ADS1298: Full-scale issue

Hi Alexander, thanks for your reply.

Input signal = i've tried with the internal test signal and connecting myself to the device to get my heart signal. The results are the same. 

My VREFP = 2.4v, i don't know if i have to change the VREF_4V bit of the config3 register, and set it to 1. 

My VREFN =-2.5v

PGA gain per channel = 12

Hi Carlos,

Can you confirm that the device is still functional using the BIOFAQ?

Can you show scope shots of your digital communication? MISO, MOSI, CS, START?

Hi Alexander.

The device allows to write and read the Registers, i can read back what i write to each register. 

I can Read the ID register, and the voltages of  VCAPs seem to agree with the values of the BIOFAQ.  

MISO is connected to DOUT

MOSI is connected to DIN

CS is a selected pin by me, i reset the pin and set it manually 

Start is no connected so a i can issue the command as needed 

Communication.docx

Hi Carlos,

Is it possible to see oscilloscope screenshots of your communication? 

Does the result change if a gain of 1 is used instead of 12? 

Can you provide the schematic and pictures of PCB layout? 

Hey.

unfortunably i have no oscilloscope in here. I debug the MCU step by step and i see directly what is the device is giving back.

with gain 1 i recieve the same data. 0x7fffff and 0x800000.

here is the schematic i'm using for the regulators and the ADS1298.

Esquemas.pdf

Hi Carlos,

Got it. For now, do not configure WCT, and other miscellaneous registers. I recommend only enabling the internal reference and necessary input channels to debug, leave all other registers default and see if that helps. 

If the device ID is being returned correctly, the power supplies should be powering up correctly. Since changing the gain or input signal does not affect the full-scale result, that shouldn't be the issue. That means it either has to be something with the reference being mis-configured, or an error in the code. 

The schematic for the reference looks fine, but if possible confirm on the PCB that the capacitors are correctly routed and going to the pins. 

If you can somehow get access to an oscilloscope, that would be great because then we can confirm that the device is outputting data correctly and it is an error with the code.  

Let me know. 

Alexander,

I've configurated the nexts registers as follow, the other registers are configurated by default by the device.  

Reg Config3 = 0x80 only enable the internal reference buffer
RegCH1Set = 0x10 only enable the firt channel with a gain of 1 

I connect myself to the device and i get sometimes only 0x7ffff back and sometimes only 0x80000 but no different values, so i tried  leaving  the cable on the air just capting the enviroment to see if changes anything, but no. I don't have a singal generator to try out with a known signal, so i don't really know how else i can try. 

Don't know if it's a code error but this is how i'm converting data to voltage, as i've seen in the "How do I convert ADC output codes to volts?" forum.

it's in Python so...

 import serial

import numpy as np

from matplotlib import pyplot as plt

dato=np.zeros(3000)

contador=1

puerto = serial.Serial('COM3',230400, timeout=1)

VREF= 2.5

GAIN= 1  

MSB= 0x800000

Samples=int(input('ingrese numero de muestras a graficar\n Muestras= '))

LSB=((2*VREF)/GAIN)/(2**23)

while contador<=samples:

    ch = puerto.read(3)

    valordato=ch[0]<<16 or ch[1]<<8 or ch[2]

    datoMSB = valordato & MSB

    if datoMSB == 0:

        data=valordato*LSB

    else:

        data=((2**24)-valordato)*LSB

    dato[contador]=data

    contador+=1

Hi Carlos,

Can you measure between VREFP and VREFN? 

In you previous post it looks like the difference is actually 4.9V?

I'm not sure if this is Python language or not, but the " ** " in LSB=((2*VREF)/GAIN)/(2**23) and  data=((2**24)-valordato)*LSB should be to the power of " ^ ".

LSB should also be LSB = (2 x VREF)/ Gain / (2^24 - 1). Notice that 1 is not subtracted from the exponent 24. 

valordato is the decimal equivalent of the number, correct?

 data=((2**24)-valordato)*LSB should be  data=(valordato-(2**24))*LSB

Hi Carlos, 

Any luck? 

Hi Alexander,

Yes, in python the "^" operator is represented by "**".

Notice that i substract the -1 on the equation since the beginning LSB=((2*VREF)/GAIN)/(2**23) as the exponent is not 24-1 but 23, in any case i'll put it following your suggestions. so it'll be  LSB = (2 *VREF)/ GAIN / (2**24 - 1).

Correct, valordato is the 24 bit channel data that i recieve from the device. 

OH ok! i corrected that so the data equation is now data=(valordato-(2**24))*LSB. 

The problem  still that i'm only getting either 0x80000 the most part or 0x7ffff some times from the ADS1298,  no other values. 

I've changed the programming, instead of sending the RDATA command i'm using the  RDATAC   following the  figure 93 of the datasheet, and it sends different values not only 0x7fffff nor 0x80000, but i'm not sure if it's working the proper way, since i'm getting as test signal this.... 

