The AliExpress bench
The main bench page priced the safe routes: vintage gear with service manuals, new gear from local distributors, a software bench with published measurements. This page covers the gray market — what AliExpress actually sells against the same Table 1-1 shopping list, held to the same numbers.
The rule that makes this page honest is a two-tier evidence standard: measured means someone independent — a forum member with a cesium standard, a reviewer with an FFT — put numbers on the actual product, and only measured gear can earn a "buy" here; claimed means the only numbers are the listing's own, and no matter how good they look, the strongest I'll say is if the claimed figure is real, it would meet the requirement. Every category below gets one of six verdicts: Buyable — measured, Promising — claimed only, Nothing credible, two honest escape hatches — Buy locally and Generic — buy anywhere — and, for the two Table 1-1 rows that aren't actually products, Not a purchase.
Audio oscillator↑ TOP
Verdict: Nothing credible.
The requirement: residual distortion under 0.1% (3-4-(2) — "necessary to use an audio oscillator with less than 0.1% distortion"), a sweep of at least 30 Hz–22 kHz for the record-equalization checks, clean spot tones from 1 kHz to 18 kHz, and output settable 14 dB below operating level (3-6-(4), 3-6-(2)).
This is the category where the cheap gear fails on measurement, not on prejudice. The DDS function generators AliExpress sells in this class — the JDS6600 family (sold as JUNTEK, Joy-iT) and the FeelTech FY6600/FY6800/FY6900 family (FeelElec, and a spread of generic storefront names, A$114–213 shipped as of July 2026, AliExpress search) — carry manufacturer THD specs of 0.5–0.8% across the audio band. That's 5–8× worse than the 0.1% floor even if the claim is true. And the one independent measurement found says it isn't: Elektor Magazine's review of the Joy-iT JDS6600 quotes the manufacturer's own "THD (Sine): < 0.8%" spec, then FFTs the review unit's sine wave in the audio range and finds "distortion remnants of just over 1%" — real performance worse than an already-failing claim. GroupDIY and EEVblog threads describe hand-picked "sweet spot" frequencies reaching lower distortion after hardware modification (new op-amps, PSU rework), but no independently verified stock unit gets under 0.1% at a normal alignment frequency.
The safe routes are on the main bench page — a used HP 339A, or the software bench.
AC voltmeter↑ TOP
Verdict: Promising — claimed only.
The requirement: reads dBm directly at 0, +4 and −10 dBm against the 700 Hz reference tone, resolves ±2 dB response readings, holds a visible maximum for azimuth nulling, and resolves the S/N noise floor tens of dB below reference (3-5-(2), 3-5-(3), 3-4-(1)).
This is the one genuinely AliExpress-native instrument on the page: a new-manufacture analog dual-needle AC millivoltmeter, the TVT-322 (same Longwei-made unit rebadged as Troneer TVT-322, Tenma variants, and the MeterTo MVT-172/MVT-172D/MVT-171 siblings). ⚡ mains — 110 V/220 V switchable, no RCM.
- Claimed spec, unverified: dB range −70 to +40 dB (−80 to +52 dBm, 0 dBm = 0.775 V), accuracy within 3% of full scale at 1 kHz, response 5 Hz–1 MHz ±10%, narrowing to 20 Hz–200 kHz ±3%, 10 MΩ input. If those figures are real, the range covers the 0/+4/−10 dBm reference points comfortably, and ±3% full scale (roughly ±0.26 dB) clears the ±2 dB resolution bar with wide margin. The listings don't state low-end accuracy separately, so the microvolt-level mic-sensitivity use case stays unconfirmed even if the mid-range claim holds.
- Price: typically A$91–169 shipped as of July 2026 (AliExpress search).
