DRV8301, DC_CAL floating, so question..,

Edward Averill

Other Parts Discussed in Thread: DRV8301

We have the first spin of our motor control board, and while things appear.. ok.. we're seeming some weirdness when we run the motor.. specifically it seems to consistently miss one Hall state (making the motor spin rough).

Our hardware guru (new hire) probed the OCTW pin and we're seeing a burst of pulses at regular periods. There's no motor shutdown, but we suspect that it's dropping one PWM cycle during this burst and so we lose that Hall state.

The new hardware guy noticed that DC_CAL is left floating and not tired to ground. We'll fix that next spin, but could that possibly cause what we're seeing? The DRB8301 has an internal pulldown so I wouldn't think this was the issue..

All help, manual references, recommended reading appreciated!

Ed Averill

over 9 years ago

0 Nicholas Oborny over 9 years ago

TI__Mastermind 21305 points

Hi Ed,

DC_CAL should be ok floating. The pull down should be sufficient. It would also only affect the current amplifiers.

Can you provide any scope shots of the bahavior on OCTW alongside the motor current?

OCTW indicates the device is entering thermal shutdown (unlikely) or you are having an overcurrent event.

What type of MOSFET are you using, what is the VDS trip level, what is the OCP mode, and what current are you trying to drive?

If the error consistently occurs for a specific hall cycle is there some issue (short, etc) in the MOSFETs or motor that only shows up during that cycle.

0 Nicholas Oborny over 9 years ago in reply to Nicholas Oborny

TI__Mastermind 21305 points

You can refer to the DRV8301 datasheet starting on page 18 for more information on the overcurrent protection.

0 Edward Averill over 9 years ago in reply to Nicholas Oborny

Genius 3100 points

Thanks! My hardware guy is in a meeting but as soon as I can I'll gather the infor and pass it on.

Ed Averill

0 Edward Averill over 9 years ago in reply to Nicholas Oborny

Genius 3100 points

Ok, I can tell you that the OCP and VDS trip levels are whatever the DRV8301 defaults are as this spin of our board has SPI issues. We're working on screenshots, current measurements may be a problem as we don't have a current probe, but we might be able to rig something up.

We have a sensing resistor of .002 ohms, all DRV values at default, do you know if this puts us out of range? (HW guy question)

Here's a shot of the DRV8301 portion of our board:

Again, thanks a TON for your help!

Ed Averill

0 Edward Averill over 9 years ago in reply to Edward Averill

Genius 3100 points

Ok, we've since replaed the source resistors with .01 ohm ones and removed the feedback resistors to take that out of the equation, AND replaced the motor, and still no joy. We're just not seeing Hall State 3 ever entered, but 1,2,4,5, and 6 have expected counts. Very strane.

0 Nicholas Oborny over 9 years ago in reply to Edward Averill

TI__Mastermind 21305 points

DRV8301 VDS default is 0.403V. That is ~200A with the MOSFETs you have (~0.002Ohm).

What state is Hall 3? Can you try enabling the MOSFETs that would be ON in this state without a load to see if there are any issues?

0 Rick Duncan over 9 years ago in reply to Nicholas Oborny

TI__Guru** 114905 points

Hi Ed,

A trick I have used in the past (assuming it is safe in your application) is to disconnect the motor phases and leave the hall sensor connectors in place.

The motor can then be spun by hand to watch the 6 commutation states, and the outputs. This can go a long way to debugging the problem.

0 Jeff Salisbury over 9 years ago in reply to Nicholas Oborny

Intellectual 345 points

We are currently spinning the motor with a drill and/or by hand to see if Ed gets 3 A, 3 B and 3 C highs with 6 stages per cycle. We aren't getting the hall counts in yet.

Thanks for the info. So with the .01 now in place we should still not be tripping over current till approx 40 A so it is not an overcurrent trip on the OCW pin we see.

0 Jeff Salisbury over 9 years ago in reply to Jeff Salisbury

Intellectual 345 points

*In evenly yet anyway. We see some states. We have verified the output of hall buffer chip into the MCU and they look good.

0 Nicholas Oborny over 9 years ago in reply to Jeff Salisbury

TI__Mastermind 21305 points

Jeff,

The Vds overcurrent protection depends on the Vds trip point set in the registers and the MOSFET rds(on). The monitor is watching the voltage drop across the MOSFET when it is ON. It is unrelated to the current sense resistor used for the amplifiers for a precise current measurement.

0 Jeff Salisbury over 9 years ago in reply to Nicholas Oborny

Intellectual 345 points

Thanks for the clarification between which .002 ohms! ok, so it thinks it is seeing 200 amps and that is why it is tripping or it wont trip till 200 and we need to cross that off the list for the source of the OCTWn fault?

0 Nicholas Oborny over 9 years ago in reply to Jeff Salisbury

TI__Mastermind 21305 points

Yes,

If you are "Current Limit" mode (default register setting) (page 18 datasheet) and OCTW is toggling then the device is entering either thermal shutdown (unlikely) or hitting the current limit trip point.

You can refer to the fault table to see how I ruled it down to these two (page 20 datasheet).

Since it appears we are hitting the current limit trip point and the limit is very high (200A!). This points to some sort of short in the half-bridges or motor.

It would be good to manual toggle each half-bridge without a load and ensure they are functioning properly (no OCTW reporting).

0 Nicholas Oborny over 9 years ago in reply to Nicholas Oborny

TI__Mastermind 21305 points

Can you also try another PCB to rule out a possible manufacturing issue?

