[ECOS] Re: stm3210e eval board adc question

[jayant biswas <biswasj at gmail dot com>]

Thanks Christophe for the patch... you are a bit of a legend!

It seems to work now, both adc1 and adc2 tests are running. I am able
to read the values from the potentiometer on the board. And I get some
random values for the BNC inputs (which are currently not connected to
anything). I still have to fix my own program so I can produce similar
results, but I think I can follow the code of the two tests to make my
code work.

I will post again if I have more questions

Thanks again!

Jayant

On Mon, May 2, 2011 at 23:42, Christophe Coutand <ccoutand@stmi.com> wrote:
> Hi Jayant,
>
> While reviewing the code I found few issues in the ADC driver. Please try with the attached code and let me know if it works for you.
>
> Christophe
>
>
> -----Original Message-----
> From: jayant biswas [mailto:biswasj@gmail.com]
> Sent: 2. mai 2011 12:16
> To: Christophe Coutand
> Cc: ecos-discuss@ecos.sourceware.org
> Subject: Re: stm3210e eval board adc question
>
> Hello again!
>
> I have tried stepping through the code for adc1.c. Stepping from line
> 124 in adc1.c, I notice that when trying to read from the channels it
> is always waiting for data to be available on /dev/adc014. This means
> that in packages/io/adc/current/src/adc.c it never reaches line 131
> (// A sample is available, transfer it to the buffer).
>
> Do you have an idea why the data is never available from the
> potentiometer on the ?STM3210E-eval board? Let me know if I can
> provide more information about my tests.
>
> Thank you,
> Jayant
>
> On Tue, Apr 19, 2011 at 14:42, jayant biswas <biswasj@gmail.com> wrote:
>> Hi Christophe.
>>
>> Thanks for the info about how to find out which channel corresponds to
>> which input. This will be useful for me.
>>
>> As for the terminal I am using cutecom on ubuntu. 38400,8,1, parity
>> none, no handshake, open for reading and writing.
>>
>> I tried running the adc1 test and that only runs this far:
>> \0x00INFO:<ADC test>
>> Testing ADC channel '/dev/adc00'
>>
>> when I force stop the program it says the following
>> (gdb) continue
>> Continuing.
>> ^C
>> Program received signal SIGINT, Interrupt.
>> 0x08003832 in Cyg_Mutex::lock (this=0x6800001c)
>> ? ?at /home/jayant/ecos/adc/ecos_install/include/cyg/kernel/sched.inl:99
>> 99 ? ? ? ? ?if( __lock == 0 ) unlock_inner(0);
>> Current language: ?auto; currently c++
>>
>> Stepping through adc2 in gdb with a breakpoint on line 181 as you
>> mentioned i get this
>> (gdb) break adc2.c:181
>> Breakpoint 1 at 0x800027e: file adc2.c, line 181.
>> (gdb) continue
>> Continuing.
>> Note: automatically using hardware breakpoints for read-only addresses.
>>
>> Breakpoint 1, adc_thread (data=<value optimized out>) at adc2.c:181
>> 181 ? ? ? ? ? ? ? ? ? ? cyg_uint32 len = sizeof(sample);
>> (gdb) info frame
>> Stack level 0, frame at 0x68004668:
>> ?pc = 0x800027e in adc_thread (adc2.c:181); saved pc 0x8002b4e
>> ?called by frame at 0x68004670
>> ?source language c.
>> ?Arglist at 0x68004628, args: data=<value optimized out>
>> ?Locals at 0x68004628, Previous frame's sp is 0x68004668
>> ?Saved registers:
>> ?r4 at 0x68004644, r5 at 0x68004648, r6 at 0x6800464c, r7 at
>> 0x68004650, r8 at 0x68004654,
>> ?r9 at 0x68004658, r10 at 0x6800465c, r11 at 0x68004660, lr at 0x68004664
>> (gdb) print len
>> $1 = <value optimized out>
>> (gdb) print sample
>> $2 = 26624
>> (gdb) s
>> cortex_m3 interrupt mask on
>> 176 ? ? ? ? ? ? for (i = 0; i < num; i++) {
>> cortex_m3 interrupt mask off
>> (gdb) print len
>> $3 = 2
>> (gdb) print sample
>> $4 = 26624
>> (gdb) s
>> cortex_m3 interrupt mask on
