C2000™ Microcontrollers

Forum Post: RE: HOW TO RUN TMS320f28069 without debug mode

Michele Fiorino — Wed, 22 May 2013 10:58:00 GMT

Dear David, your indications are very precious.

Today our DSP is booting from flash.

Now there is a little problem:

the SPLL and the sgen3D work very good in debug mode, but don't work in flash mode.

All the other part of code (PWM, ADC, SPI interface, CAn interface, ECAP etc.) work correctly.

Why?

Best reguards

Forum Post: RE: [ Config and Start CPU-Timer of DSP TMS320F28335 ]

Igor Gorbachev — Wed, 22 May 2013 09:25:00 GMT

Hello!

Sorry! So far I have not found somehing criminal in your code. But If I'd been in your shoes I would prefer to begin the development project from the very first on base http://e2e.ti.com/cfs-file.ashx/__key/communityserver-discussions-components-files/171/3326.Example_5F00_2833xCpuTimer.c. First, I would run an empty interrupt handler timer0 for 200 us and I would make sure that it works properly. And then I would be included to the your project the rest functions gradually.

Regards,

Igor

Forum Post: RE: How to use DSP/BIOS for PLC PRIME PHY layer

Igor Gorbachev — Wed, 22 May 2013 08:53:00 GMT

Hello Bin!

You read "I have to call HAL_cpuTint0Func somewhere. In PRIME PHY example, it is called in cputimer0_isr...."

So...Not so easy as it seemed at first. Indeed, I also still do not understand where to call this function HAL_cpuTint0Func...We need to think...

Regards,

Igor

Forum Post: RE: New dsp processor

Igor Gorbachev — Wed, 22 May 2013 08:08:00 GMT

Hello!

I didn't work with this board, but I would prefer to set thus:

Regards,

Igor

Forum Post: Initial rotor position detection (rotor alingment) in PMSM motor, using Hall sensors

Grigory — Wed, 22 May 2013 07:49:00 GMT

Dear All,

I`ve implemented closed loop servo design based on HVPM_sensored project. Last thing to do is to startup motor without any excess rotation looking for Index signal (to align flux like in axamples) or bumps caused by flux alingment (like in my project, I use algorithm Iq=0/Id=1 at start up to align rotor and stator flux). I need motor start to rotate immediately from stand steel position to desired position according command - like any industrial servodrive do. So I need to detect initial rotor position.

As I know initial rotor positon can be detected by:

1) Spesial algorithm with test signals. We apply test signals to phases and calculate rotor position based on responces. This takes few seconds and complicated to implement. Much literature and patents in Web. I think this is not for me, now. May be later.

2) Using Hall sensors. Servomotor that I use and plan to use has Hall sensors + incremental encoder. Moreover industrial servodrivers I examined use this tecnique to start up.
in this doc on page 4 you can find idea how to implement this tecnique - http://www.ti.com/sc/docs/general/dsp/fest99/digital_control/1xufuzhen-osu.pdf

I tried to do but did not succeeded.
As I undestand I must read Hall states and change CalibratedAngle, like this

if (HALL_U&&!HALL_V&&HALL_W) qep1.CalibratedAngle = qep1.LineEncoder*qep1.PolePairs/(2*6)+0*qep1.LineEncoder*qep1.PolePairs/6;
if (HALL_U&&!HALL_V&&!HALL_W) qep1.CalibratedAngle = qep1.LineEncoder*qep1.PolePairs/(2*6)+1*qep1.LineEncoder*qep1.PolePairs/6;
if (HALL_U&&HALL_V&&!HALL_W) qep1.CalibratedAngle = qep1.LineEncoder*qep1.PolePairs/(2*6)+2*qep1.LineEncoder*qep1.PolePairs/6;
if (!HALL_U&&HALL_V&&!HALL_W) qep1.CalibratedAngle = qep1.LineEncoder*qep1.PolePairs/(2*6)+3*qep1.LineEncoder*qep1.PolePairs/6;
if (!HALL_U&&HALL_V&&HALL_W) qep1.CalibratedAngle = qep1.LineEncoder*qep1.PolePairs/(2*6)+4*qep1.LineEncoder*qep1.PolePairs/6;
if (!HALL_U&&!HALL_V&&HALL_W) qep1.CalibratedAngle = qep1.LineEncoder*qep1.PolePairs/(2*6)+5*qep1.LineEncoder*qep1.PolePairs/6;

Does anybody have ideas how to implement initial rotor positon detection using Hall sensors?

Forum Post: RE: Concerto SPI2 interface problem

Stefan Buechner — Wed, 22 May 2013 06:54:00 GMT

Problem solved recently. It was a bit tricky, the other master (FPGA) was kept in a config-mode but did not give me control over all SPI signals, missed an remark in the user guide of the FPGA.

Another thing which I run into while searching for the problem: When I read some registers (e.g. GPIODATA or SSISR) I don't get the value I expect. I have to read the content to an variable. Just to take care of when you are debugging...

