DRV8243-Q1: Driver makes noise with duty cycle <80% and does not move the motor

Hi João,

Thank you for you post. You mentioned "The motor moves, without any noise, when the duty cycle is bigger than 80%, but if the duty cycle goes bellow 80%, the motor stops moving and the driver starts making noise.". I assume you mentioned the motor made a noise but did not spin, correct?

You also mentioned "We try to use a less powefull power supply (3.3 A) but the results were the same. We checked that the motor draws a little more than 20 A when locked, but since the power supply only allows 20A max, the voltage never goes bellow 24 V and we are using the ITRIP function to regulate the current to 11 A max, we though that we shouldn't have any problem (we are waiting for the DRV8244-Q1 (21 A) to be available on our local market).". This may be the root cause. Based on your description it looks like you are testing the BDC motor with actual load conditions. By nature BDC uses the inrush current to generate a peak torque to overcome mechanical inertia and start spinning. When you limit the peak current using ITRIP and reduce the duty cycle at the same time the motor cannot generate sufficient torque to overcome inertia. The ITRIP regulation causes noise due to chopping as well as the motor stalls with duty cycle below 80% when trying to start. The solution to overcome this issue due to mechanical inertia is to increase the ITRIP to a suitable level. You should consider DRV8244-Q1 or DRV8245-Q1 as your option if the DRV8243-Q1 results with over temperature shut down with the higher current needed. I hope this clarifies the cause for the motor behavior experienced.

Regards, Murugavel

Hi João,

Could you please elaborate the circumstances under which the motor does not run and makes noise? Thanks.

Regards, Murugavel

Hi Murugavel,

Sorry i was writting this answer. Thank you for your support.

Answering directly to your question, the motor stops moving below 80% duty cycle (pwm 25khz, 3.3V). In concern of noise, it comes from the driver itself wwhen the motor stops moving.

I try to follow your sugestion but the driver continues to make noise when I try to move the motor bellow 80%. The IPROPI goes to max and nFault pin goes low. It's a normal behavior from a driver from this size to produce a relatively loud noise or are we doing something wrong?

I forgot to mention but I have tested the driver with load and without any load and all stays the same. Before using this drive, we test the motor with a simple MOSFET P and N channel H-Bridge, and we were able to start moving it with 40% of duty cycle (no load) or 60% (with load) using a power supply capable of only 3.3 A output current.

Regards,

João

Hi João,

Sure, no worries. 

I think the fault reaction for temperature may be set to retry. This will cause constant on/off at the retry rate see specifications from the datasheet. If you set it to latch it will latch on the first fault event.

Regards, Murugavel

Hi Murugavel,

I think you are right from the perspective of the retry/latch function, as after I disable the retry function, the motor didn't move even with 100%, and when retrying, it takes about 5ms to retry again. (the picture below shows the motor output while driver is making noise).

We also saw that the IPROPI was at maximum value, does this not indicate that the there is a output short to ground? 

The question is that the temperature never goes above 35ºC (we a have a ntc resistor placed near and we also use a infrared temperature gun to measure it). Can the temperature increase and decrease so fast that we can't detect it? It's a pcb poor conception?

We try to force the driver to maximum load (way above the expected operation) at 100% duty  and it works fine for about 2 minutes (our firmware then stops the motor if the driver temperature goes above 60ºC)

Regards,

João

Hi again,

Sorry, I forgot to mention that that happens even if there is nothing connected to the motor output (exactly the same behaviour), almost like if the driver couldn't handle a pwm signal.

Regards, João

Sorry to bother you again, but I have more data that could be useful. Below there is a picture of the input pwm signal and output signal from driver to the motor. As you can see, even with a duty cycle of 80% at the input, the driver output goes always to 100%.

Hi João,

I'm glad you were able to confirm the noise issue was due to retry operation of the driver. Thanks for further details. You mentioned "I forgot to mention that that happens even if there is nothing connected to the motor output (exactly the same behaviour), almost like if the driver couldn't handle a pwm signal.". I'm unable to reproduce this issue with an EVM with this device. It appears the device on your board may have an output short to GND internally. So when the HS FET is conducting with PWM input the device shuts down. The high output on IPROPI observed also corroborates this possibility. 

You also mentioned you monitored the temperature with an NTC thermistor. It is possible the output FET can heat rapidly and shutdown and cool down fast enough for the thermal inertia of the system to not register a high temperature. This does not mean the PCB design is poor. The power FET may just heat up rapidly before all the heat is conducted and dissipated to the power pad and the pcb, sort of an internal rapid hot spot due to high current density.

You may want to replace the driver device and try again with a current limited power supply. Thanks.

Regards, Murugavel  

Hi Murugavel,

I followed your suggestion and replaced the driver device and tried again with a less powerful power supply (3A max only). The behavior stays the same as you can see bellow (it's the same without connecting or with the motor connected)

Until the 82% the driver makes noise and retries

After that, at aprox. 83% driver starts recovering as you can see below:

(the results are the same with or without the motor)

It's normal that the output never is a pwm signal?

