Close this topic...move to other place.

[ausera uirwaij]

Thank you.

[ausera uirwaij]

update...​

[Javier1]

@ausera uirwaij​ There is nothing wrong with that "jump", that jump is the BACK-EMF readings trapezoidal modulation relies on for coil position/ electrical speed calculations.

The faster your motor goes and the lower your MOSFET/IGBT resistance the higher the backEMF signal would be seen.

The "jump" is there because the second picture motor is running at lower speed for some reason. (if the haware is different and youre driving both 100% there you have it, your drive board is able to put more current into the motor than the comercial one)

 

One way of filtering Current ripple (your sistem is missmatched) is to add Filtering capacitors to match your IGBT/MOSFET with the coils

https://www.planetanalog.com/capacitor-choices-in-power-related-applications/

I bet this happens because your IGBT/MOSFET is way faster (its like slamming the door of a room instead of gently closing it) than the comertial solution youre comparing it to, and has lower ON resistance.

[ausera uirwaij]

Hi Cristiana @Cristiana SCARAMEL​ , thank you.

Not matter my 5V or the other 3.3V, the elapsed time of their rising point to trigger is not same ? the size of voltage has so big influence?

Best Regards

[ausera uirwaij]

Hello @Javier Muñoz​ ,

Thanks for your professional reply. It helps me to understand the underlying principals.

Both pictures are captured at 100% duty cycle,However, my board running at 10200 rpm lower speed than the commercial one which has 10800 rpm with the same 24V supply voltage. We have different hardware configures but the same hardware architectural. So both running at similar speed should have the similar back-EMF voltages, my board's picture seems higher back-EMF voltage than the other one from the aspect of voltage "jump".

I bet this happens because your IGBT/MOSFET is way faster (its like slamming the door of a room instead of gently closing it) than the comertial solution youre comparing it to, and has lower ON resistance.------->

my board is shipped with STM32F427x and the commercial one is MCU made by local manufacturer.

Can you explain what're ways can result in such way faster for stm32 and how to switch my MOSFETs gently? faster way causes what results？ I may agree with your words because I shut off and on MOSFETs once hall edge triggers.

Good day,

[Javier1]

@ausera uirwaij​ i think i missunderstood you

can you post pictures of both boards?

So the only thing different are the MCUs?

[ausera uirwaij]

@Javier Muñoz​  my board is F427 __>send 6 PWMs directly__>DRV8302__>send 6PWMs ___>resistors__>6MOSFETs__>three phase bldc motors.

the other one is MCU(8051 core) __>send 6 PWMs directly__>DRV IC (manufactured by local company)__>send 6PWMs ___>resistors__>6MOSFETs__>three phase bldc motors.

We has the same commutation methods: uncomplementary 6step .

But IGBT/MOSFET switching faster has any influences and I has never hear about this knowledge？

Thank you.

[Javier1]

@ausera uirwaij​ Are the 6 Mosfets identical in both boards?

>>We has the same commutation methods: uncomplementary 6step .

But wildy different microcontrollers (you use 32bits and 8051 are 8 bits), probably also very different motor control implementations.

>>But IGBT/MOSFET switching faster has any influences and I has never hear about this knowledge？

Sharper opening and closing of any electronic switch produces more current ringing (slamming the door analogy)

[ausera uirwaij]

@Javier Muñoz​  Are the 6 Mosfets identical in both boards?____>

my 6 Mosfets is ON semiconductor's and the other's is manufactured by local company, and my NMOS has higher quality.

But wildy different microcontrollers (you use 32bits and 8051 are 8 bits), probably also very different motor control implementations.---->

The other MCU is 32bits enhanced 8051. I have been confirmed that we both apply 6step commutation, but their engineer didn't tell the detail.

​Sharper opening and closing of any electronic switch produces more current ringing (slamming the door analogy) ------>

I have resistors to improve impedance to gate drive.

Do you know why there exist voltage "jump" with similar speed?

[Javier1]

@ausera uirwaij​ 

>>my 6 Mosfets is ON semiconductor's and the other's is manufactured by local company, and my NMOS has higher quality.

So, totally different systems behaving in different valid ways, both sound ok to me.

>>The other MCU is 32bits enhanced 8051. I have been confirmed that we both apply 6step commutation, but their engineer didn't tell the detail.

I have to confess i have almost 0 experience with intel mcus.

>>I have resistors to improve impedance to gate drive.

I didnt mean input side, i meant output side (slew rate), The sharper a Mosfet is able to open and close the Power side (output) the higher the current ringing will be.

>>Do you know why there exist voltage "jump" with similar speed?

I dont see that the difference between the back-emf and the driving voltage ("jump") should be a problem.

If you want to explore that option anyway, ¿could you post a picture/schematic of where are you probing your oscilloscope?, more info about both setups in general.

• The fix for current spikes at the DC side is usually huge-*** capacitors (normally electrolitic)
• The fix for current spikes at the motor phases is to create an LC filter (you create a LC circuit and tune it to filter out the coils resonance)

Or to choose a mosfet with a slower slew-rate

Or to choose a mosfet with higher Ron..... :(

https://www.semanticscholar.org/paper/Active-Damping-of-Output-%24LC%24-Filter-Resonance-for-Hatua-Jain/40b476cc10d10cad506870fad6c4baad82a87bd2/figure/0

[ausera uirwaij]

@Javier Muñoz​ thank you for your kindly reply.

I'm sorry and I got off work , I draw a picture to describe my probing position.

I probe phase current ,terminal voltage and hall sensor of phase A using oscillator showed in above posts. HA,HB and HC is 3 hall sensors.

I dont see that the difference between the back-emf and the driving voltage ("jump") should be a problem.---->

I just wonder to know how 6step commutation has lower back-emf voltage compared to my board, 120°commutation has no flux weakening like FOC after all. Is any possible suppress or discharge back-emf voltage by hardware? I think if my board has lower back-emf voltage will speed less current to counter back-emf and has higher speed like commercial one.