Shop OBEX P1 Docs P2 Docs Learn Events
STMicroelectronics Introduces A New Stepper Driver, The L6480. Oh Yea Baby!!!! — Parallax Forums

STMicroelectronics Introduces A New Stepper Driver, The L6480. Oh Yea Baby!!!!

idbruceidbruce Posts: 6,197
edited 2014-01-18 02:46 in General Discussion
Do you want massive amounts of power for your stepper motors? Well then take a peek at the attached documents. The EVAL6480H evaluation board boasts 25A R.M.S. per phase and can handle a supply voltage of 10.5 to 85 volts. This driver is specifically designed to use external mosfets, the DPAK STripFET II Power MOSFET (STD25NF10) to be more specific.

Bruce

EDIT: It may be time to reevaluate my desire to create boards from the L6208 :)
«1

Comments

  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-10-27 05:33
    25 Amps RMS! I presume that is Root Mean Square or about 70% of Peak Amps. Why the RMS? Are these intended to be driven by sine waves?

    25 amps at 18 volts would be about 1/2 horsepower - serious holding power.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2012-10-27 05:47
    idbruce wrote: »
    Do you want massive amounts of power for your stepper motors? Well then take a peek at the attached documents. The EVAL6480H evaluation board boasts 25A R.M.S. per phase and can handle a supply voltage of 10.5 to 85 volts. This driver is specifically designed to use external mosfets, the DPAK STripFET II Power MOSFET (STD25NF10) to be more specific.

    Bruce

    EDIT: It may be time to reevaluate my desire to create boards from the L6208 :)
    Thank you very mmmmmuch! I have placed my order. I will have to see how small a board I can do. 85V operation yes!
    289 x 160 - 90K
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 05:51
    Loopy

    Most, if not all stepper driver manufacturers basically give you two amperage ratings, peak and rms. I am not an electrical engineer, but as close as I can gather, from reading datasheets and comparing the various driver boards, the rms rating is close to the maximum useable amperage of the chip, providing there is proper cooling. So I would assume, that the MOSFETs can handle 25A continous current.
    The L6480 controls the phase current applying a sinusoidal voltage to motor windings. Phase current amplitude is not directly controlled but depends on phase voltage amplitude, load torque, motor electrical characteristics and rotation speed.

    Bruce
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 06:01
    @Peter
    85V operation yes!

    If only we could rectify 120VAC down to 85VDC, we could eliminate the need for a transformer :)
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2012-10-27 06:14
    idbruce wrote: »
    @Peter



    If only we could rectify 120VAC down to 85VDC, we could eliminate the need for a transformer :)
    120VAC (RMS) = 170VDC
    If only we could run them in series as 85+85 but we have 240VAC which might be a bit high.
    BTW, some manufacturers like to quote peak current combined, that is for both windings. fake fake fake. I prefer RMS per winding which is what is quoted for the L64xx series. The other thing about these chips is that you don't need big external schottky diodes or hefty current sense resistors either as you do with many other chips.
  • YanomaniYanomani Posts: 1,524
    edited 2012-10-27 06:58
    Hi Loopy Byteloose
    25 Amps RMS! I presume that is Root Mean Square or about 70% of Peak Amps. Why the RMS? Are these intended to be driven by sine waves?

    Don't be surprised, but chapter 7 (pages 35/36) of the datasheet expresses exactly what you dreamed! Internal sinusoidal voltage generation to control phase current applied to motor windings.

    Oh oh oh oh oh! It's Santa on the roof? Mint on my lips? ;)

    Yanomani
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-10-27 07:50
    I suspect the advantage of a sinusoidal power source is that there are no nasty harmonic frequencies at these rather high power outputs to interfere with radio wave communications.

    Sine waves are simply no harmonic content. Lets see .... 25amps at 85 volts is ...... 2.125 Kilowatts !!!! and one HP is 746 watts

    So about 2.85 Horsepower....! I wonder how many horsepower there are in the average human bicep.
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 08:07
    @Peter
    I have placed my order.