Regards,

João

Hi João,

Thanks for trying this and reporting back. Could you please verify if you have the correct bridge control MODE selected for your control? See below table from the datasheet.

Like I mentioned I do not see this issue with my EVM. Did you get the devices from estore.ti.com? Could you please take a close up picture of the IC to clearly show the markings on it so I can look up the version of this device? Do you have a TI DRV8243H-Q1EVM with you? 

Regards, Murugavel

Hi Murugavel,

I test both PWM mode (Hi-Z - ESP_MODE pin open) and PH/EN (ESP_MODE connected to GND) and both worked as expected until I try a PWM signal.

I don't have any EVM with me.

You can check the IC ref. below (we got the devices from mouser electronics)

Once again thank you for your support.

Regards, João

Hi João,

Thanks for the update and the information about the chip. Based on your schematic RIPROPI is 330 Ω. I used the 470 Ω in the EVM for testing. ITRIP was set to LVL2 which amounts to 1.18 V VTRIP. This would mean 1.18/470 * 3075 = 7.72 A for ITRIP setting. No current regulation will happen with < 7.72 A output current. But this current may be above the continuous current that can be supported by this device at 24 V. 

I connected a 24 V supply to the EVM with PH/EN mode selected for testing purposes. I did some scope captures for you for reference. Yellow trace is PWM input on EN/IN1, pink trace is OUT2 and blue trace is OUT1. The green trace represented motor current. One of the key factors to keep in mind is the slew rate SR setting. I used the LVL1 which set to 24.5 V/us (faster).  

Trace with no load 10 % duty 20 kHz PH/IN2 = 1:

Trace with free running motor load 10 % duty 20 kHz PH/IN2 = 1:

Trace with no load 90 % duty 20 kHz PH/IN2 = 1:

Trace with lightly loaded motor load 90 % duty 20 kHz PH/IN2 = 1:

The driver performed as expected. No over current tripping or over heating happened as expected. The schematic diagram you sent initially looks okay. Only the motor power supply bulk capacitor was missing. TI recommend at least 100 uF in addition to the 100 nF you have. You may need larger capacitors based on motor load current and the supply ripple as well as length of the wires from the power supply to the driver.

Based on the full-bridge thermal estimator which you can download from https://www.ti.com/tool/download/SLVRBI3  you can see the SR setting significantly impacts the maximum load current supported by this package at 24V.

With slowest SR setting LVL2 1.6 V/us even 1 A results in over temperature shutdown. Based on your first message this is the SR you were using. I should have caught it right away. However with no load connected this should not cause any problem. I verified this to be true with my EVM.  

With the same SR a 5 A load would cause a super hot spot and thermal shutdown.

With SR set to LVL1 24.5 V/us (faster) you can see even at 5 A load current the device may operate in the safe region assuming good PCB design.

Please set the SR to LVL2 and run the tests again. Note, the SR setting will be updated only after you power cycle or do a sleep to awake sequence after setting the required LVL.

Regards, Murugavel

Hi Murugavel,

Thanks for your answer.

I try to change the SR settings, but the results were the same.

We can't understand why our driver goes into 'protection' when it has nothing connected. 

We start suspecting that is something with our PCB layers and layout. Can you give us your opinion about that?

This is one layer (red one)

And here the other one (blue one). (we only have two layers on our pcb)

Regards, João

Hi João,

I do not expect the driver go to protection mode when no load is connected even with this two layer PCB. The current through the output FETs would be 0, so not heat generation is expected other than heat generated by the device . I did not notice anything unusual in the schematic. Could there be a short in the PCB either layout or excess solder causing this issue? I think this should be verified. Can you measure the current through the device's VM input by connecting a multimeter current measurement mode in series with it? The EVM I have measures 12.8 mA with 24V VM supply both with 0% and 100% duty cycle with 20 kHz PWM with no load connected. This is for your reference.  

In general the thermal performance of the 2-layer would not be as good as the TI EVM for this device. The Altium files for the EVM can be downloaded from the EVM webpage. With 2-layers I still expect some level of continuous output current to be supported. My recommendation would be to use the top layer with copper pour maximized from the thermal pad to dissipate heat faster. I think in your layout blue is the bottom layer that has the maximum pour which is not as efficient as it dissipates the heat conducted through the thermal via-s. If blue was your layer contacting the thermal pad please ignore this comment.

Regards, Murugavel

Hi Murugavel,

We test the PCB and the driver, and we could not find any short, but, when we use the PWM without connecting anything, we detect spikes of current (about 250mA or more).

The only thing that we see that can be causing problems on the driver may be the soldering process. I could not find the soldering profile on datasheet. Where can I find it?

Regards, João

Hi João,

Spikes of current > 250 mA alludes to something not normal likely could be a short happening with the output pins to GND. The referenced application report in the datasheet should cover this. https://www.ti.com/lit/an/slma002h/slma002h.pdf. 

Regards, Murugavel