- Sentiment: no independent bench review of this product family was found despite an extensive search — that's exactly why it stays claimed-tier. A diyAudio search snippet describes the sibling MVT-172 as usable but wanting a power-supply rework for lower noise — an informal user report, not a measurement. Platform reviews (consistency signal, not endorsement): one buyer reports "accuracy was within specifications," another received a unit with a cracked meter-glass face — a fragile-shipping theme that recurs on the frequency counter below. These boxy analog-meter products travel poorly in standard AliExpress packaging.
The verified vintage routes — HP 400-series, Leader LMV-181A — are on the main bench page.
Frequency counter↑ TOP
Verdict: Buyable — measured.
The requirement: one number — discriminate a ±0.2% window around the 3000 Hz speed-test tone, i.e. 2994–3006 Hz (3-4-(5)).
The BG7TBL FA-2 Plus (1 Hz–6 GHz, 10 MHz OCXO timebase; sold as TZT FA-2-3G Plus / FA-2-6G Plus, "by BGTBL") is the best-verified thing on this page.
- Measured: an EEVblog metrology-forum thread and M0WGF's bench review put real numbers on it: measured within 0.003 Hz of an HP 5071A cesium-standard reference after a minor trim, and separately well within the ±0.01 Hz expected deviation against a LeoBodnar GPSDO (1 ppb) over a three-hour test, with OCXO drift around 0.001 Hz/day. Those tests read a 10 MHz reference, so they verify the timebase: ≤0.01 Hz of error at 10 MHz is ≤1 ppb relative, and 1 ppb of timebase error at 3000 Hz is about 3 microhertz — several orders of magnitude inside the ±6 Hz window. Compare the main page's rejected budget candidate, the OWON XDM1041, whose rated error at 3 kHz consumes the entire window by itself.
- The honest caveat: none of the loaded sources tested it on a low-level audio-band sine wave — these counters get exercised against RF references, not tape-machine playback tones. How cleanly it triggers on the 3000 Hz tone off the flutter test tape is unverified, and it's the one thing worth bench-checking against a known tone on arrival.
- Price: typically A$188–284 shipped as of July 2026 (AliExpress search); one A$442 outlier listing excluded as non-representative.
- Sentiment: EEVblog metrology members with lab-grade references specifically prefer the FA-2 (OCXO) over the cheaper FA-1 (TCXO), finding the FA-2's timebase measurably more stable and free of the FA-1's small but real measurement bias. Platform reviews (consistency signal, not endorsement) are positive — "very accurate," "excellent functioning" — against one recurring complaint: light packaging, with one BNC socket damaged in transit.
A cheaper module, the PLJ-8LED, claims ±1 Hz at a 1-second gate — a roughly 6:1 margin if confirmed — but the one independent builder write-up found publishes no before/after numbers and flags thermal drift, so it isn't worth carrying as a second recommendation when the FA-2 clears the bar by three orders of magnitude with cesium-grade evidence.
The vintage route (HP 5300A+5301A) is on the main bench page.
Flutter meter↑ TOP
Verdict: Nothing credible.
The requirement: resolve well under 0.03% rms — the flutter test tape's own residual — against a 0.05% machine spec at 15 ips, with a selectable NAB-weighted position (3-4-(6)).
No NAB-weighted wow-and-flutter meter is manufactured new, and AliExpress doesn't change that. The category search turns up exactly one meter: an unbranded "Hot Direct Selling Measuring Wow & Flutter Meter" (⚡ mains, AC220V) at roughly A$1,388 landed as of July 2026 — no brand, no model, no published accuracy or resolution spec, no NAB-weighting claim, no reviews, and "100K+ similar items sold" template boilerplate. Not recommended at any tier or any price, let alone that one. Everything else the search returns is flutter test tapes (TCC-114, MTT-X257 class) — test media, not test instruments.
Flutter still means vintage Japanese hardware: the used Meguro MK-668 on the main bench page remains the least expensive spec-compliant route, full stop.
Wave analyzer↑ TOP
Verdict: Nothing credible.
