DRV8308EVM unstable and no lock

Dave Cawley

Other Parts Discussed in Thread: DRV8308EVM

I am trialling a DRV8308EVM to run a Papst ECA 18.08 http://www.ebmpapst.com/en/products/motors/ecigearmotors/ecigearmotorsdetail.php?pID=54331

With the motor not loaded or driving anything, I can get a good lock with minimum flutter. I am using 180 degree sine-wave and the motor frequency is 234Hz. MOD 3970 and XOR the sensors.

The problem arises in the intend application, where the motor has to run at 234Hz and is driving a flywheel 12" diameter that weighs over 4Kg. It is run through a small pulley on the motor to a flange on the flywheel via a rubber belt to give an actual speed of 33 1/3 RPM.

So in essence, it's a small motor going slowly, driving a big flywheel going extremely slowly, coupled by a small rubber belt.

After two whole days I cannot get lock, only approximate speed stability with a huge amount of flutter ! I don't need phase lock, but I do need a fairly tight frequency lock.

I am looking for a solution using this or any other device. This is not for a 1 off so it needs to be a manufacturable solution and work totally reliably in the field. Free help would be great, or if I have to, I'll pay someone to do it ? motor@time-step.com

Thoughts please ?

Thanks

Dave

over 11 years ago

0 RE over 11 years ago

TI__Mastermind 20985 points

Hi Dave, our help is free :)

I recommend not setting FGSEL to XOR, as I've seen it cause much more flutter than HALL_U (on the order of 3% versus 0.3%). If you switch to HALL_U, you'll of course also need to slow down the CLKIN frequency to 1/6th of what it currently is.

What CLKIN frequency are you using, and what are you getting on FGOUT? If you increase/decrease f_CLKIN, does f_FGOUT also increase/decrease?

What is SPEEDTH set to? Increasing that will help the lock tolerance. Also, please send me a screenshot of the 3rd GUI tab, and I'll check out all the settings. I'll also send you an email, and maybe we can get in touch over the phone.

Best regards,
RE

0 Dave Cawley over 11 years ago in reply to RE

Prodigy 30 points

Hi Ross

I'll do that in the morning. I assume you are in the USA ? If so which time zone ? I'm in the UK.

Thanks Dave

0 RE over 11 years ago in reply to Dave Cawley

TI__Mastermind 20985 points

Yes, I'm in the Central Time Zone in the U.S. If you'd like to setup a call, you can contact me by email at reisenbeis <at> ti <dotcom>.

The first thing I'd like to understand is whether the system is speed-locked, where f_CLKIN = f_FGOUT. If it's not, then I would try increasing SPEED (the open-loop gain needed to reach the target speed), and increase LOOPGAIN (DC gain), and make sure AA_SETPT is set correctly. If there's still not enough DC gain, set AA_SETPT one value lower. Also try decreasing f_CLKIN.

If speed is locked but there's lots of oscillation (flutter), SPDGAIN might be too high.

Since you have successful results with no load, and poor results with a load, if you have a way of applying an intermediate load, it might be easier to tune that and give you a better idea of what to do for the full load.