STEVAL-SPIN3202 Support

[CSzat]

How can the initially delivered hardware/software configurations for an "evaluation" product intended to market a suite of motor control product offerings be so poor?

NONE of the current ST software directly supports the "Active" product STEVAL-STEP3202.

All I am trying to do is spin a small 24 V 400-watt BLDC CNC "spindle" motor in open-loop, sensor-less, FOC-mode (plans were to add "hall feedback" later). The evaluation board suggests that with a simple hex binary load supplied in the available product software package, that the eval board should spin up a "generic" BLDC motor using the on-board RESET, SW1, and SW2 buttons for ON/OFF functionality and the on-board potentiometer for speed command.

NONE of the pre-compiled "binary" examples for "UART" and "Potentiometer", which are directly referenced in the STEVAL-SPIN3202 documentation, actually exist in the delivered software folders as described in the product's documentation.

 So much for a quick and initial evaluation BEFORE one needs to figure out this entire motor control+software SDK scenarios just to get a motor to spin.

Neither the Cube SDK, nor the ST "Motor profiler" GUI applications, support this "Active" STEVAL-SPIN3202 board.

 This removes my ability to evaluate the base electrical parameters of the motor before tuning it completely with the SDK.

 That is a huge failing on ST's product marketing team in my opinion.

 This is after all an "EVAL" board. Getting it properly configured should not require an "inside technical contact" at ST.

I have the entire software stack loaded and running (using the Keil IDE) and can configure and download the projects's hex code to the board, but the motor whines for a bit and never actually begins to "rotate" before it soon trips out with an"Over-current" fault.

No matter what I set the over-current/shunt settings to, the small BLDC motor never rotates and always faults on "over-current".

Additionally, how does the motor control "software" know about the position of the three jumpers on the board which determine whether the board itself is in "6-step" vs "Field Oriented Control"? I seem to get different faults from the Motor Control Workbench "Monitor" application depending on the installed positions of these jumpers.

But really the question is: 

**Why is this board not more directly supported in the entire suite of support software offerings?

[Claire O.]

Dear customer,

I am sorry to hear you had bad experience with our products.

STEVAL-SPIN3202 FOC SW can be found in the MC SDK : https://www.st.com/en/embedded-software/x-cube-mcsdk.html

In the MC workbench, please select new project. Then select "inverter" in the "Select Boards " section. STEVAL-SPIN3202 should appear in the list of inverter boards.

STEVAL-SPIN3202 is indeed not supported by the MC Profiler as based on stm32F0 without Floating point unit.

Best regards,

Claire

[CSzat]

Thank you for the reply.

The STEVAL-SPIN3202 "documentation" should be reviewed so as to not reference "pre-compiled" "UART" and "Potentiometer" binaries that do not actually exist as delivered in the product's software download page.

However you did not address the following question:

How does the motor control "software" know about the position of the three jumpers on the board which determine whether the board itself is in "6-step" vs "Field Oriented Control"? I seem to get different faults from the Motor Control Workbench "Monitor" application depending on the installed positions of these jumpers.

[Claire O.]

Dear customer,

Thanks for your remark about STEVAL-SPIN3202 documentation. I will forward the information to the right team.

About the last question, I am afraid the software has no way to check the status of the jumper by itself. You have to do it manually, and to Field oriented control if you want to use the MC SDK. It is normal if you get some error if you are in a 6 step position, as 6 step is not yet implemented.

Best regards,

Claire

[CSzat]

I am still perplexed by what appeared to be such a promising software-based FOC BLDC motor drive "evaluation" board". No matter what values I change ANY of the available FOC software settings to, I cannot get a small 4-pole-pair, 24VDC, 250 watt, 0.110mH, 0.328 ohm-per-winding CNC "spindle" motor to successfully ramp up.

I have spent numerous hours changing the values for start-up ramp values, switching frequency, speed feedback gains/filter frequencies, current-limit values etc., generating each of the new header files from the ST Motor Control Workbench, then Compiling+Linking+ Hex file downloading to the STEVAL-STEP3202 within the Keil IDE....

I never had any luck getting this motor up to successfully speed, even without any load on the motor output shaft. I have tried at least 100 combinations of FOC software/hardware parameters sequentially. I either get an "Overcurrent" (most popular), "FOC duration", or "Startup-failure".

Given the functionality of JTAG, and the inclusion of the ST_V2 debugger chip-set in this EVAL board, it sure would be nice to be able to change the values of each of the most-critical FOC software calculation registers in "real time" from the Motor Control Workbench so as to immediately see the effects of changing each setting to a more appropriate value, without re-generating the new "C" code from scratch each time a single parameter is changed, then having to compile the newly generated code in an IDE, and then re-downloading the resultant binary/HEX file to the evaluation board itself to be able to test the new parameters.

The Motor Control Workbench software sure seems to "imply" that this real-time register change is possible, but it doesn't seem like this actually works.

ST support is nowhere to be found.