>> 177 ? ? ? ? ? ? ? ? chan = &test_channels[i];
>> cortex_m3 interrupt mask off
>> (gdb) s
>> cortex_m3 interrupt mask on
>> 182 ? ? ? ? ? ? ? ? ? ? res = cyg_io_read(chan->handle, &sample, &len);
>> cortex_m3 interrupt mask off
>> (gdb) s
>> cortex_m3 interrupt mask on
>> 181 ? ? ? ? ? ? ? ? ? ? cyg_uint32 len = sizeof(sample);
>> cortex_m3 interrupt mask off
>> (gdb) print len
>> $5 = 0
>> (gdb) print sample
>> $6 = 26624
>> (gdb)
>>
>> Not sure what I should be looking for here. Can you please suggest how
>> I may proceed to find the problem?
>>
>> Thanks again
>> Jayant
>>
>> On Tue, Apr 19, 2011 at 09:25, Christophe Coutand <ccoutand@stmi.com> wrote:
>>> Hi Jayant,
>>>
>>> From this documentation: http://www.st.com/stonline/books/pdf/docs/14220.pdf, the potentiometer is connected to PC4.
>>>
>>> Looking at the IOs definition in: packages\hal\cortexm\stm32\var\current\include\var_io.h, you can sample PC4 using ADC1/14.
>>>
>>> #define CYGHWR_HAL_STM32_ADC12_IN14 ? ? ? ? ? ? CYGHWR_HAL_STM32_GPIO( C, 4, ?IN, ANALOG )
>>> #define CYGHWR_HAL_STM32_ADC1_IN14 ? ? ? ? ? ? ?CYGHWR_HAL_STM32_ADC12_IN14
>>>
>>> The adc2 test does not log any sampling results. It is used to check that the sampling rate is correct over a period of 10 seconds. The output of your test does not look good. The \0x09 is a tabulation that your terminal is not interpreting (which terminal do you use)? But the result is 0 for all channels, meaning that no samples are read.
>>>
>>> ----------------------------------------
>>> Samples expected after 10210 milliseconds: 10210
>>> Samples read (per channel):
>>> /dev/adc00 ? ? ? ? ?\0x09= 0
>>>
>>> Do you have a way to step into the code? I don't clearly see how the test can complete without reading samples. Line 181 is a loop waiting for samples:
>>>
>>> ?// Read & count samples
>>> ?do {
>>> ? ?cyg_uint32 len = sizeof(sample);
>>> ? ?res = cyg_io_read(chan->handle, &sample, &len);
>>> ? ?if (res == ENOERR)
>>> ? ? ? chan->count++;
>>> ?} while (res == ENOERR);
>>>
>>> You can also try the adc1 test which shall read and display the sampled data. I don't own a STM3210e boards therefore can't give it a try.
>>>
>>> Regards,
>>> Christophe
>>>
>>> -----Original Message-----
>>> From: jayant biswas [mailto:biswasj@gmail.com]
>>> Sent: 18. april 2011 16:49
>>> To: Christophe Coutand
>>> Cc: ecos-discuss@ecos.sourceware.org
>>> Subject: Re: stm3210e eval board adc question
>>>
>>> Dear Christophe!
>>>
>>> Thanks for your support. I have a feeling I am getting closer, but not
>>> quite there yet. So I appreciate your time in helping me out. Now I
>>> have been able to add the adc io, as in the screenshot you sent using
>>> only the cvs code. So I am not mixing anymore ecos-3.0 and the cvs.
>>> But when I run the tests on my stm3210e eval board I do not get the
>>> desired output. The output of adc2 test is shown below. It shows that
>>> no samples have been read. I am also not sure what to make of the
>>> \0x09 stuff before the '=' signs.
>>>
>>> My goal is to be able to read the value from the potentiometer on the
>>> board (http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATA_BRIEF/CD00278458.pdf
>>> as shown on page 2/3) and eventually also from the 3 bnc connectors.
>>> First of all, how can I know which input (BNC or Potentiometer) is
>>> connected to which device on which channel? And secondly why when I
>>> try to display values from all channels on both devices I receive
>>> nothing. One of these channels must be connected to the potentiometer
>>> no? The BNC connectors are currently left open.
>>>
>>> Hope I can make this work.
>>>
>>> Regards,
>>> Jayant
>>>
