Skip to main content
hsonderegger9
hsonderegger9
Visitor II
January 31, 2011
Question

EVAL6470 operation via SPI

  • January 31, 2011
  • 23 replies
  • 3322 views
Posted on January 31, 2011 at 09:15

EVAL6470 operation via SPI

#l6470 #l6470 #l6470 #l6470 #l6470
This topic has been closed for replies.

23 replies

infoinfo981
infoinfo981
Visitor II
May 17, 2011
Posted on May 17, 2011 at 10:20

I'm actually the third one having this problem...

Tried with putting the L6470 into StepClock mode but also with no luck except the L6470 gets bloody hot after a while...

Also I don't get the point in the datasheet why the nCS signal has to be raised after every byte transmitted....as defined in chapter 8. Later on in Figure 17 the SPI timing waveform shows no nCS being raised between bytes....

The application note AN3103 is totally unusable....not  real application note for me (o;

What do you send over SPI to initialise your L6470?

Funny thing is the motor only turns a little when I put the attached PIC18F4550 into reset (o;

flavio2
flavio2
Associate
May 17, 2011
Posted on May 17, 2011 at 10:20

Hello Harry

It seems that I have the same problem. Do you have any signals on OUT1B? I can see the PWM signal configurated by F_PWM_INT and F_PWM_DEC at pin OUT1B. The other three motor connectors work fine. Also the communication works as intended. I am able to read/write registers. But OUT1B continues pulsing even if I send a HardStop command.

So far I did not find any solution for this problem.

 

regards

Flavio

flavio2
flavio2
Associate
May 17, 2011
Posted on May 17, 2011 at 10:20

Hi all

My initialization sequence is as following

CONFIG = 0x1CA0

then I set all the motor parameters for my stepper like MIN_SPEED, MAX_SPEED, ACC, STEP_MODE and so on...

at least I send a GetStatus to clean up the error flags in the STATUS register.

After init the communication works fine but the motor do not want to turn regardless of the motor settings. I can measure the stepping pulses on the outputs but not on OUT1B.

Do you see any pulses with your oscilloscope?
infoinfo981
infoinfo981
Visitor II
May 17, 2011
Posted on May 17, 2011 at 10:20

BTW...what stepper motor are you using? What current specification?

Doesn't look good that the L6470 can handle a 3A, 3Nm stepper motor at all (o;

infoinfo981
infoinfo981
Visitor II
May 17, 2011
Posted on May 17, 2011 at 10:20

Got it to turn now...sort of (o;

The reason it didn't move was that really after every byte sent you have to raise CS to a logical '1' !

But I guess I can stop now with evaluating the L6470....it states that it can handle 3A RMS...but connecting my 3A stepper motor at a low voltage of 18V the L6470 gets really hot.....no clue where the 3A RMS come from in the datasheet.....

The sequence I used is:

- Send a HardStop to activate the bridge

- Then do a GetStatus

and after that you can do either set it to StepClock Mode and pulse STCK or

send the Move commands....

flavio2
flavio2
Associate
May 17, 2011
Posted on May 17, 2011 at 10:20

Hi there

My motor turns! I could manage the chip by lowering the supply voltage from 24V to 12V. Now the chip doesn't get too hot and the motor turns exactly where I want him to turn. But I really doubt the chip could handle a 3A stepper... Perhaps with an improved print layout including heat sink and fan. But surely not on the evaluation board.

Also my settings were not really ''L6470-friendly'' so I had to raise the acceleration value up to 2000 and lower the maximum speed to 120. Tonight I'm going to do a long duration test and I hope the chip will still work tomorrow morning :)

In conclusion, the chip doesn't seem to by ready to market and it's documentation is not really entire. I hope there will be a real application note soon... As you said AN3103 is not an application note at all. 

Here is my code to send and receive SPI data:

BYTE readSPI (void)

{

int bitNr = 7; //counter variable

BYTE returnByte = 0x00; //return byte

BYTE masks[8] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80}; //bit masks

BOOL pValue = 0; //MISO value

U2C_SingleIoWrite(m_spiHandle,SPI_MOSI,0);

//MOSI = 0, that looks like a NOP-command for the chip

U2C_SingleIoWrite(m_spiHandle,SPI_SS,0); //CS = 0

for (bitNr=7; bitNr>=0; bitNr--) //do it for each bit

{

U2C_SingleIoWrite(m_spiHandle,SPI_CLK,0); //clock = 0

U2C_SingleIoWrite(m_spiHandle,SPI_CLK,1); //clock = 1

U2C_SingleIoRead(m_spiHandle,SPI_MISO,&pValue); //read bit on MISO

pValue > 0 ? returnByte = (returnByte | masks[bitNr]) : NULL;

