AFE44x0SPO2EVM

Dear, Praveen-san.
I appreciate your valuable information.

I am having additional regarding AFE44x0SPOEVM GUI. Could you please teach me again.

1. GUI => Device Configuration => LED1 Convert/LED2 Convert.
What kind of transaction is implemented by LED1 Convert & LED2 Convert?

2. GUI => ADC Capture & Analysis => Horizontal-axis of displayed graph / Number of Average
Does "Number of Average" mean the sampling number by ADC?

3. GUI => ADC Capture & Analysis => Test Results
Does "Test Results(CSV file)" mean the data after averaging transaction by above(#2) number?

4. Following transactions are performed in which portion? AFE? MSP430? Software?
a)Averaging transaction(related to #3)?
b)Notch filter transaction?
c)FFT transaction?

Best Regards,

Dear, Praveen-san.

I am very sorry to ask you again and again.
Could you help me again.

5. Regarding 3V supply is separated for MSP430 and AFE onb EVM.
Is it no problem to combine both into one supply?

6. TPS7A4901 has low PSRR(72dB@120Hz) and Vnoise(15.4uVrms@10Hz-100kHz).
Is there any suggestion spec on power supply(LDO) if a customer replaced to the others?

Best Regards,

Hello Hideyuki-san,

Please find my answers below:

1. During LED1 convert and LED2 convert phases, the voltage sampled at Cled1 and Cled2 respectively are buffered and applied to the ADC for conversion. The conversion time duration is normally 25% of the Pulse Repetition Period (PRP count).(refer datasheet)

EVM Default PRF is 500Hz. With 4MHz clock, the PRP count is 8000 and so LED1 and LED2 Convert durations are 2000 counts minus 2 counts to account for ADC reset.

2. There are two data capture modes: finite and continuous. For finite mode, the number of samples to capture is specified by the user. In continuous mode, the number of samples received is plotted continuously until the user stops the data capture.

Horizontal axis of displayed graph corresponds to the number of samples received.

3. Yes the Device_Analysis_<date>_<record number>.xls stores the average, RMS and peak-to-peak voltage.

The raw data codes is stored under Device_Code_<date>_<record number>.xls.

The raw data in volts is stored under Device_Volts_<date>_<record number>.xls.

4. All three transactions (averaging, notch filter and FFT) are performed in the software GUI.

Hideyuki-san,

5. Three main sources of noise coupling in AFEs 1) input pins 2) reference voltage 3) supply and ground pins.  Other way of saying it is, we are always sampling input, reference, supply and ground pins.   We like to separate analog and digital supplies because digital switching noise is wide spectrum and can easily couple into our bandwidth of interest.  If the switching frequencies are multiples of our sampling frequencies, our converter can act to sub-sample.  It’s possible, single supply is good enough, but it’s something you have to prototype and test.

6. Below are few suggestions when you are looking for LDOs.

1. Can it output the voltage and current required? 
2. What’s the PSRR of the LDO at the switching frequencies of interest? For example at 8MHz clock, and PRF frequency.

Ideally the noise is below one LSB in the bandwidth of interest.  If the noise is high, you can filter LDO output. 

Best Regards.

Dear, Praveen-san.

Thank you for your teachings. 

I understood that the suppliers might be separated Digital and Analog. 

Is it no problem to combine MASP_AVCC(for MSP430) and RX_ANA_SUP(for  AFE4490)? 

Sorry again and again. 

Best Regards, 

Dear, Praveen-san. 

May I double-confirm about followings? Sorry again and again. 

2. There are two data capture modes: finite and continuous. For finite mode, the number of samples to capture is specified by the user. In continuous mode, the number of samples received is plotted continuously until the user stops the data capture.Horizontal axis of displayed graph corresponds to the number of samples received.

Q2) Is it a read-out value of Resister 30? 

3. Yes the Device_Analysis_<date>_<record number>.xls stores the average, RMS and peak-to-peak voltage.

The raw data codes is stored under Device_Code_<date>_<record number>.xls.

The raw data in volts is stored under Device_Volts_<date>_<record number>.xls.

Q3) Is it a read-out value of Resister 42, 43, 44, 45, 46, 46, 47?

4. All three transactions (averaging, notch filter and FFT) are performed in the software GUI.

Q4-1) AFE4490 has "Averager" block inside. Averaging is implemented by AFE?