    Do you have a magical supply house? Who did you order from and what do the chips cost? Or did you order the evaluation board? I see that Mouser has 4 evaluation boards on order.
  • LeonLeon Posts: 7,620
    edited 2012-10-27 08:20
    I just checked the ST web site, and they aren't available yet:

    http://www.st.com/internet/analog/product/253950.jsp

    The boards aren't available, either.
  • jmgjmg Posts: 15,173
    edited 2012-10-27 13:03
    I suspect the advantage of a sinusoidal power source is that there are no nasty harmonic frequencies at these rather high power outputs to interfere with radio wave communications.

    Sine (and the implied cosine) drive are used for Microstepping.
    Even if you do not need all the precision, the lack of cogging effects can be useful.

    This claims to have 128 Microstep ability, which is quite impressive.
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 13:08
    @jmg
    This claims to have 128 Microstep ability, which is quite impressive.

    The L6470 that Peter currently uses for his stepper driver module, also boasts 128 microsteps.

    Bruce
  • jmgjmg Posts: 15,173
    edited 2012-10-27 13:36
    idbruce wrote: »
    @jmg
    The L6470 that Peter currently uses for his stepper driver module, also boasts 128 microsteps.

    I see this also claims
    ["Thanks to a unique voltage mode driving mode which compensates for BEMF, bus voltage and motor winding variations, the microstepping of a true 1/128-step resolution is achieved"]

    BEMF and bus voltage the chip can know about, but it must be clairvoyant to know about motor winding variations ? ;)
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 18:47
    @jmg
    BEMF and bus voltage the chip can know about, but it must be clairvoyant to know about motor winding variations ? ;)

    Okay, just for the sake of discussion, let's say that one coil has an extra 1/2 turn of wire winding, wouldn't it have different electrical characteristics? Perhaps a tad more resistance, voltage drop, and perhaps current flow?

    I am really surprised there was not more commenting on this product release given the amperage and voltage ratings. The industry has been screaming for a solution like this. It is the first of any chip that I have found that simplifies the use of MOSFETs for bipolar stepper motor control with a dual h-bridge configuration and such a wide range of amperage and voltage.

    WOW Must be a tuff crowd.

    Bruce
  • jmgjmg Posts: 15,173
    edited 2012-10-27 19:50
    idbruce wrote: »
    Okay, just for the sake of discussion, let's say that one coil has an extra 1/2 turn of wire winding, wouldn't it have different electrical characteristics? Perhaps a tad more resistance, voltage drop, and perhaps current flow?

    Still tough to expect the chip to know more R is due to an extra turn (which would matter as AmpTurns is Torque) and not just a warmer winding... ?

    This issue reminds me of a imaging scan stepper design, I read about a few years back, for a high resolution printer, and there, each motor came with an EPROM - now THAT approach was able to 'compensates for motor winding variations', as well as Pole and even field non-linearities.
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 20:17
    @jmg
    This issue reminds me of a imaging scan stepper design, I read about a few years back, for a high resolution printer, and there, each motor came with an EPROM - now THAT approach was able to 'compensates for motor winding variations', as well as Pole and even field non-linearities.

    I am currently examining the EVAL6208N and considering the design of a board. I truly don't think it is too unrealistic to have a uC on board with EEPROM and perhaps SD storage for motor specific settings and running characteristics, such as ramping profiles. I can easily envision a string of stepper drivers, or individual drivers going back to a master controller, all receiving basic serial commands. The basic commands would be the number of steps, the direction of rotation, and enable/disable. Of course this would increase the cost of the board, but it might be a nice setup.
  • jmgjmg Posts: 15,173
    edited 2012-10-27 20:38
    idbruce wrote: »
    @jmg
    I am currently examining the EVAL6208N and considering the design of a board. I truly don't think it is too unrealistic to have a uC on board with EEPROM and perhaps SD storage for motor specific settings and running characteristics, such as ramping profiles. I can easily envision a string of stepper drivers, or individual drivers going back to a master controller, all receiving basic serial commands. The basic commands would be the number of steps, the direction of rotation, and enable/disable. Of course this would increase the cost of the board, but it might be a nice setup.