The requirement: bandwidth adjustable to 100 Hz, reads individual harmonics at the 0.5% level, then re-patched for the crosstalk (55–60 dB) and erasure (>65 dB) tests (3-4-(2), 3-4-(3), 3-4-(4)).
AliExpress sells nothing in this category at any price point checked. A general "audio spectrum analyzer" search returns almost entirely decorative LED music-reactive display kits (A$40–70, no measurement function), plus one genuine instrument — the HTOOL SA6 at A$127.22 as of July 2026 — which is an RF spectrum analyzer, 35 MHz–6 GHz, starting three decades above the audio band. A search for the Phonic PAA6 handheld analyzer class returned zero matching products.
The honest recommendation stays where the main page put it: the software bench for the distortion test, a QA403-class analyzer for all three tests — see the wave analyzer section there.
Multimeter↑ TOP
Verdict: Buyable — measured.
The requirement: Table 1-1 asks for 20,000 Ω/V sensitivity — a meter-loading spec from the moving-coil era (1-1). A modern DMM's fixed 10 MΩ input beats a 20,000 Ω/V analog meter on every range below 500 V — exactly the line-level ranges alignment work uses.
Both of the main page's budget picks are AliExpress staples with real independent measurements, at roughly a third to a half of what they cost landed from eBay AU.
- UNI-T UT61E — measured by an independent lygte-info.dk review: ±0.1% DC accuracy, True RMS AC, 22,000-count display. Typically A$29–87 shipped as of July 2026 (AliExpress search), against A$112–140 landed from eBay AU per the main page. "UNI-T Official Certification"-badged store listings sit alongside generic resellers at similar prices — not proof against grey stock, but a mildly reassuring signal. Watch for "UT61E Plus" listings — a different, newer variant. The main page's caveats travel with it: EEVblog flags weak input protection, and a teardown reviewer calls the CAT safety rating "demonstrably fictitious" while confirming DC accuracy "spot-on" — line-level bench work only, nothing mains-adjacent.
- ANENG AN8008 — measured by markhennessy.co.uk and cross-referenced via an EEVblog review thread quoting the lygte-info review: "easily within the specified tolerances (Except 10A)." Basic DC accuracy 0.5%, True RMS, 9,999 counts. Typically A$26–64 shipped as of July 2026 (AliExpress search). Same shape of caveat: both sources call the CAT rating unrealistic — one flatly "fake" — while describing the meter as genuinely fit for low-energy bench work, which is exactly this job.
The full multimeter treatment, including the used Fluke 87V route, is on the main bench page.
Test tapes↑ TOP
Verdict: Nothing credible — and an active warning.
The requirement: the speed/EQ/standard combinations of Table 1-2, recorded at known levels per 3-2 — the reference everything else calibrates against.
Start with the plain finding: not one AliExpress listing under any search term tried actually sells a ¼" open-reel calibration tape. The closest hit is a "Professional 3000Hz Audio Test Tape" at A$32.39 as of July 2026 — a cassette-shell speed tape for Walkman-class players, wrong format entirely (search). Searching "reel to reel calibration tape" matches fishing tackle; "open reel test tape" matches surveyor's measuring tapes. The category simply isn't stocked.
The warning is about what would happen if one did turn up: no "calibration tape" listing checked states a fluxivity, an EQ curve, or a tolerance figure. Without those numbers there is nothing to verify a machine against — no way to know whether the tape was ever checked against a reference, and no way to tell, when a reading looks wrong, whether the fault is the tape, the deck, or the meter. To be fair about the record: no Tapeheads thread actually documents someone buying a marketplace reel-to-reel calibration tape and getting burned — the one thread that looks like it, read in full, is inconclusive and its raw readings ran the other way. The argument doesn't need a burn-story: the listings' unverifiability is the argument, and even well-instrumented technicians in that thread couldn't resolve which reference to trust. Adjacent evidence points the same direction — a Tapeheads thread on counterfeit cassette-format test tapes discusses marketplace sellers passing off copies of genuine calibration tapes as new. MRL remains the only source that publishes its fluxivity, EQ and tolerance figures and stands behind them — the full Table 1-2 mapping is on the main bench page.
