Forces not working properly when using slabs too thick

[cam559]

I've been trying to calculate some properties of a metallic interface (Nb/Ta), every time I run the calculation with a high number of atoms (48 total) the forces look something like this:

`# Forces in Lattice coordinates: force Nb 0.000000000000000 0.000000000000000 28751.819314407683123 0 force Nb 0.000000000000000 0.000000000000000 46171.116472006462573 1 force Nb 0.000000000000000 -0.000000000000000 58874.832161463680677 1 force Nb 0.000000000000000 0.000000000000000 67574.246283973756363 1 force Nb 0.000000000000000 -0.000000000000000 77395.338443646745873 1 force Nb 0.000000000000000 0.000000000000000 47961.315487052823300 1 force Nb 0.000000000000000 -0.000000000000000 -17348.499521483390708 1 force Nb 0.000000000000000 0.000000000000000 -73409.661623366861022 1 force Nb 0.000000000000000 0.000000000000000 -76336.371243756817421 1 force Nb 0.000000000000000 0.000000000000000 -57708.825133711274248 1 force Nb 0.000000000000000 0.000000000000000 -58188.822850378484873 1 force Nb 0.000000000000000 -0.000000000000000 -38949.988740276858152 1 force Nb 0.000000000000000 -0.000000000000000 38979.923003945601522 1 force Nb 0.000000000000000 0.000000000000000 58187.638195714549511 1 force Nb 0.000000000000000 -0.000000000000000 57620.822109505432309 1 force Nb 0.000000000000000 0.000000000000000 76340.532879638223676 1 force Nb 0.000000000000000 0.000000000000000 73449.038382159691537 1 force Nb 0.000000000000000 0.000000000000000 17335.011084017209214 1 force Nb 0.000000000000000 -0.000000000000000 -47919.350254474564281 1 force Nb 0.000000000000000 0.000000000000000 -77368.542639859049814 1 force Nb 0.000000000000000 -0.000000000000000 -67633.017691133747576 1 force Nb 0.000000000000000 -0.000000000000000 -58900.226981341264036 1 force Nb 0.000000000000000 0.000000000000000 -46141.308120547291765 1 force Nb 0.000000000000000 0.000000000000000 -28779.295353140980296 1 force Ta 0.000000000000000 0.000000000000000 -1327.125627082751635 1 force Ta 0.000000000000000 0.000000000000000 1331.162125288017023 1 force Ta 0.000000000000000 0.000000000000000 4599.821065833293687 1 force Ta 0.000000000000000 -0.000000000000000 10064.205810928247956 1 force Ta 0.000000000000000 -0.000000000000000 9791.886937015642616 1 force Ta 0.000000000000000 -0.000000000000000 4780.769016837569325 1 force Ta 0.000000000000000 0.000000000000000 -1630.199012004271935 1 force Ta 0.000000000000000 0.000000000000000 -7593.209177953922335 1 force Ta 0.000000000000000 -0.000000000000000 -10908.208369088335530 1 force Ta 0.000000000000000 -0.000000000000000 -7605.483534517090447 1 force Ta 0.000000000000000 -0.000000000000000 -1266.316369958987934 1 force Ta 0.000000000000000 -0.000000000000000 -6587.970844243500324 1 force Ta 0.000000000000000 0.000000000000000 6790.901093227559613 1 force Ta 0.000000000000000 0.000000000000000 1284.764938337779995 1 force Ta 0.000000000000000 0.000000000000000 7618.552984461792221 1 force Ta 0.000000000000000 -0.000000000000000 10906.112431508054215 1 force Ta 0.000000000000000 0.000000000000000 7590.159416074802721 1 force Ta 0.000000000000000 0.000000000000000 1628.936370971757015 1 force Ta 0.000000000000000 0.000000000000000 -4778.446134244384666 1 force Ta 0.000000000000000 0.000000000000000 -9795.232561292594255 1 force Ta 0.000000000000000 -0.000000000000000 -10048.144259401851741 1 force Ta 0.000000000000000 0.000000000000000 -4619.416305135289804 1 force Ta 0.000000000000000 0.000000000000000 -1284.571681228442003 1 force Ta 0.000000000000000 -0.000000000000000 1050.468447986115734 1

Energy components:

Eewald = -2388.1864576352436416 EH = 100381.4164926642260980 Eloc = -49.6876873786794775 Enl = -14.6899515677200512 Exc = -801.8807122174267761 Exc_core = 331.0847044283245282 KE = 3374.7417498828672251

 Etot =   100832.7981381763529498
   TS =        0.0015957437105255

---

    F =   100832.7965424326393986`

I don't see this issue when using only 1 atomic species or when reducing the number of atoms by half. I'm not using coulomb truncation as the interface is repeating itself periodically. Any idea why is this happening?