In addition, there are numerous documentation errors. Specifically in the STEVAL-STEP3202 "User manual" document [UM2278] , the 3 jumpers which control 6-step versus FOC mode seem to be improperly documented (page 8/20). The schematic on page 17/20 would have you set the jumpers exactly OPPOSITE of what the section 2.2.1 states for selecting FOC/6-step). The FOC jumper positions shown within the "schematic" of the board appear to be correct, as FOC-mode should place the on-board, single-shunt operational amplifier into a single-shunt, differential current-sensing mode for FOC current-feedback sensing. I always receive an "Over-current" error, even though the motor is rated at 11A (RMS), 16A (peak) and the over-current settings (shunt values, amplifier gains that match the factory installed gain resistors of the STEVAL-STEP3202 etc) all seem correct.

I just don't get how this board could be so hard for so many to get a mostly "generic" BLDC motor to even "spin".

Also, as you might see, this STEVAL-STEP3202 board does not have a floating point unit, and therefore it CANNOT be used with the ST "Motor Profiler". I was so looking forward to this board speed-regulating and silencing my small 12KRPM BLDC spindle motor.

I am now looking at similar products from Cypress electronics.

In retrospect, it appears that STEVAL-ESC001V1 might have been a better choice for my small, 250 watt CNC spindle motor. It has the floating point unit, and can therefore be used with the "Motor Profiler" software. It seems to be more marketed for the "drone" industry, but it seems to be geared towards this same "small BLDC" market, and has the full BLDC system-on-a -board including the serial (UART+JTAG) , MPU and MOSFET-bridge power-section with very nice MOSFET specs.

ST drive technical support:

Any ideas?

I truly want to send my actual motor to the folks at ST, to see if they could get this drive to actually make it "SPIN" as the eval board's name implies it is supposed t be able to do.

This same BLDC spindle motor works just fine to beyond 12 KRPM using a $30 Chinese BLDC motor-drive board.

[DPell]

I totally share your frustration.

The documentation is BAD, it does not reflect well enough what found in the firmware package and the behavior of the software suite.

I have wasted two days without being able to compile an executable either on windows or linux.

I was hoping, getting an advanced board from a renowned company, to have a lot of support in experimenting and learning, but now I feel that I would have done better with a random Chinese piece of ...

Sorry for the rant, it's just that I really had high expectations and I really wanted to love this board, but now I'd just like to drive my car over it :(

[GDuon]

Hi guys,

After some frustrated time with the motor control workbench, I have abandoned it totally. On other hand, I have successfully run my Faulhaber 3274 BP4 with the firmware package and some config with the project's defined symbols to make it run with 6-step topology. As said above, the jumpers's positions should be config follow the schematic, not as page 8 stated: J11 is open and J12 is closed.

Hope this help.

[VK Verma]

Dear Claire,

I am evaluating STEVAL-SPIN3202 and used out of the box software "en.stsw-spin3202\STSW-SPIN3202-V1.0.0\stm32_cube\Projects\Multi\Examples\MotionControl\STEVAL-SPIN3202\MDK-ARM\STSPIN32F0". After compiling it with Keil-MDK it generates following error ."\Objects\Project_Potentiometer.sct(7): error: L6236E: No section matches selector - no section to be FIRST/LAST."

Can you please advise quick fix so that I can modify it with our own application requirements.

Also I have tried to build it with Motor control SDK V5.4 for this board and failed to build error free application for this board.

I am quite impressed with STSPIN32F060x product which you have just launched recently. If you can support on it then it will help us to explore new parts.

Thanks in advance !

@Claire OKU​ 

Regards,

VK Verma

[kyrreaa]

Glad and sad to see other have so much issues with this board/system.

It appears some of the documentation is fixed but I am still perplexed that they now ship a version of the code that has all features needed to test startup and log BEMF disabled, and apparantly one runs out of RAM if enabled... None of the python scrips supplied are thus usefull in any way.

I too had to fiddle about to get a decent starup ramp which I find odd the system uses at all.

For 6-step mode it is obvious now to use a initial position detection algorithm and then use integral control mode to handle timing. Anything less, like 0-cross with timer will fail to handle accelleration etc.

We are evaluating the SPIN series to see if we can offload the motor control and also use it's integrated FET drivers to replace external FET driver we are not happy with.

So far frustration over codebase is high.

One tip/wish I have here is for ST to support their actual official CubeMXIDE for these source releases. The project failed completely to import and eventually a convoluted way to install System Workbench for STM32 was followed which gave moderate success albeit no fw loading or debug that works yet.

[alessandro239955_stm1]

Unfortunately I am sharing the same frustration. The documentation is terrible.

I tried to use ST Motor Control Workbench and STMCubeIDE and the code exceeded the MCU memory. First fail.

I was finally able to compile using the ukeil. However I continued to have a Speed Feedback fault. Second fail.

I decided to try the 6steps package. I used ukeil and I got another error ""spin3202" error: L6236E: No section matches selector - no section to be FIRST/LAST". Third fail.

I also had a lot of expectation on this product, but for now it is failing miserably.