diff --git a/R/est_gmm.R b/R/est_gmm.R
index 6e0f307d..295853f0 100755
--- a/R/est_gmm.R
+++ b/R/est_gmm.R
@@ -516,7 +516,7 @@ pgmm <- function(formula, data, subset, na.action,
   yX <- yX1
 
   # Compute the first step matrices
-  if (transformation == "d")  A1 <- tcrossprod(diff(diag(1, T - TL1 + 1))) # TODO: to use FSM() as well?
+  if (transformation == "d")  A1 <- FSM(T - TL1, "G") # == tcrossprod(diff(diag(1, T - TL1 + 1))) # TODO: FSM's arg fsm not fully flexible
   if (transformation == "ld") A1 <- FSM(T - TL2, fsm)
 
   # compute the estimator
@@ -552,7 +552,8 @@ pgmm <- function(formula, data, subset, na.action,
                       function(x)
                       as.vector(x[ , 1L] - crossprod(t(x[ , -1L, drop = FALSE]), coefficients)))
   outresid <- lapply(residuals, function(x) outer(x, x))
-  
+
+  # A2 is only needed for model = "twosteps", but seems currently needed in vcov.pggm also for "onestep" model
   A2 <- mapply(function(x, y) crossprod(t(crossprod(x, y)), x), W, outresid, SIMPLIFY = FALSE)
   A2 <- Reduce("+", A2)
   minevA2 <- min(eigen(A2)$values)
@@ -560,7 +561,7 @@ pgmm <- function(formula, data, subset, na.action,
     warning("the second-step matrix is singular, a general inverse is used")
     ginv(A2)
   } else solve(A2)
-
+  
   if (model == "twosteps") {
     coef1s <- coefficients
     t.CP.WX.A2 <- t(crossprod(WX, A2))
diff --git a/R/tool_vcovG.R b/R/tool_vcovG.R
index ccad1063..7cc20c15 100644
--- a/R/tool_vcovG.R
+++ b/R/tool_vcovG.R
@@ -1233,6 +1233,7 @@ vcovHC.pgmm <- function(x, ...) {
     residuals <- x$residuals
   }
   else {
+    # transformation = "d"
     yX <- x$model
     residuals <- x$residuals
   }
@@ -1243,7 +1244,7 @@ vcovHC.pgmm <- function(x, ...) {
     warning("a general inverse is used")
     ginv(A2)
   } else solve(A2)
-  
+
   if(model == "twosteps") {
     coef1s <- x$coefficients[[1L]]
     res1s <- lapply(yX, function(x) x[ , 1L] - crossprod(t(x[ , -1L, drop = FALSE]), coef1s))
@@ -1254,7 +1255,6 @@ vcovHC.pgmm <- function(x, ...) {
     We <- Reduce("+", mapply(function(x, y) crossprod(x, y), x$W, residuals, SIMPLIFY = FALSE))
     B1 <- solve(t(WX) %*% A1 %*% WX)
     B2 <- vcov(x)
-
     vcov1s <- B1 %*% (t(WX) %*% A1 %*% SA2 %*% A1 %*% WX) %*% B1
     for (k in 2:K) {
       exk <- mapply(