Other Parts Discussed in Thread: TLV9151, TLV271
Hello Ti, Happy New Year!
I have a questions about the current sensing, bemf sensing and ground seperation in motor controller. I’ve searched through the forum, but I didn’t find exactly the answers.
I’m developing a three-phase motor controller with the DRV8301 (because I have a lot of them and newer devices not available at the moment). Maximum voltage is 50V and maximum current about 60A, it’s a four-layer board with 2oz, inner and outer layers.
In my design I’m using the TLV271 (INA 240/241, TLV9151 are not available) for low side current sensing.
I have two options how to route the differential traces from the shunt to the op-amp.
The first would have shorter distance from the shunt to the op-amp but I’m crossing several gate traces (Option 1). (I could put a filter on the input of the op-amp)
The second would be increasing the length of the differential traces and avoid crossing the gate traces (Option 2).
- How does the length of the differential traces impact the current sensing? It’s low side current sensing, the CMRR isn’t a problem I think or?
- How does crossing the gate traces affect the current sensing?
- Which option would be better for accurate current sensing? (I’m using Instaspin FOC)
- Just for interest, would inline current sensing made a different in which option to choose?
- In the picture above there is also a ground separation (green line), does that bring me anything? I will let the MCU approximately where it is on the picture so there is enough space between the DRV and MCU/Analog Section, so that the return paths shouldn’t be a problem.
- Another option would be changing the layout and placing the op-amp closer to the shunts, would this be the best option because the traces are shorter and I don’t cross any gate traces? (My only concern is here that the MCU is close to the power stage)
7. When sensing the bemf voltages (blue line) there is a low pass filter, with a fcutoff of around 300Hz. Crossing with the bemf traces the gate traces shouldn’t be much a problem, because of the filter, is this right? The crossing happens before the signal is applied to the filter and voltage divider.
8. The bemf filter (brown) has a current return path, when placing the bemf filter underneath the op-amps (place here in the picture) could this impact the current sensing?
9.The sensed voltage of the bemf filter is a pwm signal and I route them over a huge gap before applying it to the filter. The current will be very low but there is this high frequency component from the rise/fall time of the pwm signal. Can this gap cause a problem? (for example EMI?)
If you have other tips for accurate sensing, please let me know. (Or App notes and references)
Thank you in advance,
Daniel
