DRV8870: Operation stability problem

Hello Jeon,

Thank you for providing detailed description of your problem!

The first experiment to do, if you haven't done so already, is to remove the varistor between the outputs. Let's remove this variable for now and run your test again to check if issue is resolved.

I have a few other questions:

• Are you able to recreate the faulty state consistently? Or it just happens at random? 
• Is the driver and motor in a high ambient temperature environment? If so, what is the temperature

Regards,

Pablo Armet

Hello Pablo, 

It just happens randomly and unfortunately I can't reproduce the faulty state.
In the abnormal state, I checked that the input PWM signal and VDC voltage are normal.
In the abnormal state, the internal temperature of the equipment was about 25 degrees, and it may rise to 40 degrees during operation.
Are there any problems with the circuit or artwork?

Il-tae Jeon said:

There is only no power output on OUT1 and OUT2.

Post a scope capture of out1, out2, and IN1=50%.

Also, what is the peak motor current in the application (the current limit is set to about 2.7A)?

Il-tae Jeon said:

Motor:

DC40V, 35,000RPM, 180mA(no load)

Brian

This is the scope capture you requested. A video was also uploaded.
This is a capture from motor normal operation.
Unfortunately, I didn't have a current probe so I couldn't measure the motor peak current.
For reference, it works as a slow start like the video.

Looking forward to your reply.

ch1: IN1, ch2: IN2, about 50% 

OUT1, OUT2 at ablut 50% duty

IN1, IN2 Start 

OUT1, OUT2 start

Hey Jeon,

Thank you for the waveforms, I will aim to provide a response early next week.

Best,

Akshay

From the video, it seems the motor takes 4sec to accelerate while the pwm control input is ramping up in less than 1sec. Why not profile the pwm acceleration to match with the motor load to avoid placing a big stress on the driver at stalled motor current? There is no point to accelerate the pwm too fast while the motor load needs longer time to turn. In the field, how is the motor being used -- continuous running at fixed speed, or stop and go with fast acceleration?  

I suspect the driver shut down due to overheat protection.

Brian

Hey Jeon,

This device has three protection features that would disable the outputs;

1) VM UVLO

2) OCP

3) Over temp shutdown

Since VM is stable we can eliminate VM UVLO. So we are either looking at OCP or over temp shutdown. 

It would have been helpful to know the output current to see what is happening. 

Thermal shut down is also possible, so if you have a thermal camera you can try to capture the case temperature to see if it comes near 150C during operation.

As Brian mentioned slowly ramping the PWM might help limit driver stress. Also, how frequently is the issue occurring?

Best,
Akshay

Hi, Brian thanks for your kind reply.

The user can use it after setting the speed up to 35 steps.
In addition, it can be used while adjusting the speed according to the analog input signal.
Quick responsiveness is important, but if it puts stress on the driver, I'll tweak it.
In the abnormal state, it cannot start at no load, and the surface temperature of the driver is not hot at all.

Hi, Akshay thanks for your kind reply.

When I listen to the problems in the field, most motors won't start.
We haven't heard of any freezing issues while running yet.
When I tested it in the lab, the driver surface temperature never exceeded 70 degrees under normal load operation.
When the problem occurred in the field, the surface temperature of the driver was not hot at all, and the voltage across output 1,2 was 0V.
Even if I set the motor speed lower, the result was the same.
After rebooting the power, about 90% works normally. The remaining 10% worked fine after a few hours of reboot.

Hey Jeon,

1) Has this device been in use for a while and the field is now noticing issues? Or is this a new design?

2) How often is this happening in the field? How often did you notice this in your lab experiment?

3) Were you noticing these issues after removing the varistor between outputs? We want to limit the variables.

If overtemp shutdown is not happening then it could be OCP, but its  hard to tell without looking at current waveforms.

Best,
Akshay

Jeon,

Please do this when a driver no longer working:

Keep the power on,

Disconnect the motor from output to eliminate the condition for OCP,

Drive the Inx such that as you want to run the motor

capture the waveform of IN1, IN2, Out1, Out2, VM and label the signals clearly so we can understand them.

Brian

Hi, Akshay

1) We produce 500pcs each, and there were no complaints until the 2nd lot. This is a problem that occurred in the third production lot.
No hardware or program changes.
2) I checked it once in the lab, but I couldn't reproduce it. In the field, about 5 to 10% of complaints were received, and there are concerns that it will increase further.
3) I haven't tried removing the varistor yet. I'll try. However, since it is not reproducible in the lab, can we verify the improvement?

Hi, Brian

I checked the signal below with a scope in the field where the problem was found.
1) IN1, IN2 : ok
2) VM: ok
3) OUT1, OUT2 both ends voltage: 0V
4) Even after removing the motor, the OUT output voltage was 0V.
5) No driver surface heating.
6) Forward and reverse conversion: no spin
7) Normal operation after power reboot
Unfortunately, capturing scopes back in the field isn't easy.

Some members also doubt the possibility that the received parts may be defective.

Il-tae Jeon said:

1) IN1, IN2 : ok
2) VM: ok
3) OUT1, OUT2 both ends voltage: 0V
4) Even after removing the motor, the OUT output voltage was 0V.
5) No driver surface heating.
6) Forward and reverse conversion: no spin
7) Normal operation after power reboot
Unfortunately, capturing scopes back in the field isn't easy.

Well, there are only a few reasons why the driver has disable output:

* OCP: but without motor connected, so this is not,

* Under and over voltage: VM is good, so this is not,

* overheat protection: not this case

Then I don't know what else. 

I understand the hassle of capturing the waveform in the field, but without the seeing the waveform then I'm afraid there is nothing else for us here to help. To capture the waveform, I have the partable Picoscope connected to a laptop via USB and save the waveforms to the pc for document or sharing.  Keep in mind that what you see on the waveform is normal, but other people might spot the issue in them.

Brian 

Hey Jeon,

Thank you for the additional information.

If we are unable to see any of the three reasons why we get no output then it gets a bit tricky to diagnose. So having waveforms from the field will be needed to identify issues. 

What is the reason to suspect defective parts? If 2 lots of 500 each worked and issues are coming in lot 3, then it would be interesting to see if the parts were acquired from the same sources.

Best,
Akshay

Unfortunately, we haven't been able to fix the problem yet.

The V_dc supply voltage is approximately DC 43V. Since it is supplied through a transformer, what problems can it cause if the power is supplied over 45V?

I purchase new parts and produce in small quantities (200pcs).
I would like to know if there was an issue in which the DRV8870 was reported as a clone.