diff --git a/Grid/qcd/action/gauge/GaugeImplTypes.h b/Grid/qcd/action/gauge/GaugeImplTypes.h
index 8e418209..dbab4386 100644
--- a/Grid/qcd/action/gauge/GaugeImplTypes.h
+++ b/Grid/qcd/action/gauge/GaugeImplTypes.h
@@ -140,8 +140,38 @@ public:
       PokeIndex<LorentzIndex>(P, Pmu, mu);
     }
   }
+    
+    
 
-  static inline Field projectForce(Field &P) { return Ta(P); }
+  static inline Field projectForce(Field &P)
+    {
+        if (isSp2n == true)
+        {
+            P = Ta(P);
+            const int nsp = Nc / 2;
+
+            Sp<nsp>::iSp2nMatrix<Complex> gen;
+            
+
+            auto Psum = P;
+
+            Psum = Zero();
+            
+            for (int a = 0; a < Sp<nsp>::AlgebraDimension; a++)
+            {
+                Sp<nsp>::generator(a, gen);
+
+                auto coeff = 2. * trace(P * gen);
+                Psum += coeff * gen;
+            }
+            
+            return Psum;
+            
+        } else
+        {
+            return Ta(P);
+        }
+    }
 
   static inline void update_field(Field& P, Field& U, double ep){
     //static std::chrono::duration<double> diff;
diff --git a/Grid/tensors/Tensor_Ta.h b/Grid/tensors/Tensor_Ta.h
index a82fa7aa..2ee51fb4 100644
--- a/Grid/tensors/Tensor_Ta.h
+++ b/Grid/tensors/Tensor_Ta.h
@@ -137,6 +137,9 @@ accelerator_inline iMatrix<vtype,N> ProjectOnGroup(const iMatrix<vtype,N> &arg)
 
 // re-do for sp2n
 
+// Ta cannot be defined here for Sp2n because I need the generators from the Sp class
+// It is defined in gauge impl types
+
 template<class vtype> accelerator_inline iScalar<vtype> ProjectOnSpGroup(const iScalar<vtype>&r)
 {
   iScalar<vtype> ret;
@@ -217,88 +220,6 @@ accelerator_inline iMatrix<vtype,N> ProjectOnSpGroup(const iMatrix<vtype,N> &arg
   return ret;
 }
 
-
-
-
-
-
-/*
-template<class vtype,int N, typename std::enable_if< GridTypeMapper<vtype>::TensorLevel == 0 >::type * =nullptr>
-accelerator_inline iMatrix<vtype,N> ProjectOnSpGroup(const iMatrix<vtype,N> &arg)
-{
-  // need a check for the group type?
-  iMatrix<vtype,N> ret(arg);
-  vtype nrm;
-  vtype inner;
-  for(int c1=0;c1<N/2;c1++){
-
-    // Normalises row c1
-    zeroit(inner);
-    for(int c2=0;c2<N;c2++)
-      inner += innerProduct(ret._internal[c1][c2],ret._internal[c1][c2]);
-
-    nrm = sqrt(inner);
-    nrm = 1.0/nrm;
-    for(int c2=0;c2<N;c2++)
-      ret._internal[c1][c2]*= nrm;
-
-    // Compute row c1+N2/2: c1+N/2 = - \Omega c1*
-    for(int c2=0;c2<N/2;c2++)
-      ret._internal[c1+N/2][c2+N/2] = conjugate(ret._internal[c1][c2]); // lupo: c1,c2? dont we want c1 c2+N/2
-
-    for(int c2=N/2;c2<N;c2++)
-      ret._internal[c1+N/2][c2-N/2] = -conjugate(ret._internal[c1][c2]);
-
-    // Remove c1 from rows c1+1...N/2-1
-    for (int b=c1+1; b<N/2; ++b){
-      decltype(ret._internal[b][b]*ret._internal[b][b]) pr;
-      zeroit(pr);
-      for(int c=0; c<N; ++c)
-        pr += conjugate(ret._internal[c1][c])*ret._internal[b][c];
-
-      for(int c=0; c<N; ++c){
-        ret._internal[b][c] -= pr * ret._internal[c1][c];
-      }
-    }
-
-    // Remove c1+N/2 from rows c1+1...N/2-1
-    for (int b=c1+1; b<N; ++b){
-      decltype(ret._internal[b][b]*ret._internal[b][b]) pr;
-      zeroit(pr);
-      for(int c=0; c<N; ++c)
-        pr += conjugate(ret._internal[c1+N/2][c])*ret._internal[b][c];
-
-      for(int c=0; c<N; ++c){
-        ret._internal[b][c] -= pr * ret._internal[c1+N/2][c];
-      }
-    }
-  } //end c1 0, ... N/2-1 loop
-
-  // Compute the last row
-  {
-  // Normalise last row
-    int c1 = N/2-1;
-    zeroit(inner);
-    for(int c2=0;c2<N;c2++)
-      inner += innerProduct(ret._internal[c1][c2],ret._internal[c1][c2]);
-
-    nrm = sqrt(inner);
-    nrm = 1.0/nrm;
-    for(int c2=0;c2<N;c2++)
-      ret._internal[c1][c2]*= nrm;
-
-  // Compute row cN-1+N2/2: cN-1+N/2 = - \Omega cN-1*
-    for(int c2=0;c2<N/2;c2++)
-      ret._internal[c1+N/2][c2+N/2] = conjugate(ret._internal[c1][c2]);
-
-    for(int c2=N/2;c2<N;c2++)
-      ret._internal[c1+N/2][c2-N/2] = -conjugate(ret._internal[c1][c2]);
-  }
-  // assuming the determinant is ok
-  return ret;
-}*/
-
-
 NAMESPACE_END(Grid);
 
