STM CircleOS forum / ADC jitter issue ? or not ?

ThomasScherrer · 2008-02-11 15:43:54

Hello I have tested several STM32 Development boards now.
I see 5-8 counts of ADC jitter om them all,
and like to hear if anyone managed to get a nice stable ADC reading without software avaraging or software filters ?
have you managed to improve the ADC jitter by adding extra filters on VDD and VDDA ?
and what is your best result ? and how did you solve it ?

I know we have a 12 bit ADC so 5-8 bits of RMS jitter is 4-6mV no real big issue :-)
but I like to know if it is possible to get 1 bit of jitter with hardware only ???

If I measure Peak to Peak Jitter I have 18-22 counts of jitter,
I Added 100uF on VDD and 100uF on the ADC pin, now I measure only 11-13 counts of PP jitter.
Tomorrow I will try to add filters on VDDA and analog GND to see if I can improve it.

Last edited by ThomasScherrer (2008-02-11 18:27:03)

ThomasScherrer · 2008-02-12 09:38:18

now I have made a distrbution accumulator of ADC noise:
048,024,084,060,109,084,204,1959,2749,2076,17921,102344,51420,15142,2567,621,256,060,000,012,000,013,000,000,000,000
000,024,048,060,062,108,305,1824,2475,1985,17223,102194,52520,15274,2671,837,084,024,012,000,012,000,024,000,000,000
024,012,049,012,135,133,349,1915,2463,2195,17444,103269,51847,14174,2944,583,169,024,012,000,000,012,000,000,000,000
036,012,048,048,097,148,399,1823,2605,2281,18893,104277,49074,14369,2800,688,110,037,000,000,000,000,000,000,000,000
036,024,073,000,192,099,422,1692,2595,2436,20301,106302,47036,13254,2442,643,096,050,037,012,000,012,000,000,000,000
060,060,097,048,073,097,353,1973,2240,2478,17744,103912,50691,13885,3213,643,144,025,012,000,000,000,000,000,000,000
037,036,060,024,145,121,446,1638,2543,2367,18918,101630,51261,15264,2501,617,112,048,013,000,000,000,000,000,000,000
012,036,048,000,074,072,459,1798,2495,2188,18164,101381,51786,15353,3079,655,121,024,012,025,013,000,000,000,000,000
012,024,097,012,097,072,341,1571,2535,2204,17105,101116,51717,16639,3350,570,242,036,000,024,000,013,000,000,000,000
049,024,060,048,132,145,378,1783,2576,1977,18226,102037,51651,14914,2785,743,205,037,012,000,012,000,000,000,000,000
048,024,096,024,097,108,393,1758,2379,2259,17240,103612,50835,15318,2680,654,146,049,000,012,000,012,000,000,000,000
038,072,024,144,073,458,1668,2809,1994,16976,102256,52541,15351,2675,593,209,048,012,012,024,000,000,000,000,000,000

each line is one sec ! of hits into each value, sentre value is the correct ADC value,
each comma separeted position is one adc count.
great, a forum that only can handle old tiny monitor resolutions :-)
ok to get the files in full resolutuion see : www.webx.dk/stm32/

Last edited by ThomasScherrer (2008-02-12 09:47:57)

ThomasScherrer · 2008-02-12 15:05:26

more news, compared the KEIL MCBSTM32 board and
the ST board STM32F10F10X-128K-EVAL to really cool one with TFT LCD

the ST board have TONS more ADC noise, removed LCD and have now IMPRESSIVE low noise !
only +/- one LSB wow !!
so the switchmode supply on the back side of the LCD is one nasty noise source to the ADC :-)
be warned if you design with switchmodes and need a clean ADC, add extra coils and caps near the noise source to isoilate it.

Now I am investigating the KEIL board, the capacitors for VDDA and analog GND are placed too far away from IC,
this gives poor decoubling to the analog section, it is hard to fix on that board.

Last edited by ThomasScherrer (2008-02-12 15:11:51)

ThomasScherrer · 2008-02-12 15:22:12

with LCD on ST kit

no LCD on ST kit

KEIL MCBSTM32 kit

Last edited by ThomasScherrer (2008-02-12 16:22:25)

Francis · 2008-02-12 16:34:40

Obviously, the activity on the IOs is generating a lot of noise. Not very surprising. The solution seems to synchronize the use of the ADC with "low power states": no "fast output toglling", low CPU clock, etc... but is is not always compatible with a MCU standard usage.
I had the same remark for the issue with the precision of the MEMS on the STM32 Primer: we should try to limit the SPI access during the measurement.
Anyway, if you define some "good rules" for the design of the board, I will be interested by following them for the future generation of Primers. Thanks.

ThomasScherrer · 2008-02-12 18:43:11

the LCD was not initialized, and the MCU did not contain any toogling LCD port pins !
it is noise solely from LCD power converter

TroyC3 · 2008-02-14 03:41:48

Thanks for sharing this info Thomas.

BTW how did you capture the data, serial, usb? and how did you get those cool graphs.

Thanks again,
Troy

ThomasScherrer · 2008-02-14 07:56:57

We have developed a PC program called XMON
please see thread about XMON it is free to DL now

Last edited by ThomasScherrer (2008-02-14 12:35:38)

afillera · 2008-11-12 21:22:07

i need help about the ADC of the stm32f103rbt6 microcontroller...? How can i help me?

ben · 2008-12-21 01:25:24

Hi,

I recommend you to have a look at this App Note from STMicro : AN2834 - How to get the best ADC accuracy in STM32F10xxx devices
http://www.st.com/stonline/products/lit … /15067.pdf

Regards,

Ben,

# 1 2008-02-11 15:43:54 ADC jitter issue ? or not ?

ADC jitter issue ? or not ?

# 2 2008-02-12 09:38:18 ADC jitter issue ? or not ?

Re: ADC jitter issue ? or not ?

# 3 2008-02-12 15:05:26 ADC jitter issue ? or not ?

Re: ADC jitter issue ? or not ?

# 4 2008-02-12 15:22:12 ADC jitter issue ? or not ?

Re: ADC jitter issue ? or not ?

# 5 2008-02-12 16:34:40 ADC jitter issue ? or not ?

Re: ADC jitter issue ? or not ?

# 6 2008-02-12 18:43:11 ADC jitter issue ? or not ?

Re: ADC jitter issue ? or not ?

# 7 2008-02-14 03:41:48 ADC jitter issue ? or not ?

Re: ADC jitter issue ? or not ?

# 8 2008-02-14 07:56:57 ADC jitter issue ? or not ?

Re: ADC jitter issue ? or not ?

# 9 2008-11-12 21:22:07 ADC jitter issue ? or not ?

Re: ADC jitter issue ? or not ?

# 10 2008-12-21 01:25:24 ADC jitter issue ? or not ?

Re: ADC jitter issue ? or not ?

Board footer