Did my LabVIEW VI Code fry my Maple Rev 3 board?

LeafLabs Garden » Support » Maple Hardware Support

Did my LabVIEW VI Code fry my Maple Rev 3 board?

(12 posts) (3 voices)

Started 5 years ago by StephenFromNYC
Latest reply from bnewbold

StephenFromNYC
Moderator

Hello,

I am playing with a Maple board (Rev 3, Summer 2010). A great product.

I am running the Maple-IDE software version 0.0.6 on a 1 GHz Pentium III Thinkpad with 512 Meg RAM running Windows XP SP3.

I am powering the Maple using the USB cable. I have not tried using battery power. I have not moved any of the power "jumpers". Only the "USB" power pins are connected.

I do not know if I fried my Maple board yesterday when I was trying to port code which works on an Arduino Duemilanove board (ATMega328 chip). Currently, the board is not completely dead, because the blue LED flashes when I first plug in the USB cable and when I press the RESET button the Maple responds with several fast LED flashes followed by several slow LED flashes. However, at this point the LED stops flashing.

Unfortunately, the Maple-IDE does not see the board (which was on COM7) which means I cannot attempt to load a simple Blink program. The Maple board is not listed under "Tools >> Serial Port".

I tried another Windows XP SP3 system, but the second system does not recognize the Maple board (eliminating the possibility that the problem is a Maple-IDE or system configuration change). Something on the Maple board itself has been altered.

Is there anyway to "reset" the Maple board back to the factory state?

Here is more background information regarding the project:

I am very interested using Arduino/Maple boards to connect external analog sensors to a computer. I am trying to switch from the Arduino to the Maple, because the Maple has 12-bit ADC (0-4095) vs 10-bit ADC (0-1023) for the Arduino. I am also hoping the faster clock rate of the Maple should permit faster ADC.

Earlier this year I wrote a LabVIEW VI which communicates with an Arduino Duemilanove board. A LabVIEW "VI" is simply a program. Using the LabVIEW VI the user is able to select an ADC channel. The chosen ADC channel is sent to the Arduino board using LabVIEW "serial VISA" commands. Arduino reads the chosen ADC channel and then the converted value is sent back to LabVIEW which graphs the ADC value (and writes the value to a simple data file). Using an Arduino Duemilanove I am able to achieve 20 data points per second.

The LabVIEW VI is very basic. My code works even though it does not contain any handshaking.

The original Arduino code uses Serial.read(), Serial.print(), and Serial.println(). I modified the Maple code to use SerialUSB.read(), SerialUSB.print(), and SerialUSB.println().

At first, I was optimistic about the port. After starting the LabVIEW VI I could get five or ten data points sent from the Maple graphed and saved. However, at this point the LabVIEW VI would stop graphing, although the blue LED would flash indicating the Maple sketch was waiting for a command from the LabVIEW VI.

It did not make a difference if I selected "Board >> Leaflabs Maple R3 to flash" or "Board >> Leaflabs R3 Maple to RAM".

To help with the debugging, which I assumed was a handshaking problem, I minimized the Maple-IDE code. Unfortunately, this is when something happened to the Maple which prevents me from uploading any Maple sketches (such as Blink). Under "Tools >> Serial Port" the Maple is not listed.

Any ideas for how I can communicate with the Maple?

I am very new to microcontrollers.

Thanks for your help.

Stephen from NYC

Posted 5 years ago #
StephenFromNYC
Moderator

Hello again,

I found a partial solution to the problem I posted earlier today.

By placing the Maple into "perpetual boot mode" I managed to get control of the board (Rev 3; summer 2010). For more details:

http://forums.leaflabs.com/topic.php?id=76#post-407

I can upload upload the Blink program to the RAM, but I cannot get it to load into the FLASH memory.

For the moment I am not too worried about the FLASH memory, because I now have access to the board.

Does anyone have any experience with LabVIEW-Maple serial communication?

Thanks for your help.

Stephen from NYC

Posted 5 years ago #
bnewbold
Ex-Leaflabs

Hi Stephen!

