new code calibrates daqcard

diff --git a/comedi_calibrate/comedi_calibrate.c b/comedi_calibrate/comedi_calibrate.c
index 5e0d25f34ff9966efba438c72a8054dfc9782885..6ddae2869aea838e2aae147c00a8fa8d415ad2fd 100644 (file)
--- a/comedi_calibrate/comedi_calibrate.c
+++ b/comedi_calibrate/comedi_calibrate.c
@@ -33,7 +33,7 @@
 #include <stdlib.h>
 #include <string.h>
 
-#define DPRINT(level,fmt,args...) do{if(verbose>=level)printf(fmt,args);}while(0)
+#define DPRINT(level,fmt,args...) do{if(verbose>=level)printf(fmt, ## args);}while(0)
 
 #define N_CALDACS 32
 #define N_OBSERVABLES 32
@@ -84,6 +84,9 @@ double check_gain_chan(int ad_chan,int range,int cdac);
 
 int verbose = 0;
 
+/* tmep */
+void do_cal(void);
+
 void observe(void);
 void preobserve(int obs);
 void observable_dependence(int obs);
@@ -298,6 +301,10 @@ int main(int argc, char *argv[])
        reset_caldacs();
 
        observe();
+
+       do_cal();
+
+       observe();
 #if 0
        for(i=0;i<n_boards;i++){
                if(!strcmp(boards[i].name,devicename)){
@@ -374,6 +381,14 @@ void ni_setup(void)
                        CR_PACK(0,unipolar_lowgain,AREF_OTHER);
                observables[i].target = 0;
                i++;
+
+               /* unip gain */
+               observables[i].name = "ai, unipolar voltage reference, low gain";
+               observables[i].observe_insn = tmpl;
+               observables[i].observe_insn.chanspec =
+                       CR_PACK(5,unipolar_lowgain,AREF_OTHER);
+               observables[i].target = voltage_reference;
+               i++;
        }
 
        if(da_subdev>=0){
@@ -502,6 +517,95 @@ void observable_dependence(int obs)
 
 }
 
+
+void postgain_cal(int obs1, int obs2, int dac)
+{
+       double offset1,offset2;
+       linear_fit_t l;
+       double slope1,slope2;
+       double a;
+       double gain;
+       comedi_range *range1,*range2;
+
+       DPRINT(0,"postgain calibration\n");
+       preobserve(obs1);
+       check_gain_chan_x(&l,observables[obs1].observe_insn.chanspec,dac);
+       offset1=linear_fit_func_y(&l,caldacs[dac].current);
+       DPRINT(1,"obs1: [%d] offset %g\n",obs1,offset1);
+       range1 = comedi_get_range(dev,observables[obs1].observe_insn.subdev,
+               CR_CHAN(observables[obs1].observe_insn.chanspec),
+               CR_RANGE(observables[obs1].observe_insn.chanspec));
+       slope1=l.slope;
+
+       preobserve(obs2);
+       check_gain_chan_x(&l,observables[obs2].observe_insn.chanspec,dac);
+       offset2=linear_fit_func_y(&l,caldacs[dac].current);
+       DPRINT(1,"obs2: [%d] offset %g\n",obs2,offset2);
+       range2 = comedi_get_range(dev,observables[obs2].observe_insn.subdev,
+               CR_CHAN(observables[obs2].observe_insn.chanspec),
+               CR_RANGE(observables[obs2].observe_insn.chanspec));
+       slope2=l.slope;
+
+       gain = (range1->max-range1->min)/(range2->max-range2->min);
+       DPRINT(3,"range1 %g range2 %g\n", range1->max-range1->min,
+               range2->max-range2->min);
+       DPRINT(2,"gain: %g\n",gain);
+
+       DPRINT(2,"difference: %g\n",offset2-offset1);
+
+       a = (offset1-offset2)/(slope1-slope2);
+       a=caldacs[dac].current-a;
+
+       DPRINT(0,"caldac[%d] set to %g\n",dac,a);
+
+       caldacs[dac].current=rint(a);
+       update_caldac(dac);
+       usleep(100000);
+
+       if(verbose>=2){
+               preobserve(obs1);
+               measure_observable(obs1);
+               preobserve(obs2);
+               measure_observable(obs2);
+       }
+}
+
+void cal1(int obs, int dac)
+{
+       linear_fit_t l;
+       double offset;
+       double target;
+       double gain;
+       double a;
+
+       DPRINT(0,"cal1\n");
+       preobserve(obs);
+       check_gain_chan_x(&l,observables[obs].observe_insn.chanspec,dac);
+       offset=linear_fit_func_y(&l,caldacs[dac].current);
+       gain=l.slope;
+       
+       target = observables[obs].target;
+       a=caldacs[dac].current+(target-offset)/gain;
+
+       caldacs[dac].current=rint(a);
+       update_caldac(dac);
+       usleep(100000);
+
+       DPRINT(0,"caldac[%d] set to %g\n",dac,a);
+       if(verbose>=2){
+               measure_observable(obs);
+       }
+}
+
+void do_cal(void)
+{
+       // daqcard
+       postgain_cal(0,1,2);
+       cal1(1,8);
+       cal1(2,0);
+}
+
+
 double ni_get_reference(int lsb_loc,int msb_loc)
 {
        int lsb,msb;
@@ -1210,6 +1314,16 @@ void update_caldac(int i)
        int ret;
        
        DPRINT(3,"update %d %d %d\n",caldacs[i].subdev,caldacs[i].chan,caldacs[i].current);
+       if(caldacs[i].current<0){
+               DPRINT(0,"caldac set out of range (%d<0)\n",caldacs[i].current);
+               caldacs[i].current=0;
+       }
+       if(caldacs[i].current>caldacs[i].maxdata){
+               DPRINT(0,"caldac set out of range (%d>%d)\n",
+                       caldacs[i].current,caldacs[i].maxdata);
+               caldacs[i].current=caldacs[i].maxdata;
+       }
+
        ret = comedi_data_write(dev,caldacs[i].subdev,caldacs[i].chan,0,0,
                caldacs[i].current);
        if(ret<0)perror("update_caldac()");