A wagon of initial mass $M$ is moving with an initial velocity $V_o$. Now sand starts accumulating on this wagon, increasing the total mass from $M$ to $M+s$. We have to find the final velocity of the wagon. Provided is the fact that the accumulating sand does not have any initial velocity.
Let p(t) be the momentum of the wagon and the sand on it.
$$p(t)=(M+m)v \\$$ where m = m(t) is the sand accumulated on the wagon $$p(t + dt) = (M+m+dm)(v+dv)$$ Since momentum is conserved, hence: $$P_{initial} = P_{final}$$
$$p(t)=p(t + dt)$$
$$(M+m)v = (M+m+dm)(v+dv)$$
$$(M+m)v = (M+m)v+(M+m)dv+vdm+dmdv$$
$$0 = (M+m)dv+vdm+dmdv$$ Since dmdv is very small, we get $$(M+m)dv+vdm = 0$$ $$\frac {dv}{v} = -\frac{dm}{(M+m)}$$
We observe that dv is negative, making v+dv lesser than v.
Integrating, we get
$$\int_{V_o}^{V} \frac {dv}{v} = -\int_{o}^{s}\frac{dm}{(M+m)}$$ $$V = \frac{V_o M}{M+s}$$
But now if we write -dv instead of dv, and redo the calculation, then: $$p(t)=(M+m)v \\ p(t + dt) = (M+m+dm)(v-dv)\\ P_{initial} = P_{final}$$
$$p(t)=p(t + dt)$$
$$(M+m)v = (M+m+dm)(v-dv)$$
$$(M+m)v = (M+m)v-(M+m)dv+vdm-dmdv$$
$$0 = -(M+m)dv+vdm-dmdv$$ Since dmdv is very small, we get $$-(M+m)dv+vdm = 0$$ $$\frac {dv}{v} = \frac{dm}{(M+m)}$$
We see that dv is positive, making v-dv less than v.
Integrating, we get
$$\int_{V_o}^{V} \frac {dv}{v} = \int_{o}^{s}\frac{dm}{(M+m)}$$ $$V = \frac{V_o (M+s)}{M}$$
The answer in this case is inverted. What exactly is wrong in the second version of the calculation. In most calculations, changing the sign of dv doesn't affect the final answer, as it readjusts to give out the same answer in either case.
So does this mean that we should always write a change in quantity $x$ as $x+dx$, or is there anything that I specifically did wrong in the second calculation?