A self-leveling compound has to flow like a liquid for 20–30 minutes, then behave like a strong, crack-free screed for 20 years. Those two demands pull in opposite directions, and the additive package is what reconciles them: a superplasticizer for flow at low water, a very low-viscosity cellulose ether to keep sand and water together, a polymer powder for flexibility and surface strength, and a defoamer to release the air that pumping puts in. Below are the Landercoll®, Accurate® and NOVASTAR® grades we recommend for each system, with starting dosages and a troubleshooting table.
| Property | How it is judged | Typical target | Controlling additive |
|---|---|---|---|
| Flow | Ring test (e.g. Ø 30 × 50 mm ring) | 130–150 mm spread, loss ≤ 10 mm after 20 min | PCE type and dosage, cellulose ether viscosity |
| Bleeding / segregation | Visual, after 10 min at rest | No water film, no sand settlement | Low-viscosity HPMC, defoamer compatibility |
| Surface | Visual after 24 h | No pinholes, craters or laitance | Powder defoamer, water content |
| Strength class | EN 13892-2 | CT-C25-F6 to CT-C35-F7 (cement); CA-C20-F4 to CA-C30-F6 (gypsum) | Water/binder ratio set by PCE; RDP for flexural strength |
| Walkability / overlay | Site schedule | Walkable 3–4 h, tiles after 24 h, resilient flooring 48–72 h | Binder system (OPC/CAC/CSA), accelerator, PCE profile |
| Dimensional stability | Shrinkage prism, visual | No curling, no cracking | RDP, water reduction, shrinkage control |
| Wear resistance | EN 13892-3/4 (A/AR classes) | Declared when used as wearing layer | RDP dosage, low w/b |
| System | Landercoll® cellulose ether | Accurate® RDP | NOVASTAR® PCE | NOVASTAR® defoamer | Other |
|---|---|---|---|---|---|
| Cement-based underlayment (OPC + CAC + anhydrite/gypsum) | K04 (300–500 mPa·s) 0.03–0.06 % | Rigid 5011N 1.5–3 % or semi-flexible 4025N 2–3 % | 540P or 550P, or TDS-series 590P, 0.15–0.35 % | P805, 0.1–0.3 % | Tartaric acid / Li₂CO₃ for set control |
| Cement-based wearing layer, high strength | K04 0.03–0.05 % | 4025N or 5045N, 3–4 % | 540P / 550P or 590P, 0.2–0.4 % | P805, 0.15–0.3 % | PP fibre 6 mm 0.05–0.1 % |
| Gypsum-based (α-hemihydrate or anhydrite) | K04 0.03–0.06 % | 4025N or 5045N, 1–2 % | 570P (gypsum system) or 590P, 0.1–0.3 % | P805, 0.1–0.2 % | Gypsum retarder P400 / AR-P, 0.02–0.05 % |
| Rapid-setting CAC-rich repair leveler | K04 0.02–0.04 % | 5011N, 1–2 % | 590FP (early strength), 0.15–0.3 % | P805 (P807 in cold weather), 0.1–0.2 % | Lithium carbonate accelerator |
| Pumpable large-area flooring | K04 0.03–0.06 % | 5011N or 4025N, 2–3 % | 550P, or 590P + 600P (slump retention), 0.2–0.4 % total | P805 (P806 alkali-resistant on request), 0.15–0.3 % | — |
| Waterproof / wet-room leveler | K04 0.03–0.05 % | 8048H (TDS lists cement-based self-leveling mortar), 2–3 % | 540P or 590P, 0.15–0.3 % | P805, 0.15–0.3 % | — |
Dosages are percentages of the total dry mix for laboratory starting points; the technical data sheets do not prescribe dosages. Cellulose ether in self-leveling compounds is dosed ten times lower than in tile adhesive; 0.01 % steps matter. The PCE technical data sheets give 0.1–0.5 % of cementitious material as the recommended range — the mortar-based figures above fall inside it once converted.
Why these grades:
| Symptom | Likely cause | Adjustment |
|---|---|---|
| Water film and sand at the bottom after 10 min | Cellulose ether too low or absent; too much water | Add 0.01–0.02 % K04; cut water 0.5 %; check PCE overdose |
| Spread falls from 145 to 120 mm in 15 min | Cement–PCE incompatibility, high temperature, C₃A-rich cement | Replace part of the 590P with 600P (slump retention); test with the second cement source |
| Pinholes and craters | Air from mixing/pumping not released; wrong defoamer type | Raise P805 to 0.2–0.3 % or move to P806 in cement-rich mixes; check K04 dosage |
| Random cracking after 1–3 days | High water, fast drying, no polymer | Lower w/b via PCE; 2–3 % RDP; protect from draughts for 24 h |
| Curling at edges | Differential drying, low polymer | 3 % 4025N or 5045N; reduce layer thickness variation; cure |
| Dusting, weak surface | Bleeding carried fines up; low RDP | Fix bleeding first (K04), then RDP ≥ 2 % |
| Too sticky, poor self-healing of trowel marks | Cellulose ether viscosity too high (a K40 or higher grade instead of K04) | Use K04 only |
| Slow strength gain in gypsum SLC | Over-retarded; PCE overdosed | Reduce retarder to 0.02 %; keep 590P ≤ 0.3 % and verify with the actual gypsum source |
| Walk-on time too long in a repair leveler | Binder too slow for the schedule | 590FP (early-strength PCE) with lithium carbonate; check CAC share |
Detailed case studies: Preventing cracking in self-leveling compounds · Cracks after curing · Bleeding and segregation.