>>> \0x00INFO:<ADC performance test>
>>> INFO:<Opening available ADC channel
>>> s>
>>> Opened 32 ADC channels
>>> INFO:<Preparing ADC channels for test>
>>> INFO:<Starting measurement>
>>>
>>>
>>> /dev/adc00 ? ? ? ? ?\0x09= 0
>>> /dev/adc01 ? ? ? ? ?\0x09= 0
>>> /dev/adc02
>>> ? ? ? ? ?\0x09= 0
>>> /dev/adc03 ? ? ? ? ?\0x09= 0
>>> .
>>> .
>>> .
>>> /dev/adc113 ? ? ? ? \0x09= 0
>>> /dev/adc114 ? ? ? ? \0x09= 0
>>> /dev/adc115 ? ? ? ? \0x09= 0
>>> INFO:<Finished measurement>
>>>
>>>
>>> ----------------------------------------
>>> Samples expected after 10210 milliseconds: 10210
>>> Samples read (per channel):
>>> /dev/adc00 ? ? ? ? ?\0x09= 0
>>> .
>>> .
>>> .
>>>
>>> On Mon, Apr 18, 2011 at 11:47, Christophe Coutand <ccoutand@stmi.com> wrote:
>>>> Hi Jayant,
>>>>
>>>> You are using ecos-3.0 with the additional STM32 ADC package from CVS? It is usually not recommended to mix source code as it make it more difficult to get support from the mailing list.
>>>>
>>>> It is not enough to copy the directory packages\devs\adc\cortexm\stm32 in your repository, you must also include the package entry in the database (i.e. ecos.db):
>>>>
>>>> package CYGPKG_DEVS_ADC_CORTEXM_STM32 {
>>>> ? ?alias ? ? ? ? { "STM32 ADC driver" adc_stm32 }
>>>> ? ?hardware
>>>> ? ?directory ? ? devs/adc/cortexm/stm32
>>>> ? ?script ? ? ? ?adc_stm32.cdl
>>>> ? ?description "
>>>> ? ?This package provides a driver for the ADC interfaces found on the
>>>> ? ?ST STM32 microcontroller family."
>>>> }
>>>>
>>>> To build the STM32 HAL with ADC support from CVS repository, I do:
>>>> $ ecosconfig.exe new stm3210e default
>>>> $ ecosconfig.exe add io_adc
>>>>
>>>> You can configure the ADC from the GUI:
>>>> $ configtool.exe ecos.ecc
>>>>
>>>> Example:
>>>>
>>>> http://dl.dropbox.com/u/21589565/misc/stm32_adc.JPG
>>>>
>>>> Regards,
>>>> Christophe
>>>>
>>>> -----Original Message-----
>>>> From: jayant biswas [mailto:biswasj@gmail.com]
>>>> Sent: 18. april 2011 10:47
>>>> To: John Dallaway; Christophe Coutand
>>>> Cc: ecos-discuss@ecos.sourceware.org
>>>> Subject: Re: stm3210e eval board adc question
>>>>
>>>> Thank you Christophe and John for your replies.
>>>>
>>>> I did checkout the cvs repository and was able to get the right folder
>>>> i.e. packages\devs\adc\cortexm\stm32
>>>>
>>>> However, this hasn't solved my problem. I am still getting the same
>>>> output as before when calling cyg_io_lookup and cyg_io_read.
>>>>
>>>> To integrate the new packages from the cvs. I copied ?the entire
>>>> repository over my ecos-3.0 folder. Then I regenerated the ecos.ecc
>>>> file for stm3210e. I noticed here that the
>>>> CYGPKG_DEVS_ADC_CORTEXM_STM32 checkboxes were grayed out. I then
>>>> recompiled my application with the newly generated libraries.
>>>>
>>>> Please let me know if I am missing some steps. I am sure I am. Is
>>>> there a guide that I can follow that shows how to integrate the ecos
>>>> cvs packages?
>>>>
>>>> Best regards,
>>>> Jayant
>>>>
>>>> On Wed, Apr 13, 2011 at 20:11, John Dallaway <john@dallaway.org.uk> wrote:
>>>>> Hi Jayant
>>>>>
>>>>> jayant biswas wrote:
>>>>>
>>>>>> I installed ecos as indicated here
>>>>>> http://ecos.sourceware.org/getstart.html .
>>>>>
>>>>> ... so you have the eCos 3.0 release repository.
>>>>>
>>>>> The STM32 ADC driver is not present in eCos 3.0. You will have to
>>>>> checkout the eCos CVS repository in order to gain access to
>>>>> CYGPKG_DEVS_ADC_CORTEXM_STM32. Ref:
>>>>>
>>>>> ?http://ecos.sourceware.org/anoncvs.html
>>>>>
>>>>> I hope this helps...
>>>>>
>>>>> John Dallaway
>>>>> eCos maintainer
>>>>> http://www.dallaway.org.uk/john
>>>>>
>>>>
>>>
>>
>

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss