Field Oriented Control development boards

As far as I can tell there is no separate hardware JTAG port on these, just the USB on the ControlCard. So if the licensing works with this USB then we are set. Since we don't plan to build our own ControlCard and just buy them from Ti the IDE should continue to be adequate. We just replace the power board and plug in the Ti ControlCard.

If there is a hardware JTAG connection it would not be used since the kit doesn't include a JTAG interface. It might be optional for improved debugging capability.

The real question might be what are the limitations of the ControlCard USB interface? Is it able to do all the programming and debugging that we would need? They mention serial and debugging capability, but no details are provided.

I'm aware of JTAG, but have never used it. I've always been able to do without it, generally on small processors it is not required. It often offers better debugging features but there are other ways to debug.
 
The Piccolo ControlCard mentions JTAG Emulation

http://www.ti.com/tool/tmdxcncd28069

The Stellaris ControlCard hits a home run:

http://www.ti.com/tool/MDL-LM4F211CNCD

The LM4F211 controlCARD module kit contains:

MDL-LM4F211CNCD controlCARD module
Stellaris 32-bit ARM® Cortex™-M4F LM4F211-based microcontroller controlCARD DIMM form-factor module
Onboard Stellaris In-Circuit Debug Interface (ICDI)
Cables/Accessories
USB Micro-B to USB-A plug cable (for debug and serial communication)
1/2-inch blue jumper wires (for bridging power)
DK-LM4F-DRV8312 kit DVD with:
Complete documentation
Applicable source code
Schematics
Texas Instruments’ Code Composer Studio™ Integrated Development Environment (IDE)v5
Click to expand...

So at this moment it appears to me that the Piccolo may be a slightly more capable machine, but the Stellaris is taking the lead for software development tools and pricing!

Does Stellaris leap ahead and take the win?
 
Looks like with that you will just need a header on a board with the links to your throttles and powerstage outputs etc. then a powerstage and your set!
 
I am away out of my element posting here, but I have build several different JTAG interfaces which allowed me to change the software in the CPU's of SAT receivers to make one brand work exactly like another brand, thus allowing me to use decoding software that was available. This was a simple cable with 100 ohm resistors on each leg of a 9 pin serial cable. I still have much of the hardware, programmers and software if it is in any way useful? Is this serial stuff too slow? USB had too much trouble with spurious noise pickup for what we were doing.
pelle242 said:
The way the TI eval kits usually work is that it comes with a limited version of their eclips based Code Composer Studio IDE. The limitation is typicaly that it only connects to the USB JTAG interface that is on the evm. So if you just want to play around with the evm you are good. However if you want to build your own board you need to hook up an external JTAG interface and to use that you need the full blown version of the IDE.
Click to expand...
 
Hi Gordo,

Typically these JTAG interfaces are more complicated than that, but it is hard to know for certain.

I'm a bit torn between the DSP which may have slightly better performance but has a more closed architecture, tools and debugging, and the ARM that has more open tools and perhaps a little less performance in some respects.

At the end of the day we need "good enough" performance, and I suspect that either will do. If that is the case the ARM is probably the better choice, and it has built in USB debugging from what I can tell. The price is also a bit lower and the availability of free and supplied tools may open up this to a lot more people being able to work on it which could be a huge benefit for us here on ES.

I believe it is also the case that if we develop a power board for this ControlCard interface that we could support BOTH types of brain boards, in case some folks wanted to use a different one.
 
I found a TI video that describes all the product lines. It confirms what we already learned about how the products are positioned. It was a really poor video of someone just reading the slides, and I don't have the link handy, but it did make clear that the C2000 and Hercules are the flagship products with the highest clock rates, performance and features, up to over 200 Mhz and nearly 400 MFlops. It also made very clear that the 80 Mhz Stellaris ARM is a serious high end controller product that is well supported by software tools.

The nice thing is that all ControlCards are fundamentally compatible so if we design an ebike power board for a ControlCard then we can use any of them interchangeably. There may be small differences in I/O that bear further investigation for a fully compatible design.

So for top of the line the C2000 is probably it, or the safety or automotive versions for best certifiably safe control. For an open ebike project these are perhaps not the best choice.

I think I've come full circle on this due to the tool support and am again leaning toward the Stellaris ARM product. It may not be the absolute best in terms of raw features and performance, but it should be more than adequate and may be best in open tools and open support which is very important for this project. The availability of off the shelf ControlCard brain boards and development kits really makes this approach appealing.

I have not pulled the trigger on ordering one yet, but it might make a nice Christmas present. :)

Anyone else going to try this? If someone wants to try the C2000 we can still compare notes and collaborate on power board design. Sourcecode should be fairly compatible too.

