DRV8262EVM: DRV8262 IPROPI Current Sensing – Which Pins to Connect for Single H-Bridge Mode?

Hi Dennis,

Thank you for this post. 

Based on the previous post I assume there is not change with mode settings and they are as below. I'll answer you questions with this configuration. An important point about this MODE1 and MODE2 settings, these two inputs MODE1 = 1 and MODE2 = 0 must be at these logic levels at power up or prior to a sleep and wake cycle, nSLEEP = 0 to 1 transition. If this was done while the device is awake the device configuration will not change and remain in dual H-bridge mode. Please ensure this.

According to the datasheet, I have the following settings:

• MODE1 = HIGH → Single H-bridge 

• MODE2 = LOW → PH/EN interface 

Dennis Formisano said:

Is this the correct way to configure the board if I want to:

• Use STM32 to control the H-Bridge

• Manually set VREF

• Read current via IPROPI using ADC

Sure, this configuration is supported by the device.

Dennis Formisano said:

• Do I really need to connect IPROPI1 and IPROPI2 together in single H-bridge mode?

• If yes, which ones are the correct physical pins on the board that correspond to IPROPI1 and IPROPI2?

Yes this is required in single H-bridge mode. See pin out description. They're not internally connected in this mode hence this is required. If not tied together only about 1/2 the current (from one bridge) will be reflected in the RIPROPI voltage.

Bridging J19 would make this connection on the EVM.

Dennis Formisano said:

Can someone please take a look at the attached image and confirm if the jumpers look correctly set for this configuration?

I've provided this information in the referenced post towards the end. I'll reproduce the image for convenience. The jumper shorts with a red X mark must be unpopulated and jumper shorts of the headers that need to be populated are shown, J25 and J19. External VREF must be provided only to pin-1 of J7 - has no jumper short. 

In the image shared by you they all seem to be agreement although with the angle of the image I cannot say for sure. One thing I noticed was the two jumpers connecting IPROPI to the onboard MCU ADC were still populated - see below. The ADC inputs of the onboard MCU may load these pins better to remove these two jumpers. Likewise you can remove the nFAULT and the OCPM jumper shorts as well to isolate it from the onboard MCU. 

With the suggested connections I tried the EVM in Single-H mode with a BDC motor. See below scope capture, motor driven with 35 % duty cycle. Blue trace C4 is the motor current and the yellow trace C1 is the IPROPI voltage. The cursor on C4 reads V1 = 2.2852 A and the IPROPI voltage V2 = 800.81 mV. RIPROPI is 1.65 kΩ (J19 populated). The IPROPI current would be V2 / 1650 = 800.81E-3 / 1650 =  4.85E-4. Using 212 μA / A, the AIPROPI value the calculated current is 4.85E-4/212E-6 = 2.288 A which is in agreement with V1 readout 22852 A. I tested with other load currents as well and found expected readout. 

Now, here's a tricky point to keep in mind and this is due to the EVM circuit behavior. Observation: When J19 is removed the measured voltage on each IPROPI pin is same. This is expected behavior because when the J19 short is removed each IPROPI pin will have 3.3 kΩ RIPROPI on them. So even if the current is 1/2 of the total value x2 is provided by the change from 1.65 kΩ to 3.3 kΩ! This is proof that each IPROPI output is 1/2 of the total current, the current flowing via each of the two parallel H-bridges. 

Dennis Formisano said:

Despite this seemingly correct configuration, the measured values are significantly lower than expected. Here are the details of our setup and results:

• Voltage measured at A0 (IPROPI output) with a multimeter: 0.378 V

• Current shown by the power supply (CV mode): 2.21 A

• Supply voltage: 5.2 V

• Raw ADC value (12-bit) read by the STM32: approximately 300–400

See below capture with VM = 5.2 A CV power source, Trigger for load current set at 2 A. The load is a BDC motor so it has an inrush current of several amps. The cursor on the load current at trigger point V1 reads 2 A. The corresponding IPROPI voltage reads V2 = 698.27 mV. With 2 A the IPROPI current should be 2 x 212E-6 = 424E-6 A, the voltage would be 424E-6 x 1650 = 699.6 mV which closely aligns with V2 = 698.27 mV. VREF was set to 3.3 V for the above two tests. Now, if I set VREF voltage to 699 mV the load current would be regulated at 2 A which means the inrush current would not exceed 2 A. See the second capture with VREF = 0.699 V.

Note: I'd not rely on the current shown by the power supply. Instead, measure the actual load current. Especially while running with current regulation 

VREF = 699 mV.

Dennis Formisano said:

We kindly ask if there is any known issue or configuration detail we might be missing that could cause this discrepancy. Is it possible that the IPROPI current mirror is not scaling symmetrically in single H-bridge mode, or could this be related to an internal current limit despite VREF being set to 1.65 V (externally applied)?

There are no known issues that'd impact IPROPI output in Single-H configuration PH/EN mode. See below capture with VREF = 1.65 V input from an external source to the suggested J7 header pin. With VREF = 1.65 V the inrush is expected to chop at around 4.72 A and not exceed this value. To help achieve the motor exceed this level of inrush current I increased VM to 24 V and control to 100 % duty cycle. 

VIPROPI should read close to 1.65 V. Cursor V1 reads 4.76 A and V2 reads 1.69 V pretty close to the expectation, albeit, minor errors likely contributed by the ITRIP circuit offset voltages.  

I hope this helps to debug your setup further. I hope you can resolve the observed issues. Thank you.

Regards, Murugavel 

Dear Murugavel,

thank you again for your continued support and the detailed explanations.

We have now implemented the recommended setup, including the isolation of the onboard MCU by removing the corresponding jumpers, and we are supplying an external VREF via pin 1 of J7. For the IPROPI signal, we used a 1 kΩ resistor and a 10 nF capacitor to form a low-pass filter as suggested.

Unfortunately, the resulting signal looks nothing like the clean waveforms shown in your scope captures. Even with different capacitor values, we still observe significant high-frequency noise on the IPROPI signal.
The best result was achieved when connecting the oscilloscope probe directly to the IS pin on the DRV8262 we averaged 10 consecutive ADC samples, but even that signal was disturbed and far from your picture (see attached screenshot).

We are now wondering whether this might be a hardware-related issue—possibly due to grounding, layout, or interference—but we are not entirely sure what the root cause could be.

We have also attached photos of our hardware setup, and we would greatly appreciate it if you could take a quick look. Perhaps you can spot a detail we have missed or suggest improvements based on what you see.

Thank you very much in advance for your time and help!

Best regards,
Dennis Formisano

this picure is directly connected with IPROPI pin and we averaged 10 consecutive ADC samples

This picture is with the Condensator 10 nF 

this is the condensator hardware setup

The attached photo shows our current wiring of the DRV8262EVM, including the direct IPROPI connection.

Thank you very much