The root of this problem is most commonly something in your usercode which is mucking up the SerialUSB interface, which then prevents the IDE from auto-resetting the board and uploading new code.

If the status LED blinks at start that means the bootloader is intact and the fix is pretty easy: enter "perpetual bootloader" mode by holding the BUT button while tapping the RESET button. The status LED should blink continuously, after which you can release BUT and try to upload a new program. The IDE will initially complain that it can't find the Maple serial device (COM port), but the IDE will then search for the Maple DFU device and upload to that. Getting in to perpetual bootloader mode might take one or two tries but the continual blinking will let you know when you're there.

If that doesn't work for you for whatever reason there are some other things to try (using a seperate serial device) but I bet this is the problem.

This happens to me a lot when I try to use a serial port (Serial2.println()) which does not have a begin declaration (Serial2.begin(9600) or similar) in the setup() function; perhaps you have one floating somewhere in your code?

Any other "hardfault" (like a divide by zero) could also cause a hard lock, though usually this will cause the status LED to fade in and out. If you post your code I can scan through it or run it in debug mode and see where it's hanging up.

FWIW, what you're trying to do sounds pretty cool! It's amazing how much NI charges for data collection boards and this is a project that i'm definitely in. I would love to get some sort of isochronous USB channel (which has guaranteed timing and bandwidth, unlike the CDC/ACM-based SerialUSB) to do data collection and drive PWM outputs using mathematica, labview, scilab, etc.

Messing with the USB stack will be a longer term project though, it's cool you're getting something up and running! These web services may be interest as well:
http://cloud.labjack.com/
http://www.pachube.com/

Posted 5 years ago #
StephenFromNYC
Moderator

Friday, August 20, 2010

Hello,

Thanks for the help, bnewbold. I bookmarked the URLs to the two data acquisition sites.

Please let me know if I should break my comments into different threads. I believe my comments are straying from the initial "did I fry by board" focus.

Is "perpetual boot mode" (see my previous comment) the same as the "perpetual bootloader mode" which you described?

To load sketches into my Rev 3 Maple board I have been using the procedure posted on the forums by keymaster "poslathian":

http://forums.leaflabs.com/topic.php?id=76#post-407

Poslathian says "To enter perpetual bootloader mode, hit reset, and immediately following press and hold the boot button (labeled BUT). If BUT is held when the fast blinks transition to the slow blinks, then youre in perp. boot mode".

I press the RESET button followed by pressing and holding the BUT button. This usually results in a slow "continual blinking" pattern which allows me to then upload sketches.

Unfortunately, the behavior of the board is unpredictable. Occasionally, I am able to upload sketches without using the RESET/BUT trick (only when the board is correctly listed under "Tools >> Serial Port"), but usually I have to resort to using the "perpetual" mode which is inconvenient.

Is there a more permanent long term fix for this problem? In the Maple documentation did I overlook a description of different ways to resolve a Maple trapped in a hard lock?

bnewbold, you mentioned a separate serial device which might be helpful. Can you provide me a link with instructions? If I can get my board into the "perpetual bootloader mode" is the serial device you mentioned still useful?

How can Maple-IDE code be tested in "debug mode"?

When I first tried to port my Arduino code for use on the Maple I converted all the Serial.print() and Serial.println() instructions SerialUSB.print() and SerialUSB.println() instructions. My code does not have any "serial" instruction, only "serialusb". However, I will double check my code next week.

Yes, it would be great for users in many different fields if the Maple can be used as a DAQ board. The Maple (at $50) is more expensive than the $30 Arduino, but the extra $20 is worth it for users who need built-in 12-bit ADC (without soldering an external 12-bit ADC circuit). Initially, I decided to focus on data input (using ADC) because it is easier (ie, more "passive"). Bnewbold, you have anticipated my future need to use PWM outputs to turn on/off lights and other environmental controls.

My background is not in engineering, so I am completely new to implementing "isochronous USB channel". Can you recommend any resources which I can read? I can be more helpful to this project if someone points me in the right direction.