0 Jeff Salisbury over 9 years ago in reply to Nicholas Oborny

Intellectual 345 points

ok, cool. i see the 2 rows with n, n, y and dont care yet because of our spi not working. We have one last board with mods currently made up and "trying" lol to hold off smoking it as we have melted the C phase insulation a few times trying to get the hall counts.

ok so the trip set point is set by .403 x the fet rds of .002 to equal the 200 section 9.2.2.2 and the actual measurement it is comparing to the value set in 9.2.2.2 is from 8.3.4.1. Vds = Ids * Rds(on) so 201 A * .002 ohm = .4 volts meaning if there is more than .4 volts between lines SN1 and SP1, baam, trip? About to give it up for the day and attach fresh in the morning

0 Nicholas Oborny over 9 years ago in reply to Jeff Salisbury

TI__Mastermind 21305 points

Almost right,
The values are right but the measurement is taken across each MOSFET not through the current shunt amplifier. You can see this outlined in the block diagram (page 14). The HS and LS Vds monitors.

0 Jeff Salisbury over 9 years ago in reply to Nicholas Oborny

Intellectual 345 points

Sooo much clearer today! Thank you for getting us over our hump. We have a 10 amp power supply so it seemed the only logical answer to the voltage across the on fet being higher than .403 would have to be if the fet wasn't fully on. I recalled looking at a really crappy signal to the gate a while back. I just looked at section 9.2.2.1 for the gate drive average current load.
This section states a the DRV8301 can deliver up to 30 mA. Our IRFS3006 fets have a 200 NC Qg rating!!! this equals out to 54mA!!!! TIMMMYYYYYYYYYYYYYY
so the over current is related to the gates not fully being turned on.

In your opinion this would be a part that defiantly needs to be changed in the design?

0 Jeff Salisbury over 9 years ago in reply to Jeff Salisbury

Intellectual 345 points

so if we cut the switching frequency down to 20 kHz we would come put to 24mA and be ok. Is there any reason to not use the 45kHz I see listed in the docs and use a 20 kHz frequency?

0 Jeff Salisbury over 9 years ago in reply to Jeff Salisbury

Intellectual 345 points

am I correct in thinking also that since the gate of the fets has such low current draw that the 1 ohm resistor between the GL/H_A,B,C pins to the gate of the fet wont have that much difference from your 10 ohm resistor in the document? Can you help me understand this ?

0 Nicholas Oborny over 9 years ago in reply to Jeff Salisbury

TI__Mastermind 21305 points

Hi Jeff,

Yes, that is alot of gate charge to switch to quickly. The switching frequency required will mainly depend on the motor and control alogorithm

I see anywhere from 20kHz to 60kHz. Higher frequency will give you less current ripple but increases switching losses and current required from the gate driver.

You could investigate moving to a MOSFET with a lower Qg.

0 Nicholas Oborny over 9 years ago in reply to Jeff Salisbury

TI__Mastermind 21305 points

The series gate resistance will only affect the slew rate of the gate voltage, not the average current over multiple switching periods.

You will see a small difference moving from 1 ohm to 10ohm but not a large difference. Generally, lower Qg FETs require some series gate resistance to slow down the slew rate with the DRV8301.

Very fast slew rates can lead to inductive ringing when switching and EMI issues.

0 Jeff Salisbury over 9 years ago in reply to Nicholas Oborny

Intellectual 345 points

Has anyone seen anything like these signals on the gate pin of half bridge motor control N channel MOSFETS? Here is the schematic

0 Jeff Salisbury over 9 years ago in reply to Jeff Salisbury

Intellectual 345 points

0 Jeff Salisbury over 9 years ago in reply to Jeff Salisbury

Intellectual 345 points

Goals to understand:

- Why different levels

- Gap between signal bunches

- WTH C high. What is going on here

Here is the DRV8301DCA Gate High for A signal ( pin 47 to Q2 )

B high to Q4 ( Bottom signal of the two. Zoomed out on the time axis from A high previous pic )

C high to Q6 ( .....interesting) ( again bottom signal )

A B and C low fets Q3, Q5 and Q7 These all at least look the same level yet I need to understand the gapping. Suspect this is related to OCTWn pin tripping.

0 Nicholas Oborny over 9 years ago in reply to Jeff Salisbury

TI__Mastermind 21305 points

Jeff,

Are these with the motor connected?

I recommend taking things back a step and looking at each half-bridge individually with no motor connected. This is to ensure that each half-bridge can switch both high and low with no issues.

0 Jeff Salisbury over 9 years ago in reply to Jeff Salisbury

Intellectual 345 points

We found the cause of the weird C high gate pin signal. Bad connector! This eliminated the current OTWC flag and cleaned up the motor running.

0 Jeff Salisbury over 9 years ago in reply to Nicholas Oborny

Intellectual 345 points

Good call, that is what we did. With the motor unhooked and spinning the motor with a drill we say the signal clean up. Replaced the fried connector and baam!

0 Nicholas Oborny over 9 years ago in reply to Jeff Salisbury

TI__Mastermind 21305 points

Good to hear. Best of luck moving forward!

0 Kister Genesis Jimenez over 8 years ago in reply to Jeff Salisbury

Prodigy 210 points

I read that your DC_CAL is floating. When you fixed your problem, was the DC_CAL still floating?

Motor drivers

Motor drivers forum

DRV8301, DC_CAL floating, so question..,