And does anyone use a handheld spectro at receiving to approve dye lots? I’m testing ΔE ≤ 1.0 on 75D DTY to cut shade disputes — each reject costs us about $450 in freight and reslotting — and want to hear which units export clean pass/fail into ERP and what tolerances your suppliers will sign off.
We’ve been using a Konica Minolta CM‑2600d at receiving; it dumps a simple PASS/FAIL + ΔE00 CSV straight into ERP, and our suppliers signed ΔE00 ≤ 1.2 on 75D DTY under D65/10°, SCI — pushing to “ΔE ≤ 1.0 on 75D DTY” just spiked $450 rejects for us. One tip: wrap the yarn on a standard card and average 5 reads with the medium aperture; we also lock UV ‘on’ to keep OBA drift from inflating ΔE and causing false fails.
For ERP pass/fail, Datacolor CHECK 3 exports clean CSV/XML (or Bluetooth) and lets you lock a pass/fail template; we run D65/10°, SCI, 8 mm, averaging 3 reads to tame texture noise. I’d try CMC(2:1) at 0.8–1.0 instead of ΔE00=1.0 — on yarn it usually matches visual better and reduces the “is it off?” debates. What aperture and SCI/SCE are you set to, @lisa_87j?
, the $450 hits sting; our best ROI came from an X‑Rite Ci64 that tags an OK/NO‑GO + ΔE00 in a CSV the ERP ingests without mapping headaches. I’d keep your ΔE00 at 1.0 for auto‑accept but set a review band up to 1.4 for filament goods, and force a 5‑read average with the 4 mm tip to calm yarn streaks; suppliers pushed back less when we framed it that way. @lisa_87j, did you see fewer false rejects after moving to multi‑read averaging?
If you’re holding ΔE 1.0 on 75D DTY, go to the largest aperture and average 5 reads with 90° rotations — it kills the texture noise that’s causing $450 “rejects.” Also add a second light source check (e.g., TL84) as a hard gate and get suppliers to sign that; metamerism bit us more than ΔE spread. @michaelR77 are you locking barcode input for lot IDs at the device or in ERP — that one detail ended up deciding our handheld choice.
Quick tip: flip your handheld to SCE for the filament yarn — SCI kept giving ‘OK’ numbers on lots that still looked off, and that’s where our $450 hits came from. Konica Minolta CM‑26dG exports a simple CSV with a result flag and ΔE00 over USB/BLE and adds 60° gloss, which tightened acceptances when luster drifted (https://sensing.konicaminolta.us/products/cm-26dg/). Would your suppliers sign off on SCE as the agreed mode, @lisa_87j, or do you need both SCI and SCE logged?
We cut those $450 hits by locking D65/10°, UV status, and a black backer on a tight wrap card for 75D DTY — , the backer alone was swinging our numbers. Datacolor Check 3 spits a clean CSV with pass/fail plus a reason code (hue vs chroma) that our ERP eats, and I’d use CMC(2:1) with an overall delta cap of 1 so the math matches what people see. @Ravi did your suppliers sign the measurement spec itself, or are you only quoting the tolerance?