Quick question: I’m evaluating online UV254 and TOC analyzers to tighten coagulation dosing at a surface water plant with volatile NOM, weighing s::can spectrolyser vs Hach UVAS paired with a Sievers TOC. My priorities are response time (<60 s), fouling control (wipers vs periodic acid clean), and stability when trued up to weekly grab DOC/SUVA. If you’ve run these on raw water ahead of rapid mix, what kept calibration drift in check and where did maintenance become a bottleneck?
We hit <60 s by running an s::can in a 1 m/s fast-loop with a 50 µm strainer and 10‑min wiper; a quick 1% HCl rinse at shift change and a weekly 2‑point to DOC/SUVA kept drift tight. > run these on raw water ahead of rapid mix, what kept calibration drift in check and where did maintenance become We pulled from a constant‑head tap to dodge bubbles and floc carryover; Sievers TOC was solid but too slow for feedback, so UV254 drove alum and TOC just sanity‑checked — does that match your flow path?
On raw river water, our s::can only held <60 s once we moved the fast‑loop takeoff within about 2 m of the channel and added a small degasser — bubbles were our “phantom spikes.” We kept a Sievers for DOC truth but did a quick daily blank or the SUVA tie‑in wandered. @j_thomas92’s strainer note is spot‑on; in storm NTU surges, enable the spectrolyser’s solids compensation (or spec the short‑path model) or you’ll chase scatter — are your spikes tied to runoff?
I kept a Hach UVAS responsive by mounting the flow cell vertical and cracking a tiny bleed to waste so bubbles don’t sit on the window — step changes showed up in under a minute and held steady when trued to your “weekly grab DOC/SUVA.” One caveat: after alum runs the wiper didn’t remove the aluminum hydroxide film reliably, so a quick 0.5–1% acid rinse each shift made the baseline far more stable than wiper-only (same story when I tested the s::can).
But i’ve had better stability by using a “slope-only” UV→DOC fit and letting the Sievers apply a rolling bias every 30–60 min so UV drives the dose and TOC trims drift. Small caveat: big temperature swings nudge UV a bit, so add temp compensation or a short tempered bypass to keep it steady.
@emills48 Good call on the bleed; the tweak that finally stabilized ours was adding a small backpressure regulator right after the UV cell (about 2–3 psi) so microbubbles don’t sit and the wiper contacts the glass, and we trigger a 30‑s flush with filtered plant water before the morning zero — drift dropped and step changes stayed crisp; worth trying?
Quick example: we kept a spectro::lyser under 60 s by feeding it from a small constant-head overflow loop so the cell sees steady flow and pressure instead of pump ripple; that alone cut noise and drift. With the autobrush at 3-min and a 30-sec daily air blast, raw surface water stayed stable enough that the weekly DOC/SUVA true-up was just a slope touch. One caveat: after any acid clean, wait about 10–15 min before trusting the UV254 again — I’ve seen a brief “low” bias.
Locking the UV cell temperature helped more than I expected: we wrapped a small 10 W heater and held the UV cell at about 20±1°C, which knocked out drift and kept response tight (“UV254 will move a few mAU per °C”). Minor caveat: the warmer cell did speed fouling a bit, so we nudged the wiper frequency up.
On response time, I’d let UV254 run the “fast loop” and use TOC only for a slow trim — our Sievers M5310C never beat about 4–5 min, while a spectro::lyser with a wiper hit step changes in about 30–40 s. For fouling, a tiny 0.2–[redacted]/L free chlorine in the UV slipstream kept the optics clean without daily acid; caveat: it biases TOC a hair, so dose after Sievers or split the takeoff. Weekly DOC/SUVA true-up held steady when we refreshed the UV254:DOC slope monthly and kept temperature compensation on, which echoes what @fchen_1 saw about drift control.
Biggest win for us was bubble control on the spectro::lyser — added a tiny vertical vent/bubble trap right before the probe and mounted the optics facing down, which cut jitter and kept response under 30 s. Only wrinkle was winter supersaturation; a small degassing tee upstream fixed it, @emills48.
To hit your “under 60 s” target, the biggest gains for us were plumbing: keep the sidestream short (<3 m), wide-bore, and push about 1–2 L/min; response snapped into line. We also put a small 3‑way valve on the UV probe to auto‑zero with DI for 30–45 s every 8–12 h, which tightened alignment to weekly grabs more than any wiper setting. Caveat: if raw Fe spikes, add a 5 µm prefilter ahead of the sensor head or you’ll chase phantom dose swings.
If you’re truly chasing <60 s, let UV254 drive the dose; Sievers TOC is great for trending but its cycle lives in minutes, not seconds. We tightened DOC/SUVA agreement by adding a daily DI auto‑zero and temporarily freezing turbidity compensation during storm NTU spikes; only Fe/Mn bursts knocked it off. If iron’s in play @c_klein21, the s::can’s spectral correction helped more than UVAS; quick primer on UV254 interferences: https://www.epa.gov/water-research/ultraviolet-absorbance-254-uv254-monitoring.
We’ve run the s::can spectrolyser on raw surface water; the biggest stabilizer for “response time <60 s” was killing bubbles — mount the probe slightly nose-down and add an auto air-vent on the sidestream so UV doesn’t hunt (, bubbles). The wiper’s solid, but when Mn/Fe creep up we saw baseline drift; a quick weekly 1% citric flush kept DOC/SUVA trued up better than harsher acid. If algae-driven color swings are common, the Hach UVAS held zero a bit better for us, but you trade some sensitivity — do you have any entrained air upstream of the tap?