You are here

DRV8302 sensibility

20 posts / 0 new
Last post
lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33
DRV8302 sensibility

in vesc 4.xx , how can I find gain of TI Mosfet driver DRV8302 on shunts inputs ? 10 or 40 V/V?

 

skyline1970
Offline
Last seen: 3 months 6 days ago
VESC Free
Joined: 2017-09-07 22:28
Posts: 232

Gain pin of DRV8302 is at GND level, so gain = 10.

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

thanks . I thought it was programmed by the  STM32 .

 

I ask that , because I have intempestive accel/brakes at 80% duty  on flat floor near Vmax .

 I know problem with observer  become noisy at this duty and low current ( 5A/motor in my case ) , on VESC4.xx with 2 shunts ..

 so 2 solutions :

 1) I push ' sensorless ERPM ' over my Vmax  , always in sensored mode . some says it's noisy , but I use big hoverboards hubs , Kv18 and 12000 ERPMmax in 12s ..

2) I use 40V/V gain to augment sensibility on current sense , and quadruple Current Max  and braking . (functionnal current : 25A/ motor max )

3) I keep 10V/V and  modify shunts to 2 or 3mOhms for better sensibility , and double or triple Current Max  and braking .

 

I will try ..

 

skyline1970
Offline
Last seen: 3 months 6 days ago
VESC Free
Joined: 2017-09-07 22:28
Posts: 232

in fact, observer should not be related with current, it should be much more related with back emf. and at low rpms sensored ctrl, at high rpms sensorless ctrl could be adjusted 

hw solution, if your motor current is low, you may change the pos. of Gain pin to 3.3v level, and gain becames 40, this means if your current 5a actually, then 20a is read

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

thanks .

 i tried to use foc sensored over my erpmmax (~11500 full batt.)  , it solve my intempestive accel/brakes problem . but it' a little noisy and I can't reach Vmax . 

pity not to be able to use the hall sensors up to my max rpm. or 900rpm .

 you say problem is related to back emf . i have hubs hoverboard motors with these parameters detected :

141mOhms

299uH

19mWb

KP 0,3

KI 141

observer gain 2,9

 

if back EMF is not high enough, there is no solution ? 

modify observer gain ?

another value ?

find a VESC6 ? laugh

 

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

thanks .

 i tried to use foc sensored over my erpmmax (~11500 full batt.)  , it solve my intempestive accel/brakes problem . but it' a little noisy and I can't reach Vmax . 

pity not to be able to use the hall sensors up to my max rpm. or 900rpm .

 you say problem is related to back emf . i have hubs hoverboard motors with these parameters detected :

141mOhms

299uH

19mWb

KP 0,3

KI 141

observer gain 2,9

Kv ~20

 

if back EMF is not high enough, there is no solution ?

modify observer gain ?

another value ?

find a VESC6 ? laugh

 

skyline1970
Offline
Last seen: 3 months 6 days ago
VESC Free
Joined: 2017-09-07 22:28
Posts: 232

just test in bldc mode , hall sensored detection, and investigate.

For FOC to work it is very important to get the resistance and flux linkage right, at least within 5 %. You can roughly calculate the flux linkage from the kv and number of poles with the formula

60 / (sqrt(3) * poles * pi * kv) = ( Weber flux linkage)

The observer gain can be calculates roughly as 1000 / (lambda^2) where lambda is in mWb.

From the formulas, you could check the  detected and the calculated values are near. So, you could use the values from calculation for modifying, and try.

 

 

 

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

tried BLDC yesterday .. work fine .. but sorry ,too noisy I'm a Foc lover heart

I will calculate these values , with my milliohmmeter values .

 but flux linkage isn't proportional to magnets size/force ?

 

skyline1970
Offline
Last seen: 3 months 6 days ago
VESC Free
Joined: 2017-09-07 22:28
Posts: 232
lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

yes

I use these old rolls bike trainer to reproduce my foc brake/accel problem .

Safer than ride laugh

20211027_140230.jpg

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33
not being able to reproduce the phenomenon, I went for a ride ... bad idea .. nice fall and very bad shoulder / ear / wrist ..
in short. I have a log !!yes

the rest of the time my hubs are working fine. acceleration, speed, braking (even if it grumbles a little in hard braking), no engine or esc heating .. I no longer understand.frown
look at the 95% duty continue as the current reverses and the speed drops. Id and Iq also have a strange behavior.
this happens when the current drops below 10A and duty almost to the maximum...
but not on my chassis dynamometer. .angry


I'm going to put some ice ...

rpm.pngcurrent motors.pngduty.pngid.pngiq.png

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33
I also noticed an oscillation of the motors on my unbraked dynamometer, but still presenting a slight resistance. test with the 3 modes of the L / M / H remote control. no worries with the wheels in the air .. I'm looking .. can we adjust the reaction time of the closed loop somewhere ?. I am not familiar with microprocessors and their code, but I am familiar with  oscillators . here we have a loop band problem. it lacks depreciation.

yoyo3modesVX2.png


 

Icône de validation par la communauté

skyline1970
Offline
Last seen: 3 months 6 days ago
VESC Free
Joined: 2017-09-07 22:28
Posts: 232

get well soon, it is always a bad idea to test by yourself at risky situations.

it seem to be that the motor reverses rotation, which should not. And assume on flat surfaced way. I had made settings on fw 3.38 for a balance app., Segway. project which using two 3kw hub motors that worked well. Maybe downgrade fw, and try. And it is recomended to use a big bypass cap. around 22000uf near the drivers supply, maybe it simply solves the problem.

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

yes . at low current and "high" speed and duty ,vesc want to reduce something .

It reminds me of YOYO's problems on Xiaomi M365 scooters when the motor reaches its maximum physical speed.frais

 


 

Icône de validation par la communauté

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

"And it is recomended to use a big bypass cap. around 22000uf near the drivers supply, maybe it simply solves the problem"

good idea . but at ~5A battery , should the present capacitors be sufficient?

I have no problem on a steep hill at "high" current (2x30A)

but i will  try .wink

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

good news . increasing flux linkage decrease oscillation .

detected at ~16,5 , better at 18~19 .beyond that the engine begins to knock a little. I don't know which parameter influences the linkage flow, but it intervenes in the reaction time of the loop. I must go back and try .. when my shoulder is better ..laugh

skyline1970
Offline
Last seen: 3 months 6 days ago
VESC Free
Joined: 2017-09-07 22:28
Posts: 232

calculation results?

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

18mWb and gain 3 . like detected . but it often detect 16mWb.

resistance measured  ~380mOhms . detected 140mOhms.

but i've seen in MOTOR<>GENERAL>AVANCED . my DUTY CYCLE CURRENT LIMIT is at 100% . it should be at 85% for e-skate applications ..perhaps the solution.

 

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33

I've found this on help page :

Warning! For E-Skate and some other applications you now need to set DUTY CYCLE CURRENT LIMITS.

Warning! Failing to set the duty cycle current limit value to 85% on all VESC motor controllers in your system will affect your ability to keep the balance when reaching top speed! 

 

and here is my morning test .. effectively , with 85% , the current never becomes negative, as in the log of my fall. left gragh 85% , right 100%)

 

DCCLS85et100.png

 

 

and here , test with 16 and 18mWb . less oscillations at 18 . cogg at 20 .

 

FL16et18.png

lurch
Offline
Last seen: 4 months 3 weeks ago
Joined: 2021-10-05 18:56
Posts: 33
thanks for your advice . everything seems to be working fine now. I am now looking for a pilot who wants to test well, the time to repair my injuries.laugh

I still have the brakes which growl at strong current. I will open another post.