• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 29 subscribers
  • Views 389 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DRV8818: inconsistent motor speed with fixed step timing in Full-step Mode.

F. Schmuki
Prodigy 61 points
Part Number: DRV8818
Other Parts Discussed in Thread: DRV8424

Dear Forum, Dear TI Experts,

we replaced a stepper motor driver due to EOL notice.
On the legacy board, it was used a stepper motor driver from Allegro Micro Systems. A3977SED-T ( Stepper Motor Controller, 2.5A, NMOS, PQCC44 ).
Looking for alternatives with the same functionality we have found: A5977GLPTR-T ( Stepper Motor Controller, 2.8A, PDSO28 ) also from Allegro Micro Systems.

Since we always had good experiences with components from TI, we searched for an equivalent via the cross reference search and found one.
-> DRV8818PWPR

according the DataSheets a 1:1 replacement.

Our application requires a constant speed of the motor of 401.57 rpm, since we are measuring samples in a disk with fixed segments.
So if the speed of the motor varies, we miss the right time slot to measure our samples.

Tests with the new driver shows, that the motor speed is inconsistent, due to a delay in applying the next step.


Driver Settings:
VMx: 25 V
VCC: 5 V
USMx: 0 --> FullStep
SRn: GND --> synchronous rectification enabled
Vref: 3.178 V
Rsense: 0.33 Ω
Itrip: 1.2 A
step frequency: 1.34 kHz

compared a legacy board with a new one shows that the drivers are working different.
We could measure that the driver applies the next step only when the winding current is decreasing and toff time has expired or the winding current is increasing.
lower the toff time increases the chopper frequency, but the "dead time" still exists. (R= 12 kΩ / C = 1 nF --> 12us)
extending the toff time decreases the chopper frequency, "dead time" stands, (R= 100 kΩ / C = 1 nF --> 100 us)

Varying the voltage on the DECAY pin, does not have any effect!?!

According E2E Posts, the DRV8818 can only operate in slow decay mode for full step setting
https://e2e.ti.com/support/motor-drivers-group/motor-drivers/f/motor-drivers-forum/634627/drv8818-mixed-decay?tisearch=e2e-sitesearch&keymatch=DRV8818#

Searching for a solution we have mounted two diodes from the STEP input to RCA and RCB, to force the driver into PWM ON mode so that the next step is created immediately.
This solves the problem, but does not look like the right solution. Are we missing something here?

We are very appreciative of any information you may provide.

Stepper_Schematic_legacy_A3977

Stepper_Schematic_new_DRV8818
Stepper_Schematic_new_DRV8818

Comparism_Oscilloscope_Pictures
Comparism_Oscilloscope_Pictures.pdf

Stepper_Schematic_new_DRV8818_with_fix
Stepper_Schematic_new_DRV8818_with_fix

Best Regards,
Fabio

  • 0
    TI__Guru 54645 points

    Fabio,

    Let us investigate and reply back.  If this is a new design, I would recommend moving to the DRV8424.  

    https://www.ti.com/product/DRV8424

    Much newer device with integrated current sensing and smart tuning features. 

    Regards,

    Ryan

  • 0 in reply to Ryan Kehr
    Prodigy 61 points

    Hi Ryan, thanks for the prompt feedback.

    We will keep your recommendation in mind.
    For compatibility reasons we had chosen the DRV8818.

    Our instrument is a Medical Device with currently enormous demand. (COVID related)
    Due to this fact we have already manufactured and assembled 500 PCBs.

    We have been investigating the problem for several weeks and yesterday we were able to identify the root cause.
    Our customer is already in a back order situation and not happy about this situation. So there is no time for a new PCB Layout.
    As a short term solution we need a proper fix as soon as possible.

    Your support is greatly appreciated.

    Regards,
    Fabio

  • 0 in reply to F. Schmuki
    TI__Guru 56650 points

    Fabio,

    I am not sure what DRV8818 issue do you have?

    a. Only slow decay in full step mode. Why does it have to run full step?

    b."PWM ON mode so that the next step is created immediately" STEP doesn't synchronize PWM pulse? Why do we need PWM on the next STEP?

    c. TOFF setting or PWM frequency?

    d. Decay mode setting: "Varying the voltage on the DECAY pin, does not have any effect!?"

    Regards,

    Wang Li

  • 0 in reply to Wang5577
    Prodigy 61 points

    Hi Wang Li,

    a. Only slow decay in full step mode. Why does it have to run full step?

    --> not an issue, for legacy compatibility it has to run in full step mode.
    Only stumbled across it in the forum while searching for the error.

    b."PWM ON mode so that the next step is created immediately" STEP doesn't synchronize PWM pulse? Why do we need PWM on the next STEP?
    c. TOFF setting or PWM frequency?
    --> Please ignore "PWM", probably  a bad formulation from my side, referring to Figure 8 of the Datasheet.

    d. Decay mode setting: "Varying the voltage on the DECAY pin, does not have any effect!?"
    --> observed during root cause finding.

    Observed Problem:

    Lets focus on this:
    Driver Output has sometimes a delay on step execution.
    On rising edge on the STEP input the driver should execute the step immediate.
    Step_Execution_DRV8818

  • 0 in reply to F. Schmuki
    TI__Guru 54645 points

    Fabio,

    Are you able to modify the firmware and try an experiment?  If you can't modify the firmware, you can just use a pulse generator to drive the STEP pin.  

    Would like to see if a 50% duty cycle on the DRV8818 STEP pin helps with the issue.  I see you are currently running a high duty cycle with low step time - can't tell for sure how long based on scope captures.  Try changing this signal to 50% duty and see if it helps.

    Regards,

    Ryan

  • 0 in reply to Ryan Kehr
    Prodigy 61 points

    Hi Ryan,
    We are not able to modify the firmware, but we have used a pulse generator to drive the steps with approx. 50% duty.

    Steps generated from MCU:

    • f = 1.34 kHz
    • Ton = 720 us
    • Toff = 24 us

    Steps generated with Waveform Generator approx. 50% duty:

    • f = 1.34 kHz
    • Ton = 400 us
    • Toff = 346 us

     The behavior still exists.

    Regards,
    Fabio

  • +1 in reply to F. Schmuki
    TI__Guru 56650 points

    Fabio，

    Thank you for the test result. I think you have found the root cause which was related to the PWM clock. It is reasonable. When STEP rising edge is accept by the DRV8818. DRV8818 Itrip comparator could change the current setting. But, the Itrip comparator would check the current setting in the beginning of the next PWM cycle. So, the output takes action at that time.

    Also, I think the workaround solution also looks good to me. 

    Regards,

    Wang Li

  • 0 in reply to Wang5577
    Prodigy 61 points

    Hello Ryan, Hello Wang Li,

    We will go with the proposed solution as a quick fix.

    thanks for your support.