| View previous topic :: View next topic |
| Author |
Message |
JamesW
Joined: 23 Apr 2007
Posts: 91
Location: Rochester, England
|
| Measuring VBG (bandgap voltage ref) on PIC24FJ12GB204 |
 Posted: Fri Feb 02, 2018 10:15 am |
 |
|
Hi folks,
Has anyone had any success measuring the 1.2V bandgap reference using the ADC on a PIC24FJ12GB204?
I don't think I am doing anything that wrong, but I'm getting some very weird results. (CCS support are also looking into this).
Processor setup
| Code: |
#include <24FJ128GB204.h>
#device PASS_STRINGS = IN_RAM
#case
#device ADC=12
#device ICSP=1
#use delay(internal=32MHz, USB_FULL,restart_wdt)
#BIT VBGEN=getenv("BIT:VBGEN")
|
ADC Setup
| Code: |
setup_adc_ports(sAN10, VSS_VDD);
setup_adc(ADC_CLOCK_INTERNAL | ADC_TAD_MUL_0);
|
Bit where I actually read it
| Code: |
/* ENABLE THE BAND GAP VOLTAGE REFERENCE */
VBGEN = TRUE;
set_adc_channel(28);
delay_us(20);
BandGapADC = read_adc();
fprintf(DEBUG_PORT, "\r\nBAND GAP:%04u", BandGapADC);
|
.
The voltage level on my processor is 3.3437V.
By that rationale if I am reading the band gap voltage (which according to the datasheet is 1.2V) I should expect to see 35.8% of the full scale value of the A/D converter (running in 12 bit mode) when I read from channel 28. So I should expect to see a reading of 1466.
However :
With the ADC_TAD_MUL_0 set – I get a reading of 2917
With the ADC_TAD_MUL_31 set – I get a reading of 1939 instead.
Neither of these is even close to the 1466 I am expecting.
Unless of course I am being an idiot, and I'm always willing to eat humble pie if it solves the problem!
Thanks in advance
James |
|
 |
Ttelmah
Joined: 11 Mar 2010
Posts: 19540
|
|
| Posted: Fri Feb 02, 2018 1:20 pm |
|
|
I'd set the BGREQ bit as well. Otherwise the band gap may not actually be on.
As a comment, you need ACT=USB added to your clock setup if you intend to use the USB. |
|
|
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
| Posted: Fri Feb 02, 2018 2:42 pm |
|
|
Put a delay after you enable the band gap generator. See what happens.
| Quote: |
VBGEN = TRUE;
delay_ms(10); // *** Add this line
set_adc_channel(28);
delay_us(20);
BandGapADC = read_adc();
fprintf(DEBUG_PORT, "\r\nBAND GAP:%04u", BandGapADC);
|
|
|
|
JamesW
Joined: 23 Apr 2007
Posts: 91
Location: Rochester, England
|
|
| Posted: Mon Feb 05, 2018 7:58 am |
|
|
Hi Folks,
Thanks for the help as always.
I have stuck the above code changes in, and it makes absolutely no difference.
My ADC reading is 2913
However I then had an email from Richard @ ccs support, who has found this. . .
To get the correct VBG reading I had to increase the sampling time. To do that I had to use a slower ADC clock, surprisingly the internal clock was to fast, a 1ms ADC clock period with a sampling time of 31 Tad seem to work for me. For example with a clock speed of 32MHz the following ADC setup worked for me to read it correctly:
| Code: |
setup_adc(ADC_CLOCK_DIV_32 | ADC_TAD_MUL_31);
|
originally I had this setup
| Code: |
setup_adc(ADC_CLOCK_INTERNAL | ADC_TAD_MUL_31);
|
Under the new settings I get an ADC value of this....
BAND GAP:1479
Which is pretty much where I'd expect it to be.
However - I am at a loss to explain why! Any ideas?
Thanks
James |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9243
Location: Greensville,Ontario
|
|
| Posted: Mon Feb 05, 2018 8:03 am |
|
|
this
setup_adc(ADC_CLOCK_INTERNAL
most PICs 'internal' is ONLY supposed to be used for sleeping PICs..it's somewhere in the ADC section, a chart or table, that shows valid adc clock vs CPU clock...usually there's 2 or 3 valid settings you can use.
the div_31 is interesting to me... 32 makes sense, 31 is just 'odd' though I don't use that PIC so I'll assume it's a correct value
Jay |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19540
|
|
| Posted: Mon Feb 05, 2018 9:43 am |
|
|
| temtronic wrote: | this
setup_adc(ADC_CLOCK_INTERNAL
most PICs 'internal' is ONLY supposed to be used for sleeping PICs..it's somewhere in the ADC section, a chart or table, that shows valid adc clock vs CPU clock...usually there's 2 or 3 valid settings you can use.
the div_31 is interesting to me... 32 makes sense, 31 is just 'odd' though I don't use that PIC so I'll assume it's a correct value
Jay |
I'm sorry Jay, this is wrong for the PIC24's.
The internal clock is often recommended.
I originally was going to suggest increasing the divider, since some other chips in the family have the internal RC too fast without a divider, and /2 has to be selected. However I checked the data sheet, and for this chip it is specific that the internal clock is slow enough.
Looks like a case where the data sheet is wrong for this chip.... 
On several of the other chips there is advice _not_ to have a TAD_MUL of 0. However '1' is enough if you have the delay as you do.
I'd suggest:
setup_adc(ADC_CLOCK_DIV_2 | ADC_TAD_MUL_1);
Might well work (assuming you have the acquisition delay you show in your code). |
|
|
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
| Posted: Mon Feb 05, 2018 11:15 am |
|
|
There is a related errata on this PIC, though it's for reading VBG/2:
| Quote: | 2. Module: A/D Converter
Incorrect VBG/2 voltage measurement of the A/D
Converter at full speed.
When the A/D Converter is converting at full speed
(500 ksps for 10-bit and 200 ksps for 12-bit), the
A/D Converter count may not match the VBG/2 voltage.
Work around:
The A/D Converter clock should be lowered to
below 100 ksps (in 12-bit mode) to read the correct
value of the VBG/2 voltage. In 10-bit mode,
the clock must be lowered to below 200 ksps.
|
http://ww1.microchip.com/downloads/en/DeviceDoc/80000613e.pdf |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19540
|
|
| Posted: Mon Feb 05, 2018 11:43 am |
|
|
Yes. I had seen that one.
It definitely looks like one that MicroChip should be told about. |
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