| Component | Share |
|---|---|
| CEM I 52.5 R | 25 % |
| Calcium aluminate cement | 10 % |
| Anhydrite or α-gypsum | 5 % |
| Quartz sand 0.1–0.6 mm | 50 % |
| Limestone filler | 9 % |
| Landercoll® HPMC K04 | 0.04 % |
| Accurate® RDP 4025N | 2.5 % |
| NOVASTAR® PCE 590P | 0.25 % |
| NOVASTAR® defoamer P805 | 0.2 % |
| Tartaric acid / lithium carbonate | 0.05–0.15 % (set control) |
| Water | ≈ 20–22 % of dry mix |
Laboratory starting point only; the ternary binder ratio and set control must be tuned to local cement and the required walk-on time.
Self-leveling compounds expose weak additives faster than any other product: a batch of cellulose ether 20 % off viscosity shows up as bleeding on the next job. LANDU controls viscosity batch by batch in its own plants (three production bases, 75,000 t/year), supplies REACH-compliant grades under ISO 9001, and runs flow, bleeding and strength tests on customer formulations in its Jinan application laboratory. European customers are served through partners in several countries, including Greece, Spain and Italy, with 25 kg bags on pallets and 7–15 working-day factory lead times. Shelf life per TDS: K04 two years, RDP one year (use within six months of receipt recommended), PCE powder twelve months in the original packaging.
300–500 mPa·s (Landercoll® K04, the only Landercoll® grade whose TDS is written for self-leveling compounds) at 0.02–0.06 %. Tile-adhesive and putty grades (K40 and above, 18,000 mPa·s and higher) destroy flow and should not be used.
Typically 0.15–0.35 % of the dry mix for NOVASTAR® 540P / 550P or the TDS-series 590P, which corresponds to the 0.1–0.5 % of cementitious material given on the TDS. Start at 0.25 %, measure spread at 0 and 20 minutes, and adjust in 0.05 % steps. Overdosing shows up as bleeding and retarded setting.
Yes, at 1.5–3 %. Even under tiles, the polymer film gives the tensile and flexural strength that stops cracking and curling and secures the bond to the tile adhesive above it. 5011N is the economical rigid grade; 4025N adds flexibility.
Start with P805, the standard grade, at 0.1–0.3 %. P806 is the alkali-resistant version for cement-rich, high-pH binders where standard defoamers lose activity; P807 is the low-temperature version for winter production and cold sites. Our technical team sends a comparison sample for your binder.
Keep K04 at 0.03–0.06 %, use the gypsum-system PCE 570P (or 590P) at 0.1–0.3 % and confirm its water reduction on your calcium sulphate binder, add 0.02–0.05 % gypsum retarder (P400 or the alkali-resistant AR-P) for working time, and use 1–2 % semi-flexible 4025N or flexible 5045N.
Yes. Send 5 kg of your dry mix or the recipe; our laboratory reports spread at 0/10/20 min, bleeding, setting time, 1-, 7- and 28-day strength and surface appearance, with recommended adjustments.
At 0.03–0.06 % a very low-viscosity grade (Landercoll® K04, 300–500 mPa·s) holds water and fines together so the mix flows as one liquid instead of separating into a water film on top and sand at the bottom. It is a stabiliser, not a thickener: any grade above K04 makes the compound sticky and stops trowel marks from healing. Flow comes from the PCE; the HPMC only has to be enough to stop bleeding and settlement.
Yes. Landercoll® K04 is made in LANDU's own cellulose ether plants in Shandong, as are the NOVASTAR® PCE and defoamer and the Accurate® RDP on this page. Minimum order 1,000 kg per grade in 25 kg bags on 500 or 600 kg pallets; stock grades ship in 7–14 days, FOB or CFR/CIF to European ports. A 500 g – 1 kg sample with TDS and batch COA is free.
Cracks in the first 1–3 days are drying shrinkage: too much water, too fast drying, or too little polymer. Bring the water down with PCE rather than by cutting flow, add 2–3 % RDP (4025N or 5045N), and keep the floor out of draughts and direct sun for 24 hours. Cracks that follow substrate joints or sit over hollow-sounding areas are substrate problems, not additive problems. Case studies: preventing cracking in self-leveling compounds · cracks after curing.
We return a grade recommendation and dosage plan within two working days and ship samples of K04, 4025N, 590P and P805 free of charge.
LANDU — Committed to Every Gram.