Thanks again for the help. Have a restful weekend.

Stephen from NYC

Posted 5 years ago #
poslathian
LeafLabs

Stephen,
You mentioned that if you select the correct serial port from the tools menu, then the auto-reset function during programming works as expected? This should be the case... Perhaps after programming, for some reason, your selection of the board in the tools menu is getting lost, forcing you to reselect it. I agree this is annoying, Ill file a ticket for the next revision to make the selection more persistent.
When the board resets for programming, the serial port does in fact "die," probably the reason your selection setting are being lost. Ill dig in to the ide code for a fix.

Posted 5 years ago #
StephenFromNYC
Moderator

Sunday, August 22, 2010

Poslathian,

Thanks for offering to look into the IDE code, but first I wanted to clarify my problem.

I have to use the "perpetual bootloader mode" when the correct serial port is MISSING from the tools menu. It is not simply a matter of the correct serial port being unselected. If the correct serial port is listed I use it, but if it is not shown I need to use the "perpetual bootloader mode".

I do not know if the following observation is related.

Before I thought I had fried my Maple board the most reliable way to upload a sketch was to first manually compile the code (using the IDE) before hitting the IDE upload button. If I am correct, with the ARDUINO the IDE (version 017; I never upgraded to Arduino IDE version 018) recognizes when the code has been altered and will compile the code when just the IDE upload.

If I did not issue a separate compile command there were times when I was not sure if the altered code was compiled or if the unaltered version was uploaded.

I will send more observations later this week.

Thanks for your help.

Stephen from NYC

Posted 5 years ago #
StephenFromNYC
Moderator

Hello,

In my previous post I should have written, to be more clear "If I did not issue a separate compile command WITH THE MAPLE there were times when I was not sure if the altered code was compiled or if the unaltered version was uploaded."

I am still having "hardware" problems with my Maple.

When my Maple first arrived, compiling/uploading/modifying simple sketches (such as "Blinki") for my Maple was trivial. Programming the Maple seemed as reliable and predictable as programming an Arduino Duemilanove board.

My current problems started after I started to port code I wrote which permits communication between an Arduino Duemilanove board and LabVIEW, a very robust engineering programming environment by National Instruments which uses GUI icons instead of text.

The original code also works on a "Seeduino" clone.

Now, I am reluctant to invest time into debugging my LabVIEW-Maple code until I am sure my Maple is behaving correctly.

Here is a quick step-by-step description of my work this afternoon:

1) turn on Pentium III computer running Windows XP Professional SP3 (512 meg RAM)
system running with "administrator" privileges
start Maple IDE (0.0.6)
under "Tools >> Serial Port" see only see the system COM1 port (which does not surprise me since the Maple board is not connecte)

2) plug in Maple (Rev 3, summer 2010) into USB 1.1 (not 2.0) port
hear two identical sets of tones (each set has two "notes", but I forgot if the pitch of each set is high/low or low/high)
under "Tools >> Serial Port" see both COM1 and COM7

3) "Tools >> Board >> LeafLabs Maple R3 to Flash" is preselected

4) open "Examples >> Digital >> Blink" sketch

5) sketch compiles, using 12296 bytes

6) sketch uploads and LED blinks correctly (on 1 sec, off 1 sec, etc)

7) modify code, increasing each delay from 1000 to 2000 (msec).

8) try uploading (without first recompiling changes)

9) error message:

last few lines of status message:

<start of message>
Opening USB Device 0x1eaf:0x0003...
Found Runtime: [0x1eaf:0x0003] devnum=2, cfg=0, intf=0, alt=1, name="DFU Program FLASH 0x08005000"
Setting Configuration 1...
Claiming USB DFU Interface...
Setting Alternate Setting ...
Determining device status: state = dfuIDLE, status = 0
dfuIDLE, continuing
Transfer Size = 0x0400
bytes_per_hash=245
Error during download get_status
Starting download: [
<end of message>

10) board status:

blue LED is not on
"Tools >> Serial Port" shows only system COM1 port

11) Conclusion: Maple cannot be selected because it is not available

12) Enter "Perpetual bootloader mode" as follows:

Press RESET button.
During rapid flashes, press and hold the BUT button.
After several seconds, when the slow LED flashes are continuous for several seconds, release BUT button.

