L3G4200d output stability

[Picbuster]

I am facing difficulties with the zero stability ( at no movement) of this chip.
Communication via SPI to pic18lf14k22 works correct.
Who has the same experience with the L3g4200d?

 

[bigdogguru]

How are you interfacing with the L3G4200D, with a commercial breakout board or a DIY design?

Can you post a detailed schematic of the L3G4200D and all its pins?

BigDog

 

[Picbuster]

How are you interfacing with the L3G4200D, with a commercial breakout board or a DIY design?

Can you post a detailed schematic of the L3G4200D and all its pins?

BigDog

via SPI bus as being used in many of my design's.
temp and serial number are reading correctly

spec sheet at:
https://www.google.nl/url?sa=t&rct=...J4Uf4ck2RBHuAWUXiQ&bvm=bv.49478099,bs.1,d.ZWU

 

[bigdogguru]

You did not answer my question.

Are you using a commercial breakout board or have you mounted the device on your own PCB?

Post a schematic of the implementation.

The reason I ask this question is to determine whether or not the PLL Low Pass Filter has been implemented correctly, the device has proper decoupling caps and the state of the unused pins.


BigDog

 

[Picbuster]

You did not answer my question.

Are you using a commercial breakout board or have you mounted the device on your own PCB?

Post a schematic of the implementation.

The reason I ask this question is to determine whether or not the PLL Low Pass Filter has been implemented correctly, the device has proper decoupling caps and the state of the unused pins.


BigDog

Here it is I did attach schematic but not sure if you received it.
I removed all unneeded parts and functions and keep it to the parts involved.
aim: to detect rotations of 10 milli degree per second.( this should work according to ST but I ??)

sorry the copy code into this screen will add spaces.

---

Code C - [expand]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
//============  SPI int =========================    
    if  (PIR1bits.SSPIF)
    {
    Spi_rec[Spi_pointer]=SSPBUF;
              Spi_pointer++;
             PIR1bits.SSPIF=0;       // flag off
             }
//========= request receive ======= 
 
 
void R_Inclino(int P_function,int number)
{
        int i;
        Spi_pointer=0;
        Klok_sel=0;              // select chip
        Xt_SPI(P_function);  // pointer to function
        for (i=0;i<number;i++)  
        {
         Xt_SPI(0); // dummy generate clock to receive
 
         }  
        Klok_sel=1;
}
//=================== main ===========
 
 
// -------------------spi bus master ------------------
  
 
    SSPSTATbits.CKE=1; //0=Data transmitted on falling edge of SCK 
    SSPSTATbits.SMP=1; //1=SPI mode-Input data sampled at end of data output time
    SSPCON1bits.CKP=0; //0=idle clock state is low  
    SSPCON1bits.SSPM3=0; //FOSC/4 
    SSPCON1bits.SSPM2=0; //FOSC/4 
    SSPCON1bits.SSPM1=0; //FOSC/64
    SSPCON1bits.SSPM0=0; //FOSC/4  1=/16
 
    IPR1bits.SSPIP=1;
   
    PIE1bits.SSPIE=1;
    PIR1bits.SSPIF=0;
//----------------------------------------------    
    PEIE=1;                 // Enable peripheral int.
    TMR0 = 0; 
    GIE = 1; //enable Global interrupts
//================================
 
 
//========= 
    SSPCON1bits.SSPEN=1;  // start spi interface
 
 
    R_Inclino(0b10001111,1);// read Chip_ID
              Chip_ID= Spi_rec[1];
             CREN=1;
             input=0;
            Char_pointer=0;
            Address= eeprom_read(0);
   
    putchar('>');
 
 
 
    
    
             SPI_s[0]=0b00100000;
             SPI_s[1]=0b00001111;
    S_Inclino();               // controlreg1 write power, x,y,x on
              Delay10KTCYx(3500);  // wait to start l3g
 
            SPI_s[0]= 0b00100011;
            SPI_s[1]=0b01000000;//  
    S_Inclino();        // controlreg4 write
   
 
            SPI_s[0]=0b00100100;
             SPI_s[1]=0b01000000;   // fifo enable
            S_Inclino();                // controlreg5 write
   
    
 
             SPI_s[0]=0b00101110; // fifo_cntrl_reg
             SPI_s[1]=0b01100111; // 
    S_Inclino();        //  write fifo_cntrl_reg
 
           Delay10KTCYx(3500);
 
// ============== read x,y,z    loop
 
 
  signed int Zas;         
    signed int Xas;
    signed int Yas;
 
 
              R_Inclino(0b10101111,1);   // fifo src
    int status = Spi_rec[1];
     if   ((status & 128) ==128)  // if data read it else skip
         