    Sounds a good idea to make it optional, at least. Small micros are cheap.
    I was impressed they took the trouble to calibrate each motor (and I guess somehow also marked the shafts..), but if you DO want to push precision, and be more tolerant of motor choices, this is the smart way to go.
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 21:29
    In fact, this is my current decision and it is one of those darned double edged sword dilemas. By adding a uC onboard, it drives up the cost of the board and complexity, and if you don't add it, then you are eating up a bunch of IO pins on the main controller to take full advantage of the chip. For instance, to take full advantage of the L6208, it requires 6 IO pins, and that is a bunch when your dealing with a multi-axis machine. I figure two way serial communication with a uC on board is the way to go, cutting down the required pin count from the master controller to 1/3 (2 pins), but then, in order to make use of the high voltage supply going to the board already, a switching regulator would have to be added for logic supply.

    Additionally, SPI communication might even be better, but I really know nothing about it.
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 21:51
    Which brings about a question for anyone that cares to answer it. In the schematic attached below, with a known voltage of lets say 50V coming in on CN1 and R2 adjusted to output 5V on terminal 3 of JP1 and going through a 3.3V regulator, can this 3.3V be used to reliably power a Propeller chip and how efficient or inefficient would such a circuit be?

    EDIT: Here is a list of the proposed component values:
    D3 BZX79C5V1 5.1V Zener Diode
    C5 10μF/16V Capacitor
    R2 820 Ohm 0.6W Resistor (approximate)
    204 x 422 - 16K
  • jmgjmg Posts: 15,173
    edited 2012-10-27 21:59
    idbruce wrote: »
    Additionally, SPI communication might even be better, but I really know nothing about it.

    SPI is fast, but vanilla SPI slaves can quickly become a pain. The most brain dead ones have no buffering, and are byte-wide.
    A SPI port with a small FIFO and/or larger packet size makes life a lot easier.
    Some micros have a SPI mode in their UART, which usually have better buffering.

    Alternatives are i2c, which many can do to 400KHz, and even fast Async, which can get to 1MBd on some small uC.
  • kwinnkwinn Posts: 8,697
    edited 2012-10-27 22:14
    idbruce wrote: »
    @jmg

    .........

    WOW Must be a tuff crowd.

    Bruce

    More of a crowd that has had a great deal of experience with misleading marketing brochures and overstated or misleading claims.
  • jmgjmg Posts: 15,173
    edited 2012-10-27 22:15
    idbruce wrote: »
    Which brings about a question for anyone that cares to answer it. In the schematic attached below, with a known voltage of lets say 50V coming in on CN1 and R2 adjusted to output 5V on terminal 3 of JP1 and going through a 3.3V regulator, can this 3.3V be used to reliably power a Propeller chip and how efficient or inefficient would such a circuit be?

    Do you mean a current source style supply ?

    The plus side is they have a (very) high score in the robust column, and are inherently reverse and spike protected.

    One down side is not many devices actually state that they have some Max start-up current need, and an unexpected load on a pin, can compromise the power supply budget.

    Another issue, you hinted at, is they are limited to lowish currents - at 50V, even 10mA is ~500mW, and they draw that at all times.


    Where a wide operating range is needed, we've looked at parts like OnSemi's NSI50010YT1G a 10 mA 50 V Constant Current device.
    (but they lose reverse protection)
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 22:42
    Do you mean a current source style supply ?