Link to kit order page:

http://www.ti.com/tool/dk-lm4f-drv8312

Contents of the $299 Kit:

The DK-LM4F-DRV8312 is a bundle of the following components:

MDL-LM4F211CNCD controlCARD module
Stellaris 32-bit ARM® Cortex™-M4F LM4F211 microcontroller
80-MHz floating point processor core
256 K bytes Flash
32 K bytes SRAM
DRV8312 baseboard (TI-integrated 3P motor driver board with InstaSPIN™-BLDC and Sensorless Sliding Mode Observer FOC):
Supports sub-50 V and 6.5-A peak brushless motors
50-V, 3.5-A inverter drive board
24-V NEMA17 BLDC/PMSM motor
24-V, 2.5-A power supply
USB Micro-B to USB-A plug cable (for debug and serial communication)
Kit DVD:
Complete documentation
Texas Instruments’ Code Composer Studio™ Integrated Development Environment (IDE)v5
InstaSPIN™-BLDC and sensorless FOC software projects
Schematics
 
Alan said:
Anyone else going to try this? If someone wants to try the C2000 we can still compare notes and collaborate on power board design. Sourcecode should be fairly compatible too
Click to expand...

Im really interested in this. But I'm thinking just getting a control card and building the rest. There is no point in paying for a powerstage that I will never use.
I would say maybe in the next few months. I would like to see how this works for >10kw I would like to learn more about the TI stuff they kinda seem to leave some info out of it. What I meant is I have watched all the videos and with insta spin does this mean all the code is done??
I am not running away from my projects with my code and with lebowski's code and microchip. I just want to learn all I can as fast as I can.
 
I am a bit interested in the Piccollo 2000 series, as it must be a jump above the dsPIC stuff that I already know and have the development tools working. I don't want to have to drop a kilobuck however on the TI Code Composer Studio to get the compiler for the Piccollo however...

The devil is in the details on what you get, and what you still need.

One thing I was thinking about was that these cards use a memory card connector. That type of interface is good for shock and thermal, it is not so good if moisture or condensing conditions are present. This should be kept in mind if the intent is to "field" a controller with these cards. It will work in the lab, but what about out in the field in condensing conditions... just another engineering trade to consider.
 
There are a lot of questions that the kit will answer, like how easy is it to extend their code. For many applications the Insta-spin or FOC code is probably enough, just set the parameters and done, but for ebikes I'm sure that's not enough for what I want.

One advantage of getting their kit is to have a quick start initially and later a reference testbed. For example later on if there is a question about whether the ControlCard is still working correctly or not, plug it back into the development board and test it. I think this testbed is sufficient to operate an unloaded hubmotor, so testing the various software tools and making a version that estimates initial and low velocity position from hall sensors could be done with the testbed without having to build a power stage and attack multiple problems at the same time. Then when building the powerstage the software is fairly well sorted out and one can focus on the hardware issues.

The question is what software comes with the $49 ControlCard, that is important. I don't see the insta-spin mentioned with that package, so perhaps that is not included there. They mention the Code Composer Studio. So perhaps that is a lower level approach. The $299 Kit includes both CCStudio and the InstaSpin and Sensorless FOC projects.

So it appears that if you want to slug through the low level approach and deal with hardware development right up front the $49 board is a good start, but if you want the quick-start projects and hardware the $299 kit is the deal. Either approach can work.
 
Good points on the memory interconnect. Should be okay inside a sealed controller if internal condensation is managed. It is also possible to solder to those edge mount connectors if it was necessary, however there goes the plug-replacement feature.

OK on the Piccolo. Looks like a top notch choice, with some remaining questions about tool licensing. One place I saw that the boards come with software locked to the board, so as long as you are using ControlCards it might be okay software wise. When you want a new compiler you might have to buy a new Control Card but that is probably acceptable and a lot cheaper than a new full up software license. Plus if these boards work out well it would not be hard to raise funds to pay for the software.

We can still share power boards and perhaps source code at some level.

Looking forward to your results,
 
Alan, the subtlety to be aware of is Code Composer Studio is like MPLab... there is a GUI and "plug in" compilers.

They may give the GUI away free, but sell the compilers. I can't fully decipher their marketing. Could be sloppy prose, as "comes with Code Composer GUI..." I don't know if the full C compiler for that class of chip is included or not.

If you talk to a sales rep, that would be a question to ask.
 
bigmoose said:
One thing I was thinking about was that these cards use a memory card connector. That type of interface is good for shock and thermal, it is not so good if moisture or condensing conditions are present. This should be kept in mind if the intent is to "field" a controller with these cards. It will work in the lab, but what about out in the field in condensing conditions... just another engineering trade to consider.
Click to expand...
What if we got it all working then once it was not going to be removed we sprayed the connector with conformal coating while it was in place? Inside a controller will be warm during use anyway and we will want to keep it drag as possible.
 
bigmoose said:
I am a bit interested in the Piccollo 2000 series, as it must be a jump above the dsPIC stuff that I already know and have the development tools working. I don't want to have to drop a kilobuck however on the TI Code Composer Studio to get the compiler for the Piccollo however...

