DRV8306EVM: Help with initial setup / spinning

Hi Adam,

Nice to chat with you again!

Thanks for sending your picture, it gives me a good idea of what you are doing. I would guess that there is something wrong with the Hall sensor inputs to the DRV8306 causing the motor to not commutate & lock - this causes the very high current you are seeing.

Can you check a couple of points:

1. Try reversing the position of the Hall Select switch, just in case there is an error in the user guide
2. Measure the voltage on A-, B-, and C- (the Hall connections that you aren't tying the Halls to). They should be 1.65V; if they aren't something is wrong.
3. When I rotate the motor manually, each hall input (A+, B+, C+) changes between 2.4V and 0V depending on rotor position. --> That sounds very odd, the output of the Hall should pulled up to 3.3V, not 2.4V. Get Hall A to operate and measure the voltage across R17, then measure the voltage at A+. Do the same thing when the Hall releases (measure the voltage across R17, then measure the voltage at A+)
4. Can you try all different combinations of Hall sensor Inputs (white/yellow/blue)? I do often see that one particular combination works better that other.

IDRIVE and VDS jumpers

• I would recommend placing jumpers at the lowest settings to start (that would be the position near the edge of the board)

Thanks,

Matt

Thanks again Matt Hein ! Responses are below:

Can you check a couple of points:

1. Try reversing the position of the Hall Select switch, just in case there is an error in the user guide

   I was hesitant to try this first up, based on the halls being supplied with 5V and what I read in this thread re: potentially damaging the DRV8306 (not sure if this actually would be an issue though)

   1. 

2. Measure the voltage on A-, B-, and C- (the Hall connections that you aren't tying the Halls to). They should be 1.65V; if they aren't something is wrong.

   I measured 1.5V here between A-, B-, and C- and HGND - so something must be off? 

3. When I rotate the motor manually, each hall input (A+, B+, C+) changes between 2.4V and 0V depending on rotor position. --> That sounds very odd, the output of the Hall should pulled up to 3.3V, not 2.4V. Get Hall A to operate and measure the voltage across R17, then measure the voltage at A+. Do the same thing when the Hall releases (measure the voltage across R17, then measure the voltage at A+)

   When A+ = 0V then R17 = 3.3V, and when A+ = 2.5V then R17 = 0V (Note: A+ voltage - I measured this value at 2.4V yesterday but carefully checking today it seems to be 2.5V).  

4. Can you try all different combinations of Hall sensor Inputs (white/yellow/blue)? I do often see that one particular combination works better that other.

I tried all combinations as suggested. With the IDRIVE and VDS jumpers at lowest setting, I was getting the nFAULT led flagging at a certain point when turning up the motor speed. So I moved both jumpers down 2 places (third position from the edge) and tried again......and on the last combination the motor started spinning!!! I still don't quite understand why that combo worked vs. the first combo I tried? My motor actually came with a built-in controller (which I used for initial testing then removed). On the built-in control PCB the hall inputs were labelled A, B and C.....so I followed that combo in my initial setup. I thought that would need to keep that the same, but now it's seems B, A, C works? :)

Either way I'm very happy to have my motor spinning, thank you!

One thing I did notice though was running clockwise the motor sounds good/smooth and current draw matches initial testing using built-in controller....however when when running anti-clockwise, it doesn't sound great and pulls higher current. What is that a sign of?

Also what are your thoughts re: the values I provided....the halls still aren't reaching 3.3V when high, so do I need to solve that somehow? 

Thanks!

Hi Matt,

OK thanks - that's good to know I'm safe with the halls running at 5V, regardless of switch position. 

Also sorry I should have mentioned, my motor runs in both directions, and did so with the built-in controller without issue.

I've done more testing since last post, and the performance with DRV8306 is still great when running clockwise....but very rough when running counter-clockwise. Sometimes it doesn't even start in that direction, it just locks / pulls high current like it was doing previously in both directions (before I changed the hall wiring). Other times it will spin counter-clockwise, but as mentioned it sounds quite rough and pulls higher current. It's pretty strange as it's running so well clockwise! 

I have another of the same motor here, so will test using that also to see if it's the same. I also need to do some more reading re: how to set IDRIVE properly, and it sounds like a need a scope also :)

In the meantime if you have any re: the counterc-lockwise issue - please let me know!

Hi Adam,

It is interesting that then motor is not behaving as well in both directions with DRV8306 while it operates well using the other solution.

DRV8306 does *not* implement lead angle adjustment, which results in a lower efficiency. If the built-in controller has lead angle adjustment then I would expect that solution to be more efficient/smooth.

With an oscilloscope, you can check the Hall transition alignment with the motor voltage waveform (Measure HALLA, OUTA, OUTB, OUTC). Basically we would need to verify the Hall position versus the motor-generated back-EMF. Sometimes the Halls are not aligned in the way the driver is expecting.

We have other solutions with lead angle adjustment, but they are either much lower current (DRV10970), or sensorless (DRV10987). In about a month or so we will be pre-releasing a new sensored driver based on DRV8316 (at least based on our current schedule) with lead angle adjustment which would be interesting to try.