    Not sure what current source style supply is, but let's say that the supply going to CN1 is from a power supply similar to the PS-5N50R as attached below. To provide further understanding of what I want to, I have also attached a larger portion of the schematic and an excerpt from AN1451 (which is also attached). It is also worth mentioning that the 50V would be unregulated.
    The 5V voltage for logic inputs and for references (VrefA and VrefB) is obtained from R2, D3. Depending on the
    supply voltage, the value of resistor R2 should be changed in order to ensure a correct biasing of D3.
    The jumper JP1 allows choosing the 5V voltage from the internal zener diode network or pin 11 of CN5 (for example
    an external μC board can provide 5V to the evaluation board). Also CN2 connector can be used to provide
    an external 5V voltage to the board (in that case R2, D3 should be disconnected). CN2, or pin 1 of CN5,
    can also be used to provide a 5V voltage to external circuits (as, for example, the PractiSPIN ST7 board). In this
    case the current that can be drawn form the board depends on the supply voltage and on R2 value.
    I interpret this to mean that a 5V logic supply can be obtained from a much larger supply going to pin 3. If this is the case, can this be used as potential source to power a propeller chip to control the L6208 IC. And if necessary, can the values of D3 and R2 be changed to allow more current?
  • jmgjmg Posts: 15,173
    edited 2012-10-27 23:02
    idbruce wrote: »
    And if necessary, can the values of D3 and R2 be changed to allow more current?

    Yes, D3 and R2 are just a classic Zener shunt regulator.
    I call it a current source style supply, as the value of R2 is determined by the current budget, and the operating point is essentially constant current.
    eg 50V and 450ohms, will have 10mA max current, at Zener 'drop out' and an 11mA short circuit value.
    R2 dissipates 450mW in normal use, and 500mW in a SC case.

    A fish-hook in this type of design, is the start-up system current should be less then the operate current.
    Something that draws 20mA at 3V, as it ramps to 5V & 5mA, is a no-no.
  • idbruceidbruce Posts: 6,197
    edited 2012-10-27 23:24
    @jmg

    Thanks for clarifying that for me.

    Bruce
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-10-28 05:15
    idbruce wrote: »
    Which brings about a question for anyone that cares to answer it. In the schematic attached below, with a known voltage of lets say 50V coming in on CN1 and R2 adjusted to output 5V on terminal 3 of JP1 and going through a 3.3V regulator, can this 3.3V be used to reliably power a Propeller chip and how efficient or inefficient would such a circuit be?

    EDIT: Here is a list of the proposed component values:
    D3 BZX79C5V1 5.1V Zener Diode
    C5 10μF/16V Capacitor
    R2 820 Ohm 0.6W Resistor (approximate)

    Not all regulators are the same (Duh).

    From what I gather many of the reasons for LDO regulators are to accomodate the rather noisy automotive environment. The Best of these would likely enhance the stability of your high power stepper driver. In addition to having LDO for brown out episodes, the really rugged 'automotive' voltage regulators have reverse polarity protection and can seem to tolerate swings from -50VDC to +60VDC as momentary events.

    About the worse choice one can make is the ubiquious 78xxx regulator series. LDO is not about better battery use as much as it is about brown-out prevention.
  • archcomparchcomp Posts: 1
    edited 2013-04-08 05:15
    idbruce wrote: »
    In fact, this is my current decision and it is one of those darned double edged sword dilemas. By adding a uC onboard, it drives up the cost of the board and complexity, and if you don't add it, then you are eating up a bunch of IO pins on the main controller to take full advantage of the chip. For instance, to take full advantage of the L6208, it requires 6 IO pins, and that is a bunch when your dealing with a multi-axis machine. I figure two way serial communication with a uC on board is the way to go, cutting down the required pin count from the master controller to 1/3 (2 pins), but then, in order to make use of the high voltage supply going to the board already, a switching regulator would have to be added for logic supply.

    Additionally, SPI communication might even be better, but I really know nothing about it.