Stefan

Forum Post: RE: Schematics for TMDSHVPSFBKIT?

Paul Teagle — Wed, 22 May 2013 06:10:00 GMT

Hello Hrishi,

thanks for the info; its just what I was looking for.

It is just about identical with the power stage board for the UCC28950, so I guess it has been adapted for the C2000 dev kit by the developers.

rgds,

Paul T.

Forum Post: RE: Moving from old motion controller integrated DSP design.

Tai-Wook Kim — Wed, 22 May 2013 04:59:00 GMT

Dear Mr. Clearman,

Thank you for your reply.

The configuration will be 48V and ~5A peak and ~1.5A continous roughly.

The motion required is just like that of a sewing machine, forward/backward, faster/slower, stop/reverse. Also the C2000 will need to process some non-motion task (e.g. check an external safety sensor status, ring a warning buzzer, etc).

I can say that I was afraid about 'Isn't C2000 a overspec one? Is there any famous 'cheaper' solution from TI (or from other manufacturer)?'

I keep studying related things through materials and postings in this site, however, I needed a help about my basic concern on this selection of C2000.

Thank you.

Forum Post: RE: I2C data corruption! reliable SW

Thiago M — Wed, 22 May 2013 04:16:00 GMT

Decided adding TXS0104EDRG4 Voltage translator to +5V to the chips. And setting chips addresses the most different as possible - add0=add1=1 for one and add0=add1=-0 for other. That should make it more difficult for misaddressing - if that word exists :)

Couldn't find CRC code or instructions for I2C neither a working solution for stuck bus yet!

Any thoughts?

Forum Post: !!! HELP !!! can not debug with FLASH mode, but working in RAM mode.

Rafeeq MUHAMMED — Wed, 22 May 2013 03:28:00 GMT

Hello,

I am using C2000 launchpad with Digital Power Supply example project CCS v5 running. I could run debugging if "RAM" debugging mode is selected, unfortunately I need to change it to "FLASH" mode to increase code size. While changing debugging mode to "FLASH" and start "Built All" i get many error messages and stop the built. I do not understand what makes difference same code run well in RAM debugging option and not built in FLASH. I have attached error log and screen while return built error.

(Please visit the site to view this file)

Forum Post: RE: Question on F28069

Juan Gonzales — Wed, 22 May 2013 00:03:00 GMT

Hi Hrishi,

Does this have to do with DCAEVT1.sync?

(*ePWM[n+1]).DCACTL.bit.EVT1SYNCE = 1;

One interesting thing our customer found during testing, is if the event (comparator trip in this case) remains high, the DCAEVT1.sync generates a pulse every 10 clks. Is this documented anywhere? I could not readily find it in the data-sheet.

Forum Post: RE: the compiler told me symbol "DSP28x_usDelay" was undefined

BING NIE — Tue, 21 May 2013 23:58:00 GMT

the error was gone. thank u

Forum Post: TMS320F28335 Experimenter Kit Contents

Angelo Ballungay — Tue, 21 May 2013 23:21:00 GMT

Hi all,

I need a microcontroller that can sample at >2MSPS and store the data into memory. Does anybody know if I order the TMS320F28335 Experimenter Kit (http://www.ti.com/tool/tmdsdock28335), if that will be able to serve my purposes and come with everything I need, e.g. dock, microcontroller, cords, ADC/memory interface, etc...

Thank you

Forum Post: Loading program tms 320F28335

Saifuzzaman Sabuj — Tue, 21 May 2013 23:15:00 GMT

I am a new user of TMS320F28335 and trying to program it using code composer studio. I have written a file and converted to .obj format. How can i load it processor ?

Is there any good document or video to learn it ?

Forum Post: RE: DRV8312 getting hot

ChrisClearman — Tue, 21 May 2013 22:27:00 GMT

Sam,

This whole discussion is about scaling inside of a control system. We're in process of adding this discussion to the UG as it is very important to understand. There is quite a bit of adaptive scaling that happens inside of the algorithm, but it is limited based on the user set values which effect per unit scaling. The way we did this is very flexible to address a wide variety of motors, but it puts the onus on the user to make some good decisions. We need to document this decision making process better.

Essentially the discussion is around: how do you set the USER_IQ_FULL_SCALE variables (and to do this you need to understand their impact throughout the system)

The main one is USER_IQ_FULL_SCALE_VOLTAGE_V, and we are in process of documenting some checks / requirements.

