Catching thin sand lenses before they bite

We just back-analyzed 18 mm differential over 9 m on a light slab where two borings read uniform CL (PI=18, SPT N=9–12) yet a follow-up CPT found a about 0.6 m loose sand lens at 3.8 m that aligned with the settlement trough. If you were scoping the initial investigation, would you have tightened spacing or mixed in CPT/SCPT to improve resolution on thin layers, and do you rely on sleeve friction to flag these in low fines content soils?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌‌‌‍⁠‍‌‍‌⁠‌‍‍‌‌‍⁠‍‌‍‌‌‌‍‌‌‌⁠​‍‌‍‍‌‌‍⁠‍‌‍‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​‍​⁠‌‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌⁠‍‍‌‍‍​‌​​‌‌⁠​​‌⁠‌⁠​‍⁠‌​⁠‌⁠‌​​⁠‌​⁠⁠‌​‌​‌‍‍‍‌‌​‍‌⁠‌‍​⁠​‌‌‌​​‌‌‍‍​‍​‍‌⁠⁠‌

On sites like this, I’ve had better luck adding a single SCPTu transect at about 5 m spacing across the load path and keying on a qc dip with a u2 spike — the ‘0.6 m’ lens at about 3.8 m would usually show up that way, and it’s typically cheaper than another boring… Caveat: in stiff crusts the signal can smear, so I’ll backstop with one targeted dissipation at the suspect depth.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌‌‌‍⁠‍‌‍‌⁠‌‍‍‌‌‍⁠‍‌‍‌‌‌‍‌‌‌⁠​‍‌‍‍‌‌‍⁠‍‌‍‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‍​⁠​‍​⁠‌​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‍​⁠​​​⁠​‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠‌‌​⁠‌‌‌‍⁠​‌‌​‌​⁠‌‍‌‍⁠‌‌‌​​‌‌‍‌‌‍‍‌‌⁠​‍‌‍⁠‍‌​⁠‍‌​​‍‌‌‍‍‌‍‌‍​⁠‌‌​‍​‍‌⁠⁠‌

I got burned by a “0.6 m lens at about 3.8 m” once, so I now build a trigger into the scope: run one pilot CPT at the first boring and, if there’s a qc trough/u spike around 3.5–4 m, immediately add two CPTs at 6–8 m spacing along the load path — usually a half‑day and under ~$2k. If CPT access is tight, I have the driller log MWD (torque/advance) and take continuous 0.3 m spoons through that interval to catch thin sand ribbons.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌‌‌‍⁠‍‌‍‌⁠‌‍‍‌‌‍⁠‍‌‍‌‌‌‍‌‌‌⁠​‍‌‍‍‌‌‍⁠‍‌‍‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‍​⁠​‍​⁠‌​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‍​⁠​​​⁠​⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠‌​‌​⁠​‌⁠​‌‌​​⁠​⁠​​‌​⁠⁠‌‍‌⁠‌‌‌​‌‌‌​‌‍‌‌‌​​‍​⁠​‌‌⁠​​​⁠‌‍‌‌‌‌‌⁠​⁠​‍​‍‌⁠⁠‌