Thanks,

Matt

Hi Matt,

Yes it's definitely interesting re: the counter-clockwise operation. You are correct that the built in controller does implement lead angle adjustment - so I expected less efficiency / smoothness with the DRV8306. I am definitely seeing that in the clockwise direction, but it still seems acceptable for our application. Counter-clockwise is different though - it definitely seems like something is wrong there.

I've got more testing to do - so just two quick questions if that's ok, then I'll leave you alone for a while!

1. I calculated IDRIVE setting at 130mA for a 200ns TRISE (as a starting point). This was based on a Qgd value of 26nC from the CSD88584Q5DC data sheet. So I can select the correct setting (120mA or 180mA), can you please confirm if the jumper setting closest to the edge is the 15/30mA setting? (so my settings would be 3rd and 4th jumpers along). There are only 6 jumper settings (and 7 IDRIVE settings), so I'm not sure if the 'no jumper' setting is min or max. 

2. I have set up a PWM input now for speed control, and am running a test program to ramp the duty cycle up to different speeds at varying rates. When the motor is unloaded this works well (in clockwise), however when the motor is loaded - I sometimes get a bad 'stall' type conditions where the motor pulls very high current and makes a high pitched sound (the other times it works great though). What is normally a sign of?

FYI the load on the motor is a gear pump head which is used for drink dispensing. Sometimes when I get a stall it will start working again during the same run (i.e. stall briefly then start operating normally), and other times it stalls repeatedly and I need to remove the pump head and get motor going again un-loaded. 

Finally thanks for the input re: the new DRV8316 based option with lead angle adjustment. The DRV8306 is very appealing to me though, due to being integrated control with simple IO. For this application I'm actually upgrading an existing product which currently uses brushed motors driven by one of your LMD18200 chips. This has worked great for us for years now, however due to a new compliance change/issue I need to quickly move to a BLDC solution. The IO for the DRV8306 is similar to the LMD18200 so it seems like potentially a great solution which would reduce cost and complexity (as we need to interface it with an existing control board and software). Also don't worry, we would be outsourcing the final implementation to an experienced embedded systems engineer - I'm just trying to far as I can before that!

Thanks Matt!

Looks like I have IDRIVE set at 120mA then which is good, I'll try 180mA today by removing the jumper.....but guess I'll need a scope to properly determine the best final setting. 

Re: Stalling - my program ramps the duty cycle up from 0% to 65% (at 40kHz) over approx. 1 second. At 65% duty - I get the flowrate which I would expect from my pump head, and the current draw (from the lab PSU supplying VM) is 3A which is in line with what I would expect also (based on similar testing with the built-in controller).

About 8/10 times it starts well, but the other times it stalls and pulls very high current - my lab PSU tops out at 10A, and I have seen it flickering up in that range. During this time the motor makes high pitched sounds, a bit like a dying robot :)  As mentioned, sometimes the next try will work - but other times I just get repeated stalls....normally then I remove the pump head and manually rotate the shaft, then it will start again under no-load. 

During these stalls I still haven't had the nFAULT led turn on, which is interesting (note: I have VDS set at 0.4V).

I'm wondering if it's an alignment issue like you suggested for the counter-clockwise issue. Like it lands in a certain spot and struggles to commutate from there. 

Time to get a scope! 

Hi Adam,

It does sound like the motor is getting stalled out on start-up. There could be an issue with the pump starting without load, or the Hall sensor placement in the motor is abnormal. A scope would help to determine if the Halls are aligned like we are expecting.

You may want to try out other combinations of motor output placement, Hall Sensor inputs,and the DIR pin. There could possibly be a better configuration, if you have not tried them all yet.

In stall state I don't think you are pushing enough current to trip nFAULT - this would be something like ~500A based on the RDS(ON) of those MOSFETs.

Thanks,

Matt

Hi Adam,

There is a procedure to find the motor "dead point"starting on page 20 of this document. This guide is obviously not for DRV8306, but it shows how to measure where the point is. The document recommends a slower acceleration rate to overcome this.

Thanks,

Matt

Hey Matt,

Good news - I followed your suggestion "you may want to try out other combinations of motor output placement" and you were right.....a different phase wiring setup changed everything! The motor now runs in both directions and starts successfully every time - both unloaded and loaded!!

Thanks again for all your help with this :) 

There was also a significant drop in current draw (3.0A down to 2.1A under load) and less noise as well which is great....I also noticed the mosfets were getting quite hot before, but now they are much cooler.

With all these positives there was one thing though - an overall drop in flowrate. This is both compared to running clockwise in the previous phase wiring setup, and also to the built in controller. I'm now using a much higher duty cycle (98%) and while the flow is acceptable, it's still below what I would consider 'max' for this pump head.  

I know the built-in controller uses phase angle adjustment - which is more efficient, so could that drive the motor faster?

Also the phase wiring I had previously was obviously 'off' - so maybe it was commutating differently (I only had duty cycle at 65% and it was max flow and high current). 

I tried increasing VM which obviously helped (27V got me to where I would like), so that is an option.....however if there are any other parameters I should try adjusting please let me know!

Thanks again Matt

Hey Matt,

I got there over the weekend....turns out the flow issue was related to the pump head configuration, not the DRV8306 :)

I'm now getting a great flow with consistent (and fast) start ups. 

Thanks again for helping out, really happy with where I'm at now. 

Hi Adam,

Excellent! Please do reach out in the future if you have other questions!

Thanks,

Matt