DRV8711: Extreme noise production with calculated settings

Thom,

I saw you already put in "Force slow decay all times" which could give a small ripple current. But, if you can provide a winding current waveform, that may help us understand this issue.

I will recheck this thread tomorrow morning to see if anything can come to my mind.

Hi Wang,

I put it into slow decay as the speeds aren't very high, I'm targetting speeds between 1.1111 rotations per second to about 11.111 rotations per second.

Which I'd assume slow decay would work fine with.

Forcing fast decay (or auto mixed) creates a not-so-midly annoying noise, but perhaps that is due to wrong settings.

I do not have current measuring capabilities, but I could measure the 0.033O sense resistor. Last time I checked I got some useless results though, it looked more like a PWM (which I assume is due to half-stepping).

Is that what you meant?

Regards,

Thom

Edit: I'm targetting speeds between 0.111 and 1.333 rotations per second.

Hi Wang,

I've been busy attempting to find the correct settings and it appeared that microstepping + Slow/Mixed (Indexer Mode Only) gives by far the best settings. Calibrating is very painful as there are a lot of variables to tweak, but I settled for TBLANK=50 and TOFF=40 with TDECAY scaling with the speed.
For example I have a "slow" speed of 0.111 rps, the best TDECAY with this setting is actually flat 0, but the "fast" speed of 1.333 rps requires a higher decay, TDECAY = 41. Having a too low TDECAY produced a highly audible low "humming" noise, overshooting the TDECAY by too much result in a whining noise, there truly appears to be some kind of sweet spot.
All other "modes" simply do not work for me as, no matter what (practically), the noise is unbearable.

I wouldn't consider this resolved, having to manually tweak these settings for every speed is a massive chore, isn't there a better way about this?

Regards,
Thom

Thom,

First, 40us or 50us Toff time plus the time of driving the winding current to regulation point can cause the winding current ripple frequency lower than 20kHz. So, we can hear it. Would you try to reduce the Toff time to 30us?
Second, you use half step setting. Could you increase to 32-step or 64-step setting which can have a smooth motor motion and may reduce the audible noise.

Hi Wang,

I tried the following settings:

The speed is set to the lowest necessary, with these settings I can choose between a woodpecker (Force Slow Decay at all times) or a high-pitched whining noise ( all others ).

Regards,

Thom

Thom,

Since we don't have the current probe to monitor the current waveform, I just try what we think the best and cannot verify it.
1. If tBLANK is too long with a small winding inductance, that may give us a higher peak current and extend the cycle time to audible range, if you change the tBLANK to 1us, does it affect the noise?
2. In your setting, the Tdecay setting is much longer than Toff setting. Can you make Tdecay < Toff and aggressively reduce the Toff to 10us?
Do you get noise sound with auto-mix decay?

Hi Wang,

I've manually tuned the device with 12 presets.

I will pick this project up again some time in january and try to see if I can resolve it with a current probe and your advice.

Happy holidays!

Regards,

Thom