(similar in each channel)

(similar in each channel)  this with more samples.

(similar in each channel) and this when i connect myself.

i don't really see what i want to see on this. I'm sampling 500 SPS and it doesn't show what it supposed to the square signal on the Test and the heart signal with noise or something, so i don't know. 

Hi Alexander,

Yes, in python the "^" operator is represented by "**".

Notice that i substract the -1 on the equation since the beginning LSB=((2*VREF)/GAIN)/(2**23) as the exponent is not 24-1 but 23, in any case i'll put it following your suggestions. so it'll be  LSB = (2 *VREF)/ GAIN / (2**24 - 1).

Correct, valordato is the 24 bit channel data that i recieve from the device. 

OH ok! i corrected that so the data equation is now data=(valordato-(2**24))*LSB. 

The problem  still that i'm only getting either 0x80000 the most part or 0x7ffff some times from the ADS1298,  no other values. 

I've changed the programming, instead of sending the RDATA command i'm using the  RDATAC   following the  figure 93 of the datasheet, and it sends different values not only 0x7fffff nor 0x80000, but i'm not sure if it's working the proper way, since i'm getting as test signal this.... 

(similar in each channel)

(similar in each channel)  this with more samples.

(similar in each channel) and this when i connect myself.

i don't really see what i want to see on this. I'm sampling 500 SPS and it doesn't show what it supposed to the square signal on the Test and the heart signal with noise or something, so i don't know. 

Hi Carlos,

You are getting closer! 

Remember your order of operations and PEMDAS. Exponents (2^24) are processed before Subtraction. So for (2^24-1), it is really ((2^24)-1). 

Few other things... do not write 0x80 to CONFIG3. This is invalid, see page 69 for details, but notice that the second bit must be 1. Instead of 0x80, write 0xC0. 

Additionally, see that the reference voltage is actually 2.4V, not 2.5V as written in your code. This will return an invalid LSB size. 

After making these changes, can you return your LSB size just so we can confirm that it is being calculated by the program correctly? 

Hi Alexander, 

Ok i had no idea it was ((2^24)-1). i've all ready changed it. 

i've measured the voltages with this results.

VREFP=1.4v 

VREFN=-2.3v

so VREF= 3.7v  (Note: internal buffer is enable)

VCAP1 = -1.1

VCAP2 = 0.4v

VCAP3 = 4.4v

VCAP4 = 1.4v

I've set the VREF_4V bit on the Config3 to 0 so the VREF is set to 2.4, however, the voltage in the hardware is one volt down as i mentioned above. 

is this the problem? 

Hi Carlos,

This could certainly be the issue. Is it possible that the device was damaged when installing or debugging? 

Was this soldered in by hand or at a board house? 

You may want to try replacing the device to see if that helps. 

Hi Alexander, 

That certainly are really bad news, since is not the first device damaged in my attempt to run this project.  

it was soldered by hand by me actually, but i was as  carefully i could. 

is there a way where i can lift those voltages without replacing the device? i mean, i think it is no damaged because it lets me read and write reagisters and ID. And im connecting th devices as you noticed posts above. So, is there anything i could do to enhance this voltages issue? 

Hi Carlos,

It could be that just the reference buffer is damaged. You could perhaps try wiring in an external reference. 

The last possibility is that there is some error during power-up such that the reference is not being configured correctly. However I don't believe this to be the issue since the ID is returning correctly. 

Try to use an ESD strap and magnification tool if possible. 

Hi Alexander, 

I've achieved to lift the VREFP voltage so now that pin is 2.4 and the VREF=4.7v. 

When i put the test signals i get this result:

channel 1

channel2

Channel3

Certainly it has been a progress since i've lifted that voltage. Channel one seems to be failling but i dont now why, it could be damaged?, the second channel ilustrates the square test signal with noise, i guess its working?, and the third one is the best signal of all, this could have solved that. Nontheless, when i connect my self i still don't get the signal that i want for the device.

this is the code in python to get and convert data to voltage. 

VREF= 2.4
GAIN= 1
MSB= 0x800000
muestras=int(input('ingrese numero de muestras a graficar\n Muestras= '))

LSB=((2*VREF)/GAIN)/((2**24)-1)

while contador<=muestras:

ch = puerto.read(3)
valordato = (ch[0] << 16) or (ch[1] << 8) or ch[2]
datoMSB = valordato & MSB

if datoMSB == 0:
     data=valordato*LSB
else:
     data=(valordato-(2**24))*LSB
dato[contador]=data
contador+=1

And this is how i'm configurating ADS1298

Reg0x01=0x86
Reg0x02=0x00
Reg0x03=0x40 (External reference)
Reg0x05=0x10
Reg0x06=0x10
Reg0x07=0x10
Reg0x08=0x10
Reg0x14=0x00
Reg0x17=0x08
Reg0x18=0x08
Reg0x19=0xD4