 Q4-2)Is it cetainly sure that notch filter is implemented by software GUI? 

Q4-3)Is it certainly sure that FFT is implemted by software GUI?

Best Regards, 

Hello Hideyuki-san,

Please find my explanations for your questions below:

Q2) Is it a read-out value of Resister 30?
The value stored in Register 30 corresponds to the number of ADC conversions minus 1 for LED2, LED2 Ambient, LED1, LED1 Ambient channels.
During each Pulse Repetition Period (PRP), there are four ADC resets timed every 25% of the PRP. After every reset, ADC conversion begins.
Each ADC conversion time is 50us. So total ADC conversion time = (Value in Reg 30 + 1) * 50us.

At the end of 25% of the PRP, the average of the ADC conversions are stored in register 42.
At the end of 50% of the PRP, the average of the ADC conversions are stored in register 43
At the end of 75% of the PRP, the average of the ADC conversions are stored in register 44
At the end of 100% of the PRP, the average of the ADC conversions are stored in register 45 and register 42 minus register 43 is stored in register 46 and register 44 minus register 45 is stored in register 47.

(Please refer to the datasheet for more detailed explanation)

For example,
Let Reg 30 = 7 which implies Number of ADC conversions is 8.
Total ADC conversion time is 400us for each channel.
During the 4 resets in each PRP, the average of the 8 ADC conversions are stored in reg 42  through reg 45 respectively.

Let Pulse Repetition Period (PRP) = 2ms
At the end of 500us, the average of first 8 ADC converted samples (corresponding to LED2) are stored in reg 42.
At the end of 1ms, the average of second 8 ADC converted samples (corresponding to LED2 Ambient) are stored in reg 43.
At the end of 1.5ms, the average of third 8 ADC converted samples (corresponding to LED1) are stored in reg 44.
At the end of 2ms, the average of fourth 8 ADC converted samples (corresponding to LED2 Ambient) are stored in reg 45. Also reg 42 minus reg 43 is stored in reg 46 and reg 44 minus reg 45 is stored in reg 47.

With PRP = 2ms, data rate is 500 samples per second.
Each sample is the averaged ADC data stored in reg 42 through reg 47.
When you enter 1024 as the number of samples in the GUI, the GUI captures and displays approximately two seconds worth of data.

Q3) Is it a read-out value of Resister 42, 43, 44, 45, 46, 46, 47?
Yes they are the values from reg. 42 through reg 47 as explained above.

Q4-1) AFE4490 has "Averager" block inside. Averaging is implemented by AFE?
The average data in the Device_Analysis_< >.xls file is the average of the total number of samples captured. Example: average of 1024 samples.
Each sample is the averaged ADC data from the ADC Averager Block of AFE4490.

Q4-2)Is it cetainly sure that notch filter is implemented by software GUI?
Yes Notch filter is implemented by the GUI.

Q4-3)Is it certainly sure that FFT is implemted by software GUI?
Yes FFT is implemented by the GUI.

Regards

Praveen.

Dear, Praveen-san.

Thank you for your clear explanations. All is maybe cleared. 

Sorry again regarding previous question about power supplies. 

I understood that the suppliers might be separated Digital and Analog. 

Is it no problem to combine MSP_AVCC(for MSP430) and RX_ANA_SUP(for  AFE4490)? 

Both is same analog. 

Sorry again and again. 

Best Regards, 

Hello Hideyuki-san,

It depends on the performance required and what MSP430 is doing.

It is possible to combine the analog supplies as well. But this is something you have to prototype and test on your board.

Hello Praveen, 

I have a question about the axis of the Gui. In normal case the Pulsoximeter gives his values in percent. At the AFE44x0SPO2EVM GUI we get volts or codes in de y-axis. I don´t understand the context. How we get the value in volts. And can you explain me how I get the values in precent, like a normal Pulsoximeter ? 

And at least, can you explain me the difference of the 4 LED's?

Thank you!

Best Regards
Hatice Zorbulut

Hi Hatice,

Our AFE gives the digital code which represents the voltage .
The voltage information from the different LEDs are processed offline to derive the oxygen saturation in terms of percentage.
You can refer to following documents for the calculations involved. Note that this is only for reference.
www.ti.com/.../tidu475.pdf

Regards,
Prabin