DRV8424: overheating issue

Dear Slava,

Thank you for your question. Could you please verify if the testing was done using the DRV8424EVM hardware? If I understood correctly the issue was experienced with a 2.244V VREF setting for a phase current of 1.7A with a VM supply voltage of 24V, with a 1.7A max. rated stepper motor, correct?

For reproducing the mentioned issue we ran a bench setup using the EVM under maximum phase current conditions mentioned in your post, 2.5A (VREF = 3.3V), VM = 24V, Full-step mode, 300PPS, and ambient temperature was 23degC with a 3A maximum current rated stepper motor. The decay mode was at the default setting, Smart Tune Dynamic Decay, TOFF = 24us. The EVM was flipped vertical to allow free-air conditions to the bottom of the EVM. We tested the setup under the full load condition for longer than 1.5 hours with no incidences of OTSD, over temperature shutdown. We are unable to reproduce the described issue. Please see attached oscilloscope screen capture. Yellow trace is step input and green trace is one winding phase current with 1mV = 10mA current probe.

We suspect heat was not effectively removed from the driver during your test. With proper PCB thermal design we expect 1.7A RMS (for square waves as in full-step, Vrms = Vpeak) phase current setting should be comfortably driven with proper PCB thermal design suitable for the specific use case. This document https://www.ti.com/lit/an/spra953c/spra953c.pdf is a great reference read for understanding IC Package Thermal Metrics. 

Regards, Murugavel

Dear Murugavel,

Thank you for your answer. The testing was done without using the DRV8424EVM hardware. Schematics and layout are shown below. 

I note that the DVDD pin is pulled to the ground by C8. The issue wasn't experienced with VREF setting, VREF duty dycle was 68%, which is equivalent to a current of 1.7 A, VM supply voltage of 24V. Also note that the VREF voltage amplitude is 3.3 V, TOFF, ENABLE, STEP, DIR, etc - 5V.

We ran a bench setup 1.7A (VREF = 3.3V), VM = 24V, Full-step mode, no PPS, and ambient temperature was 23degC with a 1.7A maximum current rated stepper motor. The decay mode was at the default setting, Smart Tune Dynamic Decay, TOFF = 7us.

I doubt the PCB design is a problem. We use a 4-layer board with large polygons of ground. Perhaps the above notes are the cause of our problem?

Regards, Vladimir.

Vladimir,

Since I designed the DRV8424EVM hardware. Let me jump in. The footprint shows no enough thermal vias underneath the device thermal pad. Please use EVM footprint as reference. https://www.ti.com/lit/zip/sloc356

For power loss, would you set TOFF HiZ or 330kohm to ground?  So, we can have 24us or 32us TOFF. That can reduce the switching loss. it can help us estimate the DRV8424 thermal resistance in your PCB. If the thermal resistance is too much off from the datasheet spec, we need to update the device footprint and check the thermal via (is it a blind via to block solder paste flowing in?)

Regards,

Wang Li

Hi Vladimir,

I will refer just to power losses and output power.

1. I think that losses in datasheet were calculated for Rds at 25C and 30kHz, at higher junction temperature Rds is higher so the conduction

losses, switching losses are proportional to switching frequency. Conduction losses are not the only ones that rise with temperature but

the major ones.

2. Real (active) output power for single phase would be equal Irms x Urms x cos(phi), it looks well only for motors with sinusoidal current

and voltage like induction motors powered from electrical grid. For stepper motor and PWM driver it is more complex. Current is not sinusoidal

at full step, voltage is PWM waveform with variable duty cycle not to mention power coefficient. Current probe and oscilloscope with some

power analysis function would be needed to measure driver output power.

If you can tell something more about "2*Idr.out*Udr.out" in your equation then we can discuss if it is correct.

Regards,

Grzegorz