The slow LED flashes continue.

13) board status:

"Tools >> Serial Port" still shows only system COM1 port

14) upload (without first recompiling changes)

15) upload not successful

16) board status:

blue LED is not on
"Tools >> Serial Port" shows only system COM1 port

17) Conclusion: Maple cannot be selected because it is not available

18) Enter "Perpetual bootloader mode" as described above (set 12)

19) sketch does not compile

20) error message:

last few lines of status message (same as before; step 9):

<start of message>
Opening USB Device 0x1eaf:0x0003...
Found Runtime: [0x1eaf:0x0003] devnum=2, cfg=0, intf=0, alt=1, name="DFU Program FLASH 0x08005000"
Setting Configuration 1...
Claiming USB DFU Interface...
Setting Alternate Setting ...
Determining device status: state = dfuIDLE, status = 0
dfuIDLE, continuing
Transfer Size = 0x0400
bytes_per_hash=245
Error during download get_status
Starting download: [
<end of message>

21) board status:

blue LED is not on
"Tools >> Serial Port" shows only system COM1 port

22) sketch uploads and LED blinks correctly (on 2 sec, off 2 sec, etc)

23) under "Tools >> Serial Port" see both COM1 and COM7

24) modify code, decreasing each delay from 2000 to 1000 (msec).

25) sketch compiles

26) sketch does not upload

27) error message:

last few lines of status message (same as before; step 9):

<start of message>
Opening USB Device 0x1eaf:0x0003...
Found Runtime: [0x1eaf:0x0003] devnum=2, cfg=0, intf=0, alt=1, name="DFU Program FLASH 0x08005000"
Setting Configuration 1...
Claiming USB DFU Interface...
Setting Alternate Setting ...
Determining device status: state = dfuIDLE, status = 0
dfuIDLE, continuing
Transfer Size = 0x0400
bytes_per_hash=245
Error during download get_status
Starting download: [
<end of message>

28) unplug board from USB 1.1 port
hear one set of tones (high/low)

29) plug board back into USB 1.1 port
hear two set of tones (low/high)

30) under "Tools >> Serial Port" see both COM1 and COM7
LED is blinking (on 2 sec, off 2 sec, etc) so changes made between steps 24 and 26 did not reach Maple

Conclusion:

The COM7 port is often lost from the list, not simply deselected.
Even when the board appears as COM7 under "Tools >> Serial Port" I cannot upload sketches.
I can upload only using the "perpetual bootloader mode".

Any ideas?

Today was the first day I speculated that the problem might be related to USB 1.1/2.0 issues.

Thanks for your help.

Posted 5 years ago #
poslathian
LeafLabs

Thanks for the detailed info! This gives us a lot to work with. First some info dumpage:
http://leaflabs.com/docs/maple-bootloader/ The take home message there is that the DFU mode (when maple accepts new code during programming) and the serial port mode (the COM port) do not run concurrently. When you reset maple, the serial port SHOULD be lost, and the DFU mode should come up. Then after programming, the board should auto reset, sit in dfu mode for about 3 seconds, then exit to "user mode" (where it runs your code) and bring the serial port back up. Thus, if the board is blinking away running YOUR code, you should see the serial port. If its doing its bootloader thing (such as after reset for a few seconds) you should not see the serial port. On some OS's weve gotten reports that they have to re-select the serial port every time. This is annoying. I agree, were going to patch in some sort of fix for the next IDE release. For the auto-reset feature to work (one-click programming, without infinite bootloader mode) the Maple must be running in user mode (running your blinky or whatever) and the IDE must have the serial port for Maple selected (in your case COM7). You should not need to run with admin privileges.

