DRV8350F: SPI Signal Problem

Hi Brian,

Thank you for posting to the Motor Driver’s forum!

I will aim to provide a response by the end of next week.

Best,

~Alicia 

Hi Brian,

Regarding your first question about why there is a response on SDO when sending a write command on SDI, what you are seeing is the data that is currently stored in the register that is being accessed.

Regarding your second question, what is being done differently in the second image that the output from SDO looks so different from the first?

Best,

~Alicia

Thanks. For the question 2, it has been solved as customer found it's caused by the Levels shift enabling signal. 

For the previous question 1, RD thinks when SDI is in process to set the registers of 0x07h - 0x02h by writing values, the SDO should be no any response. But, you can see there are response on SDO while SDI is in writing the data. It could not affect the performance, but we're curios how come.  

There are following questions. 

1. If SDx and SNx are not used, how to deal well with these? Floating or connected to GND?

2.  On the below waveform, while high/low side FET is in switching zone, the SOA pin could see either spike voltage or drops to 0V. Why? How to improve it?

The switching frequency is 10KHz. 

CH1: SOA CH2: Gate_Low side FET CH3: Voltage sensing on Shunt CH4: Gate_High Side

Hi Brian,

Brian Wang0928 said:

RD thinks when SDI is in process to set the registers of 0x07h - 0x02h by writing values, the SDO should be no any response.

Regardless of whether you are doing a read/write, you will receive a response from the SDO as briefly mentioned in the datasheet below:

Brian Wang0928 said:

1. If SDx and SNx are not used, how to deal well with these? Floating or connected to GND?

Sections 5. Unused phases, spinning brushed motors and 6. Unused CSA of the following FAQ describe what to to in this case: [FAQ] Quick Guide to Debugging Common Issues in BLDC motor drivers

Brian Wang0928 said:

2.  On the below waveform, while high/low side FET is in switching zone, the SOA pin could see either spike voltage or drops to 0V. Why? How to improve it?

I will aim to provide a response on this question by the middle of next week.

Best,

~Alicia

Customer has tried to adjust the IDRIVE to be smaller, it improves the behavior as waveform I offered before. 

Two questions here:

1. When IDRIVE is adjusted to be smaller, is there side effect? Should we need to adjust switching frequency to be smaller?

2. One question, when frequency is setting at 10KHz, is there suggestion of the IDRIVE value?

Thanks. 

Regards

Brian

Customer has improved the spike voltage already. But there is still a issue as below waveform. the SOx is pulling to ground during switching. is there any effective way to solve it? 

Hi Brian,

I will aim to provide a response by the middle of next week.

Best,

~Alicia

Could you help on this inquiry! Very thanks. 

Hi Brian,

Would it be possible to get a waveform monitoring the following channels: GHx, GLx, SPx, and SOx?

 What could be happening is that there may be some negative transience occurring that may be causing the behavior the customer is seeing.

Best,

~Alicia

Customer found the behavior difference is on the filter design. 

On the EVM board, there is no resistor in series connected. In customer board, it put two 10 Ohm for differential signals. 

After removing off the ER214/215/216/217, the problem is gone. Could you help to comment how to explain this behavior(Sensing voltage dropping to 0V)?

Hi Brian,

I am glad the customer was able to figure out what was causing the problem.

I will aim to provide a response by the end of next week.

Best,

~Alicia

Got it. Looking forward your comments. thanks. 

Hi Brian,

When you say the customer removed the ER214/215/216/217 resistors, do you mean they just depopulated it and left it open or did they replace it with a 0ohm resistor?

Also would it be possible to see the snippet of the schematic which includes the connection to the DRV?

Best,

~Alicia

Hi Brian,

What are the values of ER210/211/248/242?

Best,

~Alicia

ER210&211：0.5mohm

ER248&ER242：10kohm

Pls help to offer comments. Very thanks. 

Regards

Brian

Hi Brian,

Are the GX/GY nets tied to the gate of the MOSFET? If so, after looking further into the schematic, it appears that the customer uses ER242/ER248 to essentially tie the gate of the FET to the high-side of the shunt resistor.

I would be curious as to why the customer has chosen to design the system in such a way?

Best,

~Alicia

Customer mentioned that the Resistor located between Gate and source pin of the MOSFETs are common used in that team. The purpose of this way is to enhance the noise immunity against the FETs been influenced by noise. However, it's first time that the current sensing topology is using low side current sensing for cost down e-bike application. The general ebike is using high side current sensing. 

Is there any problem when these resistors are used between gate and source pin of the MOSFETs? Is that related to the SOx dropping problem?

Regards

Brian 

add up some measurements with ER214/ER215/ER216/ER/217 and ER242/ER248

Please find waveform with different configurations。 as a result, the SOx voltage dropping out is happened when ER214/215/216/217 are 10 Ohm. it seems the issue has no relation with ER242/248. Pls help to clarify it how the filter could affect the behavior of SOx.   

Probe: Yellow: SOx, Red: SOx after pass though a TVS. 

Diagram 1:ER214/215/216/217：10 ohm ；ER242/248 ：NC

Diagram 2: ER214/215/216/217：0 ohm ；ER242/248 ：10k ohm

Diagram 3:ER214/215/216/217：0 ohm ；ER242/248 ：NC

EC217/EC218 = 1000pF

Regards

Brian

Hi Brian,

Let me look into this in more detail and get back to you by the end of the week.

Regards,

Anthony 

Hi Brian,

My apologies for the delay in getting back to you. I don't see an issue with the customer using a GLx to SPx pulldown resistor for the gate of the low side MOSFET. 

Could the customer see if keeping the 10 ohm resistors and removing the EC217 and EC218 changes the behavior at all? Also, could the customer probe the voltage differential between PAGND and PDGND to see if there is a significant noise between the grounds during MOSFET switching? 

Are the MOSFETs on the same board as the driver? Or is there a separate board for the MOSFETs? I didn't see any MOSFETs in the schematic provided. 

Regards,

Anthony Lodi