Spring scales↑ TOP
Verdict: Promising — claimed only.
The requirement: brake tension 100–275 g (2-2), tape tension 40–77 g (2-3), pinch-roller pressure 6.4–7.0 lb on a 0–10 lb scale (2-4).
A measured tier is realistically impossible here — nobody independently bench-tests a two-to-thirty-dollar spring gauge — so everything below is claimed-only, with a home check doing the work a lab test does elsewhere.
- Physics-lab tubular spring dynamometers (sold generically as "Newton Force Meter," "Spring Scale Dynamometer") — stock capacities from roughly 100 g to 2,500 g full scale; a 100 g or 250 g unit would nominally cover the 40–77 g tape-tension band and the 100–275 g brake bands, if the claimed capacity is real — no listing checked publishes a graduation spacing or tolerance, so there's no saying from the listing how finely it resolves. Typically A$1.45–25.99 shipped as of July 2026 (AliExpress search). No independent review of the class exists.
- SHANHR 5 kg digital force gauge — 0–5 kg / 0–11 lb claimed capacity covers the 0–10 lb pinch-roller scale with margin if real; no accuracy or resolution figure anywhere on the product page. A$32.99 shipped as of July 2026 (AliExpress search) against the main page's A$281.60 Pesola pick for the same role — though the Pesola is a mechanical drag-pointer scale from an established instrument maker and this is an unverified digital gauge from a generic storefront. Platform reviews (4.8★, 24 reviews, attributed as such): buyers comparing it favourably to name-brand trigger-pull gauges — everyday consistency signals, not a bench measurement.
The honest workaround: verify at home before trusting it. Check any gauge like this against known weights — stacked coins of known mass, or a cross-check against a kitchen scale — before reading inside the tight 40–77 g band. The manual's tolerance bands are wide enough that a rough home check catches a badly wrong gauge; it can't certify a marginal one.
The Pesola picks are on the main bench page.
Stroboscope↑ TOP
Verdict: Promising — claimed only.
The requirement: none found — no procedure in the chapters researched actually uses a stroboscope, the same honest gap the main page documents; 2-5 adjusts the servo from the frequency-counter speed measurement instead. So there's no numeric spec to verify anything against here either.
What AliExpress offers is the same laser-tachometer route the main page already recommends via the Jaycar QM1449 — same claimed-accuracy class, notably cheaper:
- UNI-T UT373 mini laser tachometer — claimed ±(0.04%+2) accuracy over 10.0–9999.9 RPM, per Uni-Trend's own product page — not independently bench-tested, hence claimed-tier. The Jaycar QM1449's claim is ±(0.05%+1): same generic non-contact design sold under multiple badges. Typically A$31.99–32.79 shipped as of July 2026 (AliExpress search — a "UNI-T Factory Tools Store" listing carries a "Certified Original" badge) against the Jaycar unit's A$87.95. Platform reviews (4.8★, 30 reviews, attributed as such) include a usability note — "you need to get used to it to catch the reflected beam" — not an accuracy complaint.
- HABOTEST HT670/HT671/HT673 — claimed ±(0.05%+1) per the HT670's own manual, A$1.79–38.39 shipped as of July 2026 depending on model and promotion (same search). One disclosure worth passing on: an AliExpress-hosted "article" comparing the HT671 against a Fluke reference turned out to be self-disclosed AI-generated marketing content — excluded from evidence here, and flagged so nobody mistakes it for a real comparison.
The tachometer method and its Tapeheads corroboration are on the main bench page.
Head demagnetizer↑ TOP
Verdict: Nothing credible.
The requirement: no field-strength spec — a hand-held wand used on heads and guides every 8 hours, with procedural cautions (1-2-(2)).