Ex: USER_IQ_FULL_SCALE_VOLTAGE_V must meet the following

> USER_MOTOR_RATED_FLUX * USER_IQ_FULL_SCALE_FREQ_Hz [assuming this is your real Hz limit in all operating modes]

Satisfies USER_IQ_FULL_SCALE_VOLTAGE_V / EST_HZ < USER_MOTOR_RATED_FLUX [we are looking to clarify exactly how much smaller

as a trade-off with Rs Full Scale = USER_IQ_FULL_SCALE_VOLTAGE_V / USER_IQ_FULL_SCALE_CURRENT_A

You can tell that you need an IQ_V:
- larger than your Bemf produced (dominate feature for large flux machines, non-dominant for small flux machines) so Flux calculations don't overflow
- but you want it as small as possible to maximize your Rs resolution (your current is pretty set by your motor, always keep it as small as needed)
- and in the proper range to satisfy/maxmize flux resolution

Q: Presumably I need to also ensure that IQ_VOLTAGE is greater than my supply voltage?

A: Not necessarily. There is some head-room and you can have a ratio of >1 for the Vbus / IQ_V. Though in most cases the IQ_V will be higher than Vbus based on the other criteria that need to be met. Hope that doesn't confuse you.

Q: But it doesn't necessarily need to be greater than maximum ADC voltage?

A: Technically it doesn't, again because of a similar reason as the Vbus, but I believe the recommendation will probably be that for your own custom design there is no reason that you should have IQ_V < ADC_V, if so you designed your voltage scaling incorrectly. With the kits the ADC_V is fixed but there are times with certain motors we certainly MUST set the IQ_V < ADC_V to meet the above criteria.

Forum Post: RE: TMS320F2812 QEP peripheral

David M. Alter — Tue, 21 May 2013 20:50:00 GMT

[quote user="Alan Blumenstein"]

Oder way to say it is

v = EvaRegs.T2CNT/EvaRegs.T2PR ?

[/quote]

No, that is not correct. The 'T' in the equation is the elapsed time between QEP readings. That is not the same thing as the T2PR value. Actually, you want T2PR to be set to 0xFFFF in order for the T2CNT to roll over at 0xFFFF (I assume T2 is holding your QEP count). That way you can perform modulo-16 arithmetic when you compute x(t)-x(t-T).

I mean, think about this. Suppose x(t)=0x0001, and x(t-T)=0xFFFF. It means that x(t)-x(t-T) = 0x0002.

Take a look at that appnote SPRA363.

Forum Post: RE: to determine start and the end of one period of a sinusoidal waveform

Igor Gorbachev — Tue, 21 May 2013 20:47:00 GMT

Hello!

In addition the zero crossing method you can apply FFT-method also. But this method is not very convenient because it requires more RAM and it requires also knowledge of the approximate waveform max.frequency estimates for the correct choice of the sampling frequency. By the way If you want to calculate the full power, you just need to multiply the voltage by the current. But to calculate the active power is necessary to determine the phase shift between current and voltage and multiply full power by the cosine of the phase shift.

Regards,

Igor

Forum Post: RE: Warning of IQmath.Lib

David M. Alter — Tue, 21 May 2013 20:03:00 GMT

Fanx,

As Chris said, it is not important. It means that the object files in the library were compiled with an older compiler. The older compiler did not put certain compatibility checking information into the object module (for example, floating point code vs. fixed point code. Just an example). We tend to add additional information into the compiler output for new compilers, as such information is needed. Without it, the tools cannot perform certain compatibility checking for us.

The fix would be for us to recompile the IQmath library using a newer compiler. You always run the risk then that the library might not be compatible with code generated from an older compiler should someone be using an older compiler.

Regards,

David

Forum Post: RE: IQmath operation

Manish Bhardwaj — Tue, 21 May 2013 19:59:00 GMT

yes

Forum Post: RE: hvkit InstaSPIN Motor

ChrisClearman — Tue, 21 May 2013 19:23:00 GMT

a- what may the error be?

There is no working connection from your PC USB to the JTAG of the Piccolo device on the controlCARD. I can duplicate this error simply by not connecting the USB cable into the controlCARD.

b- M3-j5 is Populated or not?

POPULATED

and main-j9 is Populated or not?

NOT

2- when i need to work from ccs v5 and did connection and tried to do 'test connection" also an error massage and it is attaiched with this post in txt file? can you help me for that? are the same hw Set-up as work from GUI?

Yes, same HW set-up for all TMDSCNCD28069MISO and HVKT r1p1 (yellow), using GUI Composer or CCS.

3- what is the mening by ls_d and ls_q? for ACI?

stator inductance (Ls) is made of of two parts, called Ls_d and Ls_q. In motors with high saliency the difference is quite large. For induction motors and most synchronous motors a single "average" Ls works just fine. For highly salient motors you can compensate for this difference (even during run time) to achieve even better FAST tracking for enahanced performance and use in more efficient control systems.

When you plug in USB from PC to your controlCARD please check your Windows Device Manager and insure that a new USB port is found by your PC.

there should be two, called

TI XDS100 Channel A

TI XDS100 Channel B

and under Ports you should see USB Serial Port (COM##) - this is the UART chanel provided by XDS100/FTDI chip on the controlCARD.