 #endif
diff --git a/tests/sp2n/Test_hmc_Sp_pureGaugeWilson.cc b/tests/sp2n/Test_hmc_Sp_pureGaugeWilson.cc
new file mode 100644
index 00000000..64049ec2
--- /dev/null
+++ b/tests/sp2n/Test_hmc_Sp_pureGaugeWilson.cc
@@ -0,0 +1,100 @@
+/*************************************************************************************
+
+Grid physics library, www.github.com/paboyle/Grid
+
+Source file: ./tests/Test_hmc_WilsonFermionGauge.cc
+
+Copyright (C) 2015
+
+Author: Peter Boyle <pabobyle@ph.ed.ac.uk>
+Author: neo <cossu@post.kek.jp>
+Author: Guido Cossu <guido.cossu@ed.ac.uk>
+
+This program is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License along
+with this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+See the full license in the file "LICENSE" in the top level distribution
+directory
+*************************************************************************************/
+/*  END LEGAL */
+#include <Grid/Grid.h>
+
+int main(int argc, char **argv) 
+{
+  using namespace Grid;
+
+  Grid_init(&argc, &argv);
+  GridLogLayout();
+
+  typedef GenericSp2nHMCRunner<MinimumNorm2> HMCWrapper;
+  HMCWrapper TheHMC;
+
+
+  TheHMC.Resources.AddFourDimGrid("gauge");
+
+
+
+  // Checkpointer definition
+  CheckpointerParameters CPparams;  
+  CPparams.config_prefix = "ckpoint_lat";
+  CPparams.rng_prefix = "ckpoint_rng";
+  CPparams.saveInterval = 5;
+  CPparams.format = "IEEE64BIG";
+  
+  TheHMC.Resources.LoadNerscCheckpointer(CPparams);
+
+  RNGModuleParameters RNGpar;
+  RNGpar.serial_seeds = "12 22 32 42 52";
+  RNGpar.parallel_seeds = "76 77 87 79 70";
+  TheHMC.Resources.SetRNGSeeds(RNGpar);
+
+  // Construct observables
+  // here there is too much indirection 
+  typedef PlaquetteMod<HMCWrapper::ImplPolicy> PlaqObs;
+  typedef TopologicalChargeMod<HMCWrapper::ImplPolicy> QObs;
+  TheHMC.Resources.AddObservable<PlaqObs>();
+  TopologyObsParameters TopParams;
+  TopParams.interval = 5;
+  TopParams.do_smearing = true;
+  TopParams.Smearing.steps = 200;
+  TopParams.Smearing.step_size = 0.01;
+  TopParams.Smearing.meas_interval = 50;
+  TopParams.Smearing.maxTau = 2.0; 
+  TheHMC.Resources.AddObservable<QObs>(TopParams);
+  //////////////////////////////////////////////
+
+  /////////////////////////////////////////////////////////////
+  // Collect actions, here use more encapsulation
+  // need wrappers of the fermionic classes 
+  // that have a complex construction
+  // standard
+  RealD beta = 8.0 ;
+  SymplWilsonGaugeActionR Waction(beta);
+  
+  ActionLevel<HMCWrapper::Field> Level1(1);
+  Level1.push_back(&Waction);
+  //Level1.push_back(WGMod.getPtr());
+  TheHMC.TheAction.push_back(Level1);
+  /////////////////////////////////////////////////////////////
+
+  // HMC parameters are serialisable 
+  TheHMC.Parameters.MD.MDsteps = 10;
+  TheHMC.Parameters.MD.trajL   = 1.0;
+
+  TheHMC.ReadCommandLine(argc, argv); // these can be parameters from file
+  TheHMC.Run();  // no smearing
+
+  Grid_finalize();
+
+} // main