The mains-wand form factor genuinely exists on AliExpress — just never for this job. Every plausible search returns the wrong tool, three ways:
- Cassette-shell combo cleaner/demagnetizer kits (A$1.45–3.84 as of July 2026, search) — insert into a cassette transport like a cassette; there's no wand or tip that could reach an open-reel head stack. Wrong format.
- CRT/kinescope degausser wands ⚡ mains, 220 V (A$9.24–23.19 as of July 2026, search) — the right principle, the wrong field geometry. A professional restorer on Tapeheads puts it plainly: a CRT needs "a very large field size... You can't use a tape degausser for a CRT because the field is very small in comparison" — and the mismatch runs in reverse: a CRT wand's wide, low-density field is the wrong tool for the small, controllable field a head needs at its pole tips. No listing states a field-strength figure to check against a head-safe range anyway.
- Watch demagnetizers ⚡ mains (A$8.89 as of July 2026, search) — an enclosed slot sized for a wristwatch, no wand at all. Wrong form factor.
The R.B. Annis Han-D-Mag — still in production, AU plug — remains the real answer, on the main bench page.
Consumables and small tools↑ TOP
The rest of Table 1-1, grouped — short answers, honestly labelled, down to the two filters and the oil below.
Head cleaner & alcohol
Verdict: Buy locally. International carriers' restrictions on flammable liquids mean AliExpress doesn't actually sell bulk isopropyl or head-cleaner solvent — every search result is individually wrapped alcohol prep pads (A$1.45–12.59 as of July 2026, search) — and hardware-store IPA or metho is cheaper and faster than importing either. Cleaning guidance is on the main bench page.
Swabs, vacuum, Allen keys & technician's tools
Verdict: Generic — buy anywhere. Ordinary products at ordinary prices; AliExpress is as good a source as any. As of July 2026: cotton swabs in bulk packs, A$1.45–9.14 (search); mini/handheld vacuums, A$9.17–44.85 (search); metric hex-key sets, A$6.24–11.69 (search — the manual never states hex sizes, so any standard metric set); precision screwdriver/nut-driver sets, A$3.35–60.76 (search).
Test cables
Verdict: Promising — claimed only — by nature: nobody independently bench-tests a A$1.45 cable's shielding, so platform-review themes are the closest thing to sentiment this category gets. Generic balanced XLR male-to-female leads run A$1.45–27.48 shipped as of July 2026 depending on length and finish, with ¼"/TRS and adapter variants at A$2.23–9.81 (AliExpress search). Platform reviews on the top-selling XLR listing (4.8★, 2,406 reviews, attributed as such, not endorsement) show a consistent functional-but-basic theme — "they are not top quality but they work very well" — with clean-signal reports from condenser-mic use and, notably, no failed-connector or broken-strain-relief reports among the reviews checked. The machine side is XLR for line in/out and ¼" for the mic inputs, per the main bench page.
Bandpass filter & ASA noise filter
Verdict: Not a purchase. Table 1-1 lists these as two separate line items, but neither is a shopping trip. The bandpass filter row looks stale — no procedure in any chapter researched ever calls for a standalone bandpass filter, only the wave analyzer's own front end set to 100 Hz for the distortion test. The ASA noise filter was never a shelf product even in 1976 — the manual shows the circuit in Fig. 3-1 and tells you to build it in a small aluminum box. No AliExpress search applies to either. The DIY routes, including the modern software-based equivalents, are on the main bench page's filters section.
Lubricating oil
Verdict: Nothing credible. AliExpress doesn't sell a traceable Mobil DTE Oil Medium (ISO 46) either. Searching the product name directly turns up zero oil listings — "DTE" on AliExpress is dominated by an unrelated dental-scaler product line also sold under that badge, so the results page is wall-to-wall ultrasonic dental tools. Broadening the search to "ISO VG 46 hydraulic oil" clears that collision and surfaces exactly one genuine industrial lubricant: a Klüber synthetic gearbox oil, A$170.39 shipped as of July 2026 (AliExpress search) — the wrong oil family (turbine/circulating oil is what the manual actually calls for), sold across a broad ISO VG 22–1500 span with nothing pinning the bottle to grade 46, and, like every option found, packaged at an industrial scale nowhere near "2 or 3 drops" twice a year. This mirrors rather than solves the gap the main bench page already found on the Australian-retail side: this exact grade, in a hobbyist pack size, isn't stocked anywhere checked.