    Hello,
    Does this means, this controller does not have the flexibility of being used as plain Pulse / Dir type of controller? There is one pulse input pin, but for changing the direction, as I understand, it needs to use SPI command to change the internal registers. Also a UPC is required to initialise the parameters like microstepping and other settings as per user requirements. So, is it that this controller can't be used as an independent stepper motor driver, with a flexibility of making once in lifetime parameter setting and controlling by pulse / dire, which is a standard these days for Motion Controllers. Why on earth ST can't offer such a flexibility, by allowing SW pin for Direction and SPI pins for Microstepping setting inputs.!!
    Please clarify if I am mistaken.
    Regards,
    M.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2013-04-08 07:41
    archcomp wrote: »
    Hello,
    Does this means, this controller does not have the flexibility of being used as plain Pulse / Dir type of controller? There is one pulse input pin, but for changing the direction, as I understand, it needs to use SPI command to change the internal registers. Also a UPC is required to initialise the parameters like microstepping and other settings as per user requirements. So, is it that this controller can't be used as an independent stepper motor driver, with a flexibility of making once in lifetime parameter setting and controlling by pulse / dire, which is a standard these days for Motion Controllers. Why on earth ST can't offer such a flexibility, by allowing SW pin for Direction and SPI pins for Microstepping setting inputs.!!
    Please clarify if I am mistaken.
    Regards,
    M.

    archcomp, clearly you have an interest in steppers but I can't understand what the problem is with the L6480 that you seem to indicate. It took me a few seconds to realize that UPC must refer to a microcontroller which is always a given in any system. ST have allowed for flexibility by bringing out the step clock pin whereas all the parameters are loaded in serially just like this message that you are reading has been loaded into your computer serially, that's the way things work. Which world is it that we would want a stepper chip to operate without a micro? I understand OTS motion controllers have step and direction pins and DIP switches etc but even they are a combination of a stepper chip and a microcontroller.

    Remember, the L6480 is a tiny chip the size of a fingernail, it's not a module.

    BTW, I've been using them.
  • dandumitdandumit Posts: 1
    edited 2013-04-09 02:03
    BTW, I've been using them.

    PEter, can you give us some feedback using it ? Some pictures of your boards ?

    BR
    Daniel
  • Mark_TMark_T Posts: 1,981
    edited 2013-04-09 12:19
    archcomp wrote: »
    Hello,
    Does this means, this controller does not have the flexibility of being used as plain Pulse / Dir type of controller? There is one pulse input pin, but for changing the direction, as I understand, it needs to use SPI command to change the internal registers. Also a UPC is required to initialise the parameters like microstepping and other settings as per user requirements. So, is it that this controller can't be used as an independent stepper motor driver, with a flexibility of making once in lifetime parameter setting and controlling by pulse / dire, which is a standard these days for Motion Controllers. Why on earth ST can't offer such a flexibility, by allowing SW pin for Direction and SPI pins for Microstepping setting inputs.!!
    Please clarify if I am mistaken.
    Regards,
    M.

    Absolutely, I was just about to make that point - without step/direction inputs its a waste, once you have a uController you might as
    well do most of the work there - only need some comparators and MOSFET drivers, which are probably cheaper than this new chip will
    be...
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2013-04-10 06:11
    Mark_T wrote: »
    Absolutely, I was just about to make that point - without step/direction inputs its a waste, once you have a uController you might as
    well do most of the work there - only need some comparators and MOSFET drivers, which are probably cheaper than this new chip will
    be...

    Guys, I can't believe what I'm hearing. You mean that you have to have a physical direction pin that can be set high or low which would be from a micro obviously yet the same micro can't bang a couple more bits into that same pin to set coil current among may other parameters which would normally require dozens of I/O pins? As for "may as well get the micro to do the work" it is an easy statement to make but is the micro up to the task in real-time and are you prepared to write all the motion control software to accomplish that? This family of stepper chips are very versatile and I have run them in dumb step clock mode as well as relied on them for full motion control and software can treat the direction pin just like an I/O if it wants to as the SPI driver just belts those control bits out, an operation that is transparent to the application.

    I myself with my long experience could be forgiven for wanting things "traditional" yet I welcome a sensible and practical way of handling things. So the problem is not with the chip then, it lies elsewhere....
Sign In or Register to comment.