DRV8301 EVM High Current Kit External Current Sense Amp Vs 8301 Native Current Sense Amp

Hi Mehran

Could you give more description on "control the motor the same way as ..." ? Could you give the partly schematic and test waveform of the false condition. Also please give more details of your application such as VM voltage and phase current rating. In what condition you think there are problems with the control?

We have a newly release design of DRV8301 and using two internal amps and one external amp. The performance is very well. You can refer to the schmatic and PCB design of that kit. All desige files can be downloaded from below page.

http://www.ti.com/tool/boostxl-drv8301

We will check if there is any special reason other than promoting to use external amps in the DRV8301-69M-KIT kit.

Thanks.

Wilson Zuo

Motor Application Team

The external 3rd amps is just used in over modulation region to use more of the bus voltage and get higher speeds.

If you are seeing issues at lower speeds it will be something in the hardware design, layout, or software configuration issue.  You may want to post over on the InstaSPIN forum.

I am also curious about the need for the external amps in lieu of the internal DRV8301 current sense amps on the DRV8301 high current EVM.  Why are there external amplifiers duplicating the function of the DRV8301's built-in amps for phase A and B?  Is there some deficiency in the internal amplifiers?  I know that there is no internal amp for phase C so an external one is desirable for that.  And why are the outputs of the external and internal units tied together on page 6 of the "D3 Engineering - TI - DRV8301/DRV8302 EVM - High Current" schematic?  That schematic indicates that the S01 and S02 signals are present on page 5 as well which would imply that they are passed to the controlcard, but I don't see them referenced on that page.  The configuration is a bit confusing to me and I can't find a good explanation for it.

Max,

To be able to do lower leg current shunt sensing throughout the entire PWM modulation range you need to be able to ping-pong between all three phases.  So just having 2 shunt current sensors means you can only do about 87% duty cycle. By adding the 3rd and doing current reconstruction and clever sample triggering you can move to full 100% duty sinewave and then full 100% duty trapezoidal modulation. More duty cycle = more average voltage = higher speeds.

The DRV8301 EVM was designed some time ago. For the controlSUITE versions of software we use the 2-shunt approach and the on-board PGAs (Rev C was first release), but we added the external 3x OPA for our MotorWare InstaSPIN-FOC & -MOTION solutions so we could do proper testing and control (hardware revised to Rev D).

After experimentation we realized that with proper layout we could use the 2 on-chip PGAs and add just a single off-chip for the third sample. We did this on the new BOOSTXL-DRV8301 inverter.

Chris,

I appreciate the need for 3 current shunt measurements, I just didn't understand why the need for all 3 to be handled by external op-amps.  Looks like you responded to that in the latter part of your post.  When you say, "with proper layout", it sounds like you might have some critical tips as to how a 2 on-chip+1 external configuration should be implemented successfully?  They would probably be relevant to the original question and my own interest.

Hi Max,

We don't have a whole lot of data on this, just a few general tips.

You definitely want to match the internal component selection around the onboard amplifiers as close as you can (GAIN, BIAS, etc and utilize fairly low tolerance components (1%)). As Chris mentioned you can check out the BOOSTXL-DRV8301 for our implementations of a 3-phase sense using the two onboard and 1 external.

One of the key layout tips is to maintain a differential connection for each of the 3 sense pairs and minimize trace lengths as best as possible.

With these two tips most of the other factors can be taken care of with an offset calibration.

Chris might have some additional comments.

well stated, Nick

Thanks Nick and Chris.  Those are pretty straightforward items but I will keep an eye on them.