//if read bit is 1, set bit in return byte

}

U2C_SingleIoWrite(m_spiHandle,SPI_SS,1); //CS = 1

return returnByte; //return the byte read

}

void writeSPI( BYTE data )

{

int bitNr = 7; //counter variable

BYTE masks[8] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80}; //bit masks

U2C_SingleIoWrite(m_spiHandle,SPI_SS,0); //Chip Select = 0

for (bitNr=7; bitNr>=0; bitNr--) //do it once for each bit

{

U2C_SingleIoWrite(m_spiHandle,SPI_CLK,0); //clock = 0

U2C_SingleIoWrite(m_spiHandle,SPI_MOSI,(data&masks[bitNr])>>bitNr);

//write bit to MOSI-pin

U2C_SingleIoWrite(m_spiHandle,SPI_CLK,1); //clock = 1

}

U2C_SingleIoWrite(m_spiHandle,SPI_SS,1);

//after the byte is transmitted, raise CS

}

infoinfo981
infoinfo981
Visitor II
May 17, 2011
Posted on May 17, 2011 at 10:20

Just killed my EVAL6470 tonight (o;

Time to design with some good old fashioned devices then....

Unless ST brings out some real application note on the L6470 instead of some

screenshots of their GUI software....but I hardly doubt that (o;

infoinfo981
infoinfo981
Visitor II
May 17, 2011
Posted on May 17, 2011 at 10:20

Weird chip this is (o;

When I send a hard stop first and set the Kvl for Hold later the motor current is 0.25A with a value of 0x08....setting Kval before a HardStop the current is much lower (0.04A)...

Also when I send let's say a Move command with 200 full steps it either turns

slowly indefinitively....or it just hums and does the acc/decceleration noise....

Maybe the SPI communication isn't that perfect as it might not detect the commands rights (short humming sound versus indef. move)...

Or it doesn't like the 5V SPI jumper settings....

Enrico Poli
Enrico Poli
ST Employee
May 17, 2011
Posted on May 17, 2011 at 10:20

Dear Mr. Perucchi, Dear Mr. Klingler,

All the issues described into this topic could be related to two causes:

- SPI communication

- BEMF compensation

The SPI interface of L6470 basically is an 8 bit shift register enabled by nCS signal. The device will interpreter the byte into shift register when the nCS line rises. This behavior allows connecting more L6470 in chain, a single SPI interface can drive mote device at a time (daisy-chaining).

The L6470 logic needs 800 ns to interpreter the received byte, so the nCS line should be kept high for at least 800 ns between one byte (or group of byte if a daisy-chain is implemented) and the next one.

The L6470 drives the motor using a voltage mode approach. It is an open loop control implementing a feed forward compensation of the motor BEMF.

A correct configuration of the BEMF compensation parameters is mandatory to make the motor run. A poor compensation can cause overcurrent events, motor stall (low current), strong vibrations and device overheating.

The best way to find the setup fitting your application is to use the dSPIN evaluation software (you can find it in L6470 page on st.com). This application includes a dedicated tool for the evaluation of the BEMF compensation parameters starting from the application characteristics.

The tool returns a good starting point: monitoring the phase current and fine tuning the setup values the optimal compensation can be achieved.

About the L6470 current limit. The output current of all power electronic devices is limited by two constrains: electrical overstress and power dissipation (thermal limit).

The L6470 circuitry can actually drive 3 Arms for each phase; this is its electrical limit.

The thermal limit depends on the ability of the application to dissipate the power generated into the device (thermal resistance junction-to-ambient). This value is related to board layout (dissipating area, number of layers, material, etc.), device package, additional cooling systems (heaters, fans, etc.), ambient temperature and humidity, …

Cause its limited size, the Rthj-a of the EVAL6470 is about 40°C/W. It is a big value, far from the one that can be obtained using an accurate thermal design.

Finally, I want to highlight that the AN3103 is not the L6470 one, it is the description of the dSPIN evaluation environment only. The official AN of the product is under development.

Anyway you can always refer to our customer support (this forum included) for any issue related to the device.

Kind Regards,

Enrico Poli

Application Engineer

Motion Control Team

flavio2
flavio2
Associate
May 17, 2011
Posted on May 17, 2011 at 10:20

Ok, the motor turns now. I had a little mistake in the calculation of the register values ;o(

But I have the same problem as you have, Richard. The chip gets damn hot. I don't think the Chip is able to handle that high currents (24V with KVAL = 19 over 1.56 ohms = 1.14A). Perhaps the cooling on the evaluation board is not that ideal. If I have time today, I'll do some tests with additional cooling.

Show more replies
Terms & ConditionsAccessibility statement