cancel
Showing results for 
Search instead for 
Did you mean: 

PSoC 4 MCU

Anonymous
Not applicable

I have a custom hardware setup that is reading a current sensor and a pressure sensor. If I read the ADC input with a multimeter I see 2.5V from both sensors but in code the ADC is returning 3.085V and 2.984.

I've attached my code. I'm running a very simple Fake RTOS to sample the sensor and print info out to the UART.

Do you guys see anything wrong with the setup?

Joe

0 Likes
1 Solution
Anonymous
Not applicable

I figured out the remainder of my issue.

I was treating the return of ADC_CountsTo_MVotls() as an uint16 and then attempting to translate that value to a mV value. ADC_CountsTo_MVotls() returns a int16 and the compiler treats int16 values a two's complement numbers. So (int16)2500 mv = (2'sComp) 0x94C

All I had to do was change my variable type be a int16, remove my math that was not needed and BAM! it works!

Thank you Bob and Evgeniy for your help!

Joe

View solution in original post

0 Likes
11 Replies
JoMe_264151
Expert II

Which voltage is your PSoC running on?

Another point might be: the writing to UART may take longer than the 1/200 second, so it may be disturbed by the interrupt. Give the ADC highest prio and the UART more time: you are sending about 80 chars at 9600 chars/s which would take about 8ms whie the window is set to 5ms

Bob

0 Likes
Anonymous
Not applicable

Hi Bob,

Thank you for taking a look at my stuff.

The PSoC 4 is running on 5V and I've set the internal voltages to 5V.

I set the priority of the ADC task to the highest and my sensor sampling interrupt to one step lower. I've set the print priority to the lowest priority and the ADC reading is still off.

I double checked the settings on my UART. The baud rate is set to 115200 so the message should be really quick. If you look at the code I also set it up so that the UART only gets written to every 10 times the DEBUG_PRINT_INFO is called. I was worried about exactly what you are hypothesising.

Joe

0 Likes
JoMe_264151
Expert II

Baud rate of 115200 is a bit rate. Divide by 12 gives 9600 chars/s. Calculate for yourself how long it takes to transmit your 80 chars.

There are APIs to calculate counts to mv adjustable for each channel which you use. Are the values of v_water_ticks and v_current_ticks wrong?

Keep in mind that debugging (stop) will give wrong ADC results when resuming without a complete conversion cycle.

Found one more: Avoid multiple calls to ADC_StartConvert(). ADC is configured as free running.

Bob

0 Likes
Anonymous
Not applicable

Hi Bob,

I've removed the multiple calls to ADC_StartConvert(). I've also changed the prints to be:

      sprintf(msg,"W:0x%x\r\n",v_water_ticks );

and

      sprintf(msg,"C:0x%x\r\n",v_current_ticks );

To minimize the time needed in printing.

The ticks are the values that are off. I'm still seeing 0x980ish and 0x9D0ish. They should both be sitting at 0x800ish ticks which is equates to 2.5V.  Thank you again for taking a look at this.

Do you see anything else I'm doing wrong?

Joe

0 Likes
JoMe_264151
Expert II

Can you please provide us with an actual workspace archive.

Bob

0 Likes
Anonymous
Not applicable

Hi Bob,

Here is the full workspace. I'm using PSoc Creator 4.2 so you may have to upgrade to view the stuff. There is a second project in the workspace that is a modified version of the ADC_SAR_Seq_DieTemp_PSoc4 project. I'm currently attempting to modify that project to use 5V and my sensor pins.

Joe

0 Likes
JoMe_264151
Expert II

This is not a complete workspace archive, please use

Creator->File->Create Workspace Bundle (minimal)

and attach the resulting file.

Bob

0 Likes
Anonymous
Not applicable

Hi Bob,

Here you go.

I've been playing with the PSOC 4 test project and it is giving me the correct values. I'm going to see if I can port the project over to my project.

Joe

0 Likes
EvPa_264126
Valued Contributor II

For what you divide by 0.819 in line 73

2.5/0.819 = 3.05

I filed 1.463V and saw the v_water_ticks = 1469mV

and yet Vref select = VDDA gives an additional error

VDDA may not be exactly 5.00V

0 Likes
EvPa_264126
Valued Contributor II

if   Single Ended negative input = Vref

it seems there is something wrong

ADC.png

0 Likes
Anonymous
Not applicable

I figured out the remainder of my issue.

I was treating the return of ADC_CountsTo_MVotls() as an uint16 and then attempting to translate that value to a mV value. ADC_CountsTo_MVotls() returns a int16 and the compiler treats int16 values a two's complement numbers. So (int16)2500 mv = (2'sComp) 0x94C

All I had to do was change my variable type be a int16, remove my math that was not needed and BAM! it works!

Thank you Bob and Evgeniy for your help!

Joe

View solution in original post

0 Likes