On to your next problem, seen in steps 27, 20, and 9. This is concerning, and is not the expected behavior. In your examples, you are having failures when uploading to flash. Does it fail for ram too? This very well may be some weirdness with USB. When the upload fails like this (which is strange, ill try and bring up an SP3 machine and recreate the problem) its possible the Maple is getting stuck in bootloader mode, and thus causing your program not to run and thus the serial port not to come up. Its also possible that its not really failing, but just saying it did, the message reporting in the IDE is a bit buggy, sometimes I dont get the whole terminal dump, and it stops in the middle, even though the programming operation continued. You can tell that the programming operation completed if the board resets (by itself) after programming.

Posted 5 years ago #
StephenFromNYC
Moderator

August 24, 2010

Hello,

1)
Yes, I was aware the serial port functions and uploading steps are separate. Is there anything in my description (which I hope helps other users with similar problems) which suggests I was expecting to see both COM7 (how my Maple board connects) and the uploading at the same time (working concurrently)?

2)
I just succeeded in using the upload button in the IDE to upload the Blink sketch to "Tools >> Board >> LeafLabs R3 Maple to RAM". The sketch functioned correctly and I could still see "Tools >> Serial Port >> COM7".

Then I modified the sketch by increasing the two delay statements to 2000 (ms). The modified sketch uploaded via the IDE button and worked correctly (without requiring that I first click the compile button). I could still see COM7. Good news so far.

However, when I tried switching to "Tools >> Board >> LeafLabs Maple R3 to FLASH" the uploading crashed with the same message described in my previous post (steps 27, 20, and 9).

The LED is not flashing, so the sketch is not running. The problem is not a buggy message system.

Is there a problem with my FLASH memory?

3)
This is what I did next:

Select "Tools >> Board >> LeafLabs R3 Maple to RAM".
Place board into "perpetual bootloader mode".
Press the IDE upload button (sketch still using two 1000 ms delays).
Blink sketch functions correctly and I see COM7!
I increased the two sketch delays to 2000 ms, the modified sketch uploads and run, and I still see COM7.
Partial success.

4)
Conclusion and questions:

Using the Blink sketch the board is behaving if I use Maple RAM, but not if I use Maple FLASH memory.

When my Maple first arrived, I believe I successfully uploaded the Blink sketch to both the FLASH and RAM. It has been a few weeks, so I cannot be totally sure if the FLASH worked before I started to port my Arduino/LabVIEW code to the Maple.

How can I test the Maple FLASH memory?

Can an external application (such as my LabVIEW code) which contacts the Maple (using COM7), causing the Maple to perform an analog read, after which the Maple returns the ADC value to LabVIEW using SerialUSB.print() and SerialUSB.println() affect STM32 FLASH memory?

I am new to using microcontrollers and I am looking forward to learning how to use the Atmel and STM32 spec sheets to learn how to program the Arduino/Maple boards.

However, I am reluctant to put a lot of time into porting my working Arduino/LabVIEW projects to the Maple until I am confident that my program bugs are caused by my mistakes and not the Maple software or a problem

Thanks!

Stephen from NYC

Posted 5 years ago #
StephenFromNYC
Moderator

Hello,

The partial success I indicated in my previous post was short lived.

I turned off my computer after I sent my last message.

I just booted my system, started the Maple IDE, and inserted the Maple Rev 3 board, expecting to see a blinking LED and COM7 under "Tools >> Serial Port".

However, COM7 is not listed and the LED is not blinking.

I have never used PWM to vary the intensity of the LED, but my current observations are what I expect to see with using PWM to controll LEDs. The LED intensity first gradually increases in intensity before gradually decreasing in intensity. Then the LED goes off before repeating the cycle.

I put the board into perpetual bootloader mode. I uploaded the Blink sketch which worked as expected. Now COM7 is listed (and selected). After modifying the sketch to use 2000 ms delays and uploading the LED blinks on/off with a 2 sec intervals. COM7 is still listed (and still selected).

The solutions to my Maple problems do not seem to be reliable from day to day or after rebooting. I feel frustrated.

Thanks!

