — An AI-driven content testing site focused on real tool reviews
🔗 Recommended Tools
These are carefully selected tools. Using our affiliate links supports us to keep producing quality content:
As an Amazon Associate, I earn from qualifying purchases at no extra cost to you.