The devil is in the details on what you get, and what you still need.
...
Click to expand...

You raise a good question. How does the performance of the Stellaris compare to the DsPIC?

Processing power - resolution and rate;

ADC sample timing control, resolution, accuracy and rate;

PWM resolution and rate;

Toolset quality;

Example quality and coverage;

Documentation;

C/C++ implementation quality;

Hardware Modularity (availability of commercial brain boards at reasonable pricing);

Other important items?
 
Alan, I didn't build the spread sheet table to compare the two directly. That will be on the list of things to do.

My impression skim reading is that TI had the nifty "brain boards" and I don't see those from Microchip. The TI Piccollo variants are faster, but I haven't screened the new dsPIC's. The TI demo firmware loads are more sophisticated. From when I used Code Composer 8 or 10 years ago, the compiler and the GUI are a wash. The TI ADC is faster and has more trigger modes.

As I age I just find the learning curve getting steeper and steeper for a totally new micro/dsp/software composer suite/etc. I have to factor that in with a significant weight when making design decisions.
 
I was looking into the PIC32 processors which are based on MIPS which was a strong technology at one point. But as far as I could tell MicroChip doesn't put motor controllers into their 32 bit processors?

The learning curve is less of an issue withe "C" than it was with assembly language, but there is still stuff to learn from the IDE to the dialect of "C".
 
Looks like Infineon is doing a kit as well with FOC cababilities, http://www.infineon.com/cms/en/product/applications/Motor_Control_Drives/Motor_Drives_Application_Kits/atv_bldc_motor_drive_kit_12V.html

Video http://www.infineon.com/cms/en/product/applications/Motor_Control_Drives/BLCD_CD.html
 
For those of you interested in a good C++ compiler and microcontroller for motor control usage: AVR 32's with Avrstudio 6 is my reccomendation. Check out the application notes Atmel has released.
 
Alan B said:
Have you used AVR32's in motor control?

I have one AVR32 board but it is not set up for motors.
Click to expand...

Yes. Atmel have some nice development kits - even though some are a bit pricy. I wrote them a nice email and got mine for free ;)
 
So I emailed TI and this is there responce.

Hello Arlin,

Thank you for contacting TI Applications Support. I sincerely apologize for the delay in my correspondence.



Based on the information that you have provided, your application may be a bit outside of what our kits can handle.



The largest kit that we have is the TMDSHVMTRPFCKIT which, at best, is rated for a power output of 1.5KW. From the information that you have provided, your solution will need to be able to deliver, potentially, several KW of power which is far beyond the capabilities of the kits that we have to offer.



The kits that we offer are designed to fit the most common applications from a “baseline” perspective in that they are created to serve as starting points for what is considered the most popular of applications. For these kits we offer example code that works with motors that we provide but this code should not be considered “production ready” code as nearly all examples contain functionality for debug purposes which can introduce operational overhead and is not typically included in production level code.



Essentially, these kits are designed for customers to evaluate a particular device line such as the TMS320F28xx device series for a particular application type (PMSM, BLDC, BDC motor control) and use the kits to fine-tune the hardware design for their own end-equipment design and to assist in familiarizing the developer with the code development aspects of motor control using closed-loop control principles.



If you have any questions or concerns, please feel free to contact us.



Regards,


Michael Stevens
TI Applications Support
Americas Customer Support Center
512-434-1560
Click to expand...

I sent them another one to see what the compiler cost is and about code examples and to see about brain board examples.
PS
Don't worry Lebowski I'm no jumping ship just curious! :)
 
Alan B said:
Thanks for sharing your feedback from TI. Not surprising.

No one implied that their kits would be enough to run an ebike motor under load. I think it would run one unloaded, but a power board would be needed for full power. However their brain boards could be used with a new power board on an ebike.
Click to expand...
I think one of the kits was 60v and some decent amps and actualy quite good for an ebike. I was actualy asking about higher power levels or partail kits.
So lets see what they say about the code and compiler.
 
I just found a piccolo F2805x c2000 experimenters kit it looks like it might work for a brain board solution. http://www.ti.com/tool/tmdxdock28055
$115 not bad for someone who wants a partial diy start at things.
 
You must log in or register to reply here.

Similar threads

Burtie
Field Oriented Control using dsPic33
3 4 5 6 7
Replies
173
Views
42,248
jogi
J
Back
Top