The verdict matrix↑ TOP
Every category, one line each. Prices are shipped bands as of July 2026; evidence tier per the rules in the intro.
| Category | Verdict | Best candidate | Evidence | Typical price |
|---|---|---|---|---|
| Audio oscillator | Nothing credible | — (JDS6600/FY6600 class rejected) | Measured: >1% THD vs a 0.1% requirement | A$114–213 (rejected class) |
| AC voltmeter | Promising — claimed only | TVT-322 dual-needle millivoltmeter ⚡ | Claimed spec would meet; no independent test found | A$91–169 |
| Frequency counter | Buyable — measured | BG7TBL FA-2 Plus (OCXO) | Measured vs cesium and GPSDO references | A$188–284 |
| Flutter meter | Nothing credible | — (one unbranded, spec-free listing) | Nothing to verify | ≈A$1,388 landed |
| Wave analyzer | Nothing credible | — (category not stocked) | LED novelties and RF-band gear only | — |
| Multimeter | Buyable — measured | UNI-T UT61E; ANENG AN8008 | Independent reviews verify accuracy | A$29–87; A$26–64 |
| Test tapes | Nothing credible | — (category not stocked; listings state no fluxivity/EQ/tolerance) | Unverifiable by construction | — |
| Spring scales | Promising — claimed only | Tube dynamometers; SHANHR digital gauge | Claimed capacities; verify at home | A$1.45–33 |
| Stroboscope | Promising — claimed only | UNI-T UT373 laser tachometer | Claimed spec, same class as the Jaycar pick | A$32–33 |
| Head demagnetizer | Nothing credible | — (wrong tool, three ways) | Sourced field-geometry rejection | — |
| Head cleaner & alcohol | Buy locally | hardware-store IPA/metho | Flammable liquids don't ship | — |
| Swabs, vacuum, Allen keys & technician's tools | Generic — buy anywhere | any ordinary listing | Nothing to verify | A$1.45–61 |
| Test cables | Promising — claimed only | generic XLR / ¼" leads | Platform-review themes only | A$1.45–28 |
| Bandpass filter & ASA noise filter | Not a purchase | — (stale table row; DIY circuit) | n/a — no market category exists | — |
| Lubricating oil | Nothing credible | — (Klüber gearbox oil rejected; wrong family) | Listing-observation: name collision, wrong oil family, no viable pack size | — |
What an all-AliExpress bench gets you↑ TOP
Less than you'd hope, and the shape of what it can't do is the real finding. The gray market delivers exactly two measured buys: the FA-2 Plus frequency counter (A$188–284) and a budget multimeter (UT61E or AN8008, A$26–87) — call it roughly A$215–370 for the pair, as of July 2026. Add the claimed-tier layer if you accept the verification burden: the TVT-322 millivoltmeter (A$91–169, ⚡ mains, check it against a known reference), spring gauges verified against known weights (under A$35), a laser tachometer (about A$32), and cables and consumables for pocket change.
What no amount of AliExpress money buys: an oscillator that passes the 0.1% distortion bar — the cheap DDS generators fail it on measurement; a wave analyzer of any kind; a NAB-weighted flutter meter — that still means vintage Japanese hardware like the used Meguro; a head demagnetizer that isn't the wrong tool; a traceable ISO 46 turbine oil in a sane pack size; and above all, test tapes. The reference tape is the one purchase where an unverifiable source defeats the entire point, and MRL remains the only maker that publishes its numbers and stands behind them. For all of those, the main bench page has the routes that actually check out.