DRV8825

Hi Huiqi,

I guess this post is a continuation of previous thread by Jian.

I would check if DIR signal meets its timing requirements (Figure 1. Timing Diagram), I would also check if DIR/STEP signals are clean. Could you provide oscilloscope waveforms for DIR and STEP signals during direction change, 2 cases for forward->reverse transition and 2 cases for reverse->forward transition.

If you could also do the same for both phase currents.

Regards,

Grzegorz

Hi Huiqui,

Thank you for posting to the forum.

Please allow us 24 hours to research your questions and get back to you with a reply.

Yeah, you are right.

That thread is closed, but the root cause is not clear yet. We still want to make some investigations on this issue.

The DIR/STEP signals are clean, and the STEP signal width is bigger than the minimum limit.

The current at the beginning of a move is shown below.  

The high level pulse duration of the STEP signal is shown below, 14.3us in this case.

Hi Huiqi,

I would check if input signals meet all other timing requirements from Figure.1 not just STEP signal duration.

Waveforms of phase currents at the beginning of a move show that current is turned on just at the start of a move. By experience I know that it is a good practice to turned on phase currents some 50-100ms before giving STEP signals, at least it gives good results for Nema 24 motors.

Grzegorz

Hi Grzegorz:

It is very difficult for us to measure enable and direction signal because there is no test point on the board. So I can't provide those signals to you.

Meanwhile, the enable signal is always on. The move in our test is continuous forward/backward movement, the working current will be kept for 100ms after one move is complete, the movement is started within 100ms after previous move is complete, so the current will not drop down to the holding current in our test, so the phase current is still working current.

The other thing is the problem does not occur within short time, it happened after long time run, 24+ hour or more than several days. The problem was only found on some of the devices in one production batch, while the problem did not exist in another production batch, i.e. there is production batch difference.

Best Regards,

Huiqi

Huiqi,

Can you please send us a clear picture of the top side marking of the devices that you are having issues with and a picture of the devices that have no issue?

Step timing looks to be long enough at 14.3us.  Is it always that long?  

Finally, have you tried replacing a mishaving part on the board with one from the production batch that has no issue?  This would tell us if the issue is following the PCB or device in terms of issues.  

Regards,

Ryan

Hi  Ryan:

Please check the top side marking of DRV8825 shown below.

As to the duration of the STEP signal, it is always longer than 14.3us.

We replaced the DRV8825 several times, and at least 3 batches worked, but another 1 batch did not work.

Best Regards

Huiqi

Hi Ryan:

To investigate the problem, we purchased the DRV8825 from 2 sources, the chips from one channel is imported from Dallas, while the other is from Singapore. But the mark on the top the chips seems to be the same. Check the photos below.

The question is if they are same production batch that manufactured by the same factory.

From Singapore:

From Dallas:

Best Regards,

Huiqi

Huiqi,

To DRV8825, step loss means the indexer doesn't move to next step current setting with one step pulse. I didn't see any waveform to show that.

"Meanwhile, the enable signal is always on. The move in our test is continuous forward/backward movement, the working current will be kept for 100ms after one move is complete, the movement is started within 100ms after previous move is complete, so the current will not drop down to the holding current in our test, so the phase current is still working current."

To debug this, can we use the abnormal sample to only run in a signal direction and compare it with normal samples? So, we can check if it happens with direction change.

If it happens with direction change, we can zoom in the waveform to check DIR, STEP and current waveform to confirm it?

DRV8825 has nHOME pin to indicate the home position. What application needs to keep all the samples having the same current phase for continuous forward/backward movement long time? If that is the case, can we use nHome pin to resolve the issue?

Regards,
Wang Li

Hi Ryan:

Do you have any findings on the mark?

Best Regards,

Huiqi

Hi Wang Li:

In our tests, we found the phase current signal at the end of a move changed after long time running, while it is same in the devices that working correctly. Please check the pictures on the top of this page.

The problem can happen in any places, it has not direct relation with the direction change.

As to the nHOME usage, can you elaborate how to use it in our case.

Thanks

Huiqi

Hello, Huiqi,

"The problem can happen in any places, it has not direct relation with the direction change."

Does that mean you have verified the phase current signal at the end of a move changed after long time running, even the DIR pin voltage not change?

nHome pin logic low indicates the home state of step table (45 degree angle, see datasheet "Table 2. Relative Current and Step Directions"). So, we really want to make sure the motor have the same phase current signal after long run. We can check nHome signal and compare it with STEP pulses number to make sure no step loss. If the host controller finds nHome not matching STEP pluses, send extra pulse to make nHome pin low.  and then recount STEP pulse number.