       {  
            R_Inclino(0b11101000,6);  // get 6 bytes
            signed int  Xas_L = Spi_rec[1];
        signed int Xas_H = Spi_rec[2];
            Xas =((Xas_H << 8) + Xas_L);  
         
 
            signed int Yas_L = Spi_rec[3];
        signed int Yas_H = Spi_rec[4];
            Yas = ((Yas_H << 8) + Yas_L); 
          
            signed int Zas_L = Spi_rec[5];
        signed int Zas_H = Spi_rec[6];
            Zas = ((Zas_H << 8) + Zas_L);  
            
       if (Xas_out != Xas) // for debug purposes only
       {        
        printf("X:%d Y:%d Z:%d         \r",Xas,Yas,Zas);
       }
            Xas_out=Xas;
            Yas_out=Yas;
            Zas_out=Zas;
     }
 
                        Spi_pointer=0;

---

 

[mtwieg]

I'm pretty sure that getting a 0.01 dps measurement isn't going to be feasible with this, or any other mems gyro. You're talking about a rotational period of once per hour. Also the built-in highpass filter will probably mask such slow movements completely, and anything left over will be swamped with noise.

Also you never actually explained what you mean by "stability." Are you referring to drift or noise or something else?

 

[Picbuster]

I'm pretty sure that getting a 0.01 dps measurement isn't going to be feasible with this, or any other mems gyro. You're talking about a rotational period of once per hour. Also the built-in highpass filter will probably mask such slow movements completely, and anything left over will be swamped with noise.

Also you never actually explained what you mean by "stability." Are you referring to drift or noise or something else?

That's what I believe so ST salesman is wrong.
I talk about noise value popping at a delta value about 200. (raw data). All axis.

Did you receive drawing? ( I was not sure this edabord up-load mechanism is new to me.)

 

[bigdogguru]

Here it is I did attach schematic but not sure if you received it.

No, the schematic appears not to have been uploaded.

If you can post a link/url from which I can obtain it, I will upload it and post it for you.

I removed all unneeded parts and functions and keep it to the parts involved.

Hopefully, the PLL filter did not fall into the category of "unneeded parts."

aim: to detect rotations of 10 milli degree per second.( this should work according to ST but I ??)

Frankly, I agree with mtwieg, after reviewing the L3G4200D Datasheet it is doubtful that your objective is obtainable.




BigDog

 

[mtwieg]

That's what I believe so ST salesman is wrong.

Well there's a big difference between "possible" and "feasible." The thing is sensitive to changes as small as 10mdps, but actually being able to measure such a thing with useful accuracy is a completely different issue.

Also don't take technical advice from the sales department.

I talk about noise value popping at a delta value about 200. (raw data). All axis.

The datasheet specs the noise at 0.03dps/√Hz, so if your bandwidth is high then that much measured noise may be reasonable. The bandwidth of the measurement can be programmed in the sensor, or you could digitally filter the data yourself to reduce the noise.

 

[Picbuster]

Well there's a big difference between "possible" and "feasible." The thing is sensitive to changes as small as 10mdps, but actually being able to measure such a thing with useful accuracy is a completely different issue.

Also don't take technical advice from the sales department.
The datasheet specs the noise at 0.03dps/√Hz, so if your bandwidth is high then that much measured noise may be reasonable. The bandwidth of the measurement can be programmed in the sensor, or you could digitally filter the data yourself to reduce the noise.

View attachment l34200d_test 001.jpg here it is

 

[mtwieg]

The schematic looks fine.

Just an example noise calculation for you: if you've configured the measurement bandwidth to be high (max seems to be 110Hz) then you'll get 0.3146 dps rms noise, or about 1.89 dps of peak to peak noise on your measurement. With max sensitivity that's 216 LSBs of peak to peak noise on the digital result. Is that consistent with what you're seeing? Note that number doesn't include offset in the measurement. Table 4 in the datasheet says that it can be +/- 10dps, which is another big problem for you.

Also make sure the thing is put on a surface where it won't feel any vibration at all.

Exactly what kind of motion are you trying to detect? A tilt sensor based on accelerometers may work better than a gyro.

 

[Picbuster]

The schematic looks fine.

Just an example noise calculation for you: if you've configured the measurement bandwidth to be high (max seems to be 110Hz) then you'll get 0.3146 dps rms noise, or about 1.89 dps of peak to peak noise on your measurement. With max sensitivity that's 216 LSBs of peak to peak noise on the digital result. Is that consistent with what you're seeing? Note that number doesn't include offset in the measurement. Table 4 in the datasheet says that it can be +/- 10dps, which is another big problem for you.

Also make sure the thing is put on a surface where it won't feel any vibration at all.

Exactly what kind of motion are you trying to detect? A tilt sensor based on accelerometers may work better than a gyro.

I want to follow a slow stream running down. (compare it with a snail climbing into a tree. It goes randomly in a spiral or just not)
having all angular movements as function of the Z position we are cable to reconstruct the road.

This will work with a 'normal' gyroscope; however I am forced into a size of 10 x 45mm this with my components easy to fulfil.

Next problem: I detected is an overflow on the fifo this should not happen hence an internal mechanism in the chip should avoid this.

anyway I will sit back in the corner cry for 10.51 minutes and start scratching my head to find an other solution.

thank you for your time.