Stephen from NYC

Posted 5 years ago #
StephenFromNYC
Moderator

August 30, 2010

Hello-

I am still having unpredictable results uploading sketches to FLASH memory.

To trouble-shoot I am using the standard Blink sketch (and alternating between 1 and 3 second blink delays to make it easy for me to determine if sequences of upload succeed or fail).

For the last few days I have been unable to upload any sketches to FLASH. This did not matter to much since I was able to upload to RAM.

Today, I placed the Maple into perpetual bootloader mode and I was successful at uploading the Blink sketch to FLASH.

After editing the blinking interval I tried upload via the IDE button.

I got the "No such file or directory" error message (see end of message). The directory exists, but the files do not.

In the attached error message I showed the full line for the adc.c.o file, but to save space I edited the lines for the other files.

Thanks!

Stephen from NYC

Maple board: Rev 3 (summer 2010)
Maple IDE: 0.6.0

Processor: Pentium III (1 GHz)
USB ports: 1.1

OS: Windows XP Professional (32-bit)
SP: Service Pack 3

Java type: JDK and JRE
Java version: 1.6.0_20 (yes, I know that update 21 is available)

<start of diagnostics>

Going to build using 'make' (ARM)
Compiling core...
Compiling any libs...
Compiling the sketch...
running linker asynchronously...

arm-none-eabi-gcc: C:\DOCUME~1\%username%\LOCALS~1\Temp\build8983615117869570609.tmp\adc.c.o: No such file or directory

descriptors.c.o: No such file or directory
exc.c.o: No such file or directory
exti.c.o: No such file or directory
ext_interrupts.c.o: No such file or directory
flash.c.o: No such file or directory
gpio.c.o: No such file or directory
nvic.c.o: No such file or directory
pwm.c.o: No such file or directory
rcc.c.o: No such file or directory
spi.c.o: No such file or directory
syscalls.c.o: No such file or directory
systick.c.o: No such file or directory
time.c.o: No such file or directory
timers.c.o: No such file or directory
usart.c.o: No such file or directory
usb.c.o: No such file or directory
usb_callbacks.c.o: No such file or directory
usb_core.c.o: No such file or directory
usb_hardware.c.o: No such file or directory
usb_init.c.o: No such file or directory
usb_int.c.o: No such file or directory

<end of diagnostics>

Posted 5 years ago #
bnewbold
Ex-Leaflabs

First, sorry not to have replied to this thread for so long!

I have seen the "No such file or directory" problem once or twice on various platforms but have trouble reproducing it; it's noted in the Troubleshooting page: http://leaflabs.com/docs/maple/troubleshooting/#compiler

I think the root of this problem is a bug with temporary directory creation code added to the Arduino IDE recently (but before we forked off). It is probably not related to your uploading woes.

Your "throbbing led" problems mentioned is our "assertion failed" indicator and probably means that a HardFault has occurred. This is distinct from an infinite loop or looping interrupt problem and could be caused by a divide by zero, bad pointer, etc.

The DFU upload problem (where it fails at "Starting download: [") is almost certainly our fault in some way. Getting all the various USB modes to work in harmony cross platform was a very delicate balancing act and there may remain issues despite long hours of testing on many machines. We got a junker laptop that we will set up XP SP3 on in in the next week or two and can rule out that type of problem then. The possibility that your FLASH has become corrupted is intriguing. I think bootloader corruption or issues during FLASH writes describe your problems very well, but we don't have any code to check for that yet.

We're sorry that you've had so much trouble reliably uploading programs and that it has taken us so long to debug them. We can ship you a new, carefully tested board to try if you agree to send your current one back eventually so we can figure out what's wrong with it. You could also wait another week or two until we have an SP3 system up and running to test with. Lastly, we will be traveling to NYC the weekend of Sept 25th (for the OSHW summit and Maker Faire) and would be happy to meet with you in person to debug any issues you're having then, though hopefully you'll be up and running by then!

Posted 5 years ago #

RSS feed for this topic

Reply

You must log in to post.