Regards,

Wang Li

Hi Wang Li:

Yeah, to our understanding, the phase current signal changed over long-time running, while the DIR voltage has not obvious change.

The number of STEP signal is correct. 

Best Regards,

Huiqi

Hi Wang Li and Ryan:

As to the two questions I asked before, can you give me a reply? (Please check the questions posted 2 days ago.) Thanks in advance.

1. As to the top side marking from the malfunctioning board and functional board, do you have any findings?

2. Are those 2 batches of DRV8825 from different purchasing channels, but with same top side marking, the same batch manufactured by one factory?

Best Regards,

Huiqi

Hi Wang Li

Question 1, the top side marking of malfunctioning board and functional board, please check the picture below. The green board is the functional one, and the black one is malfunctioning one.

Question 2, do you mean they are the same batch from specific factory?

As to the nHome signal, let us have a look.

Best Regards,

Huiqi

Huiqi,

Wang is handling this post and will respond soon.

Regards,

Ryan

Huiqi,

1. The DRV8825PWPR top marking should follow the following format. Green board sample date code should be 2/2021. Black board sample date code should be 5/2020 or 6/2020. I cannot find anything wrong from the top marking.

2. The green board and black board samples have different date code and different ASSY LOT code. So, they should be come from different lot or batch. 

If DRV8825 STEP pin holding time (timing for 2 and 3 in the following picture) is longer than 2 internal clock cycles (typical 1.25us/cycle), DRV8825 could recognize the "STEP" signal and indexer moves to the next step setting. Since the STEP signal holding time is 14.xus, I would think the issue may be caused by the STEP noise. So, besides the nHome workaround solution, would you try to add a small capacitor on STEP pin to see if it can help?

Regards,

Wang Li

Hi Wang Li:

Thanks for the reply.

Maybe I did not describe the other question clearly. It is about 2 batches of chips purchased from 2 sources, but with the same top marking. The snapshot of that question is shown below, and you can find the original question in previous conversation.

Best Regards,

Huiqi Li

Huiqi,

The above Dallas sample and Singapore sample have the exactly same top marking. I would think they are same production batch that manufactured by the same factory.

I would think adding a small capacitor (1nF) on STEP pin could be a solution. Please try it if possible.

Regards,

Wang Li

Hi Wang:

As to your suggestion, we have some discussions. We think it is unlikely that the step loss is caused by the noise on the STEP signal. The reasons are:

1. The step loss problem is found in the early of this year, but the product has been manufactured for almost 1 year, and no similar problem is found before that.

2. After the problem is found, we did some tests to check the DRV8825. The test is to change the DRV8825 only (by another batch), with the same board and same firmware, and no problem showed up. (NOTE, there are 3+ batches of DRV8825 which can work correctly in our test, and no step loss found.)

3. Based on our test with logical analyzer and oscilloscope, no loss of STEP signal was found and the width of STEP signal is almost more than 14.xus.

Based on the information listed above, we think the STEP signal will not be the root cause of this problem. What is your opinion, and what else can cause this problem from your perspective? Thanks.

Best Regards,

Huiqi Li

Huiqi,

For 1,2, the device performance difference could be caused by lot-by-lot variation. I didn't see this issue before and cannot be 100% sure what happens. If want to check the device getting damage or not, please send it to our quality engineer.

3. 14.xus is longer enough. But, if a noise break in the middle, I am not sure DRV8825 would account more STEP pulses or not. That is why I want to add a small capacitor to verify it.

When we implement nHome solution, we can use it to catch the missing step or extra steps. Before we have nHome solution, I would try to add a small capacitor on STEP pin.

Regards,

Wang LI

Hi Wang Li:

There is another question about the decay mode. No step loss is found in mixed decay mode, but that is not the case for fast decay mode.

Can you elaborate how to determine the decay mode, like the running speed, stepper model, and etc.? The stepper we used is MS17HD2P4150.

Best Regards,

Huiqi Li

Hi Wang Li:

Last time you told us "If want to check the device getting damage or not, please send it to our quality engineer." We think it might be helpful to let TI's quality engineer examine the chips. Please let me know how to contact with your quality engineer send the chips to him/her. Thank you in advance.

The other question is current DRV8825 we purchased from 2 sources, Singapore and Dallas worked correctly. As you said, they are the same batch manufactured by same factory. To guarantee the quality of the product we manufactured, we hope to purchase this batch of DRV8825 in the next few months, so that we do not need to worry that we may encounter the step loss problem again, can you tell us how to do that?

Best Regards,

Huiqi Li