Custom Hardware Troubles

Hi all,

After a day of trouble shooting I now have something more substantive that hopefully someone can help me with. The PWMs are not firing because they are tripped out by TZ2 & 3 caused by a fault reported by DRV8301...Evidently, DRV8301 Fault does not result in a system fault (no ctrl_state_fault is flagged).

I did find a solder bridge from Fault to Gain/SDO, but clearing it has made no difference at all - maybe it damaged the DRV8301....

Because OCTW is not reporting a fault, I think there are only a handful of possible causes for the problem;

I was thinking that this means the fault must be PVDD/DVDD/GVDD/buck under/over voltage.. -I have gone through and checked this and with the exception of GVDD(which is internal) everything checks out....

I should mention that the fault LED (same setup as BOOSTDRV8301 - PNP transistor) is very dim which made me think that somethings wrong with the hardware... At power up, the Fault pin voltage is 57mV (very dim led) and directly after programming it sits at 3.3V--- led off. *Then when I hit play, its back at 57mV and a very dim LED....

I can't figure out the 57mV, makes no sense that the voltage floats up a bit unless the DRV8301 is damaged?

I'm looking into SPI now, but the error doesn't seem to be reported....

Could someone tell me if my understanding of DRV8301 faults is correct (re Fault low but OCTW high).

Thanks

Hi Patrick,

If you ENABLE the DRV8301 manually and try to toggle the half bridge manually does the DRV8301 respond?

Is the DRV8301 responding at all to the SPI request? You mentioned that you checked the values of the regulators and they seemed ok? GVDD is pin that you can check. Pull ENABLE high and report on the values of each regulator.

What kind of pull up value is on the DRV8301 nFAULT pin?

One to check is if the PowerPAD has been soldered (you mentioned that you had seen some solder issues). The DRV8301 makes an electrical connection to GND through the PowerPAD.

Thanks Nick,

I bit the bullet and replaced the DRV8301 on my board.... I now have a working custom instaSPIN ESC :)

80RPM-~6kRPM and about 20% less juice than my 6-step controllers! 

-I'm doing up a board for 'real' fabrication now and was wondering if you think there is an advantage to having the differential signal tracks the same length Going back over a course that covered this, it seems that this is standard practice with differential signals going into a summing amp but I can't see any reason for it on this board (considering DRV8301BOOST) --- wanted to get your thoughts.

If you're interested:

*SPI was behaving, definitely acknowledging MCU SPI requests.
*10K pull up on Fault.
*When fault triggered, it floated up a bit(~50mV).
*Yes power pad was definitely soldered correctly - I had the buck outputting ~5.5V when I first began populating the board(before I populated the MCU so it didn't matter too much) and it turned out to be the power pad floating- first foray into solder paste and I hadn't heated it up enough.

*When I replaced the DRV8301 - I tested the board out on a current limited PSU and had the current limit set extremely low. Low enough to have the DRV8301 fault on PVDD under voltage - fault LED flicked on full brightness-Fault pin no longer floating up a bit...

-Don't know the internals well enough to know why the fault would float? Could also have fixed something with the hot air when I replaced the DRV8301 - not sure....

Thanks.

*Attached a video of spinning a motor (impressively slowly - considering no encoders) using custom board for anyone who might be interested.

Great to here Patrick :)

1. Differential sense is a key layout trick to utilize. The amplifiers will be measuring signals in the millivolts and any miss match in line impedence or noise injected into the line will cause errors in the measurements. The "closer" to true differential connections the traces are the better the results. There of course can be some trade offs if you have some room in your measurements. Another key is to keep these traces away from noisy signals when possible.

2. The FAULT pin is open drain. If the DRV8301 is having issues it's behavoir can be unpredictable.

3. I highly recommend looking at the BOOSTXL-DRV8301 layout and using it as a reference for your layout. Once you get to higher currents, proper routing and placement will be key.

Patrick,

You made my day!  Love seeing our idea of InstaSPIN - empowering everyone with premium, high efficiency motor control - come to fruition.  If you would like any promotion of your work I I would love for you to create a post / blog, etc. regarding your experience. 

PS - great taste in music!

Thanks Chris & Nick...

A few more questions if you don't mind me not starting a new forum/

Nick, regarding the differential signals, I have done my best to ensure track lengths are the same - short of adding wiggly lines.

At most there out by <1mm - this doesn't seem like much to me, but as I have no experience with 'proper' board layout I thought I should get your thoughts before sending gerbers. - Would you be happy with this? -Board is 15A continuous - 0-16V.

ChrisClearman said:

Patrick,

 If you would like any promotion of your work I I would love for you to create a post / blog, etc. regarding your experience. 

PS - great taste in music!

 

Canceled their Australian tour :(

I can definitely do that, currently writing up a 'technical report' and thesis for the project, but we also have to do a poster which could be good for this - development process/results/etc... 3-4 Weeks and I will pull out the pertinent information and post my experiences with your amazing rat catcher and truly outstanding support (E2E) ;)

I'm hoping to demo this on a quad (time permitting - still have to finish I2C for my board and also for PX4), absolutely ideal for multicopters - efficiency and response (supervisor is going to use it on a submersible vehicle - commercial ESCs need a gear box to get the low speed required - thought you would be interested in yet another use for instaSPIN). I'm sending my board to a board house to get it made properly for this - need to populate at least four boards which starts to get pricey when buying MCUs and Gate drivers in small quantities... Talking to my supervisor & lab techs about this today - they said that free samples are often given out for projects like this? I'm skeptical since it's undergrad but I figured it can't hurt to ask. - Tried the free samples button but it requires a company and office phone, figured asking directly had a higher chance that falsifying details.

Thanks again.

3*5 cm board (smallest I could get without adding >2 layers)... Will be more than happy to release eagle files once I figure out if the layout is any good.

A few samples for things like this is no problem.  Just submit as you were and use your University information.